ŽŵƉƌĞŚĞŶƐŝǀĞůĞĐƚƌŽŵĂŐŶĞƚŝĐ^ŽůƵƚŝŽŶƐ

User’s Manual

Suite 6.1

July 2011

Copyright 1998 – 2011: EM Software & Systems-S.A. (Pty) Ltd

32 Techno Avenue, Technopark, Stellenbosch, 7600, South Africa
Tel: +27-21-880-1880, Fax: +27-21-880-1936
E-Mail: [email protected]
WWW: http://www.feko.info
CONTENTS i

Contents

I The FEKO Suite

1 Introduction to the FEKO Suite 1-1

1.1 FEKO Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
1.1.1 FEKO solution engine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
Method of Moments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2
MLFMM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-3
Adaptive Cross Approximation (ACA) . . . . . . . . . . . . . . . . . . 1-4
Uniform Theory of Diffraction . . . . . . . . . . . . . . . . . . . . . . . 1-4
Geometrical Optics (ray launching) . . . . . . . . . . . . . . . . . . . 1-4
Physical Optics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-4
Large Element Physical Optics . . . . . . . . . . . . . . . . . . . . . . 1-4
Finite Element Method . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-5
General non-radiating networks . . . . . . . . . . . . . . . . . . . . . 1-5
Periodic boundaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-5
1.1.2 FEKO Suite components . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-5
1.2 Platforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-7
1.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-7
1.4 Changes in this release . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-7
1.5 Contacting your distributor or EMSS . . . . . . . . . . . . . . . . . . . . . . . . . 1-8

2 General modelling guidelines 2-1

2.1 Program flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1
2.2 Modelling and meshing guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1
2.2.1 Definitions and terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1
2.2.2 Meshing guidelines regarding element sizes . . . . . . . . . . . . . . . . 2-2
2.2.3 Meshing guidelines regarding connectivity . . . . . . . . . . . . . . . . . 2-4
2.3 Dielectric solids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-6
2.4 Checking the validity of the results . . . . . . . . . . . . . . . . . . . . . . . . . . 2-7

II Working with CADFEKO

3 Introduction to CADFEKO 3-1

3.1 CADFEKO overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
3.1.1 CADFEKO Getting started page . . . . . . . . . . . . . . . . . . . . . . . . 3-1
3.1.2 CADFEKO window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
3.1.3 The CADFEKO ribbon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3

July 2011 FEKO User’s Manual

CONTENTS ii

3.1.4 CADFEKO files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4

3.1.5 Preferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4
3D View rendering and colours . . . . . . . . . . . . . . . . . . . . . . 3-4
Messages and log files . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5
3.1.6 Checking for updates to the FEKO Suite . . . . . . . . . . . . . . . . . . 3-5
3.1.7 Batch Meshing - CADFEKO from a command line . . . . . . . . . . . . . 3-6
3.2 The 3D view . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7
3.2.1 3D view interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7
Rotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7
Zooming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7
Panning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7
Zoom to selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7
Snapping to special points . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
3.2.2 Manipulating items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
Selecting items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
Deleting items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-9
Point entry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-10
3.2.3 Display options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-10
View modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-10
Cutplanes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-11
Selective viewing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-12
Colouring options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-12
Element visibility controls . . . . . . . . . . . . . . . . . . . . . . . . . 3-13
Named points, annotations, solution entities and settings . . . . . . 3-13
3.2.4 View settings (View tab) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-13
Viewing angle and distance . . . . . . . . . . . . . . . . . . . . . . . . 3-13
Vertical Z-lock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-14
View undo and redo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-14
3.3 The model or contents tree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-14
3.3.1 Icons displayed in the model tree and details tree . . . . . . . . . . . . . 3-15
3.3.2 Medium icons in the details tree . . . . . . . . . . . . . . . . . . . . . . . 3-20
3.4 Project tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-20
3.4.1 Distance between points . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-20
3.4.2 Angle between points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-21
3.4.3 Calculator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-21
3.4.4 Notes editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-21
3.4.5 Archiving models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-22
3.5 Short-cut keys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-22

4 Creating and meshing geometry in CADFEKO 4-1

July 2011 FEKO User’s Manual

CONTENTS iii

4.1 Model geometry properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1

4.1.1 Setting units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1
4.1.2 Geometry extents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1
4.2 Variables and named points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-2
4.2.1 Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-2
4.2.2 Named points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-4
4.3 Creating geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-4
4.3.1 General geometry primitive dialog behaviour . . . . . . . . . . . . . . . 4-5
4.3.2 The workplane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-6
Coordinates set in a given workplane . . . . . . . . . . . . . . . . . . 4-7
4.3.3 Creating solid primitives . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-7
4.3.4 Creating surface primitives . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-10
4.3.5 Creating curve primitives . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-13
4.3.6 Import points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-18
4.4 Operations on geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-19
4.4.1 Boolean operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-20
The Union operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-21
The Intersect operation . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-21
The Subtract operation . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-21
4.4.2 Operations on parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-21
Projection and imprinting points . . . . . . . . . . . . . . . . . . . . . 4-21
Split parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-22
Exploding parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-22
Simplify tool (removing detail) . . . . . . . . . . . . . . . . . . . . . 4-23
Stitch sheet parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-24
4.4.3 Creation operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-24
Spinning and sweeping parts . . . . . . . . . . . . . . . . . . . . . . . 4-24
Loft surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-26
4.4.4 Administrative operations . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-26
Reversing face normals . . . . . . . . . . . . . . . . . . . . . . . . . . 4-26
Assemblies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-27
Copy objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-27
Special copy options . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-28
Collapse - removing the creation history of a part . . . . . . . . . . . 4-28
Re-evaluating the geometry . . . . . . . . . . . . . . . . . . . . . . . . 4-28
4.5 Transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-29
4.5.1 Transformations on geometry parts . . . . . . . . . . . . . . . . . . . . . 4-29
Rotate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-29
Translate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-30

July 2011 FEKO User’s Manual

CONTENTS iv

Mirror . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-30
Scale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-30
Align . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-30
4.5.2 Transformations on mesh parts . . . . . . . . . . . . . . . . . . . . . . . . 4-31
4.5.3 Multiple (transformed) copies . . . . . . . . . . . . . . . . . . . . . . . . . 4-31
4.6 Cable modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-31
4.6.1 Cable cross sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-32
4.6.2 Cable shield . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-33
4.6.3 Cable path . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-34
4.6.4 Requesting a cable harness request . . . . . . . . . . . . . . . . . . . . . . 4-36
The Cable tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-36
The Excitations and loads tab . . . . . . . . . . . . . . . . . . . . . . . 4-37
4.7 Geometry validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-37
4.8 Creating meshes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-38
4.8.1 Automatic meshing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-39
Wires . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-39
Faces and edges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-39
Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-40
4.8.2 Custom meshing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-40
4.8.3 Advanced meshing options . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-40
4.8.4 Specifying local mesh sizes . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-41
4.8.5 Editing mesh vertices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-42
4.8.6 Creating mesh triangles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-43
4.8.7 Reversing normals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-43
4.9 Mesh refinement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-44
4.9.1 Point refinement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-44
4.9.2 Polyline refinement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-44
4.10 Mesh manipulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-45
4.10.1 Mesh element radius . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-45
4.10.2 Relabelling mesh elements . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-45
4.10.3 Merging coincident vertices . . . . . . . . . . . . . . . . . . . . . . . . . . 4-46
4.10.4 Merging elements (removing vertices) . . . . . . . . . . . . . . . . . . . . 4-46
4.10.5 Merging mesh parts (union for meshes) . . . . . . . . . . . . . . . . . . . 4-47
4.10.6 Remove duplicate and collapsed elements . . . . . . . . . . . . . . . . . 4-47
4.11 Mesh validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-48
4.11.1 Mesh information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-48
4.11.2 Finding problems in the mesh . . . . . . . . . . . . . . . . . . . . . . . . . 4-48
4.11.3 View Distorted and Oversized mesh elements . . . . . . . . . . . . . . . 4-48
4.11.4 View free mesh edges and segment nodes . . . . . . . . . . . . . . . . . 4-49

July 2011 FEKO User’s Manual

CONTENTS v

4.11.5 Find intersecting mesh triangles . . . . . . . . . . . . . . . . . . . . . . . 4-49

5 Importing and exporting 5-1

5.1 Importing and exporting geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1
5.1.1 Importing CADFEKO models . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1
5.1.2 Working with Parasolid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-2
Import of Parasolid models . . . . . . . . . . . . . . . . . . . . . . . . 5-2
Export of Parasolid models . . . . . . . . . . . . . . . . . . . . . . . . . 5-2
5.1.3 General CAD formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-3
Importing general CAD . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-3
Importing AutoCAD DXF . . . . . . . . . . . . . . . . . . . . . . . . . . 5-3
Exporting general CAD formats . . . . . . . . . . . . . . . . . . . . . 5-5
5.1.4 Faults on geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-5
5.2 Importing meshes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-5
5.2.1 Importing *.fek files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-5
5.2.2 Importing general mesh formats . . . . . . . . . . . . . . . . . . . . . . . 5-6
5.3 Exporting meshes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-6
5.4 Printing and image export . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-6

6 Defining the solution configuration in CADFEKO 6-1

6.1 Using dielectric and lossy media . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1
6.1.1 Defining media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1
6.1.2 Setting material properties on geometry . . . . . . . . . . . . . . . . . . 6-10
6.1.3 Setting properties on mesh elements . . . . . . . . . . . . . . . . . . . . . 6-15
6.1.4 Displaying dielectric media, coatings and thin sheets . . . . . . . . . . . 6-17
6.1.5 Displaying the model view, simulation mesh and overlay . . . . . . . . 6-17
6.2 Setting the frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-17
6.3 Setting the total source power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-19
6.4 General networks and ideal transmission lines . . . . . . . . . . . . . . . . . . . 6-20
6.4.1 Ideal transmission lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-20
6.4.2 General networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-21
6.4.3 Connecting, exciting and loading general networks . . . . . . . . . . . . 6-22
6.5 Sources independent of the geometry . . . . . . . . . . . . . . . . . . . . . . . . . 6-23
6.5.1 Plane waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-23
6.5.2 Electric and magnetic point dipoles . . . . . . . . . . . . . . . . . . . . . 6-24
6.5.3 Point sources with a specified pattern . . . . . . . . . . . . . . . . . . . . 6-25
6.5.4 Impressed spherical mode excitation . . . . . . . . . . . . . . . . . . . . . 6-27
6.5.5 Impressed aperture excitation . . . . . . . . . . . . . . . . . . . . . . . . . 6-29
6.5.6 Impressed current . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-30
6.6 Setting ports, sources and loads on geometry . . . . . . . . . . . . . . . . . . . . 6-31

July 2011 FEKO User’s Manual

CONTENTS vi

6.6.1 Ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-31

Wire ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-33
Edge ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-34
Microstrip ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-36
Waveguide ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-37
FEM line ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-39
FEM modal ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-41
6.6.2 Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-42
Voltage sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-42
Current sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-43
Waveguide excitations . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-44
FEM modal excitations . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-45
6.6.3 Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-45
6.7 Infinite planes — Planar Green’s functions and ground planes . . . . . . . . . . 6-46
6.8 Symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-49
6.9 Periodic boundary conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-50
6.10 Numerical Green’s function (NGF) . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-51
6.11 Solution control and special solver settings . . . . . . . . . . . . . . . . . . . . . . 6-52
6.11.1 Setting solution parameters on faces . . . . . . . . . . . . . . . . . . . . . 6-56
6.11.2 Viewing special solution parameters on faces . . . . . . . . . . . . . . . 6-57
6.12 Requesting results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-57
6.12.1 Currents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-58
6.12.2 S-parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-59
6.12.3 Transmission/reflection coefficients . . . . . . . . . . . . . . . . . . . . . 6-60
6.12.4 Far fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-61
6.12.5 Near fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-64
6.12.6 SAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-65
6.12.7 Cable harness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-66
6.12.8 Ideal receiving antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-67
6.13 Error estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-68
6.14 Mesh manipulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-69
6.15 CEM validation of the CADFEKO model . . . . . . . . . . . . . . . . . . . . . . . . 6-69
6.16 Running the FEKO solver components . . . . . . . . . . . . . . . . . . . . . . . . . 6-69
6.16.1 Component parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-70
6.17 Additional solution configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-74
6.18 Working with CADFEKO models in EDITFEKO . . . . . . . . . . . . . . . . . . . . 6-74
6.18.1 Disabling the CADFEKO solution configuration . . . . . . . . . . . . . . 6-75
6.18.2 Setting units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-75
6.18.3 Referencing elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-75

July 2011 FEKO User’s Manual

CONTENTS vii

6.18.4 Using variables and named points in EDITFEKO . . . . . . . . . . . . . . 6-76

6.18.5 Setting dielectric parameters . . . . . . . . . . . . . . . . . . . . . . . . . 6-76

7 Setting up optimisation in CADFEKO 7-1

7.1 Defining optimisation searches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1
7.1.1 Setting the optimisation method and stopping criteria . . . . . . . . . . 7-2
7.1.2 Defining optimisation parameters . . . . . . . . . . . . . . . . . . . . . . 7-3
7.1.3 Defining optimisation masks . . . . . . . . . . . . . . . . . . . . . . . . . . 7-5
7.1.4 Setting up optimisation Goals . . . . . . . . . . . . . . . . . . . . . . . . . 7-7
7.2 Optimisation Goal types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-11
7.2.1 The Impedance Goal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-11
7.2.2 The Near field Goal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-12
7.2.3 The Far field Goal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-14
7.2.4 The S-parameter Goal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-15
7.2.5 The SAR Goal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-17
7.2.6 The Power Goal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-17
7.2.7 The Receiving antenna Goal . . . . . . . . . . . . . . . . . . . . . . . . . . 7-17
7.2.8 The Transmission/reflection Goal . . . . . . . . . . . . . . . . . . . . . . . 7-18
7.3 The global Goal: combining and weighting of multiple Goals . . . . . . . . . . 7-19
7.3.1 Goal weighting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-20
7.3.2 Goal combination tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-20
7.4 Specifying special optimisation solver settings . . . . . . . . . . . . . . . . . . . . 7-20

III Working with POSTFEKO

8 Introduction to POSTFEKO 8-1

8.1 Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1
8.2 Start page . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-3
8.3 Application menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-4
8.4 Ribbon tabs: tools and controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-5
8.4.1 Default tabs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-5
8.4.2 Context-sensitive tabs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-6
3D View context-sensitive tabs . . . . . . . . . . . . . . . . . . . . . . 8-6
2D View context-sensitive tabs . . . . . . . . . . . . . . . . . . . . . . 8-6
8.5 System operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-7
8.5.1 Application Button and quick access toolbar . . . . . . . . . . . . . . . . 8-7
8.5.2 Save and load project sessions . . . . . . . . . . . . . . . . . . . . . . . . . 8-7
8.5.3 Preferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-7
8.5.4 Animation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-8

July 2011 FEKO User’s Manual

CONTENTS viii

8.5.5 Import and export features . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-9

Importing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-9
Exporting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-10
8.5.6 Storing data and using stored data . . . . . . . . . . . . . . . . . . . . . . 8-11
8.5.7 Math scripting using Lua in POSTFEKO . . . . . . . . . . . . . . . . . . . 8-11

9 Using POSTFEKO 9-1

9.1 Launching POSTFEKO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-1
9.2 Typical user scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-1
9.2.1 Managing projects/models . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-1
9.2.2 Add results to a view . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-2
9.2.3 Convert graph types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-3
9.2.4 Using 2D graphs (Cartesian, Smith and polar) . . . . . . . . . . . . . . . 9-4
The Display tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-4
The Trace tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-5
The Measure tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-6
The Format tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-8
9.2.5 Using 3D views . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-9
The Display tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-9
The Mesh tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-12
The Result tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-15
The Animate tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-18
9.2.6 Using legends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-19
9.2.7 Use the project browser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-21
9.2.8 Manually specify axis properties, ranges and captions . . . . . . . . . . 9-21
9.2.9 Using math traces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-23
9.2.10 Using math scripts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-23
9.2.11 Generating reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-25
Exporting and copying data, images and animations . . . . . . . . . 9-25
Quick reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-25
Generating a report from a template . . . . . . . . . . . . . . . . . . . 9-26
9.3 Errors and warnings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-27
9.3.1 Warning icon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-27
9.3.2 POSTFEKO workflow error and warning dialog . . . . . . . . . . . . . . 9-27
9.3.3 Error and warning logging dialog . . . . . . . . . . . . . . . . . . . . . . . 9-28
9.4 Getting Help . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-28

10 Scripting guide for POSTFEKO 10-1

10.1 The interface between POSTFEKO and the scripting environment . . . . . . . . 10-1
10.2 Pulling data into the scripting environment . . . . . . . . . . . . . . . . . . . . . 10-2

July 2011 FEKO User’s Manual

CONTENTS ix

10.3 The structure of a dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-2

10.3.1 Axes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-4
10.3.2 Quantities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-5
10.3.3 Metadata . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-6
10.4 Processing and modifying a dataset . . . . . . . . . . . . . . . . . . . . . . . . . . 10-6
10.4.1 Indexing a single element in a dataset . . . . . . . . . . . . . . . . . . . . 10-6
10.4.2 Iterating through all elements in a dataset . . . . . . . . . . . . . . . . . 10-7
10.5 Creating a custom dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-7
10.6 Structures for internal dataset types . . . . . . . . . . . . . . . . . . . . . . . . . . 10-8
10.6.1 Near field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-9
10.6.2 Far field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-11
Directivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-12
Gain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-12
Realised Gain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-12
10.6.3 Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-13
10.6.4 Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-14
10.6.5 Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-15
10.6.6 S-Parameter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-16
10.6.7 Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-17
10.6.8 Transmission and reflection coefficients . . . . . . . . . . . . . . . . . . . 10-18
10.6.9 Specific Absorption Rates (SAR) . . . . . . . . . . . . . . . . . . . . . . . 10-19

IV Working with EDITFEKO

11 Introduction to EDITFEKO 11-1

11.1 Toolbars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-2
11.1.1 File control toolbar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-2
11.1.2 The edit and window toolbars . . . . . . . . . . . . . . . . . . . . . . . . . 11-3
11.2 Menu items in EDITFEKO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-3
11.2.1 File menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-3
11.2.2 Edit menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-3
11.2.3 Accessing Geometry Control card definition panels . . . . . . . . . . . . 11-4
11.2.4 Run menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-4
11.2.5 Window menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-5
11.2.6 Help menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-5
11.3 Preferences in EDITFEKO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-5
11.3.1 Appearance tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-5
11.3.2 Options tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-6
11.4 Inserting and editing input cards and variables . . . . . . . . . . . . . . . . . . . 11-6

July 2011 FEKO User’s Manual

CONTENTS x

11.4.1 Generating input cards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-6

11.4.2 Variable editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-8
11.5 Important keystrokes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-9

12 PREFEKO language and concepts 12-1

12.1 PREFEKO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-1
12.2 Structure of the *.pre input file . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-1
12.2.1 Order of the cards in the *.pre file . . . . . . . . . . . . . . . . . . . . . 12-1
12.2.2 Format of cards in the *.pre file . . . . . . . . . . . . . . . . . . . . . . . 12-1
12.3 Guidelines for creating geometry in EDITFEKO . . . . . . . . . . . . . . . . . . . 12-3
12.3.1 Meshing guidelines regarding connectivity . . . . . . . . . . . . . . . . . 12-3
12.3.2 Reducing GUI mesh sizes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-4
12.4 Usage and concept of labels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-4
12.5 Symbolic variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-6
12.6 Built-in functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-7
12.7 FOR/NEXT loops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-12
12.8 IF/ELSE/ENDIF constructs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-13
12.9 EXIT command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-14
12.10 PRINT commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-14
12.11 Symbolic node names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-15

13 Geometry cards 13-1

13.1 ** card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-3
13.2 BL card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-4
13.3 BP card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-6
13.4 BQ card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-8
13.5 BT card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-10
13.6 CB card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-12
13.7 CL card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-14
13.8 CN card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-16
13.9 DD card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-17
13.10 DK card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-18
13.11 DP card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-20
13.12 DZ card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-21
13.13 EG card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-23
13.14 EL card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-25
13.15 FM card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-27
13.16 FO card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-29
13.17 FP card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-30
13.18 HC card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-31

July 2011 FEKO User’s Manual

CONTENTS xi

13.19 HE card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-32

13.20 HP card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-34
13.21 HY card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-35
13.22 IN card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-36
13.23 IP card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-55
13.24 KA card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-56
13.25 KK card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-57
13.26 KL card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-61
13.27 KR card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-62
13.28 KU card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-64
13.29 LA card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-66
13.30 MB card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-67
13.31 ME card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-68
13.32 NU card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-71
13.33 PB card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-73
13.34 PE card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-75
13.35 PH card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-76
13.36 PM card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-78
13.37 PO card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-80
13.38 PY card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-84
13.39 QT card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-86
13.40 QU card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-87
13.41 RM card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-89
13.42 SF card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-93
13.43 SY card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-95
13.44 TG card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-96
13.45 TO card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-99
13.46 TP card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-101
13.47 UT card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-103
13.48 UZ card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-107
13.49 VS card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-108
13.50 WA card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-111
13.51 WG card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-112
13.52 WR card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-113
13.53 ZY card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-114

14 Control cards 14-1

14.1 ** card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-5
14.2 Ax Cards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-6
14.3 A0 card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-9

July 2011 FEKO User’s Manual

CONTENTS xii

14.4 A1 card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-12

14.5 A2 card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-13
14.6 A3 card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-15
14.7 A4 card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-17
14.8 A5 card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-19
14.9 A6 card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-20
14.10 A7 card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-21
14.11 AB card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-22
14.12 AC card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-23
14.13 AE card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-25
14.14 AF card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-28
14.15 AI card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-29
14.16 AK card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-31
14.17 AN card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-34
14.18 AP card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-35
14.19 AR card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-41
14.20 AS card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-46
14.21 AV card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-52
14.22 AW card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-54
14.23 BO card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-61
14.24 CA card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-62
14.25 CD card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-67
14.26 CF card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-71
14.27 CG card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-73
14.28 CM card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-77
14.29 CO card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-78
14.30 CS card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-80
14.31 DA card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-82
14.31.1 General File Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-83
14.31.2 *.efe / *.hfe file: Electric / Magnetic near fields . . . . . . . . . . . . . . 14-84
14.31.3 *.ffe file: Far fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-87
14.31.4 *.ol file: Surface charge density . . . . . . . . . . . . . . . . . . . . . . . . 14-89
14.31.5 *.os file: Surface current density . . . . . . . . . . . . . . . . . . . . . . . 14-90
14.31.6 Other supported file formats . . . . . . . . . . . . . . . . . . . . . . . . . . 14-92
14.32 DI card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-93
14.33 DL card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-97
14.34 EE card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-99
14.35 EN card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-100
14.36 FE card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-101

July 2011 FEKO User’s Manual

CONTENTS xiii

14.37 FF card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-106

14.38 FR card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-109
14.39 GF card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-112
14.40 L2 card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-118
14.41 L4 card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-119
14.42 LC card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-120
14.43 LD card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-121
14.44 LE card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-122
14.45 LF card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-124
14.46 LN card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-125
14.47 LP card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-126
14.48 LS card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-127
14.49 LZ card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-128
14.50 NW card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-129
14.51 OF card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-132
14.52 OS card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-133
14.53 PP card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-135
14.54 PS card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-137
14.55 PW card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-139
14.56 RA card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-143
14.57 SA card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-145
14.58 SH card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-147
14.59 SK card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-150
14.60 SP card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-155
14.61 TL card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-157
14.62 TR card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-160
14.63 WD card . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-161

V Working with SECFEKO

15 The program SECFEKO 15-1

15.1 Introduction to SECFEKO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-1
15.2 Displaying licence information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-1
15.3 Managing floating licences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-3
15.3.1 Concept of a preferred licence . . . . . . . . . . . . . . . . . . . . . . . . . 15-4
15.3.2 Managing floating licence servers . . . . . . . . . . . . . . . . . . . . . . . 15-5
15.3.3 Allow or deny usage of certain floating licences . . . . . . . . . . . . . . 15-5
15.4 Determining machine codes and creating licence requests . . . . . . . . . . . . 15-6
15.5 SECFEKO command line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-6

July 2011 FEKO User’s Manual

CONTENTS xiv

VI Working with QUEUEFEKO

16 Introduction to QUEUEFEKO 16-1

16.1 QUEUEFEKO system overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-1
16.2 Creating and extracting packages . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-1
16.2.1 Package configuration files . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-2
16.2.2 Including package files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-2
16.2.3 FEKO options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-3
16.2.4 CLUSTER options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-3
16.2.5 Using encryption (optional) . . . . . . . . . . . . . . . . . . . . . . . . . . 16-4
16.2.6 Generate package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-5
16.2.7 Adding the package to the execution queue . . . . . . . . . . . . . . . . 16-5
16.2.8 Extract package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-5
16.2.9 Decryption of package (if required) . . . . . . . . . . . . . . . . . . . . . 16-6
16.2.10 View results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-6
16.3 Setting preferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-6

VII Preprocessing and the FEKO solver

17 The preprocessor PREFEKO 17-1

17.1 Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-1
17.2 Running PREFEKO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-1

18 The FEKO solution kernel 18-1

18.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-1
18.2 Running the FEKO kernel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-1
18.2.1 Running the sequential version . . . . . . . . . . . . . . . . . . . . . . . . 18-1
18.2.2 Running the parallel version . . . . . . . . . . . . . . . . . . . . . . . . . . 18-2
18.2.3 Running on a remote host . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-6
General settings and usage . . . . . . . . . . . . . . . . . . . . . . . . . 18-6

19 Description of the output file of FEKO 19-1

19.1 Geometric data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-1
19.2 Excitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-5
19.3 Currents and charges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-8
19.4 Finite conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-10
19.5 Near fields and SAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-11
19.6 Far fields and receiving antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-12
19.7 S-parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-16
19.8 Computation time and peak memory . . . . . . . . . . . . . . . . . . . . . . . . . 19-17

July 2011 FEKO User’s Manual

CONTENTS xv

VIII FEKO Utilities

20 The optimiser OPTFEKO 20-1

20.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-1
20.2 The optimisation method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-1
20.2.1 Simplex (Nelder-Mead) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-2
20.2.2 Particle swarm optimisation (PSO) . . . . . . . . . . . . . . . . . . . . . . 20-5
20.2.3 Genetic algorithm (GA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-8
20.2.4 Grid search . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-10
20.3 Sensitivity analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-12
20.4 Running OPTFEKO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-13

21 The program ADAPTFEKO 21-1

21.1 Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-1
21.2 Running ADAPTFEKO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-1
21.3 The *.pre input file for adaptive sampling . . . . . . . . . . . . . . . . . . . . . . 21-1

22 The program TIMEFEKO 22-1

22.1 Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-1
22.2 The *.pre input file for TIMEFEKO . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-1
22.3 The *.tim input file . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-2
22.3.1 Defining the pulse form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-2
22.3.2 Definition of the frequency block . . . . . . . . . . . . . . . . . . . . . . . 22-5
22.3.3 Definitions of the normalisation . . . . . . . . . . . . . . . . . . . . . . . . 22-6
22.3.4 Definition of the excitation output . . . . . . . . . . . . . . . . . . . . . . 22-6
22.3.5 Definition of a time point . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-6
22.4 Running TIMEFEKO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-7
22.5 TIMEFEKO output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-7

23 The FEKO software updater 23-1

23.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-1
23.2 GUI update utility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-1
23.3 Command-line update utility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-2
23.4 Proxy settings - advanced . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-3
23.5 Creating a local update repository . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-3

IX Appendix

24 Advanced modelling and solution control 24-1

24.1 Utilisation of symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-1

July 2011 FEKO User’s Manual

CONTENTS xvi

24.1.1 Geometric symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-1

24.1.2 Electric symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-1
24.1.3 Magnetic symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-2
24.1.4 Example of the application of symmetry . . . . . . . . . . . . . . . . . . 24-2
24.1.5 Special enforcement of symmetry: Even – odd method . . . . . . . . . 24-3
24.2 Dynamic memory management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-4
24.2.1 Telling FEKO how much memory can be used . . . . . . . . . . . . . . . 24-4
24.2.2 Variables that are automatically set correctly . . . . . . . . . . . . . . . . 24-5
24.3 Environment variables and registry keys . . . . . . . . . . . . . . . . . . . . . . . 24-7

25 Summary of files 25-1

26 SPICE3f5 general structure, conventions and syntax 26-1

27 CADFEKO geometry faults 27-1

28 Copyright notices and acknowledgements 28-1

28.1 Copyright to Voronoi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-1
28.2 Copyright to Berkeley Spice 3f5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-1
28.3 Copyright to SuperLU . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-2
28.4 Copyright of libcurl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-2
28.5 Qwt project usage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-3
28.6 HOOPS and Parasolid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-3
28.7 MeshSim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-3
28.8 FFmpeg (export video files) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-3
28.9 Microsoft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-3
28.10 Copyright of libXML2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-4

X Index

Index I-1

July 2011 FEKO User’s Manual

Part I

The FEKO Suite

INTRODUCTION TO THE FEKO SUITE 1-1

1 Introduction to the FEKO Suite

The name FEKO is an abbreviation derived from the German phrase FEldberechnung bei Körpern
mit beliebiger Oberfläche. (Field computations involving bodies of arbitrary shape.) As the name
suggests, FEKO can be used for various types of electromagnetic field analyses involving objects
of arbitrary shapes.

1.1 FEKO Overview

FEKO is a software Suite intended for the analysis of a wide range of electromagnetic problems.
Applications include EMC analysis, antenna design, microstrip antennas and circuits, dielectric
media, scattering analysis and many more. The kernel provides a comprehensive set of powerful
computational methods and has been extended for the analysis of thin dielectric sheets, multiple
homogeneous dielectric bodies and planar stratified media. Figure 1-1 illustrates some of the
numerical analysis techniques available in FEKO and the types of problems for which they are
intended.

Figure 1-1: Illustration of the numerical analysis techniques in FEKO.

1.1.1 FEKO solution engine

The Method of Moments (MoM) technique forms the basis of the FEKO solver. Other techniques
such as the Multilevel Fast Multipole Method (MLFMM), the Finite Element Method (FEM) Uni-
form Theory of Defraction (UTD), Geometrical optics (ray launching) and Physical Optics (PO)
have been implemented to allow the solving of electrically large problems and inhomogeneous
dielectric bodies of arbitrary shape. Special approximations and acceleration techniques are
available for problems of specific types.
FEKO provides for parallel processing usage on a range of workstations, servers and clusters. The
performance for each platform, operating system and deployment method has been optimised
for the delivery of accurate and timely results.

July 2011 FEKO User’s Manual

INTRODUCTION TO THE FEKO SUITE 1-2

Method of Moments

The core of the program FEKO is based on the Method of Moments (MoM). The MoM is a full
wave solution of Maxwell’s integral equations in the frequency domain. An advantage of the
MoM is that it is a “source method” meaning that only the structure in question is discretised,
not free space as with “field methods”. Boundary conditions do not have to be set and memory
requirements scale proportional to the geometry in question and the required solution frequency.
The following special extensions have been included in FEKO’s MoM formulation to enable the
modelling of magnetic and dielectric media.

Surface Equivalence Principle (SEP): The SEP introduces equivalent electric and magnetic cur-
rents on the surface of a closed dielectric body. The surface of such bodies can be arbitrarily
shaped and is discretized using triangles.

Volume Equivalence Principle (VEP): The VEP allows the creation of dielectric bodies from
cuboids (in EDITFEKO) or tetrahedra. More basis functions are typically required than for
the SEP, but neighbouring cuboids or tetrahedra may have differing electric and magnetic
properties.
The volume equivalence principle is associated with a volume mesh and general usability
is inhibited by the order O(N 2..3 ) memory and CPU-time scaling with number unknowns N .
There are however special cases where the VEP is advantageous over the SEP or the
FEM/MoM:

• The formulation is stable at low frequencies

• It also displays good stability and convergence properties for an iterative solution with
the MLFMM
• It is well- suited to inhomogeneous, thin dielectric bodies

Note that the VEP is not supported together with dielectric solution methods (SEP, FEM,
VEP with cuboids, special Green’s fucntions) and periodic boundary conditions.

Planar Green’s Functions for Multilayered Media: Multilayered dielectric media may be mod-
elled with Green’s functions, e.g. substrates for microstrip architecture. The special Green’s
function formulation implements 2D infinite planes with finite thickness to handle each
layer of the dielectric. Conducting surfaces and wires inside the dielectric layers have to
be discretized, but not the dielectric layers themselves. Metallic surfaces and wires can be
arbitrarily oriented in the media and are allowed to cross multiple layers. (Calculations
using Green’s functions are accelerated by using interpolation tables.)

Thin Dielectric Sheets: Multiple layers of thin dielectric and anisotropic sheets can be analysed
as a single layer in FEKO. Typical applications are the analysis of radome covered antennas
and windscreens of automobiles.

Dielectrically Coated Wires: FEKO implements two methods for the modelling of dielectric and
magnetic coatings on wires:

• Popovic’s formulation modifies the radius of the metallic wire core to change the ca-
pacitive loading on the wire, while simultaneously adding a corresponding inductive
load. The method is restricted in that the loss tangent of the layer has to be identical
to the loss tangent of the surrounding medium.

July 2011 FEKO User’s Manual

INTRODUCTION TO THE FEKO SUITE 1-3

• Pure dielectric layers (i.e. relative permeability of the layer equals that of the sur-
rounding medium) should be modelled with the equivalence theorem where the effect
of the dielectric layering is accounted for by a volume polarisation current. The only
restriction on the method is that the layering may not be magnetic.

Real Ground: Real ground can be modelled with the reflection coefficient approximation or the
exact Sommerfeld formulation.

Windscreen: Multiple windscreen antennas on multiple glass definitions can be analysed in one
model. This solution method is much faster than modelling the windscreen, antennas and
layers using any other technique.

Planar Green’s function aperture A planar Green’s function slot or aperture interface is discre-
tised to yield a more efficient solution. Simulations using this method are much faster than
discretising the finite size ground plane that surrounds a slot or aperture, since the number
of triangles can be greatly reduces.

MLFMM

The MLFMM is an alternative formulation of the technology behind the MoM and is applicable
to much larger structures than the MoM, making full-wave current-based solutions of electrically
large structures a possibility. This fact implies that it can be applied to most large models that
were previously treated with the MoM without having to change the mesh.
The agreement between the MoM and MLFMM is that basis functions model the interaction be-
tween all triangles. The MLFMM differs from the MoM in that it groups basis functions and
computes the interaction between groups of basis functions, rather than between individual ba-
sis functions. FEKO employs a boxing algorithm that encloses the entire computational space
in a single box at the highest level, dividing this box in three dimensions into a maximum of
eight child cubes and repeating the process iteratively until the side length of each child cube
is approximately a quarter wavelength at the lowest level. Only populated cubes are stored at
each level, forming an efficient tree-like data structure. In the MoM framework the MLFMM is
implemented through a process of aggregation, translation and disaggregation of the different
levels.
The MoM treats each of N basis functions in isolation, thus resulting in an N 2 scaling of memory
requirements (to store the impedance matrix) and N 3 in CPU-time (to solve the linear set of
equations). It is thus clear that processing requirements for MoM solutions scale rapidly with
increasing problem size. The MLFMM formulation’smore efficient
2 treatment of the same problem
results in N ∗ log(N ) scaling in memory and N ∗ log(N ) in CPU time. In real applications
this reduction in solution requirements can range to orders of magnitude. Significant effort has
also been invested in improving the parallel MLFMM formulation to achieve exceptionally high
efficiency when distributing a simulation over multiple processors.
Note that the MLFMM can now be applied in the hybrid FEM/MoM framework in FEKO to reduce
computational resource requirements associated with the MoM part.

July 2011 FEKO User’s Manual

INTRODUCTION TO THE FEKO SUITE 1-4

Adaptive Cross Approximation (ACA)

The ACA is a fast method similar to the MLFMM but is also applicable to low frequency problems
or when using a special Green’s function. It approximates the impedance matrix by constructing
a sparse H-matrix (only a few selected elements are computed).

Uniform Theory of Diffraction

FEKO hybridises the current-based accurate MoM with the UTD in the truest sense of the word
with the coupling between the MoM and UTD being maintained in the solution, i.e. modifying
the interaction matrix and ensuring accuracy. A practical example would be a changing input
impedance of a dipole treated with the MoM, in close proximity to a large structure treated
with the UTD. Frequency does not influence the memory resources required for UTD treatment
of a structure as only points of reflection from surfaces and diffraction from edges or corners
are considered without meshing the structure. Edge and corner diffraction, double diffraction
and creeping waves (cylinders) are taken into account. Insight into the propagation of rays are
provided in POSTFEKO during post processing. Currently the numerical formulation of the UTD
only allows it to be applied to flat polygonal plates with minimum edge length in the order of a
wavelength or to single cylinders. The UTD is thus quite well suited to the analysis of ships at
radar or electronic wave frequencies, but not well suited to the analysis of complex objects with
curved surfaces, e.g. automobiles.

Geometrical Optics (ray launching)

The Geometrical optics (ray launching) is a ray-based method intended for the consideration of
electrically large dielectric and perfect electrically conducting structures in applications like the
analysis of lens antennas. The GO method is hybridised with the MoM in a similar fashion to the
UTD. The GO method in FEKO employs ray-launching and transmission, reflection and refraction
theory to model the interaction between the dielectric region and the MoM.

Physical Optics

PO is formulated for use in instances where electrically very large structures are modelled. PO
is an asymptotic high frequency numerical method of the same nature as the UTD. Users will
typically attempt a solution with the MoM at first and when they realise that the structure is
electrically too large to solve with their available resources (platform memory, time) they will
turn to one of the asymptotic high frequency techniques.

Large Element Physical Optics

Large element PO is formulated for use in instance where electrically very large smooth structures
are modelled. This method is only to be used when there are no discontinuities in the incident
field (e.g. if the incident field closely represents a point source). Large element PO is similar to
PO in that it is an asymptotic high frequency numerical method of the same nature as the UTD.

July 2011 FEKO User’s Manual

INTRODUCTION TO THE FEKO SUITE 1-5

The high frequency large element physical optics method is applicable for large smooth areas
when calculating near and far fields. If the large element PO is used together with the MoM
method, the MoM and PO regions are to be decoupled. Note that the same options that are
available for PO are also valid for large element PO.

Finite Element Method

The FEM is applicable to the modelling of electrically large or inhomogeneous dielectric bodies,
which are not efficiently solvable with FEKO’s extensions to the MoM. The FEM is a volume mesh-
ing technique that employs tetrahedra to accurately mesh arbitrarily shaped volumes where the
dielectric properties may vary between neighbouring tetrahedra. FEM modelling is advantages
in these instances because FEM solution matrices are sparse, where MoM matrices are densely
populated, making FEM matrices significantly more scalable with an increase in frequency.
The MoM/FEM hybridisation features full coupling between metallic wires and surfaces in the
MoM region and heterogeneous dielectric bodies in the FEM region. The MoM part of the solu-
tion is calculated first, which results in equivalent magnetic and electric currents that form the
radiation boundary of the FEM region. This hybrid technique makes use of the strengths of both
the MoM and the FEM in the following ways:

• The MoM is used for the efficient modelling of open boundary radiating structures where
no 3D space discretisation is required.

• The FEM is used for the efficient modelling of inhomogeneous dielectric bodies in terms of
field distributions inside the volume.

General non-radiating networks

General networks (defined using network parameter matrices) as well as ideal non-radiating
transmission lines may be used in FEKO simulations. These non-radiating networks may be
interconnected (cascaded) and excited or loaded directly at the ports. The voltages and currents
at the ports of these ideal representations of networks may interact with currents and voltages
on parts of the model that are solved using other solution methods, though no radiation-based
coupling is taken into account.

Periodic boundaries

Large, equally-spaced periodic structures may be simulated in FEKO using an infinite periodic
boundary approach. This approach may be used to provide an accurate accelerated solution for
many applications like frequency selective surface analysis and large array analysis.

1.1.2 FEKO Suite components

The graphical user interface consists of the components:

July 2011 FEKO User’s Manual

INTRODUCTION TO THE FEKO SUITE 1-6

• CADFEKO is used to create and mesh the geometry and to specify the solution settings and
calculation requirements in a graphical environment.

• EDITFEKO is used to construct advanced models (both the geometry and solution require-
ments) using a high level scripting language which includes repetitive FOR loops and con-
ditional IF–ELSE statements.

• POSTFEKO reads results form binary output files (*.bof) and can display the results on
2D graphs or in combination with the geometry in 3D views. POSTFEKO is also used to
visualise optimisation results during and after optimisation, as well as the meshed geometry
of the FEKO model, with excitations, field requests points etc. before the actual FEKO run.

• QUEUEFEKO facilitates the creation of packages which can be transported to remote cluster
machines where the package is placed in an execution queue (such as PBS).

• FEKO_UPDATE is a command line tool that can be used to check if updates are available
from a master (internet) or local repository. The FEKO GUI update tool is an interactive
application that allows the user to set preferences regarding the automatic download of
updates.

• SECFEKO_GUI is a visualisation of the FEKO licence manager. See SECFEKO for more
details on the licence management tool.

Other components that form part of the FEKO Suite do not provide a graphical interface. These
are concerned with the analysis and solution of the electromagnetic problem as defined in the
GUI components, or the maintenance and administration of the Suite. Components are launched
indirectly from the GUI components, but may also be launched from a command line. The
solution components are fully supported by a large range of platforms.

• PREFEKO processes the model and prepares the input file (*.fek) for the FEKO solution
kernel.

• FEKO is the actual solver code. The ASCII (*.out) and binary (*.bof) output files gener-
ated by FEKO contain all the solution information.

• OPTFEKO is a tool that is used for the optimisation of a FEKO model according to specific
requirements. OPTFEKO calls the FEKO solver as required during optimisation.

• TIMEFEKO provides a Fourier-transform based time-domain analysis mechanism for FEKO.

TIMEFEKO calls the FEKO solver as required during the solution process.

• ADAPTFEKO is used in the generation of continuous adaptively sampled results. ADAPT-

FEKO is called as required by the FEKO kernel when continuously sampled results are
required.

• CADFEKO_BATCH is a command line tool that can be used to modify variable values in a
CADFEKO model file from a command-line interface without launching the CADFEKO GUI.

• SECFEKO is the FEKO licence manager and shows all the licences in the specified licence
file (secfeko.dat) for node locked licences or connects to the floating licence servers and
retrieves licence information.

July 2011 FEKO User’s Manual

INTRODUCTION TO THE FEKO SUITE 1-7

1.2 Platforms

The FEKO kernel components are available on PC’s and a wide variety of workstations. The GUI
components CADFEKO, EDITFEKO, and POSTFEKO are available on PC’s running MS Windows
or Linux. All pre- and post processing must thus be performed on a PC, while the actual com-
putationally intensive field calculations can be performed on a workstation, parallel cluster or
on the PC itself as required. FEKO includes a remote launching facility to make such a remote
execution easy to use from within the GUI running on the PC.

1.3 Examples

First time users are advised to view the Demo Example video provided with the FEKO installation
and work through the Getting Started Manual (additional demonstration videos are available
on the FEKO website1 ). The Getting Started Manual gives a basic introduction to the different
components of the FEKO Suite and introduces the basic workflow required for effective FEKO
use by helping the user complete a set of step-by-step projects. It is also recommended that new
users read the introductory sections, and the chapter on modelling guidelines (see section 2) in
this document carefully.
Various simple FEKO examples that show the application of a selection of features are discussed
in the ExamplesGuide.pdf document. More examples, models and specific information may be
obtained on the FEKO website.
Advanced examples based primarily on geometry creation and solution control using scripting in
the EDITFEKO interface may be found in the FEKO installation directory under:
examples/Miscellaneous/ScriptingExamples
The examples are explained and some results are presented in ScriptingExamples.pdf guide
that may be found in the same directory.

1.4 Changes in this release

Changes to the functionality of the code in this release with respect to the previous release of
September 2010 (Suite 6.0) are indicated by adding a column in the margin. The changes are
indicated in two ways:
Sections that have changed from those in the previous version of the manual.
Sections that were newly added to this version of the manual.
1
www.feko.info

July 2011 FEKO User’s Manual

INTRODUCTION TO THE FEKO SUITE 1-8

1.5 Contacting your distributor or EMSS

You can find the distributor for your region at

http://www.feko.info/contact.htm
Alternatively, for technical questions, please send an email to
[email protected] for North America
[email protected] for Europe
[email protected] for all other regions
or, for activation codes and licence queries, to
[email protected] for North America
[email protected] for Europe
[email protected] for all other regions

July 2011 FEKO User’s Manual

GENERAL MODELLING GUIDELINES 2-1

2 General modelling guidelines

2.1 Program flow

Models are generally constructed in CADFEKO. The model information is saved to the *.cfx file
and the workspace layout to the *.cfs file. Next the user runs PREFEKO — launched from the
CADFEKO Run menu — which processes the *.cfm and *.pre files (created automatically by
CADFEKO during the save operation) and generates the *.fek file. The *.fek file is the input
to the solution kernel, FEKO. (When running FEKO from the CADFEKO Run menu, PREFEKO is
executed automatically if the user has not yet done so.) The FEKO output is stored in the binary
*.bof file from which the results can be viewed in POSTFEKO. The results are also stored in the
*.out file.
Where an optimisation has been defined in CADFEKO, the relevant optimisation information is
written to the *.opt file and optimisation-specific visualisation quantities are saved to a spe-
cial POSTFEKO graph file (*.pfg). The user may launch the optimiser (OPTFEKO) from the
CADFEKO Run menu. The optimiser will automatically launch the other FEKO Suite components
as required during the optimisation process. During the optimisation process, general iteration-
specific results as well as optimisation process-specific information can be viewed in POSTFEKO.
Once the optimisation process has completed, the optimum model as well as a full set of results
are available for viewing in POSTFEKO.
For advanced models, the user may elect to edit the *.pre file. This allows using the scripting
commands and provides complete control of the solution process. (See the section working with
CADFEKO models in EDITFEKO (see section 6.18) and the advanced discussion on program flow
when using EDITFEKO (see section 12.1)).
If the *.pre file is manually edited, it is a good idea to validate the model in POSTFEKO before
executing FEKO or OPTFEKO.

2.2 Modelling and meshing guidelines

2.2.1 Definitions and terms

Below are a number of definitions that are used frequently in this manual:
Segment : A short section of a wire (short in comparison with the wavelength).
Cuboid : A volume element used to model dielectric and magnetic solids with the volume
equivalence method in the MoM. It has 90◦ corners similar to a cube, but does not need to
have equal side lengths. (Cuboid mesh elements can only be created in EDITFEKO.)
Tetrahedron : 3D tetrahedral shaped volume element used for discritisation of dielectric regions
to be solved with the FEM.
Polygon : A planar surface element with straight edge boundaries. This can be a primitive poly-
gon (which will be subdivided into triangles) or a polygonal plate (which is not discretised
and will be solved with the UTD).
Vertex : Any end point of a mesh segment or corner of a mesh element (triangle, tetrahedron,
etc.)

July 2011 FEKO User’s Manual

GENERAL MODELLING GUIDELINES 2-2

Node : The point where two segments are joined is called a node. One basis function is assigned
to each node.
Edge : In the geometry an edge is any free curve (these are also-called wires) or any boundary
curve of a surface. All free curves are discretised into wire-segments during meshing.
When used in connection with triangular mesh elements on a surface, the term edge refers
to the common line between two adjacent triangles. (If three triangles share two vertices,
there are two edges associated with these triangles.) If the surface is a metal, then one
basis function is assigned to each edge. If the surface is a dielectric to be solved with the
surface current method, then two basis functions are assigned to each edge, one for the
equivalent electric current density and one for the equivalent magnetic current density. A
free edge belongs to only one triangle. Unless this is in a ground plane or on a polygonal
plate, no current flows across this edge.
Connection point A connection point is where a segment is joined to a triangle. The end of
the segment is connected to the vertex of the triangle. A basis function is assigned to each
connection point.

Conducting surfaces are subdivided into triangles, and wires into segments. For dielectrics, there
are a number of possibilities (see section 2.3). Using the surface current method the surface of
the dielectric solid is subdivided into triangles, whereas with the volume current method, solid
dielectrics are subdivided into cuboids. For the FEM, the mesh is based on tetrahedral volume
elements. Thin dielectric sheets are meshed into triangles located along the middle of the sheet.
For structures that employ special solution methods (UTD, PO and GO - ray launching) special
meshing rules apply. Metal faces that employ the UTD approximation are not meshed. Metal
or dielectric surfaces that employ the PO approximation use a triangular meshing similar to the
standard meshing.
The meshing for surfaces to be considered using Geometrical optics (ray launching) is a triangular
mesh, but the required mesh size should in general be frequency-independent. For GO surfaces,
the largest possible mesh size should be used that will provide an acceptable representation of
the surface geometry.

2.2.2 Meshing guidelines regarding element sizes

During meshing the following general rules should be adhered to:

Segments

• The segment length l should be smaller than a tenth of the free space wavelength.

• Note also that the segment current flows only in the axial direction. Thus segments should
not be too short relative to the wire radius. Ideally the segment length should be at least
four times the radius.

• When modelling a conductive surface by means of a wire grid, the radius should be chosen
so that the wire area in one direction is approximately the same as the area of the original
surface. This leads to
l
r ≈ , (2-1)
2π

July 2011 FEKO User’s Manual

GENERAL MODELLING GUIDELINES 2-3

where r is the radius and l is the segment length which should be about a tenth of a
wavelength.

Surface mesh elements

• For MoM and PO mesh elements, the area A of each triangular element should be smaller
2
than λ70 . For triangles that are approximately equilateral this means the side length s should
λ
be shorter than approximately (5...6) . Depending on the geometry and the required accuracy,
more or less triangles may be needed. If the memory constraints allow it, an edge length
λ
of (8...10) is preferred.

• For large element PO mesh elements, the area A of each triangular element should be
smaller than 2λ2 when near field requests are present. The allowed edge length for when
near field requests are present should be 2λ. If only far field requests are present, the trian-
gles only need to represent the geometry accurately, regardless of frequency or wavelength.

• For dielectric surfaces to be solved using the GO (ray launching) method, the maximum
triangular mesh element size should be chosen such that the geometry of the surface is
well represented. For this method, the mesh size should be chosen independent of the
solution frequency, and should be purely a function of the accuracy of the geometrical
representation.

• The edge length of dielectric cuboids has to be small in comparison with the wavelength in
the dielectric as well as the skin depth
r
2
δ= . (2-2)
ωµσ

Due to the staircase approximation used in representing the model geometry, a mesh size
of at least the minimum between a tenth of the wavelength and a tenth of the skin depth
is recommended.

Volume mesh elements

When meshing the FEM region into tetrahedral volume elements, the element size (edge length
of the tetrahedra) should be about a fifth of the wavelength inside of the dielectric medium in
question. For the elements right on the FEM/MoM interface, a finer element size of about a tenth
of the medium wavelength is recommended. The reason that a coarser mesh can be employed
inside the medium is that higher order basis functions are employed for the FEM solution.
Sometimes the overall memory requirement for a solution may be reduced by adding a small air
region or buffer around the actual dielectric object. This region is also meshed into tetrahedral
elements (i.e. the buffer region is also solved with the FEM). Since the wavelength in air is larger
than in the dielectric, larger tetrahedral elements can then be used in the buffer regions and
at the FEM/MoM interface. This reduces the memory requirement for the FEM/MoM coupling
arrays. This memory reduction is typically much higher than the additional memory required to
solve the additional tetrahedral elements added in the air buffer zone.

July 2011 FEKO User’s Manual

GENERAL MODELLING GUIDELINES 2-4

In some cases accurate modelling of the geometry requires significantly finer mesh elements than
specified by the guidelines above. (For low frequencies in particular, the segmentation rule of
a tenth of a wavelength is often much too coarse to yield a reasonable representation of the
geometry.) One example of where finer discretisation may be required is where a wire runs
parallel to a conducting plate. If the wire is closer than a tenth of a wavelength to the plate,
the size of the triangles in the direction orthogonal to the wire should be similar to the distance
from the wire to the plate in order to give an accurate representation of the surface charge
distribution. Another case where finer discretisation may be required is on waveguide ports,
where the mesh size must be small enough to capture the field distribution of the highest mode
which is included in the modal expansion of the port. More information may be found in the
discussion of waveguide ports (see section 14.22).
If the segmentation rules are not adhered to, the errors and warnings listed in Table 2-1 will be
reported by the FEKO kernel.
Table 2-1: Segmentation warnings and errors.

Description Warning Error

Ratio of the segment length to the wavelength l > 0.3λ l > 0.5λ
Ratio of the segment radius to the segment length r > 0.3l r > 1.0l
1 2 1 2
Ratio of the triangle area to the wavelength squared A > 30 λ A > 10 λ
Ratio of the triangle area to the wavelength squared for large A > 2λ2 A > 6λ2
element PO (if near fields present)
Ratio of wire radius to the triangle edge length at a r ≥ 3.33l r ≥ 5l
connection point
Ratio of the cuboid edge length to the wavelength l > 14 λ l > 12 λ
Ratio of the cuboid edge length to the skin depth l > 15 δ l > 13 δ
Ratio of the tetrahedral face area to the wavelength squared A > 0.047λ2 A > 0.433λ2
(inner mesh elements)
Ratio of the tetrahedral face area to the wavelength squared A > 0.033λ2 A > 0.108λ2
(boundary surface mesh elements)
1 1
Ratio of the area of the triangle on a waveguide port to the A> 30
T2 A> 10
T2
smallest modal period squared

2.2.3 Meshing guidelines regarding connectivity

The MoM

FEKO approximates the current in terms of basis functions associated with edges, nodes and
connection points as defined above (see section 2.2.1). To ensure electrical connectivity, triangles
must therefore share an edge. Similarly, segments must connect to other segments at nodes or to
mesh triangles at vertices. Some examples are shown in Figure 2-1.

The UTD and GO

For the special case, where the UTD method is applied to unmeshed polygon plates, there is no
defined connectivity between plates that share an edge.

July 2011 FEKO User’s Manual

GENERAL MODELLING GUIDELINES 2-5

Figure 2-1: Example of mesh connectivity: unconnected at the top, connected at the bottom

For GO, the connectivity should generally be maintained, though this is not required. When
meshing the faces of a GO region, it is good practice to ensure that the meshes on faces that
touch do align to achieve a good geometric representation.

The FEM

When meshing dielectric solids into tetrahedral elements for the FEM, the faces of adjacent tetra-
hedra must match. In addition, when modelling conducting surfaces in or on the FEM region,
the metallic triangles must match the faces of the tetrahedral volume elements. See Figure 2-2.

Figure 2-2: Example of FEM element connectivity: invalid left, correct right

General

CADFEKO generally enforces meshing rules regarding connectivity for each part. Therefore,
connected items should be unioned (see section 4.4.1) together before meshing them. It is,
however, still important that the model is of decent quality. If, for example, a wire is attached to
a surface, but due to numerical error it is more than the model tolerance (see section 4.1.2) away
from the actual surface CADFEKO will not create a vertex in the surface mesh at the attachment
point. (It is possible to union two objects that do not touch each other at all.) The wire will then
not be considered electrically connected to the surface during the solution phase.
The FEKO kernel has several checks built in and will give errors if, for example, the mesh contains
overlapping triangles. CADFEKO also has several checks which the user can perform before trying
to solve the model (see section 4.7).

July 2011 FEKO User’s Manual

GENERAL MODELLING GUIDELINES 2-6

2.3 Dielectric solids

There are numerous ways to model dielectric objects in FEKO, three of which apply to arbitrarily
shaped bodies:
• Surface equivalence principle:
Method of Moments (MoM) generally uses the surface equivalence principle for modelling
of dielectric bodies. In this method, interfaces between different homogeneous regions
are subdivided into a surface mesh using triangular elements. Basis functions are applied
to these elements for the equivalent electric and equivalent magnetic surface currents.
Boundary conditions result through the use of equivalent sources.
For models constructed in CADFEKO this method is used by default on all dielectric regions
(see section 6.1.2) that are not explicitly meshed into tetrahedral volume elements (see
section 4.8).
When working with models in EDITFEKO, medium regions are defined with the ME card
(see section 13.31). This uses the normal vector of the triangles to distinguish the respective
dielectric media on either side of the mesh elements. It is generally advisable to check the
respective media regions in POSTFEKO before running the FEKO solver.

• Volume equivalence principle:

MoM can also be applied with the volume equivalence principle. Here the volume is subdi-
vided into cuboidal elements in EDITFEKO and tetrahedral elements in CADFEKO. In prin-
ciple, each element can be assigned a different material property. Inside the element the
polarisation current is unknown. Normally a volume would have many more unknowns
than a surface mesh, such that this method would require more memory. However, this
technique is very suitable for thin sheets and is also very stable for low frequencies. The for-
mulation is very stable, and thus when using MLFMM the number of iterations is typically
small. In EDITFEKO is accessed using the DK (see section 13.10), DZ (see section 13.12)
or QU (see section 13.40).

• Finite element method:

As an alternative to the MoM, the Finite Element Method (FEM) is also available in FEKO.
It requires that 3D volumes are discretised into tetrahedral elements. FEM matrices are
sparse, as opposed to the MoM, and the memory requirement for a FEM volume mesh is
much less than a MoM volume mesh of the same model. This method is automatically used
for regions in the model that contain tetrahedral elements.

In CADFEKO models, the medium properties are specified when defining the dielectric medium
(see section 6.1.1). In EDITFEKO, medium properties are specified with the DI card (see sec-
tion 14.32).
For the surface equivalence principle, it is possible to define metallic triangles on the surfaces
and triangles and segments within the dielectric regions. With the volume equivalence principle,
there must be a small space between the cuboids and any conducting triangles on the surface of
the dielectric. Metallic triangles can also be located inside FEM regions, but they have to coincide
with tetrahedral surfaces.
Two additional methods that are available for dielectric bodies are Physical optics and Geo-
metrical optics (ray launching). These are approximate methods with special limitations and
application methods.

July 2011 FEKO User’s Manual

GENERAL MODELLING GUIDELINES 2-7

The model geometry (e.g. metallic wires and surfaces) does not necessarily have to be embedded
in free space. In CADFEKO the properties of the free space (see section 6.1) medium may be
changed. In EDITFEKO the EG, DI and GF cards can be used to specify the material parameters
of the surrounding medium.

Special dielectrics and infinite planes

Apart from the general formulations applicable to dielectrics, there are a number of special meth-
ods to account for dielectric sheets and coatings as well as special dielectric bodies in FEKO:

• Thin dielectric sheets:

The volume equivalence principle is applied and the resulting equivalent currents approxi-
mated by a surface current.

• Dielectric coatings:
Metallic wires or triangular surface patches can have a thin dielectric coating.

• Dielectric half-space e.g. ground surface:

In this case the reflection coefficient method is used. This is activated by using infinite
planes (see section 6.7) under Solution in CADFEKO or the BO card (see section 14.23) in
EDITFEKO.

• Spheres consisting of one or more dielectric layers:

A special Green’s function can be activated by using the GF card (see section 14.39) in
EDITFEKO.

• Planar multilayer substrate:

A multilayer planar substrate (with or without a perfectly conducting ground planes at
the top and bottom) is added to the model using infinite planes (see section 6.7) under
Solution in CADFEKO or the GF card (see section 14.39) in EDITFEKO.

• Windscreen:
The active windcsreen antenna elements can be activated by using the WA card (see sec-
tion 13.50) in EDITFEKO. The dielectric windscreen reference plane is defined by the WR
card (see section 13.52) in EDITFEKO. Dielectric properties of the glass layers can be de-
fined by the WD card (see section 14.63).

In CADFEKO, thin dielectric sheets and coatings are applied to faces (see section 6.1.2). In
EDITFEKO, thin dielectric sheets are defined by applying the SK card (see section 14.59) and
coatings by applying the CO card (see section 14.29) to triangle labels.

2.4 Checking the validity of the results

Once a calculation has been completed with FEKO, the results have to be checked or confirmed.
There are a number of ways of doing this:

• Comparison with exact results, if these are available.

• Comparison with results that have been published in the literature.

July 2011 FEKO User’s Manual

GENERAL MODELLING GUIDELINES 2-8

• Comparison with results generated using another computational method.

• Comparison with measured results.
• Plausibility, e.g. negative real input impedances do not exist.

If these possibilities are not available, then the following process may be tried:

• After a calculation with FEKO, repeat it with a finer mesh. The number of elements should
be at least 1.5 times greater than with the initial calculation. If there is a large difference in
the results, then the results cannot be considered correct. In this case the model should be
refined, either by improving the meshing, or by consideration of other factors that may in-
fluence the results, for example the validity of the techniques used. Any warnings provided
by the FEKO kernel during the solution phase should be carefully considered as these are
often an indication of the source of inaccuracies in results.
• Perform a power balance check. The power fed into an antenna through the source must
be equal to the sum of the radiated power and any losses in the antenna material. The
radiated power is automatically calculated for the specified sector if the required far field
calculation has two or more angles in each angular direction, while the losses in materials
are always calculated. These values can be extracted from the *.out file and used to
confirm the power balance.

July 2011 FEKO User’s Manual

Part II

Working with CADFEKO

INTRODUCTION TO CADFEKO 3-1

3 Introduction to CADFEKO

CADFEKO is the component of the FEKO Suite that facilitates the creation and set up of FEKO
models in a graphical or CAD environment. This involves defining and meshing of geometry
as well as specifying the electromagnetic parameters and solution configuration. CADFEKO also
makes provision for the definition and launching of an optimisation process on the defined model.
CADFEKO supports parametric model construction. If the model is constructed using variables,
the entire model can be modified by changing the values of these variables. This is used, for
example, to adapt the size of an antenna to a required frequency or to make provision for model
changes during optimisation. CADFEKO maintains the construction history so that, for example,
a union operation is automatically updated if any of the individual objects involved in the union
are modified afterwards.
Similarly, the user defines media properties and applies them to the relevant parts of the model.
All of these properties can then be modified by changing the applicable medium parameters.

3.1 CADFEKO overview

3.1.1 CADFEKO Getting started page

When starting CADFEKO, the CADFEKO start page will be displayed, giving quick access to Create
a new model, Open an existing model and a list of recently opened models. Links to the PDF’s for
the FEKO suite are also available here along with FEKO introduction videos. It is recommended
that these videos are watched by first time users before attempting this example.

3.1.2 CADFEKO window

The various main elements and terminology of the CADFEKO window will be briefly described.
These terminology will be used in the examples to follow.

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-2

1. Quick access toolbar These items give the user quick access to controls such as New model,
Open model, Save model, undo and redo actions (grouped at the left side of the toolbar)
as well as launching the FEKO solver, POSTFEKO (for the display of the results obtained by
the FEKO solver), EDITFEKO and PREFEKO (grouped at the right side of the toolbar, next
to the help button [7] and called Application Launcher).

2. Ribbon The ribbon contains the application menu, default tabs, contextual tabs and contex-
tual commands.

3. Model tree The model tree contains context menus to manage parametric variables, cre-
ation of named points, workplanes, defining media, defining cables, mesh refinement,
non-radiating networks, adding of ports and excitations, setting the frequency, requesting
calculations and setting up optimisation runs.

4. Details tree The details tree contains the geometry object details (edges, faces and regions).
Custom solution and mesh settings may be set here.

5. Active statusbar The active statusbar gives the user quick access to general display settings,
tools and selection method and type.

6. 3D view The 3D view enables the user to visualise the geometry and solution settings (such
as far field requests, etc.). Additional visualisations such as cutplanes and symmetry can
also be displayed.

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-3

7. Help The Help button gives the user quick access to the FEKO manuals. Context sensitive
help is available in all FEKO Suite GUI components by pressing <F1> at any time.

8. Notes view The notes view can be used to document a model. Additional comments, expla-
nations or descriptions can be added.

3.1.3 The CADFEKO ribbon

The CADFEKO ribbon consists of several elements. Please take note of the terminology as it will
be used extensively in the examples to follow.

1. Application menu The application menu contains the following commands: New model,
Open model, Save as..., Archive, Import, Export, Check for updates, Rendering options
and Preferences.

2. Default tabs The default tabs are always visible and contain general commands.

3. Contextual tabs The contextual tabs display context sensitive tabs with commands relevant
to the selected view (3D view or schematic view). A coloured tab marker bar above the
tabs indicates the current context.

4. Group of commands Similar actions or commands are contained in a group.

5. Dialog launcher Clicking on the dialog launcher will launch a dialog with additional settings
that relate to that group.

A message window at the bottom displays messages about user interaction such as geometry
creation, meshing, source configuration, etc. It also provides details regarding warnings and error
messages. Errors and warnings in the message window will provide links to the corresponding
geometry objects in the details tree which resulted in the error or warning. When clicking on the
given link in the message window, the respective geometry object will be selected in the model
tree, the 3D view and in the details tree.

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-4

3.1.4 CADFEKO files

Only the *.cfx file is required to open the model, but when saving the model, CADFEKO au-
tomatically saves the *.cfs file — containing the workspace (views, cut planes, etc.) — and
the *.cfm and *.pre files — used when solving models (see section 6.16). When optimisation
settings have been defined in the CADFEKO model, CADFEKO will also automatically create/save
the *.opt and the *.pfg files that contain the relevant information used (in conjunction with
the *.pre and *.cfx files) during optimisation.

3.1.5 Preferences

The settings anchor on the application menu provides a number dialogs that allows the user to
customise CADFEKO by setting default preferences.
Figure 3-1 shows the Default settings dialog. A variety of options can be set, from default model
unit settings to display settings.

Figure 3-1: The Default settings dialog.

3D View rendering and colours

The Rendering mode group shows the algorithm used to remove hidden lines. These settings can
be configured manually and this may improve the display accuracy, but it can have a significant
impact on the rendering performance and memory usage. Hardware rendering (Hardware z-
buffering) is only enabled when available on the user’s system. The sliders change the face
displacement for geometry and meshes. The face displacement provides a trade-off between
edges appearing broken and supposedly hidden lines being visible. The relative positions of the
sliders also determine what is visible if both the geometry and mesh are displayed. Different
settings may be required for different view directions. These options are stored in the CADFEKO
configuration file.

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-5

Figure 3-2: Rendering options dialog

The Colours dialog allows the user to change the default colours that CADFEKO uses in the 3D
view. The Reset to defaults button restores the default colours. Figure 3-3 displays the currently
available options with the default colours used by CADFEKO.

Figure 3-3: The Colours dialog

Messages and log files

In addition to the messages and errors in the message window, CADFEKO creates a text log
file for each session in the logs subdirectory of the FEKO_USER_HOME. If CADFEKO encounters
an internal error, the log file for the session is copied to CADFEKO.ERROR.LOG and the current
model to CADFEKO.ERROR.CFX. If these files exist, they will be overwritten. If possible, these
error files should be included when reporting any problem to FEKO support.

3.1.6 Checking for updates to the FEKO Suite

Select Check for updates on the application menu and click Check for updates on the FEKO
update dialog to poll the FEKO website for update information. This operation checks if any
updates are available compared to the installed components. It does not send any information to
the web-site.
If automatic updates have been enabled on the Settings tab, CADFEKO polls the web-site each
time a GUI component is launched if the specified interval between updates have elapsed. Up-
dates can also be downloaded from a local repository for cluster machines or where internet
access is not possible. For more information see section 23.

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-6

3.1.7 Batch Meshing - changing the model variable from a command line

A stand-alone batch meshing tool can be called from a command line. This tool makes provision
for the re-evaluation and meshing of an existing CADFEKO model (after reassigning specific
variable values) without launching the graphical user interface.
This functionality is used by OPTFEKO during the optimisation process, and also provides for the
manipulation of CADFEKO models and the creation of meshes from CADFEKO models directly
from a third-party tool, or a command line script. The required variables and values are included
directly in the command line string.
The CADFEKO batch-mesher is launched using the command:
cadfeko_batch <filename> [options]

Where <filename> is an existing CADFEKO model (*.cfx) (with or without the extension).
The path may be included in the <filename> string. After re-evaluation and meshing, the
modified CADFEKO model will be saved, replacing the existing *.cfx file as well as the *.cfm,
*.pre , *.opt and *.pfg files (if the solution configuration is deactivated in the CADFEKO
model, or if the *.pre file has been edited outside of CADFEKO, then the *.pre file will not be
overwritten).
If the batch-mesher is called with no variable assignments in the options, then the indicated file
will simply be re-meshed and saved without any changes to the variable values in the model. This
will create (or overwrite existing) *.pre, *.opt, *.cfm and *.pfg files (where applicable).
If the new variable values indicated in the command line cause an error during re-evaluation or
meshing, the changes will be aborted (no changes are made to the model) and an error reported.
If suspect entities are found in the CADFEKO model after re-evaluation, the re-evaluation and
meshing will be completed, and the new model and mesh created, but an error will also be
reported (this error will be reported for all suspect items, even if they were not introduced due
to changes made by the batch-mesher).
If multiple assignments are included in the command line options for the same variable, then
the last value assigned to that variable will be used for that variable before model re-evaluation,
meshing and saving.
The command line options are:

−−version Output only the version information to the command line and then terminate. No file
name is required to use this option.

-#var=value Allows variables to be assigned new values before re-evaluation and meshing. Multiple
variables may be included, but each variable must be indicated in the same way (i.e. to
set variables ‘a’ and ‘b’ to 1, the options should contain ...-#a=1 -#b=1...).

−−run-from-gui This uses a special execution mode for the graphical user interface. In this mode,
additional information regarding the progress of each phase of the model re-evaluation
and meshing is included in the screen output.

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-7

3.2 The 3D view

3.2.1 3D view interaction

The 3D views are used to display and interact with the model. CADFEKO distinguishes between
a click operation (mouse down and up at approximately the same location) and click and drag
(the mouse is moved a minimum distance while the button remains down). Left clicking is
used for selection (see section 3.2.2) and point entry (see section 3.2.2). Right clicking opens a
context-sensitive menu with operations for the current selection.

Rotation

The model is rotated by clicking and dragging the mouse.

Zooming

It can be zoomed by pressing <Shift> while clicking and dragging the mouse (upward movement
zooms in, while downward movement zooms out; horizontal movement is ignored). Rolling
the mouse wheel also zooms. (Pressing <Shift> while rolling the mouse wheel slows down
zooming.)
CADFEKO tries to zoom in / out on the object below the mouse pointer, hence rolling the mouse
wheel while the cursor is not in the centre of the screen will also have a panning effect. After
a zoom operation the model will rotate around a point that is more or less in the middle of
the screen and at a distance approximately equal to that between the current viewpoint and the
model geometry.

Panning

The display can be panned (moved inside the display window) by clicking and dragging while
pressing the <Ctrl> key or clicking and dragging with the middle mouse button.

Zoom to selection

When searching for specific parts or elements in the geometry or mesh, the Zoom to selection
tool may be used. This tool is available in the context menu of any mesh or geometry element in
the model or details tree. The Zoom to selection tool will modify the panning and zoom factor of
the current 3D window such that the selected elements are centred in the view. The rotation of
the view is not adjusted - elements in the selection may therefore be hidden behind other mesh
or geometry. These hidden elements can easily be identified by manually rotating the model
without changing the selection, or by setting the model opacity.

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-8

Snapping to special points

CADFEKO uses a powerful auto-snapping algorithm. Snapping enables the selection of special
points from the 3D view. When pressing <Ctrl><Shift> and hovering over the model with the
mouse cursor, special snapping points near the mouse cursor will be indicated by blue dots. The
active snapping point is indicated by a black dot and can be selected by pressing <Ctrl><Shift>
and clicking on the current point. The special snapping points include: named points, geometry
points, geometry face centre, geometry edge centre, mesh vertices and the grid.
When snapping to the workplane, a small grid will be displayed at the mouse cursor. When two
objects overlap, pressing <Ctrl><Shift> and repeatedly pressing <Tab> may be used to toggle
between their respective snapping points.
When snapping is used to align a new workplane, it should be noted that the history of the
starting point and the route followed to the destination point (for example the orientation of an
edge), affects the orientation of the workplane.
To specify how points will be selected from the 3D view when using mouse-click based point
entry (see section 3.2.2), right click on the 3D view and select Snapping settings. The Snapping
settings dialog is shown in Figure 3-4.

Figure 3-4: The snapping settings dialog

3.2.2 Manipulating items

Selecting items

Items are selected by clicking either in the tree or on the item representation in any 3D view.
Selected items are highlighted in the tree and in the 3D view. When parent objects are selected
in the tree, CADFEKO displays wire-frame outlines of these objects in the 3D view. These items
are not themselves part of the model, but it is useful to be able to distinguish the different parents
in a composite part.
Three selection methods are supported.

Single selection: Select the single item under the mouse cursor.

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-9

Rectangle selection: Select all items in a rectangle. The user has to click once to place the first corner of
the rectangle, then move the mouse cursor and select a second time to indicate the second
point in the rectangle.

Polygon selection: A polygon of selection area can be defined using successive clicks. All elements inside
the polygonal area will be selected.

The selection type is set to Auto selection by default, but the user can set the selection type
manually on the Tool tab or the status bar. Auto selection cycles through the available selection
types.
The Undo selection and Redo selection buttons on the Tools tab allow reverting back / forward
to the previous / next selection states, respectively. These states are stored in a seperate undo /
redo list. This list is cleared each time the model is changed (for example, when adding / deleting
/ modifying geometry or variables). Note that just like the model manipulations, selections are
independent of 3D views (i.e. they are shared between them).
Pressing <Ctrl> while selecting items, in the tree or a 3D view, will add them to the current
selection (or remove them if they are selected). Pressing <Shift> while clicking on items in
the tree will select everything of the same type in the range from the first item of the current
selection. Since the order of items is not as apparent in the 3D view, pressing <Shift> while
selecting will merely add / remove the new item.
All selected items must be of the same type, i.e. it is impossible to select a part and some of its
faces at the same time. However, simultaneous selection of faces, regions and edges belonging
to different parts is allowed.

Deleting items

CADFEKO does not allow deletion of objects that are being used by other items - this is to main-
tain model consistency. For example, it is not possible to delete a variable that is used in the
definition of a medium or to delete a medium that is applied to some part of the geometry. When
a delete operation fails, CADFEKO will provide an indication of where the item is used. That item
must then first be removed or edited so that it no longer depends on the original item that is to
be deleted.
Note that sometimes the dependency may be indirect. For example, consider a variable used
to set a local mesh size on Face1 of Cuboid1. The cuboid is then unioned with another and
subtracted from some other object. If the user now tries to delete the variable, the message
indicates the full tree path to the face (Subtract1.Union1.Cuboid1.Face1), but this entity is no
longer accessible. At this point it is not possible to remove the local mesh size setting except if
the original objects are copied (see section 4.4.4) out of the tree and the last few creation steps
are redone. Thus, all geometry dependent settings which are likely to change should be applied
as late as possible in the model definition process.
A similar situation can occur when a solution configuration is defined in CADFEKO and then
disabled (see section 6.18) after the user has edited the *.pre file. The user can no longer edit
or remove any of these settings, but CADFEKO must keep the model consistent since the user can
at any time enable the solution again. This can be a problem if the user wants to, for example,
modify a variable that is also used in the rest of the model, but this change cannot be made since

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-10

it would result in an invalid state in the solution configuration. The only way around this is to
re-enable the solution configuration (CADFEKO automatically renames the edited *.pre file),
make the necessary change and then manually revert to the renamed *.pre file.

Point entry

The input dialogs of some geometry related and result request related fields allow the entry of
field values based on <Ctrl><Shift> + mouse click in the 3D view or tree (for variables and
named points). If such point entry fields have focus, their background colour will be yellow.
By clicking on the 3D view, the coordinates of the selected point are entered into the field and
the focus is shifted to the next field. This allows the spatial definition or editing of geometry or
solution requests based on a series of clicks on the 3D view (or tree) and one click on Create in
the dialog. For some one-dimensional input fields (such as the radius of the sphere primitive),
a value is calculated based on the distance between the specified point and the coordinates or
values already entered into other fields. Where an input field specifies a direction vector, the
vector is considered to extend from the relevant origin to the selected point. If <Ctrl><Shift> is
pressed while moving the mouse over the screen (without clicking), the fields values are updated
continuously. This is called Preview mode. The active fields show the values that would be
entered if the mouse was clicked at that position. The values in the active fields are displayed
using an italic font to indicate that preview mode is active.

Locking and preview of point entry fields

Fields which accept multiple values from point entry have Lock buttons next to them. If such a
button is toggled on (down), that field will maintain its current value, and will not be updated
based on point selection. The lock mechanism is used when there is a need, for example, to
click on points in the workplane without modifying the value in the n-direction. Different named
points may be individually referenced in the three component fields of a point. Each component
is then determined by projecting the specified point along the required dimension. The locking
mechanism provides a useful method to realise such multi-point references.

3.2.3 Display options

View modes

Two view modes and a overlay option are supported in CADFEKO that allows the user to inspect
the model.
Model view: The model (geometry and mesh) is displayed. This is the default display mode and should
be used for model creation and manipulation.

Simulation mesh view: The simulation mesh, the mesh that the kernel will use, is displayed. This option
is useful when the mesh needs to be inspected.

Overlay: The overlay options shows a bit of the view mode that is not selected. In model view mode,
the simulation mesh edges will be displayed, but not the faces. In simulation mesh mode,
the model will be displayed semi-transparent so that the model and simulation mesh can be
compared.

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-11

Cutplanes

The cutplane dialog for a specific view can be opened by clicking on the Specify cutplanes toolbar
button or by right-clicking in the 3D view and selecting Cutplanes. The first cutplane is created
automatically when the dialog is opened (shown in Figure 3-5). Additional cutplanes can be
added or deleted using the Add and Remove buttons. Unchecking the Active item on any of
the cutplanes deactivates that cutplane without losing its settings. The Flip button reverses the
operation of the cutplane, hiding the visible region, and showing the invisible region.

Figure 3-5: The Cutplane dialog

The cutplane is specified in a similar manner to the workplane (see section 4.3.2), using an
origin and two vectors. These fields all use the standard point entry (see section 3.2.2). Thus
the definition of the cutplane may be interactively modified. Before applying any changes to a
cutplane, a plane preview is shown in all 3D windows (and will accept point entry from any 3D
view), but when applied, the cutplane applies only to the current window.
The cutplane is updated with each change in the dialog. Closing the dialog with Cancel reverts
to the previous state (i.e. when the Apply or OK button was last clicked).
If a solid is cut, while the current mode selection (see section 3.2.2) is Select geometry parts, a
surface (in the colour of the internal region of the solid) is displayed in the plane of the cut where
the solid is intersected, even when the geometry is not coloured by medium (see section 3.2.3).
Such surfaces cannot be selected, and clicking on them will select geometry that lies behind
the cutplane. When the selection mode is anything other than Select geometry parts, CADFEKO
displays cut solids in a similar fashion to shells (showing only the actual faces) allowing visible
selection-access to the internal faces and edges.
The cutplane icon (on the right hand side of the status bar) is activated if the view under the
mouse cursor has any active cutplanes. Like the snap options, this icon applies to the view under
the mouse cursor, and not the active view.

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-12

Selective viewing

Where items are obscured by other items in the 3D view, it can be difficult to set up cutplanes to
view these items. In addition, 3D view representations of items such as field calculation requests
can cause a general clutter of the view window. In order to tailor views, specific items (geometry,
mesh or solution entity items) can be hidden selectively. Hidden items are removed from all 3D
views, but are still part of the model and will be exported to the *.cfm and *.pre files and
will be considered in any computation process. (Meshing a hidden object will result in a visible
mesh part — even if a hidden mesh part with that name existed beforehand.) Hidden items are
indicated with grey icons in the model tree.
Items can be shown or hidden by selecting them in the 3D view or tree and choosing Show/Hide
from the pop-up menu, or by choosing View → Show/Hide selected items from the main menu.
If the selection contains visible and hidden items, selecting Show / Hide toggles the hidden state
of each of the selected item. Selecting View → Show all in the main menu shows all hidden
items (geometry, meshes and solution entities) in the model, independent of their current state.
The Show all geometry and Show all meshes options show all geometry and mesh parts respec-
tively. Ports can be shown/hidden at the port level or the level of the individual mesh/geometry
instances.
Note that the visibility buttons on the 3D view toolbar (see section 3.2.4) apply to all items of
that type in the selected view only and take priority over the shown/hidden status of all entities
of that type, i.e. if the visibility is off in the 3D view, selecting Show all will not influence the
display.
Item viewing properties are not saved as part of the model or session and if the model is saved
and reloaded, all items will become shown (visible) again.

Colouring options

The available colouring options are described below.

Colour by element normal: All parts are drawn with the same colour. The two sides of faces
are coloured differently to indicate the normal direction of the faces. On the geometry, the
normal side is coloured green, while the reverse side is coloured red. On the mesh, the
normal side of each element is coloured blue, while the reverse side is coloured brown.

Colour by region medium: Each region medium is indicated according to the media list under
Media in the model tree (see section 6.1 and section 6.1.2). In the mesh view, surface
mesh elements are coloured on each side according to the medium on that side of the face.
For example, when viewing the mesh of a dielectric/metallic object, the entire object will
have the free space colour when viewed from outside, but the view from inside (utilising
a cutplane or after some elements of the region boundary have been deleted) will be the
colour of the dielectric/metallic medium of the inside region. In the geometry view, regions
are, however, displayed using the colour of the internal medium (whether viewed from
outside or inside the region). If the display of the segment radii and coatings are activated
on wire mesh elements, these are coloured according to the core medium or the layered
medium defined as coating for that wire respectively.

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-13

Colour by face medium: The faces are displayed according to the material type of each face.
There is no differentiation of the colouring according to the direction from which faces are
viewed on the mesh. In the mesh view, the display of segment radii is automatically acti-
vated for wire elements in the mesh and these are coloured according to the core medium.
(The segment radii display may be manually deactivated if required, in which case no spe-
cific colouring will be shown for wire elements in the mesh.)

Face normal medium: The faces are displayed according to the material colour on the two sides
of the face. As an example, an object in free space will have the colour of free space (red
by default) on the outside of the object.

Faces and wires can also have coatings applied to them. The colour of the coating is used by
default, but can be disabled by toggling the Coatings button in the Style group.

Element visibility controls

The model and simulation mesh visibility controls are controlled using the controls in the Model
visibility and Simulation mesh groups. These controls allow the user to specify how the elements
should be rendered (edges, faces, etc.).

Named points, annotations, solution entities and settings

The visibility of named points (see section 4.2.2), requested results (see section 6.12) that have
a representation in the 3D view, port annotations as well as symmetry settings (see section 6.8)
and periodic boundary conditions (see section 6.9) can be controlled.
When Show/hide port annotations is on, a tag will be displayed at each active port, indicating the
port name, and the names of the excitations and loads that are applied at that point. Annotations
are shown for both the geometry and mesh manifestations of each port, and are only displayed
if the ports are shown.
The visibility of Symmetry and Periodic boundary conditions can be set using the toggle buttons,
but these settings are overridden when the Symmetry or Periodic boundary condition settings
dialogs are open, in which case the relevant preview will always be shown.

3.2.4 View settings (View tab)

Viewing angle and distance

The Transform view dialog (shown in Figure 3-6) opens with the current view settings repre-
sented by the position of the Origin (defined as the centre of the view), the spherical angular
coordinates θ and ϕ that specify the View direction and the current radial Zoom distance. See
section 3.2.4 for an indication of the icon used for the View transformation dialog. These values
are updated each time the 3D view is changed. This means that rotating, moving or zooming
while the transform view dialog is open causes the values in the relevant fields to be updated.

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-14

Figure 3-6: Transform view dialog

Vertical Z-lock

The Vertical Z axis button controls the rotation of the model in such a way that the z-axis remains
vertical on the screen (i.e. locking it in place). The next two buttons zoom to the full model extent
or to a selected window, respectively. (To zoom to a window after clicking this button, click and
drag the mouse, without releasing the mouse button, to define the window.) The Restore default
view button resets the view angles to the standard position.

View undo and redo

Undo view action and Redo view action buttons (these can also accessed with the short-cut
keys <Alt><←> and <Alt><→>) are also provided. These apply to all view manipulations in
the current view. View manipulations history is stored in a separate list from the one used to
store model manipulations. View operations can therefore be undone independent of geometry
modifications. The Undo / Redo buttons are disabled if there are no additional undo / redo
operations available in the view transformations history.

3.3 The model or contents tree

The model tree provides a complete representation of the current model.

The top (or contents) tree (shown in Figure 3-7), contains the lists of variables, named points,
workplanes and media, as well as an overview of the geometry, meshes, non-radiating general
networks, transmission lines, ports, solution requests and the optimisation configuration. Note
that the model tree is sorted alpha-numerically within each branch.
The entire hierarchy of geometry creation can be seen in the Geometry branch of the contents
tree. Where objects are derived from existing ones (for example, the individual objects used in
a Boolean operation or the original object before a split operation), the original (parent) objects
are removed from the top level of the model and listed as sub-branches under the new object in

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-15

Figure 3-7: The model tree or contents tree in CADFEKO

the model tree. The term part is used for highest-level items. These can be at the root level under
Geometry or in the top level of an assembly and are the objects that are visible in the 3D view.
Right clicking on any entry in the model tree will open an appropriate pop-up menu. Double
clicking on an item in the model tree will display the Properties for that item (except for some
headings, in which case a new item is added).
All branches in the geometry branch of the model tree below the selected item can be expanded /
collapsed by pressing <*> or selecting Collapse / Expand all from the pop-up menu. Collapsing
an entry will show all the direct parents only. To hide the direct parents (but keep the rest of the
state) use the <-> key. Likewise the <+> key expands only the current level. The model tree
may also be navigated using the arrows on the keyboard.
When specific items are hidden (see section 3.2.3) they are displayed with greyed icons in the
model tree.
The bottom (or details) tree shows detailed information (faces, edges, regions, transforms, etc.)
of a single part, when selected in the contents tree (the selected part is highlighted, labelled with
a bold italic font). The lists of faces, edges and regions only apply to top-level parts — these
items are not listed for parent objects in sub-branches of the geometry tree.
The tree area can be hidden by dragging the slider all the way to the left. When the tree area is
hidden, it can be shown by dragging the slider to the right or by selecting View → Tree from the
CADFEKO menu.

3.3.1 Icons displayed in the model tree and details tree

Icons next to each item in the trees indicate their type, media, the application of transforms,
mesh properties, etc. (The default names also indicate the type, but after renaming, for example,
Assembly1 to Antenna, the assembly indicator icon become very useful.) These icons are listed.
Primitive icons:

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-16

– Cuboid

– Flare (cut-off pyramid)

– Sphere

– Cylinder

– Cone

– Polygon

– Rectangle

– Ellipse

– Parabolic surface

– Line

– Polyline curve

– Analytical curve

– Elliptical arc

– Parabolic arc

– Hyperbolic arc

– Bézier curve

– Spline curve through points

– Helix or spiral

– Imported (Parasolid) CAD body

– Surface body (e.g. created with face copy or explode)

– Curve (edge / wire) body (e.g. created with edge copy or explode)

– Part that has been converted to a primitive

– Mesh part

Medium icons:

– Create dielectric medium

– Create metallic medium

– Create layered dielectric

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-17

– Create impedance sheet

– Create windscreen medium

Geometry operator icons:

– Boolean union

– Boolean intersection

– Boolean subtraction

– Split

– Object created by spinning / rotating parent

– Object created by sweeping / extruding parent

– Object created by sweeping along a path

– Loft (ruled surface) created between two edges

– This geometry contains edges created by projecting other items onto it

– This geometry contains imprinted points

– Object created by stitching sheet parts

– Simplified geometry (redundant items removed)

– This is an assembly

Indicator icons:

– This item has been transformed

– This part contains dielectric regions

– This face lies on a dielectric region (shown in details tree)

– This item is suspect — it could not be mapped

– The indicated part/region/face/edge contains faults.

– Local mesh properties set on regions, faces or edges (shown in details tree)

– Local wire radius (shown in details tree)

– This icon indicates the target from which an object was subtracted from.

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-18

Geometry transforms:

– Translation (shown in details tree)

– Rotation (shown in details tree)

– Align (shown in details tree)

– Scale (shown in details tree)

– Mirror (shown in details tree)

Network icons:

– Add general network

– Add transmission line

Ports, loads and excitations:

– Wire port

– Edge port

– Microstrip port

– Waveguide port

– FEM modal port

– Line port

– Complex impedance load

– Parallel point circuit load

– Series circuit load

– Voltage source

– Current source (can be applied to a line port in a FEM region)

– Waveguide mode excitation

– FEM modal excitation

– Electric point source

– Magnetic point source

– Point source with a specified pattern

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-19

– Plane wave excitation

– Aperture field excitation

– Impressed current excitation

– Spherical mode excitation

– Frequency specification

– Power setting

Calculation requests:

– Current request

– Far field request

– Near field request

– SAR request

– S-parameter request

– Error estimation request

– Request to perform cable analysis

– Request to calculate the power received by an ideal antenna

Optimisation icons:

– Add an impedance optimisation goal

– Add a near field optimisation goal

– Add a far field optimisation goal

– Add a S-parameter optimisation goal

– Add a SAR optimisation goal

Icons used in the details tree heading:

– For wires and surfaces, the core medium; for tetrahedra, the medium

– The layered medium applied as coating

– For surfaces, the medium on the normal side; for wires, the surrounding medium

– Only used for surfaces; the medium on the back side

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-20

3.3.2 Medium icons in the details tree

In addition to the indicator icons, showing special properties of faces, regions, edges and parts,
medium information is shown using specifically coloured blocks in additional columns next to
the entries in the details tree (shown in Figure 3-8). The colour used for each medium can be
set by right clicking on the medium in the Media list and choosing Change display colour. The
colours indicated in the details tree coincide with the colourings that will be applied in the 3D
view when choosing one of the colour-by-media options (see section 3.2.3).

Figure 3-8: The Details tree in CADFEKO

For geometry regions there is a single column indicating the region medium (PEC, free space or
one of the defined dielectrics).
For geometry faces there are two columns — the first indicates the face medium (PEC, one of the
metallic media, a Layered dielectric, if it is a thin dielectric sheet or an Impedance sheet). The
second indicates the Layered dielectric if a coating is applied to the face.
For wires there are also two columns. The first indicates the wire medium (PEC or one of the
metallic media) and the second the coating.
For tetrahedral meshes, there is only one column which indicates the medium (one of the defined
dielectrics).
For surface meshes there are the two columns (of the same type as for geometry faces) and two
additional columns which indicate the medium (PEC, free space or one of the defined dielectrics)
on the front and back respectively.
For segment meshes there is, in addition to the two columns displayed for wires, a column
indicating the surrounding medium.

3.4 Project tools

3.4.1 Distance between points

The Measure distance tool can be opened by clicking the Measure distance icon on the Tools tab.
The absolute distance between two points together with the relative x, y and z distance can be
determined by pressing <Ctrl><Shift> and clicking with the mouse cursor on any two respective
points in the model. The Measure distance dialog is shown in Figure 3-9.

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-21

Figure 3-9: The measure distance dialog

3.4.2 Angle between points

The Measure angle tool can be opened by clicking the Measure angle icon on the Tools tab. The
between between three points can be determined by holding down <Ctrl><Shift> and clicking
with the mouse cursor on any three points in the model.

3.4.3 Calculator

The Calculator allows the evaluation and testing of variable and named point based expressions
without modifying any model entries. The calculator is located on the Tools tab on the ribbon.
The format of the result can be controlled. Scientific uses exponential notation, for example,
0.01 becomes 1.0e-2. Engineering format is similar to Scientific, except that the exponent part
is always a multiple of 3. Thus 0.01 becomes 10.0e-3. Decimal uses a fixed notation without
an exponent. This is not recommended for numerically small numbers, for example, with 5
decimals, 1.0e-6 becomes 0.000001 which means all information is lost. The Decimals field gives
the number of digits after the decimal point.

3.4.4 Notes editor

CADFEKO provides a rich text editor tool (shown in Figure 3-10) in which the user can add
comments to a model. The editor is opened from the standard toolbar or by toggling the Notes
editor icon on the View tab. The notes view opens in its own window, allowing usage of multiple
monitors.
Unlike the 3D views, there is only one notes editor.
The toolbar at the top of the noted editor provides buttons to Clear and Print the notes, for Undo
/ Redo changes as well as the standard Cut, Copy and Paste commands. Multiple undo / redo
commands are supported. There is no facility to Cancel any changes to the text, but this may be
achieved with the Undo command.

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-22

Figure 3-10: The notes editor

The editor allows a choice of font name, font size, font type (bold, italic, underlined) and text
colour as well as setting the justification. If no notes have been previously saved with the model,
then a basic template is loaded (as shown in Figure 3-10).
The contents of the notes editor is also written to the *.pre file for users that need to create the
model in CADFEKO and then change the model using EDITFEKO. The contents are written tot he
top of the *.pre file as a series of comments.

3.4.5 Archiving models

When working with complex models, it is sometimes useful to create a backup of the model in
a specific state that may be returned to later, before continuing. CADFEKO provides a facility
to store and retrieve different versions of the project files. Select Archive → Archive model
from the application menu and enter a comment to go with this version of the model files.
The current model is saved and copied to the archive directory. Note that external data files,
such as the pattern data used for patterned point sources (see section 6.5.3) and ideal receiving
antennas (see section 6.12.8) which can change independently and are not considered part of
the CADFEKO model are not included in the archive.
Select Archive → Revert from archive from the application menu to open the Revert model dialog
which lists the archived versions with their time stamps and comments. Select a version and click
Revert to revert to that version. Note that the revert operation cannot be undone and the model
files at the time of the revert operation are overwritten by those retrieved from the archive. A
model should therefore usually be archived before reverting to a previously archived version.
Finally selecting Archive → Delete from archive allows deleting archived versions to save disk
space. Multiple archives can be selected and deleted simultaneously.

3.5 Short-cut keys

The keyboard short-cut keys that are available in CADFEKO are listed in Table 3-1 and Table 3-2.

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-23

Table 3-1: Short-cut keys (running other FEKO components)

Short-cut key Description

<Alt><1> Run EDITFEKO
<Alt><2> Run PREFEKO
<Alt><3> Run POSTFEKO
<Alt><4> Run FEKO
<Alt><6> Run OPTFEKO
<Alt><7> Run TIMEFEKO

Table 3-2: CADFEKO short-cut keys

Short-cut key Description

<F1> Context-sensitive help for the dialog / window that has focus.
<F2> Rename selected item.
<F9> Create workplane.
<Del> Delete selected items.
<Shift><Ins> Paste clipboard text.
<Ctrl><Ins> Copy selected text.
<Shift><Del> Cut selected text.
<Ctrl><A> Select all items of the current selection type.
<Ctrl><C> Copy selected text/image.
<Ctrl><E> Export image.
<Ctrl><K> For mesh parts or solution items which allow multiple instances,
create copies of the selected items. For geometry items (at any
level) create new root-level parts as copies of the selected items.
<Ctrl><M> Create mesh.
<Ctrl><Shift><M> Quick mesh - mesh with the current settings without requiring
confirmation.
<Ctrl><N> Create new model.
<Ctrl><3> Create new 3D view.
<Ctrl><O> Open model.
<Ctrl><P> Print
<Ctrl><S> Save model.
<Ctrl><V> Paste.
<Ctrl><X> Cut selected text.
<Ctrl><Y> Redo model creation / modification.
<Ctrl><Z> Undo model creation / modification.
<Alt><←> Undo view manipulation.
<Alt><→> Redo view manipulation.

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-24

Table 3-3: CADFEKO tree and message window short-cut keys

Short-cut key Description

<Alt><T> Open/closes the tree
<Alt><W> Open/closes the message window

Table 3-4: CADFEKO create solids short-cut keys

Short-cut key Description

<C,1> Create a cuboid
<C,2> Create a flare
<C,3> Create a sphere
<C,4> Create a cylinder
<C,5> Create a cone

Table 3-5: CADFEKO create surfaces short-cut keys

Short-cut key Description

<S,1> Create a polygon
<S,2> Create a rectangle
<S,3> Create an ellipse
<S,4> Create a paraboloid
<S,5> Create a NURBS

Table 3-6: CADFEKO create curves/arcs short-cut keys

Short-cut key Description

<V,1> Create a straight line
<V,2> Create a polyline
<V,3> Create a fitted spline
<V,5> Create a Bezier curve
<V,6> Create an analytical curve
<A,1> Create an elliptic arc
<A,2> Create a parabolic arc
<A,3> Create a hyberbolic arc
<A,4> Create a helix

July 2011 FEKO User’s Manual

INTRODUCTION TO CADFEKO 3-25

Table 3-7: CADFEKO modify short-cut keys

Short-cut key Description

<U> Union the selected parts
<B,1> Subtract selected object from another object
<B,2> Intersect the selected geometries
<B,3> Split selected items along a plane
<B,4> Stitch selected face parts together
<E,1> Spin selected items around a specified axis
<E,2> Sweep selected items in a specified direction
<E,3> Sweep selected item along a path
<E,4> Connect two curve parts to form a loft surface

Table 3-8: CADFEKO view short-cut keys

Short-cut key Description

<F5> Zoom to extents
<0> Restore view
<8> Top view
<2> Bottom view
<5> Front view
<Ctrl><5> Back view
<4> Left view
<6> Right view

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-1

4 Creating and meshing geometry in CADFEKO

In this chapter the various methods and concepts related to the creation and manipulation of
geometry and meshes in CADFEKO are discussed.

4.1 Model geometry properties

4.1.1 Setting units

To set the unit used for all distances and dimensions in CADFEKO, click on the Model unit button
on the Construct tab. Besides the standard units provided in the dialog, the user can specify an
arbitrary unit conversion factor with respect to metres. For example, the model unit should be
1 × 10−6 when working in micrometres. The unit is displayed on the status bar (see section 3.1).
Changing the unit does not directly modify any numbers specified in CADFEKO, but rather the
interpretation of all numbers (created before and after the unit change) — hence the unit can be
changed at any time during construction of the model.

4.1.2 Geometry extents

Operations on geometry (such as checking if two points are at the same physical location) require
a numeric tolerance. This tolerance is dependent on the model size. For example, microwave
structures may require dimensions that differ by only few micrometers, but this level of accuracy
is clearly not required for studies of antennas on large ships. CADFEKO therefore allows the
specification of the model size — click on the Extents icon on the Consruct tab. The model extent
is the same in all coordinate directions and therefore only one value needs to be specified. This is
the Maximum coordinate which gives the largest offset — in either direction — along any of the
three axes. For example, if the Maximum coordinate is 500, the entire geometry must fit inside a
1000x1000x1000 box centred at the origin. The size of the extents are specified in the CADFEKO
model unit (see section 4.1.1).
The tolerance on the model dimensions can be determined by taking the Maximum coordinate
divided by 5 × 108 . If coordinate values differ by more than this amount, they are considered
unique. Note, however, that values are only guaranteed to be considered identical if the distance
between them is less than a hundredth of the model tolerance. (Between these values is a range
where the uniqueness and connectivity cannot be guaranteed.)
The default value for the geometry extents is 5E+02 (500). It is recommend that this value be
used except if the model is large (for example, when modelling an automobile in mm model
units), or small with a requirement for high dimension-accuracy (accurate CAD model and / or
mesh). For extent settings other than the default value of 500, exported Parasolid models will
not be in the same unit used in CADFEKO (see section 5.1.2.)
The extents box limitation applies to all geometry. For example, an intersection between two
spheres cannot be performed if either sphere exceeds the size box even if the resultant intersec-
tion would be within the extents bondary. CADFEKO will display an error if the geometry exceeds
the specified size. In this case, the extents can be changed without having to close the dialog that
caused the error and the operation re-attempted.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-2

4.2 Variables and named points

4.2.1 Variables

CADFEKO supports fully parametric models. Thus most input fields can be specified using vari-
ables or mathematical expressions such as 1+sqrt(x), where x is a user defined variable. The
expressions are stored as part of the model and when a variable is changed, all values and items
depending on that variable are re-evaluated and updated. It is therefore very easy to adapt a
model using variables and a little planning, for example, to modify the geometry based on the
frequency. Since it is possible to edit geometry objects, variables can be introduced after con-
structing the model.
While the geometry is fully parametric, the mesh is not (storing expressions for mesh vertices
would have significant implications on memory requirements).
In general, operations at the part level are fully parametric, but operations that apply to in-
dividual elements are not. When working with mesh elements, most input fields still accept
expressions (for example, adding +1 to a coordinate of a vertex will move that vertex by one
unit). These expressions, however, are not maintained, but converted to numerical values when
the operation is executed.

Defining and editing variables

Variables may be added by double clicking on the Variables entry in the tree or selecting Model
→ Add variable from the main menu. The first character of a variable name must be alphabetic
(‘a’ through ‘z’ or ‘A’ through ‘Z’) or the underscore; the remaining characters may be numeric (‘0’
through ‘9’).
Variable names are case insensitive, i.e. ‘a’ and ‘A’ are treated as the same character. The Ex-
pression defines the value of the variable and may be a simple number (such as 1.23) or a
mathematical expression — which may use round brackets, the operators +, -, *, \, ˆ (exponen-
tial notation), as well as the functions listed in Table 4-1 and Table 4-2 and may reference other
defined variables.
If there is an error in the expression, an error message will be displayed in the message window
when the Create or Apply button is clicked. The variable can only be added/modified when the
expression is valid. A short comment for the variable can be added to the Comment field. Note
that the adding of comments for variables are optional. Predefined variables will already contain
a comment. The Evaluate button evaluates and tests the validity of the expression without closing
the dialog. The result is maintained until the next time the expression is evaluated and is not
updated automatically when the expression is changed.
Variables that have already been created can be changed by double clicking on the variable itself
(or by right clicking and selecting Properties). CADFEKO will display a circular dependencies
error if a variable is changed in such a way that it depends on itself. All items with defined
dependencies on the variable will be updated automatically. If this results in invalid geometry
(for example, a sphere radius that becomes zero or an intersection or split that becomes empty),
an error is written to the message window and the variable modification will not be accepted. The
Modify variable dialog may be left open while editing the item which gave the error. Thereafter
focus reverts back to the Modify variable dialog. Variables can be deleted via the context menu.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-3

Table 4-1: Trigonometric functions supported in CADFEKO expressions

sin trigonometric functions (arguments expected in radians)

cos
tan
cot
arcsin trigonometric inverse functions (results given in radians)
arccos
arctan
arccot
atan2 atan2(y,x) yields arctan(y/x) in the range -π. . .π
sinh hyperbolic functions
cosh
tanh

Table 4-2: Mathematical functions supported in CADFEKO expressions

fmod fmod(a,b) returns the remainder of the division a/b

deg converts radians to degrees
rad converts degrees to radians
sinh hyperbolic functions
cosh
tanh
log logarithm to base 10
ln natural logarithm
exp exponential function
sqrt square root
abs absolute value
step step(x) is 1 when x>0; otherwise it is 0
ceil rounded upwards
floor rounded downwards
min min(a,b) or max(a,b) gives the minimum / maximum of the two arguments
max

The delete operation will fail for all variables that are still referenced in any other part of the
model (for example, in defining other variables or geometry). A variable or multiple variables
may also be copied. Select the variables, right click and select Copy from the context menu. The
copy of the variable will be indicated by appending _1 to the variable’s original name.
The variables in Table 4-3 are pre-defined as part of a new model. These variables may be deleted
and / or modified just like any other variable and only have an effect if the user explicitly refers
to them.
Multiple variables may also be modified before re-evaluating by right clicking on Variables in the
tree and selecting Modify multiple variables. The Modify multiple variables dialog is shown in
Figure 4-1.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-4

Table 4-3: Predefined variables

c0 the speed of light in free space in m/sec

eps0 the permittivity of free space in F/m
mu0 the permeability of free space in H/m
pi the mathematical constant π (Ludolph’s number)
zf0 the characteristic impedance of free space in Ohm

Figure 4-1: Modify multiple variables dialog

4.2.2 Named points

Named points can be created by double clicking on the Named points entry in the tree, by right
clicking it and selecting Add named point, or by selecting Add point on the Construct tab. Named
points may be copied by right clicking and selecting Copy from the context menu. The copy of
the named point will be indicated by appending _1 to the point’s original name.
Named points are effectively vector variables and are typically used to create geometry that
should change when the point is modified (see section 3.2.2). The Add named point dialog
contains input fields for the three coordinates of the point in the global coordinate system.
The x, y and z components of a point may be accessed using a dot followed by the required
component, for example, ‘P.x 0 is the x component of point P. Points may also be constructed
using the ‘pt’ command. For example, the expression ‘pt(1,1,1) + pt(2,1,1)’ will result in a point
definition of (3,2,2). Only the subtract and add operations are allowed between two points
(the corresponding components are added / subtracted) and a point may be multiplied with or
divided by a scalar. The ‘abs’ function may also be applied to points. This function gives the
distance from the origin to the point, i.e. the vector length if the point is interpreted as a scalar
value. Currently no other operations may be applied to points.

4.3 Creating geometry

CADFEKO provides various tools for the creation of basic geometric forms from which a more
complex geometry can be constructed. These basic forms are referred to as ‘primitives’. Geometry
primitives can be categorised into three groups.

1. Solid primitives

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-5

2. Surface primitives and

3. Curve primitives

The basic concepts involved in the creation of geometry primitives are common to all primitive
types.

4.3.1 General geometry primitive dialog behaviour

A typical primitive creation dialog (the Create cuboid dialog) is shown in Figure 4-2.

Figure 4-2: The geometry tab of the Create cuboid dialog

Workplane: All geometry objects have a fixed orientation according to their workplane. Note that any
modifications to a primitive’s workplane, will only affect that specific geometry.

Definition methods: Many geometry primitives support creation using different parameters. The easiest
definition depends on the information available for the geometry. Each definition is also
accompanied by an image that illustrates the dimensions required for a complete and unique
definition of the primitive.

Points and dimensions: All coordinate and dimension fields support general point entry (see section 3.2.2).
All required dimensions are constrained (for example, the width specified for a cuboid prim-
itive must be non-zero). If a value is applied that does not meet the specific constraints,
CADFEKO will display an error in the message window and highlight the offending field
during creation or when application of changes is attempted. Where negative values are ac-
ceptable (most dimension and angle fields), they are interpreted in the opposite direction.
(If a value of 1 implies a length in the direction of the positive u-axis, then a value of -1 will
be valid, and interpreted as a length in the direction of the negative u-axis). For an angle
field, if 10 implies rotation in the positive φ direction, then a value of -10 is interpreted as
rotation in the negative φ direction.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-6

Label: The Label field allows entering the name of the new object. This is visible in the model
tree and can be changed at any time after creation. The Label may be chosen freely, but
two identically labelled objects are not allowed in the same level of the tree (irrespective of
type). Note that labels used in FEKO are considered case insensitive.

Create / Close: A new object is created when the Create button on the creation dialog is clicked. The
dialog can be closed by clicking Close or by pressing <Esc>.

4.3.2 The workplane

In CADFEKO, there are three predefined workplanes namely: Global XY, Global XZ and Global
YZ. Any of these workplanes can be at any time be set as the default workplane by right-clicking
on the respective workplane in the contents tree and selecting Set/Unset as default from the
pop-up menu. After a workplane is set as the default workplane, it will be indicated by the text,
[Default] in the model tree. The default workplane will be used as the starting workplane for the
dialogs concerning primitive creation, operators and transforms. If necessary, the workplane can
be modified on any Workplane tab for primitive creation, operators and transforms. (The easiest
is to use point entry - <Ctrl><Shift> + click on workplane in the tree.)
The workplane is determined by the orientation of the axis which is displayed as a blue X/U,
green Y/V and red Z/N tangential lines. When an axis is obscured by geometry, it will still be
visible as a dotted line in their respective colour.
The workplane display can be switched off by clicking Show/Hide workplane on the 3D view
toolbar. This action will also hide the dynamic grid. The grid is dynamic as the tick marks are
determined by the amount to which the model is zoomed in or out. When the model is zoomed
out, the interval between the tick marks will be increased. An extra option is available under the
Show/Hide workplane to hide the tick marks on the dynamic grid.
The workplane is specified in terms of an Origin which defines the position, and the U vector and
V vector directions which define the orientation of the workplane. The workplane can also be
rotated around the U, V or N axis. The angle (in degrees) for rotation can be entered underneath
the icons (shown in Figure 4-3). Note that these buttons perform actions on the values entered
in U and V.
A workplane can be easily modified by snapping to existing faces, edges and wires (see section
3.2.1).
Only the directions of the vectors are relevant, i.e. using the vector (2,0,0) is the same as using
(1,0,0). When calculating the v-axis, the component of V vector parallel to the U vector is
removed to ensure that the two axes are orthogonal. Thus, the V vector need not be specified
exactly orthogonal to the U vector, though it may not be specified exactly parallel to U vector.
The n-axis is normal to the uv-plane and in the direction u × v.
The Origin, U vector and V vector fields may be specified using the <Ctrl><Shift> (see sec-
tion 3.2.2) and clicking with the mouse. This means that the current workplane may be used to
enter values which will define the settings of a new workplane. (Pressing <Ctrl><Shift> and
clicking on the view while any of the Origin, U vector or V vector fields has focus will modify its
values.)

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-7

Figure 4-3: The workplane tab on the Create cuboid dialog

Coordinates set in a given workplane

All primitive geometry objects and solution entities with spatial properties in CADFEKO are de-
fined in the default workplane environment, unless the default workplane was modified on the
workplane tab. When a workplane is defined (see section 4.3.2), the coordinate system is set in
the defined workplane environment.

4.3.3 Creating solid primitives

The Create solid group on the Construct tab contains tools for the creation of cuboids, flares
(a possibly-truncated pyramid), spheres, cylinders and cones. (A typical solid primitive dialog
is shown in Figure 4-2.) Solid primitives are by default perfect electric conductors, but can
be changed to dielectric or shell structures. This is done by setting region properties (see sec-
tion 6.1.2).
For cuboid primitives, all faces lie in the principle planes of the indicated coordinate system. For
flares, the base is aligned with the uv-axes and the height is in the direction of the n-axis.
For cylinders and cones, a number of special definition options are available in order to aid
flexibility in the creation of complex geometries. The various primitive definition options are
depicted and described below. If the Radius field for a sphere or cylinder is specified using a
point, the surface of the sphere or the extended cylinder will pass through that point. For a cone
the radius is determined as the distance between the origin of the local coordinate system and
the projection of the point in the local uv-plane. To specify a sharp tipped cone, set the Top radius
field to ‘0’.
When creating any of the solid primitives, surface primitives or line primitives, default values are
added to the geometry dimension fields. The default values are dependant on the zoom level in

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-8

the 3D view and are only given to enable a preliminary preview of the primitive in the 3D view.
The exceptions to these default values are the polygon and fitted spline.

Definition options for solid geometry objects

Cuboid - Method 1

Base corner (C): One corner of the cuboid.

Width (W ): The cuboid dimension in the u-axis direction.

Depth (D): The cuboid dimension in the v-axis direction.

Height (H): The cuboid dimension in the n-axis direction.

Cuboid - Method 2

Base centre (C): The centre of the base of the cuboid.

Width (W ): The cuboid dimension in the u-axis direction.

Depth (D): The cuboid dimension in the v-axis direction.

Height (H): The cuboid dimension in the n-axis direction.

Flare - Method 1

Base centre (C): The centre of the flare base.

Bottom width (Wb ): The width of the base in the u-axis direction.

Bottom depth (D b ): The depth of the base in the v-axis direction.

Height (H): The height of the flare, in the n-axis direction.

Top width (Wt ): The width of the top in the u-axis direction.

Top depth (D t ): The depth of the top in the v-axis direction.

Flare - Method 2

Base corner (C): A corner of the flare base.

Bottom width (Wb ): The width of the base in the u-axis direction.

Bottom depth (D b ): The depth of the base in the v-axis direction.

Height (H): The height of the flare, in the n-axis direction.

Top width (Wt ): The width of the top in the u-axis direction.

Top depth (D t ): The depth of the top in the v-axis direction.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-9

Flare - Method 3

Base corner (C b ): A corner of the flare base.

Top corner (C t ): A corner point of the flare top.

Bottom width (Wb ): The width of the base in the u-axis direction.

Bottom depth (D b ): The depth of the base in the v-axis direction.

Flare - Method 4

Base centre (C b ): The centre of the flare base.

Bottom width (Wb ): The width of the base in the u-axis direction.

Bottom depth (D b ): The depth of the base in the v-axis direction.

Height (H): The height of the flare, in the n-axis direction.

Flare angle (AU): The angle of the flare from the u-n plane.

Flare angle (AV ): The angle of the flare from the v-n plane.

Sphere - Method 1

n Centre (C): The centre point of the sphere.

v C
u
Radius (R): The radius of the sphere.

Sphere - Method 2 (ellipsoid)

Centre (C): The centre point of the sphere.

Rn
n Radius (Ru ): The radius of the ellipsoid in the U dimension.
v C
u
Rv Radius (R v ): The radius of the ellipsoid in the V dimension.
Ru

Radius (R n ): The radius of the ellipsoid in the N dimension.

Cylinder - method 1

Base centre (B): The centre of the cylinder base.

Radius (R): Cylinder radius (parallel to the uv plane).

Height (H): Cylinder height in the n-direction measured from B.

Cylinder - method 2

Base centre (B): The centre of the cylinder base.

Top centre (T ): The centre of the cylinder top.

Radius (R): The cylinder radius (perpendicular to the line from B to T ).

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-10

Cone - method 1

Base centre (B): The centre of the base of the cone.

Base radius (R b ): The radius of the cone base (parallel to the uv plane).

Height (H): The cone height in the n-direction measured from B.

Top radius (R t ): The radius of the cone top (parallel to the uv plane).

Cone - method 2

Base centre (B): The centre of the base of the cone.

Top centre (T ): The centre of the top of the cone.

Base radius (R b ): The radius of the cone base (perpendicular to the line
from B to T ).

Top radius (R t ): The radius of the cone top (perpendicular to the line from
B to T ).

Cone - method 3

Base centre (B): The centre of the base of the cone.

Base radius (R b ): The radius of the cone base (parallel to the uv plane).

Height (H): The cone height in the n-direction.

Flare angle (A): The growth angle measured from the n-axis.

Cone - method 4

Base centre (B): The centre of the base of the cone.

Top centre (T ): The centre of the top of the cone.

Base radius (R b ): The radius of the cone base (perpendicular to the line
from B to T ).

Flare angle (A): The growth angle measured from the line defined be-
tween B to T .

4.3.4 Creating surface primitives

Flat polygons, elliptical disks and paraboloid surfaces are supported as surface primitive types.
More complex surfaces may be created using the loft, spin, sweep or path sweep (see sec-
tion 4.4.3) operations on line primitives (see section 4.3.5).
The polygon creation dialog is shown in Figure 4-4. All of the specified points must lie in a plane
— the simplest way to ensure this is to align the workplane with the plane of the polygon and
ensure that all n-coordinate values are exactly equal (usually zero).

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-11

Figure 4-4: The dialog used to create polygons

Clicking the Add button adds an additional corner to the list directly below the one that has focus
or, if no corner has focus, at the end of the list. The polygon is created by connecting the lines in
the specified order and the edges are not allowed to cross. The corner with focus is represented
in the 3D display with a blue square. If the last coordinate is entered using the mouse (see
section 3.2.2), CADFEKO automatically adds another corner and moves the focus to the added
corner.
The polygon primitive provides an Import points option (see section 4.3.6) to populate the corner
point definitions based on an external file.
The Reverse normal button reorders the points in such a way that the normal vector (determined
in a mathematically positive sense from the direction of the edges) is reversed. When Create is
clicked, all empty points are removed automatically, and the surface is created.
The definition options for the different surface primitives are depicted and described in the sec-
tion that follows.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-12

Definition options for surface primitive objects

Polygon

Corner 1 (C1 ): The first corner of the polygon.

Corner 2 (C2 ): The second corner of the polygon.

...Corner n (Cn ): Additional corners of the polygon (an arbitrary number).

All must be in the same plane. The polygon is closed by con-
necting the last point to C1 .

Rectangle - Method 1

Base corner (C): A corner of the rectangle.

Width (W ): The width of the rectangle.

Depth (D): The depth of the rectangle.

Rectangle - Method 2

Base centre (C): The centre of the rectangle.

Width (W ): The width of the rectangle.

Depth (D): The depth of the rectangle.

Ellipse

Centre point (C): The centre of the ellipse.

Radius (Ru ): The radius (half of the axis length) in the u-axis direction.

Radius (R v ): The radius (half of the axis length) in the v-axis direction.

Paraboloid

Centre point (C): The apex of the paraboloid.

Radius (R): The radius of the paraboloid aperture, parallel to the uv-plane.

Focal depth (F ): The Focal depth of the paraboloid is the distance F from
the centre point (C) to the focal point. If this is negative, the
paraboloid is oriented towards the or -n axis. The focal depth
is related to the height (h) by

R2
F= .
4H
with the height (H) defined as the distance from the centre
point (C) to the centre of the aperture of the paraboloid.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-13

NURBS

Specify the order of the Bezier curves: The degree of the Bézier curve in the
û-direction and v̂-direction.

Point position: The position of each surface control point is specified.

Weight: The weight of each surface control points is specified.

4.3.5 Creating curve primitives

Curve primitives in CADFEKO can be used either as building blocks for constructing geometry (for
example, using the spin, sweep or loft operations) or as free-standing conducting wires. In ad-
dition to the general curves resulting from intersections between faces of shell bodies, CADFEKO
supports a number of simple primitives. These include single straight lines, polylines, analytical
curves, fitted splines, Bézier curves, elliptical arcs, hyperbolic arcs, parabolic arcs and helices /
spirals. The creation of curve primitives is very similar to the creation of solids and surfaces.
The polyline and fitted spline primitives make use of a list of points. These points may be man-
ually specified, or imported from an external text file (see section 4.3.6). Here, as for polygon
surfaces, new fields are created if the last field is entered with the mouse, and blank points are
removed before constructing the geometry.
The helix primitive may be used to create conical spirals by setting the top and bottom radii to
different values, or flat spirals by setting the height to zero. The direction of rotation of the helix
can be specified using the Left handed checkbox. When this box is checked, a left-handed spiral
will be generated, when it is un-checked, a right-handed spiral will be created. The handedness
of the spiral is defined based in the direction from the start point to the end point of the spiral.
The definition options for the different line primitives are depicted and described below.

Definition options for line primitive objects

Line

Start point (P1 ): The starting point of the line.

End point (P2 ): The end point of the line.

Polyline

Corner 1 (C1 ): The starting point of the polyline.

Corner 2 (C2 ): The second point of the polyline.

...Corner n (Cn ): Additional points in the polyline. There may be an arbi-

trary number of points.

July 2011 FEKO User’s Manual

 CREATING AND MESHING GEOMETRY IN CADFEKO 4-14
n
v
u
z
Analytical curve - Cartesian
P[x(t),y(t),z(t)]

x n
Parametric interval: Interval over which the analytical curve is parametri-
z
y z P[u(t),v(t),n(t)] cally defined.
v
Cartesian description: The description of the curve using the Cartesian co-
y u ordinate system. The u, v and n dimensions as a function of
x
variable t.

Analytical curve - Spherical

H

B
n Rb Parametric interval: Interval over which the analytical curve is parametri-
P[r(t), (t), (t)]
cally defined.
r
v
Spherical description: The Spherical description of the curve in the r, θ and
u
φ dimensions as a function of variable t.

Analytical curve - Cylindrical

z n Parametric interval: Interval over which the analytical curve is parametri-

P[ (t), (t),n(t)]
cally defined.
v
y
Cylindrical description: The Cylindrical description of the analytical curve
x u
in the ρ, φ and n dimensions as a function of variable t.

Elliptic arc - method 1

Centre point (C): The centre of the ellipse on which the arc lies.

Radius (RU): The radius (half of the axis length) in the u-axis direction of
the ellipse on which the arc lies.

Radius (RV ): The radius (half of the axis length) in the v-axis direction of
the ellipse on which the arc lies.

Start angle (A0): The angle, from the positive u-axis direction where the
arc begins.

End angle (A1): The angle, from the positive u-axis direction where the
arc ends.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-15

Elliptic arc - method 2 (V Major axis direction)

Aperture centre (C): The centre of the aperture formed by the elliptical arc
section.

Depth (D): The distance from the aperture centre point to the apex of the
elliptical arc section.

Aperture radius (R): The radius of the aperture of the elliptic arc.

Eccentricity: The eccentricity of the ellipse on which the elliptical arc sec-
tion lies. The eccentricity must be less than 1 to specify a valid
ellipse.

Conditions for creating a valid elliptic arc:

r
R2
if R ≤ D : 1− ≤ε<1 (4-1)
D2
where D is denoted by the depth, R by the aperture radius R
and ε the eccentricity.

Elliptic arc - method 2 (U Major axis direction)

Aperture centre (C): The centre of the aperture formed by the elliptical arc
section.

Depth (D): The distance from the aperture centre point to the apex of the
elliptical arc section.
C
D R
Aperture radius (R): The radius of the aperture of the elliptic arc.

R Eccentricity: The eccentricity of the ellipse on which the elliptical arc sec-
C D
v tion lies. The eccentricity must be less than 1 to specify a valid
y u
x ellipse.

Conditions for creating a valid elliptic arc:

r
D2
if R ≥ D : 0 ≤ ε ≤ 1− (4-2)
R2
where D is denoted by the depth, R by the aperture radius R
and ε the eccentricity.

Parabolic arc - method 1

Base centre (C): The centre of the parabola on which the arc lies.

Focal depth (F ): The focal depth of the parabola.

Radius (R): The radius of the aperture of the parabola.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-16

Parabolic arc - method 2

Base centre (C): The centre of the parabola on which the arc lies.

Radius (R): The radius of the aperture of the parabola.

Depth (D): The distance from the apex of the parabola to the centre of the
aperture.

Parabolic arc - method 3

Aperture centre (C): The aperture centre of the parabolic arc section.

Radius (R): The radius of the aperture of the parabola.

Depth (D): The distance from the apex of the parabola to the centre of the
aperture.

Hyperbolic arc - method 1

Base centre (C): The centre of the hyperbola on which the arc lies.

Depth (D): The distance from the apex of the hyperbola to the centre of
the arc aperture.

Radius (R): The radius of the aperture of the hyperbola.

Eccentricity: The eccentricity of the hyperbola on which the hyperbolic arc

section lies.

Conditions to create a valid hyperbolic arc:

r
R2
1≤ε≤ 1+ (4-3)
D2
where D is denoted by the depth, R by the aperture radius R
and ε the eccentricity.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-17

Hyperbolic arc - method 2

Aperture centre (C): The centre of the aperture formed by the hyperbolic
arc.

Depth (D): The distance from the apex of the hyperbola to the centre of
the arc aperture.

Radius (R): The radius of the aperture of the hyperbola.

Eccentricity: The eccentricity of the hyperbola on which the hyperbolic arc

section lies. The eccentricity must be greater than one to spec-
ify a valid hyperbola. Note that not all values greater than one
specifies a valid hyperbola.

Conditions to create a valid hyperbolic arc:

r
R2
1≤ε≤ 1+ (4-4)
D2
where D is denoted by the depth, R by the aperture radius R
and ε the eccentricity.

Bezier curve

Corner 1 (C1 ): The starting point of the curve.

Corner 2 (C2 ): The first control point of the Bézier curve (the curve does
not necessarily pass through this point).

Corner 3 (C3 ): The second control point of the Bézier curve (the curve
does not necessarily pass through this point).

Corner 4 (C4 ): The end point of the curve.

Fitted spline

Point 1 (P1 ): The starting point of the curve.

Point 2 (P2 ): The second point through which the spline curve will pass.

...Point n (Pn ): The additional points through which the spline must pass.
There may be an arbitrary number of points.

Helix - method 1

Origin (C): The centre point of the helix base.

Base radius (R b ): The radius of the helix base (parallel to the uv-plane).

End radius (R t ): The radius of the helix top (parallel to the uv-plane.

Height (H): The height of the helix, in the n-axis direction.

Turns (N ): The number of turns for the helix (the rotation direction is
chosen based on the Left handed checkbox option).

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-18

Helix - method 2

Origin (C): The centre point of the helix base.

Radius (R): The radius of the helix (parallel to the uv-plane).

Pitch angle (A): The angle formed between the tangent of the curve and
the uv-plane — constant along the length of the helix.

Turns (N ): The number of turns for the helix (the rotation direction is
chosen based on the Left handed checkbox option).

Helix - method 3

Origin (C): The centre point of the helix base.

Radius (R): The radius of the helix (parallel to the uv-plane).

Height (H): The height of the helix, in the n-axis direction.

Pitch angle (A): The angle formed between the tangent of the curve and
the uv-plane — constant along the length of the helix. (In this
method, the number of turns is related to the height (H) and
pitch angle (A) chosen).

Note that the local wire radius (see section 4.8.4) is specified on an edge (as listed in the details
tree) rather than on the curve primitive directly. Specifying the radius for a wire curve (any free
edge that is not a face border) that was created by geometry operations (such as the intersection
of two crossing faces) is done in the same way.

4.3.6 Import points

Some primitives (such as the polygon, polyline or fitted spline), as well as the cable path defini-
tion for cable analysis and the points list definition for the imprint points tool, are defined based
on a list of coordinate points. These points may be entered manually (using the keyboard or
based on mouse clicks) or imported from an external file that contains a list of the coordinates.
In the creation dialogs of primitives that support the importing of point definitions from an ex-
ternal file, an Import points button is available (Figure 4-5 shows the Import points button on
the create polyline dialog).
If the Import points button is clicked, the Import points dialog, shown in Figure 4-6, is opened to
allow file location and data format selection.
A scaling factor may be applied to the coordinates as specified in the import file before they are
used (this may be required, for example, to convert to the current model unit setting). This
scaling is linearly applied to all coordinates, and cannot be specified only in a specific coordinate
direction.

File format for imported points

The coordinates may be specified using either a comma, tab or space delimited file format. The
format must be set on the Import points dialog. The expected file format comprises three columns

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-19

Figure 4-5: Create polyline dialog

Figure 4-6: Import points dialog

(for each of the three Cartesian coordinates) separated by the specified delimiter, with each
coordinate set on a separate line. Any errors in the file format will be reported when the file
selection is applied and the points are imported. Errors in the individual imported values will
be reported during geometry/solution request creation or when the changes are applied. This
allows the addition/removal or editing of specific points before application.
It is very important to note that CADFEKO does not maintain any form of linkage to files specified
as the source for point coordinate definitions. During the import process, the values represented
in the source file are simply used to automatically populate the relevant dialog fields. In order to
reflect changes made in the source file, the point import process will have to be repeated.

4.4 Operations on geometry

Complex geometry is created from a sequence of operations starting with simple primitives. This
sequence is represented in the tree (see section 3.3). It is possible to select and modify objects at
any level in this tree, or change variables upon which the geometry depends. If parent items are
selected, CADFEKO displays a temporary wire-frame representation of all selected items.
All items higher up in the geometry tree are re-evaluated each time an item is changed. During

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-20

this process CADFEKO may not be able to maintain the identities of some regions / faces /
edges, for example, where multiple faces are derived from the same original face during Boolean
operations. These items are then marked suspect (indicated by a question mark next to the
representation of the items in the details tree — see section 3.3.1) as a warning that properties
set may not have been maintained. The reason that an item has become suspect is displayed if the
mouse pointer is moved over the item in the tree. After ensuring that the properties are correct
or have been corrected, the suspect setting on the relevant regions / faces / edges may be reset
using Set not suspect from the pop-up menu. Some solution configuration items (for example
ports) that depend on the model will also be marked suspect if the model changes in such a way
that they are no longer valid. These must also be checked and the suspect setting removed. It is
advisable to resolve all suspect items before launching the FEKO solver or optimiser, as the loss of
certain properties on the model geometry may change the electromagnetic problem description
dramatically and have a large effect on computed results.
Any combination of objects or the variables that they depend on can be selected and modified.
All operations are available from the Construct and Transform tabs.
The geometry modification operations are listed below.

Boolean operations Union, Subtract and Intersect

Operations on parts Project, Imprint points, Split, Explode, Simplify, Stitch sheet parts

Creation operations Spin, Loft, Sweep, Path sweep

Property operations Reverse face normals and Properties (which opens an edit dialog similar
to the create dialog.)

Administrative operations Rename(<F2>), Copy(<Ctrl+K>), Delete(<Del>)

In addition to these operations, transformations (see section 4.5) (Scale, Mirror, Rotate, Trans-
late) can be applied to geometry objects.

4.4.1 Boolean operations

The Construct tab provides access to the Boolean operations Union, Subtract and Intersect. The
part/s to which the operation will apply must be selected before requesting a Boolean operation.
The boolean operations are inactive when no parts are selected. (Surfaces and wire bodies are
also considered parts.) Boolean operations cannot be applied to parents or the regions / faces /
edges of parts.
When adjacent objects in a model are slightly misaligned, there may be faces and / or edges
that have very small overlapping (supposedly separate objects) or non-overlapping (supposedly
connected objects) sections as shown in Figure 4-7. If these sections are of the same order as
the model tolerance, Boolean operations involving these objects may fail, or result in very small
faces or gaps. The snap facility can be used to ensure that these items line up correctly. Similar
problems can occur when an item is split very close to a boundary.
Note that is now possible to add and remove items from an union and other boolean operators
without having to copy the contents out to the root and recreating the boolean operation.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-21

Figure 4-7: Slightly misaligned surfaces showing a short overlapping edge as well as a short non-
overlapping edge.

The Union operation

Union combines all selected parts. In CADFEKO the Union operation is used to define connec-
tivity between parts. Parts that touch, but are not unioned are considered not to be physically
connected. When meshing, incorrect connectivity may result in invalid meshes (overlapping or
misaligned meshes) that will generate errors during the FEKO solution.

The Intersect operation

For Intersect, the result is the common part of all the selected parts.

The Subtract operation

For Subtract, all the selected parts are subtracted from a final part, which must be selected when
prompted for. When only one part is selected, the only Boolean operation that can be launched is
Subtract. After the selected parts have been subtracted from the target part, the target part (the
part which was subtracted from) will be indicated by a T in the tree.

Figure 4-8: The T indicating the target part in the subtract operation.

Note that regions are taken into account during the subtract operation. Subtracting a shell struc-
ture (a free space region - implying an empty/hollow closed structure) may generate new regions,
faces or edges on intersecting geometry, even though no geometry will be removed. Subtracting
a solid structure (a region that is not set to free space) will result in geometry being removed
from any intersecting parts.

4.4.2 Operations on parts

Projection and imprinting points

In some cases it is desirable to create specific points, edges or faces on a given geometry. For
example, to provide accurate and valid attachment points for other structures on the model,

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-22

to manually force a finer mesh size along a curve or to create conducting patches on dielectric
objects.
The Imprint points operation allows the placement of specified points on the selected part. Points
can only be imprinted on one part at a time. The Imprint points dialog allows a list of points to
be specified using standard point entry (see section 3.2.2) or based on a list of points imported
from an external file (see section 3.2.2) in either global or local coordinates. The specified points
are projected onto the closest point on the selected part — either on a face or an edge. Points
may not be imprinted on top of existing points. The imprint operation creates a new entry in the
tree (allowing access to the part without points), but uses the name of the parent object, as there
is only one parent for this case. In CADFEKO 1 it was possible to specify vertices on the Edge
properties dialog. This is no longer the case — such points must be created using the Imprint
points operation.
Multiple parts can also be projected onto another part by selecting them, activating the Project
operation and then selecting a target part to project onto. This sequence is similar to the subtract
operation. All the projected parts appear as parents of the resulting part, i.e. they are no longer
present as individual parts of the model. All of the edges of the selected parts are projected onto
the faces of the target part. Any part (curves, surfaces and solids) can be projected onto any
other part containing faces. (Spheres do not have edges, hence projecting a sphere has no effect
other than removing it from the model.) Where the projected edges form closed paths, new faces
will be created.
The projection direction is determined based on the normals of the faces of the target. Projecting
edges onto convex curved target faces will thus tend to reduce their size and / or perspective.
Convex surfaces may also “shadow” other surfaces in that all points on the projected edge may
project onto the curved face even though it may seem as projection onto another face should also
result. Edges are only projected onto the normal side of faces.
Edges where the projection crosses itself or turns back on itself are not allowed.

Split parts

Selecting the Split button on the Construct tab will open a dialog where the split plane can be
specified by means of an origin, a plane (UV, UN or VN) and the angle of rotation around the
axes.
The split operation creates two new parts (named Split_back. . . and Split_front. . ., where the
Front side of the split is in the direction of the positive n axis of the split plane) for each selected
part. The two halves of each split are derived from independent copies of the original part — if
parent items are modified for one half, the other half remains unchanged and has to be modified
separately. If synchronised changing is required, the original part must be created using variables.
The split plane used in a split operation may be modified by selecting the new part and selecting
Properties from the pop-up menu.

Exploding parts

Selecting Explode from the Transform tab will explode all selected geometry parts. Separate
new surface parts are created for each face and free edge of the original parts. The new parts
represent a snapshot of the geometry at the time it was exploded — they are not parametric.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-23

Simplify tool (removing detail)

Redundant faces and edges may be deleted by selecting them and pressing the <Del> key or
selecting Delete from the pop-up menu. Faces can only be redundant if they have the same
medium (i.e. metal, free space or the same dielectric medium — see section 6.1) on both sides.
When a face separating an internal free space region from the outside free space is deleted, the
internal region is merged with the outside. Since the outside medium is free space, faces can
only be removed from closed regions if the internal medium is set to Free space.
These items are permanently removed, but it is possible to copy the original object (see sec-
tion 4.4.4) in which the faces/edges still exist out of the tree to restore them. If a component or
variable is changed in such a way that the object is re-evaluated, suspect faces (see section 4.4)
will not be removed and some deleted faces/edges may reappear.
Note that edges are not redundant if the normals of the faces on either side of the edge are in
opposite directions. For this reason the front and back sides of faces are coloured differently in
the Default view and it is possible to reverse the face normals (see section 4.4.4) of these faces.
The normals of mesh triangle elements created in CADFEKO are in the same direction as the faces
they originate from.
CADFEKO provides a tool for the automatic removal of redundant faces and edges. Select Sim-
plify from the Transform tab to open the Simplify geometry dialog.

Figure 4-9: The simplify dialog

This operation allows removing specific types of items from the selected part/s. The simplified ge-
ometry will be electromagnetically the same as the original, but may not have the same meshing
constraints. If, for example, an imprinted point is removed, there will no longer be a guaranteed
mesh vertex at this location. (Note that imprinted points that are not attached to any edges are
considered redundant.) Faces cannot be deleted unless the regions they separate can be merged.
The same applies to edges on the boundaries of faces and geometry points at the ends of edges.
By default the simplify operation does not remove redundant regions, faces or edges on which
local mesh properties (see section 4.8.4) are set. To remove them, the various Keep . . . with local

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-24

mesh sizes options must be unchecked. For example, consider the union of two spheres of the
same dielectric medium shown in Figure 4-10, case (a). A local mesh size is set on Region1. If the
union is simplified with Keep regions with local mesh sizes checked, the result is as shown in case
(b). Since the region contains local properties, it cannot be removed and the face between it and
the centre region can also not be removed. If Keep regions with local mesh sizes is unchecked,
the result is as shown in case (c).

Figure 4-10: Illustrations of the simplify operation

Stitch sheet parts

It is possible that geometry that should be connected, is not, or there are small sections that
overlap. This is most often due to problems encountered with tolerances during geometry import
and should not be encountered for geometry created in CADFEKO. These geometry elements
(sheet parts only) can be stitched together using the Stitch tool. Sheet parts that are within the
specified tolerance is then considered to be connected and meshed correctly.
Note that the stitch tool can lead to a very strange (seemingly contradictory) display of geometry
in CADFEKO. This happens due to the tolerance of some edges, faces and nodes being large (as
specified in the Stitch tool) and thus they can be displayed anywhere within the tolerance area.
The mesh will not have this strange display since the mesh elements have a very small tolerance.
In some circumstances, the Stitch tool may be used as a replacement for the Union operation.
Generally the Stitch operation is faster and more efficient than the Union operation, but it may
only be employed for connected sheet parts.

4.4.3 Creation operations

Spinning and sweeping parts

Selecting Sweep, Path sweep or Spin on the Construct tab sweeps the selected parts along a
specified vector or path (these operations are also often referred to as ‘extrude’) or spins (rotates)
them around a specified axis, respectively. Spin and sweep operations can only be applied to
parts.
The sweep / spin operation is applied separately to each of the selected parts. If multiple parts
are swept / spun, the resulting new parts have no defined relation to each other based on the
spin / sweep operation, i.e. the defined operation is applied to each item independently and the
spin / sweep parameters can be modified independently afterwards. The Path sweep tool cannot
be applied to multiple parts.
Only parts that contain edges and / or faces (not solids or closed regions) may be spun / swept
/ path swept. For surface bodies, the body must have a single boundary which does not close

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-25

on itself and no edge may be attached to more than two faces (for example, the T-plate union in
Figure 4-11 is not an acceptable part for a spin / sweep / path sweep operation as the connecting
edge borders three faces). Sweeping or spinning a curve results in a surface, while applying these
operations to a surface results in a solid.

Edge bordering
3 faces

Figure 4-11: T-plate example of an edge bordering three faces so that it cannot be swept / spun

When the Sweep operation is selected, the Sweep geometry dialog allows specification of the
sweep vector in terms of a start and end point (the path is taken as the straight line between
these two points). These fields accept standard point entry (see section 3.2.2). It is not possible
to sweep bodies in a direction that is tangential to any of its edges or in the plane of its faces.
The Path sweep operation is only available when a part that can be swept is selected. The Path
sweep dialog is shown in Figure 4-12. The tool requests that the path for the sweep operation be
selected. Any part consisting of only free edges and curves that form a joined, non-overlapping
path may be chosen as a sweep path. The sweep may be adjusted by assigning a Twist angle (in
degrees) and a Scale factor. These factors specify a scaling and rotation to be applied along the
sweep path, beginning with unity scaling and 0 degree rotation and linearly increasing, ending at
the specified angle and scale factor at the end of the sweep path. Normal or Parallel sweeps may
be selected. The Parallel sweep maintains the orientation of the part during the sweep process,
while a Normal sweep maintains the angle between the sweep object and the sweep path.
When a sweep path is selected, CADFEKO automatically determines the most appropriate sweep
direction. This can, however be changed by selecting the Start from other end of path checkbox.

Figure 4-12: The Path sweep dialog

The Spin geometry dialog allows specification of the axis of rotation (an origin and a direction)
and a rotation angle. The angle is taken in the mathematically positive sense around the specified
axis and is specified in degrees. (In CADFEKO radian angles are only used in the arguments and
results of the trigonometric functions and their inverses.) The Set to . . . axis buttons allow quick
selection of common axes. Note that the local axis buttons (U,V,N) are only available if all selected
parts have the same local coordinate system.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-26

The normal directions of the faces that result from a spin / sweep operation depend on the
direction of the curves to which the operation is applied.
There are a number of restrictions on the relation between the part and the spin axis. No free
edges may be coincident with the axis, nor are they allowed to intersect the axis at any point
other than at the ends of the curve. For full spins (360◦ ) of parts with faces, the spin axis may
not intersect the face at only one point, unless it is at an end point of an edge. The axis may be
coincident with any edge of the sheet, provided that the entire edge lies on the axis. In addition,
no edge may touch the axis tangentially as shown in Figure 4-13 — not even if the edge is broken
at this point. (Again, this is allowed IF the entire edge lies on the axis.)

Tangential touch point

Axis

Figure 4-13: Example of a surface with an edge that just touches the axis of rotation

The sweep and spin operators may also be applied with an arbitrary orientation by making use
of workplanes. This is set on the Workplane tab of the Sweep and Spin dialogs.

Loft surfaces

The Loft operation forms a smooth surface by connecting two curve parts with straight lines.
(This is sometimes called a ruled surface, but it is not a faceted model — unless one or both
curves are polylines). The curve bodies involved in the loft may consist of a number of edges,
for example, a polyline. The two lofted curves must, however, have an equal number of edges
(the loft can be considered to connect each pair of edges). If this is not the case, points can
be imprinted (see section 4.4.2) on one of the curves to satisfy this requirement. Curves with
more than two edges joining at one point cannot be used in a loft operation. While the loft tool
is active, the 3D view preview shows how the curves will be connected. The Loft dialog allows
reversing the start and end points of one curve in the case where the loft between the two curves
is not indicated in the desired direction.

4.4.4 Administrative operations

Reversing face normals

Face normals can be reversed by selecting Reverse normals from the right-click pop-up menu in
the details tree representation of the face/s or by choosing Reverse normals from the Transform
tab. (The normals of all selected faces — in the tree and 3D view — are reversed, even though
only one part’s faces are shown in the details tree at any given time.)
Currently, it is not possible to reverse the normal on a body with a single, closed face, for example,
a sphere. If such a normal must be reversed, the simplest option is to project a circular line onto

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-27

the body to create two open faces. The normals of these faces can then be reversed and the
separating edge deleted or removed using the Simplify operation (see below).

Assemblies

Assemblies are used to organise the geometry and mesh parts. They become part of the geometry
structure and are shown as the first level under Geometry in the tree. To create a new assembly,
select the required items, right click in the tree or 3D view and select Assembly → Create. Only
parts (displayed objects) can be added to an assembly. (Mesh and geometry parts cannot be
added to the same assembly.) Items are moved between assemblies by selecting Assembly →
Move to → . . . in the pop-up menu. Similarly Assembly → Move out moves the selected items
back to the root level in the tree. An item cannot be in two assemblies at the same time.
Deleting an assembly removes the entire assembly and its contents from the model. Similarly,
selecting an item in an assembly and pressing <Del> will remove it from the model. (It is very
different from Assembly → Move out.) Selecting Disassemble on an assembly moves all of the
items inside it back to the root level and deletes the assembly — in effect removing the assembly
without deleting the contents.
Operations (Boolean, transforms, splitting, etc.) are applied to the individual items. If multi-
parent operations are applied to items in one assembly, the result is also in that assembly. Where
such operations are applied to items from more than one assembly, the result is placed at the root
level.
Assembly names are part of the full label / name, for example, Assembly1.Union4.Face12. This
is the reference required in EDITFEKO to reference elements on this specific face. Note that all
names in CADFEKO are case insensitive, i.e. Face1 and face1 are the same.

Copy objects

Geometry objects (including parents of objects that are shown in the tree) can be copied to new
root-level parts by selecting the items and selecting Copy from the Transform tab (or using the
short-cut <Ctrl><K>). For example, if a spherical shell is created by subtracting one sphere
from another (which will remove both original spheres from the model and list them as parents
of the new part created by the Boolean operation), the inner sphere can be copied to create a
new root-level part which fits inside the shell. Copied items are created in the root level, even
if the original items form part of an assembly. It is also possible to make multiple transformed
copies of parts (see section 4.5).
The new part is completely independent of the existing one and neither will change if the other
is modified. If it is desired that they change together, the original object should be created using
variables as all field expressions used in the primitive definition are maintained during the copy
operation.
It is also possible to select faces or edges of parts and select Copy. In this case new parts are made
for each selected item. This allows, for example, an edge to be copied from a complex object and
used to loft to another edge. Note that these copies are snapshots of the model when the copy is
made — they are not linked to the parent object in any way and are not parametric.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-28

Special copy options

In addition to the basic copy functionality, CADFEKO provides a number of special copy options.
These special copy options are available on the Transform tab and Copy special → Copy. . . from
the context menu.
The Copy and translate, Copy and rotate and Copy and mirror options allow the inclusion of a
transformation in the copy operation.
The Copy original operation can be used to retrieve faces and edges that have been deleted from
a part. The copy of a part generated using the Copy original operation will include all of these
removed faces and edges.

Collapse - removing the creation history of a part

CADFEKO stores the entire creation history of every part allowing the user to modify any point
in the creation history. While extremely powerful this requires a significant amount of memory
and in some cases also a significant amount of processing time.
For example, if a small component of a union operation is modified, CADFEKO needs to recreate
the parent parts in order to re-execute the union. Since these are not stored at every level, it
means constructing them again from the lowest level up. However, quite often a large part of
the model will never change. (It is, for example, unlikely that the model of a specific motor
vehicle will be modified to another model, but it is common to place different antennas on such
a vehicle.) Hence it should not be required to re-evaluate the vehicle each time a small part of
the geometry is changed.
CADFEKO provides the option to convert a part to a primitive using the Collapse tool. This
removes the entire creation history and stores the model as is. If this part is then used in other
operations it does not have to be re-evaluated since the ‘primitive’ part is available immediately.
This can save a significant amount of processing time and also reduce the possibility that edges /
faces or regions cannot be mapped and will become suspect (see section 4.4).
The Collapse operation can be undone, but once the model is saved and reloaded the creation
tree can only be accessed using a previously saved or archived copy (see section 3.4.5) of the
model which still contains the tree information.

Re-evaluating the geometry

CADFEKO uses advanced mapping algorithms to keep track of individual items when geometry is
modified. Since this information was not available in models created using CADFEKO 1, 1.1 or 2,
the user is given the option to re-evaluate the geometry tree in order to generate this information
when opening in a newer CADFEKO version. This re-evaluation option is automatically launched
when an older model is loaded for the first time in CADFEKO, but the user may initiate a re-
evaluation at any time by selecting Re-evaluate from the Transform tab.
Note that all the mapping information is not available in some models created in earlier CADFEKO
versions, thus it may not be possible to map all items during the re-evaluation. These items are
then marked suspect (see section 4.4). All properties (such as local mesh sizes) set on such

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-29

suspect items will be lost during re-evaluation. In addition, faces or edges that were deleted,
may re-appear if they cannot be successfully mapped. The user therefore needs to evaluate his
situation carefully. If the model geometry is likely to change in the future, it is recommended
that the geometry be re-evaluated and that all suspect items resolved before making any changes
to the model or setting any additional properties.
If the model geometry will not change, it may be better to keep the existing model and all its
settings. Note, however, that this model should not be changed in any way that requires re-
evaluation of the tree. For example if the model contains a motor vehicle with a monopole
antenna unioned to it, changing the length of the monopole would trigger an automatic re-
evaluation of the tree as re-execution of the union operation between the vehicle and the mod-
ified monopole is required to maintain the model consistency. (For memory reasons, CADFEKO
does not store the entire model in each level of the tree.) Complex geometry in a CADFEKO model
created in an early CADFEKO version that is not re-evaluated should in general be converted to
a primitive form (see section 4.4.4).
Depending on the complexity of the model, this operation may be a quite lengthy operation. It is
possible to cancel this operation, but the cancel process may also require a significant amount of
time.
It is possible that during geometry re-evaluation, faults will be identified in the model that were
not previously displayed. This is particularly true of models built in previous versions of CAD-
FEKO, or models containing imported geometry. These faults identified during re-evaluation may
provide additional information that may be useful in repairing a model.

4.5 Transformations

Geometry objects or mesh parts can be transformed by selecting them and selecting the transfrom
from the Transform tab. The Transforms entry in the details tree contains a list (in creation
order) of all the operations applied to each object. These may be edited (by double clicking
on the specific operation) or deleted (by pressing the <Del> key). A transform may be applied
to a selection of multiple objects, but it is in fact applied separately to each object. Therefore,
changing the transformation afterwards requires selecting and editing the transform of each
object separately. A transformation can, of course, be defined in terms of variables. In this
case, changing the variable value will modify all the items to which the transform was applied
accordingly.
The transforms now also make use of transforms which enable the application of the transforms
with arbitrary orientation.

4.5.1 Transformations on geometry parts

Rotate

The Rotate operation requires an origin, axis direction, workplane and a rotation angle (in de-
grees). The angle is measured in a mathematically positive sense around the axis. The dialog,
shown in Figure 4-14, also allows the modification of the workplane on the Workplane tab.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-30

Translate

The Translate dialog, shown in Figure 4-14, requires a start and end point from which it calculates
the translation distance and direction.

Mirror

The Mirror operation requires an origin and a plane. This is specified via the dialog shown in
Figure 4-14.

Scale

The Scale transformation requires an origin (set in the global workplane) and a scale factor as
shown in Figure 4-15. Scaling is done around the origin specified on the scaling dialog, not
around the centre or origin of the object or the model.

Figure 4-14: The rotate, translate and mirror dialog

Align

The Align dialog (shown in Figure 4-15) requires an origin and an axis direction for both the
source and destination workplane. This transformation may be used to easily place objects on
structures as it performs a rotate and a translate as a single transformation.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-31

Figure 4-15: The scale and align dialog

4.5.2 Transformations on mesh parts

Mesh parts can be transformed in a similar manner to geometry parts. The transform applies to
the entire part and may be edited or removed. If a vertex of a transformed mesh is edited, the
transformed coordinates are displayed and user input is assumed to be in the transformed space.
If the transform is modified, the effect of the modification is also applied to the edited vertex.

4.5.3 Multiple (transformed) copies

Parts (geometry or mesh) may be copied and transformed multiple times via a single command.
Select Copy special from the pop-up menu or Copy from the Transform tab. A special transform
dialog that contains a Number of copies field is launched. This will make the specified number
of copies of each selected part.
If a rotation is specified, the n th new part will be rotated through n times the specified rotation
angle. For translation, the n th new part will be translated by n times the specified translation
distance. For the mirror operation only one copy is allowed. (A second copy would be back on
the original and can be achieved with the standard Copy option.)

4.6 Cable modelling

Cables with arbitrary cable cross sections and shields may be included in the CADFEKO model.
These user defined cables may also be excited or loaded directly and a calculation may be re-
quested by adding a Cable harness request to the model.
The workflow for when defining cable modelling in CADFEKO is as follows:

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-32

• Define the cable type cross section (Single conductor, Ribbon, Specified coax, Predefined
coax and Cable bundle). The defined cable type will be shown in the contents tree under
Cables. If the cable is enclosed by a shield, a shield must be defined and its label added
to the cable cross section definition. The defined shield will be shown in the contents tree
under Shields.

• Define the cable path. The defined cable path will be shown in the contents tree under
Cables and indicated in blue in the 3D view.

• Request a cable harness calculation by defining the solution method, whether the cable is
irradiating/radiating and adding excitations and loads to the cable. The calculation request
will be shown in the contents tree under Solution→Requests.

4.6.1 Cable cross sections

The following cable cross sections may be defined, Single conductor, Ribbon, Specified coax,
Predefined coax and Cable bundle. These are available on the Construct tab, Cables group.

Figure 4-16: The Cable cross sections button

Figure 4-17: The Create single conductor and the Create ribbon cable dialog

Create single conductor: Core The core of the single conductor is specified by defining the metal of the
core and its radius.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-33

Insulation layer (coating) If the With insulation option is checked, the core
is covered by a coating with the medium given by Medium and the thick-
ness specified by Thickness.

Create ribbon cable: Topology The ribbon cable topology is specified by the Number of cores, Core
spacing (Centre to centre) and the Return path (core number).
Core The core of a single conductor is specified by defining the metal of the
core and its radius.
Core insulating layer (coating) A core insulating layer is specified by Medium
and Thickness.

Create coaxial cable: Core The core of the single conductor is specified by defining the metal of the
core and its radius.
Core insulating layers Core insulating layers may be added by clicking on
the Add button. The Core insulating layer is defined by the Medium and
the Thickness.
Shield The shield enclosing the cable.

Add a predefined coaxial cable: A predefined cable type can be selected. Note that predefined cables are
currently limited to frequencies from 10 kHz to 500 MHz.

Create bundle: Cables contained in bundle A cable bundle may be created containing arbi-
trary cable cross sections. Defined cable cross sections may be added to
the cable bundle by pressing on the Add button.
Shield type The following options are available Not shielded, embedded in
background medium; Not shielded, embedded in a dielectric and Shielded,
dielectric filled.
Insulation medium The insulation medium may be defined
Outer radius When the option Compute automatically is selected, the outer
radius of the cable bundle is automatically calculated.
Shield The shield enclosing the coaxial cable.

4.6.2 Cable shield

The following cable shields may be defined in CADFEKO, Braided (Kley), Solid (Schelkunhoff)
and Define shield properties. These are available on the Construct tab, Cables group

Figure 4-18: The Cable shield button

Braided (Kley)

Weave: The number of carriers in the weave and its weave angle must be specified by the user.

Insulation layer (coating): When the Apply coating option is selected, the cable is given a coating spec-
ified by Medium with a thickness specified by Thickness.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-34

Figure 4-19: The Create cable shield dialog

Braid-fixing materials: When the option Apply braid-fixing material is selected, the Inside braid-fixing
material may be specified. If a coating has not been specified, an Outside braid-fixing mate-
rial may also be specified.

Filaments: The number of filaments, its diameter and the metallic medium to be used for the filaments
are specified.

Solid (Schelkunhoff)

Solid shield: The Shield metal and the Thickness of the solid shield is required to be be specified by the
user.

Insulation layer (coating): When the Apply coating option is selected, the cable is given a coating spec-
ified by Medium with a thickness specified by Thickness.

Define shield properties

Solid shield: The Shield metal and the Thickness of the solid shield is required to be be specified by the
user.

Insulation layer (coating): When the Apply coating option is selected, the cable is given a coating spec-
ified by Medium with a thickness specified by Thickness.

Frequency dependent properties: The frequency dependent properties may be defined for the cable or
loaded from a file. See the SH card (see section 14.58) for an XML example regarding the
file format when importing cable properties from file.

4.6.3 Cable path

A cable path may be defined by specifying the points or importing the cable path from a file. To
define a cable path, select the Construct tab, Cables group.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-35

Figure 4-20: The Cable path button

Figure 4-21: The Create cable path dialog

Cable path tab

The path of the cable is specified as a series of straight line sections. Corner points can be
inserted or removed from the points list using the Add and Remove buttons. If point entry (see
section 3.2.2) is used to specify the last point in the list a new point is added automatically. Point
coordinates may also be imported from an external text file using the Import points from Nastran
file option (see section 4.3.6). The cable path may cross itself (or other cable analysis paths), but
there is no electrical contact at these locations.

Connector at start: The label to specify the start of the cable path. This connector label uniquely identi-
fies the connection of a cable path section.

Connector at end: The label to specify the end of the cable path. This connector label uniquely identifies
the connection of a cable path section.

Label: The label of the cable path.

Advanced tab

Sampling point density: The cable path section will be subdivided into small segments for the compu-
tation of the induced currents and voltages. The default sizes can be changed by selecting

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-36

Specify maximum separation distance and specifying a value for the Maximum separation
distance.

4.6.4 Requesting a cable harness request

A cable harness request may be included in the CADFEKO model to calculate the coupling into a
cable near a large conducting surface (see section 6.12.7).

Figure 4-22: The Cable harness button

The Cable tab

Figure 4-23: The Request cable harness dialog

Cable path and type: The labels of the defined cable path and the defined cable cross section type are
required to be specified.

Cable type: Irradiating When this option is selected, only the effect of external fields coupling onto a
cable will be considered
Radiating When this option is selected, the effect of the currents radiating in the cables
will also be considered

Solution method for outer cable problem: Multiconductor transmission line (MTL) When this option
is selected, the model will be solved with the multiconductor transmission line the-
ory which is also hybridised with the MoM or the MLFMM. The cable path should be
within λ5 of the conducting surface.
Method of moments (MoM), only for shielded cables When this option is selected the
model will be solved by means of the hybrid MoM/MTL solver. Any arbitrary cable
path may be defined.

Label: The label of the cable harness request.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-37

The Excitations and loads tab

Excitations and loads may be added to the cables configuration by clicking on the Add excitation
and Add load button on the Excitations and loads tab on the Request cable harness dialog (shown
in Figure 4-23).

Figure 4-24: The Create excitation and Create load dialogs

Connector: The connector label uniquely identifies the connection of a cable path section. These con-
nector labels are specified when a cable path is defined.

Positive/Negative pin: The positive and negative pins between which the excitation or load is to be
placed.

Adding excitations or loads to a cable requires knowledge of the conductor to cable connector
pin relation. See the AK card (see section 14.16) regarding the pin numbering.

4.7 Geometry validation

CADFEKO provides tools to perform basic consistency checks on the geometry. These tools may
be applied globally (to all geometry in the model) or locally (only to the current selection — not
available if the selection is empty).
Since properly connected meshes are only guaranteed for single parts, it is important to ensure
that different parts do not clash. Parts clash if there is any contact between them or if one part is
completely inside another. Select the Mesh tab and click on the Clashing geometry button.

Figure 4-25: The Clashing geometry button

Any parts containing clashing geometry will be selected and listed in the message window. Find
→ Find in whole model → Unconnected geometry edges and Find → Find in selection → Uncon-
nected geometry edges finds and selects all the edges that are free (not attached to any face) or

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-38

attached to only one face within the indicated scope. If this operation selects an edge that seems
to be connected to multiple faces, it means that there is more than one edge at this location and
that the faces do not make electrical contact. This usually indicates a problem in the model.

4.8 Creating meshes

A mesh is a discretised representation of a geometry model. All geometry that is used during
a simulation must first be meshed before a solution can be obtained. The accuracy of results
depend greatly on generating a mesh that is appropriate for the problem being solved and as
such, generating a correct mesh is very important.

Figure 4-26: The Create mesh button

The mesh dialog can be opened by clicking on the button on the Mesh tab on the Create mesh
button (see Figure 4-26) or by pressing <Ctrl><M>. If all of the mesh settings have already
been verified, the user can also run a Quick mesh that skips the dialog and performs the mesh
without further dialogs or prompts. To run a Quick mesh, press <Ctrl><Shift><M>.

Figure 4-27: The Create mesh dialog

Figure 4-27 shows the Create mesh dialog. The various mesh settings can be set, including what
must be meshed and how finely it must be meshed. Advanced features such as ignoring geometry
and curvature refinement settings can be found on the Advanced tab. It is possible to indicate
if all parts or only the currently selected parts in the geometry should be meshed. Unlike other
dialogs, the selection can be changed while the dialog is open. This allows, for example, meshing
different parts with different mesh sizes without closing the dialog.
The Wire segment radius field specifies the radius which is applied to all wires for which no local
radius is specified. This value must be specified when wires are present irrespective of any other
mesh settings.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-39

4.8.1 Automatic meshing

Automatic meshing takes several model properties into account and generates a mesh that is ap-
propriate for that configuration. Some local refinement may still be necessary (see section 4.8.4),
but in most cases the automatic meshes will give a reasonable mesh. Meshes are generally built
relative to the wavelength of an electromagnetic wave in the medium of propagation. Each
solution method has different requirements and sometimes the model itself will influence the
accuracy of results. The following model properties are considered:

Frequency: The shortest wavelength corresponds to the highest simulation frequency. Note then that the
frequency range for the simulation must be set before automatic meshes can be generated.

Solution method: Depending on the solution method being used to solve the problem, different mesh
requirements may be needed. For example, a FEM model requires settings for tetrahedra,
a MoM solution requires settings for triangles and wires, while a hybrid solution needs to
take both into account.

Dielectric properties: The dielectric properties of the media in the model will affect the propagation
speed of an electromagnetic wave in a medium, which will in turn affect the wavelength.
Dielectric media are taken into account in all cases except in the case where infinite layers
are being used. In these cases, local refinement must be applied.

Geometry curvature: Even in cases where a finer mesh isn’t required for accurate solution results, it may
still be required to accurately model aspects of the geometry. Automatic meshes will attempt
to reasonably conform to the original geometry, the settings of which can be modified on
the Advanced tab.

Local mesh refinement settings may still be applied to individual components in a model. Any
local mesh refinement settings will be respected, meaning that a user’s setting will never be
overwritten.

Wires

For wires, the wavelength (λ) is determined based on the maximum simulation frequency and
the surrounding medium.

Table 4-4: Automatic meshing for wires.

Type Fine Normal Coarse

1 1 1
30
λ 20
λ 12
λ

Faces and edges

If any of the bounding regions have a user-defined local mesh size, then the face will inherit the
smallest of these local mesh sizes. If a face is the bounding face of a FEM region, then the first
order automatic mesh size of that region will be used. If a face is the bounding face of a VEP
region, then the mesh size of that region will be used. In all other cases the smallest wavelength
for the bounding media is determined.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-40

Table 4-5: Automatic meshing for faces and edges.

Type Fine Normal Coarse

1 1 1
Method of Moments (MoM) 16
λ 12
λ 8
λ
1 1 1
Physical Optics (PO) 10
λ 8 6
λ
Large Element PO (LE-PO) λ λ λ
Geometric Optics (GO) λ λ λ

Regions

For regions, the wavelength (λ) is determined based on the maximum simulation frequency and
the medium properties of the region. The specified lengths will be applied to the tetrahedral edge
lengths.

Table 4-6: Automatic meshing for regions.

Type Fine Normal Coarse

1 1 1
st
Finite Element Method (1 order FEM) 15
λ 10
λ 6
λ
1 1 1
Finite Element Method (2nd order FEM) 12
λ 8
λ 6
λ
1 1 1
Volume Equivalence Principle (VEP) 16
λ 12
λ 8
λ

4.8.2 Custom meshing

For finer control over the simulation mesh, the lengths of the edges of triangles and tetrahedra
as well as the lengths of wire segments can be specified. This type of control can be used if the
automatic meshing does not provide an appropriate mesh.

Triangle edge length: The Triangle edge length field specifies the default mesh size for the edges of
triangles. Note that some edges may be as much as 30% longer than the specified length.
(See section 4.11.1.)

Wire segment length: The Wire segment length field specifies the maximum length for wire segments.
All edges that do not form the boundary of a face are assumed to be conducting wires and
meshed into segments.

Tetrahedral edge length: The Tetrahedral edge length field specifies the default mesh size for the edges
of tetrahedral. Note that some edges may be as much as 30% longer than the specified
length. (See section 4.11.1.)

4.8.3 Advanced meshing options

The following options are available on the Advanced tab of the Create mesh dialog (shown in
Figure 4-27).

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-41

Hadling of small geometry features: This group allows special treatment of small geometry details. The
Geometry smaller than field specifies the limit of what is considered a small feature as a
percentage of the size of the part that it belongs to.
The Default behaviour is to mesh these structures normally. This will result in an accurate
representation of the geometry, potentially using very small elements. Optimise is useful
where the geometry has long narrow slivers or faces that are close together. If this option
is selected, CADFEKO will try to improve the mesh quality of small features. Finally, it
is possible to Ignore small features. Small details are then ignored at a possible cost of
accurate geometry representation. This option may at times allow the meshing of faces that
otherwise cannot be meshed with the default settings. Note that ignoring small features
does not work for closed edges. Such edges can, however, be divided by imprinting points
(see section 4.4.2).

Enable mesh smoothing: If Enable mesh smoothing is checked, an additional smoothing algorithm is
applied. This will result in a better quality mesh, but will increase meshing time.

Mesh quality: The Mesh size growth rate controls how quickly the mesh size changes. Fast allows an
abrupt jump from small to large elements, while for Slow each triangle will not be more
than twice the size of any other triangle that it is connected to.

Curved geometry approximation settings: The Refinement factor indicates how closely a mesh must con-
form to the geometry. For geometries with many small details, a setting of Fine will require
smaller triangles which will reduce simulation performance and perhaps even give errors.
To limit the size of these elements, the Minimum element size can be set as a percentage of
the average edge length on that part.
The option to allow elongated triangles means that the resulting mesh will make use of
long, thin triangles where required. This could lead to a reduction in the number of mesh
elements, depending on the type of geometry being meshed.

In general, the mesh size is not allowed to be smaller than the maximum coordinate divided by
1×108 . This is the limit of the numerical precision of the geometry. For very small models (which
require correspondingly small mesh sizes), the geometry extents (see section 4.1.2) should be
decreased. The mesh size is also not allowed to be larger than four times the extents. (For more
information regarding the required and suggested mesh sizes for different mesh elements, see
section 2.2.2.)

4.8.4 Specifying local mesh sizes

Sometimes an accurate solution requires a very fine mesh in some part of the model. Additionally,
when meshing curved structures, a finer mesh may be required for an accurate representation of
the geometry. Rather than mesh the entire model at this size, the user may specify a local mesh
size for these parts.
The sizes specified on the Create mesh dialog are used on all items which do not have a local
mesh size.
For regions, faces or edges, local mesh parameters can be specified in the Mesh size group on the
respective Properties dialogs. This group contains a Local mesh size field which must be checked
if a local mesh size is to be set. An icon will be displayed next to all items in the details tree on
which local mesh parameters have been set to indicate this.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-42

If a region that has a local mesh size set on it is meshed into tetrahedra, that size will also be
applicable to the bordering faces. An additional finer mesh size can, of course, be specified on
these faces. (The minimum of all applicable local mesh sizes are used.) Likewise, setting a mesh
size on a face also sets that size on its bounding edges. If a finer mesh size is specified on, for
example, an edge of a face, then the triangles of the face will adhere to this length along the
edge, even though the rest of the face may have a much larger mesh size.
The final mesh size on any item is the minimum of all local sizes applicable to it. As an extreme
example, if an edge borders a face which is the boundary of a region and all three items have
a local mesh size, the final mesh size on the edge is the minimum of all three these sizes. On
faces, the local mesh size may be larger than the global mesh size. Note, however, that if no local
mesh size is specified on the bordering edges, the global mesh size applies to them. The triangles
on the edge of the surface will then have the global mesh size. For regions, the local mesh size
is only used when meshing the region into tetrahedra. If the local mesh size on a region is set,
then volume meshing of that region still will occur at the smaller of the specified local and global
mesh sizes.
For wires, a local wire radius may be specified on the Edge properties dialog. Note that wires are
displayed as lines in the 3D view and that the radius is not represented. The radius for meshed
segments are displayed by default . An icon next to the edge in the details tree indicates that a
local wire radius has been set.

Multiple selections

When launched, the properties dialog shows the current state of the properties of the selected
items. If some selected items have a local mesh size and others not, or if the specified size is
different, the Local mesh size item is in a ternary (i.e. not defined) state. (Thus the field is
referred to as a tri-state field.) If this is the case, the parameters of the different elements in the
selection are not updated when clicking OK or Apply. This field can, however, still be checked or
unchecked to set (override) the properties on all selected items. Note that it is possible to cycle
through all three states by clicking on the field.

4.8.5 Editing mesh vertices

Figure 4-28: The Mesh vertex selection button

A mesh vertex can be modified by selecting the vertex on the unlinked simulation mesh (mesh
vertex must be visible and mesh vertex selection on). Right-click and select Properties from the
popup menu. This opens the Modify vertex dialog which contains the coordinates of the vertex
(in global coordinates). The input fields accept point entry (see section 3.2.2) and expressions,
but these are converted to numerical values once the dialog is applied and closed. The vertex
point can be moved by a given amount (rather than to a specific position) by adding the required
offset to each field. For example, changing X from 1.2 to 1.2+0.5 moves the point 0.5 units in
the x direction.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-43

4.8.6 Creating mesh triangles

Sometimes manual mesh fixing is required. This may be because the mesh contains a number
of holes or because some bad elements were selected and deleted. Unlike deleting vertices (see
section 4.10.4), deleting elements leaves a hole in the mesh surface.
For this reason, mesh triangles can be manually constructed by selecting the Mesh tab and click-
ing on the Create triangle button.

Figure 4-29: The Create triangle button

This operation is only allowed if the selection contains a single mesh label (in which case the new
element is added to this label), or a single mesh part (in which case the new element is added to
a new label created under this part). The Create mesh triangle dialog allows specification of the
three corners of the triangle. Since these three fields are standard point entry (see section 3.2.2)
fields, all the different snap options (see section 3.2.1) are available.
Manually constructed triangles may be “moved” into an existing label. Select all the elements of
the existing label as well as all the new elements. (If the elements were selected by label, click
the Select mesh element button to convert the selection to individual elements.) Now right click
on the selected elements and select Rename. This will merge all the selected elements to one
label which can then be renamed. If all elements of a label are renamed, that label is removed —
the original label name can then be used for the newly created label. Alternatively, mesh labels
may be selected, and the Merge tool applied from the right-click menu. This will merge all of the
selected mesh elements into a single mesh label within the same mesh part.

4.8.7 Reversing normals

Some features in FEKO make use of the normal vectors of elements. For example, dielectrics are
specified by giving the medium on each side of the bordering triangles. For this purpose, the
normal side of a triangle or polygon is defined as the side towards which the normal vector is
pointing. (If the start point of the normal vector is located on the triangle, the normal side is
the one closest to the arrowhead.) In CADFEKO the normal sides of mesh elements are coloured
blue, while the backs (away from the normal) are brown when the Face normal medium display
option is selected (Display options, Colour menu button).

Normal side
Back

Figure 4-30: The normal and back sides of a triangle

When the selection contains triangles or UTD polygonal plates (or labels or parts that contain
these elements), selecting the Transform tab and clicking on the Reverse normals button will
invert the normals of all the selected triangles / polygons.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-44

Figure 4-31: The Reverse normals button

Normal vectors can be shown — as lines with a red tip — per 3D view by activating the Show /
hide normals button on the 3D view toolbar.
Reversing the normals of mesh triangles on dielectric boundaries can adversely affect the consis-
tency of the model. CADFEKO tries to keep the model consistent. For example, if the normals of
all triangles in a given label are reversed, CADFEKO swaps the dielectric media on the two sides.
However, if only some triangles are modified, CADFEKO cannot ensure consistency and it is up
to the user to ensure that the model remains consistent.

4.9 Mesh refinement

CADFEKO provides the functionality for mesh refinement around a point and a polyline. Note
that display of these meshing rules may be enabled/disabled by clicking on the Display options
context tab and clicking on the Meshing rules button.

Figure 4-32: The Meshing rule button

4.9.1 Point refinement

If a point refinement is required in the mesh, select the Mesh tab and click on the Point refinement
button.

Figure 4-33: The Point refinement button

4.9.2 Polyline refinement

If polyline refinement is required in the mesh, select the Mesh tab and click on the Point refine-
ment button.

Figure 4-34: The Polyline refinement button

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-45

4.10 Mesh manipulation

4.10.1 Mesh element radius

If the selection contains individual segments or mesh labels containing segments, the segment
radii can be changed by right clicking on the selection and selecting Properties. If the selected
items have different values, the Radius field will initially be blank. This is the ternary state —
other items on this dialog can then be changed without modifying the radii of all the selected
elements. If a value is entered in this field, the radius of all selected items will be set to this value.
Mesh segments can be displayed with their actual radius in the individual 3D views. Click the
down arrow next to the Show/hide mesh icon on the 3D view toolbar (see section 3.2.4) and
ensure that Show segment radii is checked. (Selecting this item when it is checked will uncheck
it.) A representation of coatings that have been applied to mesh segments can also be shown by
activating the Show segment coating option.
When changing the mesh viewing mode to the Colour by face medium mode (see section 3.2.3),
the radii of all visible mesh segments is automatically shown (the Show segment radii option is
automatically activated). This display may be manually deactivated. In a similar manner, when
changing the mesh viewing mode to the Colour by coating mode, the coating of all visible mesh
segments is automatically shown.

4.10.2 Relabelling mesh elements

Any selection of mesh elements can be relabelled provided they are of the same type and belong
to the same mesh part. Since the full label is constructed as Assembly.Part.Label, it is not possible
to relabel elements on different parts. Note that the individual elements represented in the 3D
view, not the labels in the tree, must be selected. (If the selection is done by mesh part or mesh
label, switching to element selection automatically selects all elements belonging to the selected
parts / label — see section 3.2.2). If the original elements did not have the same settings (media,
using PO, etc.) the new label is marked suspect as CADFEKO cannot automatically determine
which of the settings to use.
Multiple mesh labels labels can also be merged (grouped under the same name) by simply select-
ing the mesh elements in the details tree and clicking on Merge in the right-click pop-up menu.
Merging of elements is also only possible within in a single mesh part.
Labels are used to reference specific elements when setting certain properties. To specify unique
properties on an element or collection of elements, those elements need to be relabelled. As an
example, consider a simple wire monopole attached to a plate, created as Union1. The resulting
mesh contains two labels: Face1 for the triangles on the plate and Wire5 for the segments on
the monopole. If, for example, only the bottom half of the dipole is coated with a dielectric,
only these segments must be relabelled before applying the coating. (A better approach to this
example would be to split the original wire — by imprinting a point — and apply the coating to
the resulting lower edge.)
When the last element with any label is removed (deleted or relabelled), the label is removed.
Thus if a selection containing all elements from more than one label is renamed, the total number
of labels is reduced.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-46

Care must be taken when remeshing parts where the mesh has been relabelled. Since the entire
mesh part is replaced by the new mesh, the label created by relabelling the elements is lost. If
the example above is remeshed, there will again be only the labels Face1 and Wire5. Therefore,
relabelling elements should be delayed till as late as possible in the model preparation process.

4.10.3 Merging coincident vertices

Connectivity in FEKO relies on vertices of adjacent mesh elements being within a small tolerance
of each other. CADFEKO allows merging all vertices within a user-specified tolerance. To do this,
select one or more mesh parts and select the Mesh tab and click on Merge vertices to open the
Merge coincident vertices dialog.

Figure 4-35: The Merge vertices button

Here a Tolerance can be specified. Any two points separated by less than this distance are then
merged to the coordinates of one of the original vertices (as opposed to taking the average
position) resulting in a single verex.
If the Snap to geometry points or Snap to named points fields are checked, mesh vertices lying
within the specified tolerance of these points will be snapped to them. For example, if a named
point lies between two mesh vertices that are less than the specified tolerance from each other
they will be merged at this point. If the snap options are checked, vertices lying within the
tolerance from geometry or named points will be included in the number of merged vertices
listed in the message window — even if they originally coincided exactly with the points.
Merging points can lead to degenerate (collapsed) triangles. CADFEKO tries to avoid this by
giving a warning if the tolerance is large relative to the mesh sizes. It also automatically removes
all degenerate elements after merging the vertices.

4.10.4 Merging elements (removing vertices)

The automatic mesh removal is meant primarily to merge vertices that should be the same, but
differ due to numeric tolerances. If the points are further apart, it may be required to control
which point is moved and which point remains fixed. It is also often required to remove specific
elements where the distance between points may be large compared to the triangle size elsewhere
in the model.
This is done by deleting vertices. This is a two-step process, similar to the subtract operation. First
select all the vertices which must be removed and press <Del>. In the second step, a final vertex
must be selected when prompted. All the originally selected vertices are then merged with the
final vertex. CADFEKO automatically removes all elements that become collapsed (degenerate)
during this operation.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-47

4.10.5 Merging mesh parts (union for meshes)

When mesh parts have been imported as separate mesh parts, it is not possible investigate mesh
connectivity or add edge port between these parts. Meshes can be merged into a single mesh by
selecting the mesh parts in the model tree and selecting the Merge option from the context menu
(right-click on the mesh parts) or merge the meshes by selecting the Mesh tab and clicking on
Merge meshes.

Figure 4-36: The Merge meshes button

4.10.6 Remove duplicate and collapsed elements

Duplicate elements

Duplicate elements within a mesh part can be deleted automatically. Select the required mesh
parts and select the Mesh tab and click on the Remove duplicates button.(Generally duplicate
elements should only occur in imported meshes or where CADFEKO meshes were edited manu-
ally.)

Figure 4-37: The Remove duplicates button

If duplicate elements have the same label, CADFEKO deletes all but one. If the elements do not
have the same label, it may be important to delete a specific element. This is controlled with the
Remove duplicate mesh elements dialog shown in Figure 4-38. The items are ordered by pressing
<Ctrl> and dragging the numbers in the left-hand column. For each set of duplicates an element
whose label is highest on this list is retained and all others are deleted. (There could be more
than two identical elements.)

Figure 4-38: The remove duplicate mesh elements dialog.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-48

Collapsed elements

Selecting the Mesh tab and clicking on the Remove collapsed button will delete all degenerate
elements (elements where two or more nodes coincide).

Figure 4-39: The Remove collapsed button

4.11 Mesh validation

4.11.1 Mesh information

Selecting Mesh tab and clicking on the Info button shows a histogram of the edge length distri-
bution (as well as the average edge length and the standard deviation of the edge lengths) and
the number of the different elements in the current mesh selection. (The selection may contain
mesh parts, labels or elements.)

Figure 4-40: The Info button

The mesh info can also be displayed by right-clicking on Meshes in the tree and selecting Info
from the pop-up menu. Alternatively, the mesh info can be displayed by double clicking on
the meshed model in the tree. The spread gives an indication of the quality of the mesh and
shows how many edges are longer than a desired length. The numbers of the different elements
determine the problem size and influence the required run time.

4.11.2 Finding problems in the mesh

Since the accuracy of the FEKO results depends on the quality of the mesh, CADFEKO provides
some mesh validation tools. These tools may be applied globally (to all mesh parts in the model)
or locally (only to the currently selected mesh parts - only available if the current selection is not
empty).

4.11.3 View Distorted and Oversized mesh elements

Selecting the Mesh model and clicking on the Distorted elements button checks the current se-
lection (mesh parts, labels or elements) for distorted elements.
After the operation is applied, the identified inappropriate elements are selected and listed in the
message window. The selection Undo and Redo operations can be used to see which elements
were selected before and after the test.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-49

Figure 4-41: The Distorted elements button

Distorted mesh elements are specified in terms of the minimum internal angle. In an ideal mesh
all internal angles are 60◦ (all angles in CADFEKO — except the arguments of trigonometric
functions used in expressions — are in degrees). If any of the three angles in a mesh element
are very small, the element is a sliver. Such elements can be removed by deleting vertices (see
section 4.10.4).

Figure 4-42: The Oversized elements button

Oversized elements are found based on the maximum edge length.

4.11.4 View free mesh edges and segment nodes

Another very important issue regarding mesh quality is connectivity. In addition to the tests on
the geometry, CADFEKO provides tools for finding unconnected edges and segment vertices. This
is done by selecting Find → Find in whole model/Find in selection → Free mesh triangle edges
or Find →Find in whole model/Find in selection → Free segment nodes from the main menu.
These operations test only the selected elements, and will not report free edges / vertices if the
element is connected to another unselected element.
Triangle edges are considered to be free if they are not attached to other triangles. (An edge is
only connected to a triangle if both end points coincide with corner points of the triangle.) Here
the two vertices at the ends of the edge are selected. These vertices may then be merged, moved
or modified as required to make the model consistent. Again, free edges may be perfectly normal
such as on the edge of a finite boundary.
Using the Find → Find in whole model/Find in selection → Free segment nodes tool, all vertices
which are not attached to another segment or a triangle corner within the indicated scope are
selected. These vertices may be valid, for example, at the ends of wires. However, if two such
points lie very close together, they should be merged.

4.11.5 Find intersecting mesh triangles

Intersecting mesh triangles indicates all triangles that intersect, but are not connected. This is
useful for identifying areas of poor mesh quality - particularly in imported meshes. This tool
highlights all mesh elements that overlap (entirely or partially) or cut each-other and are not
connected at the point of intersection. After the tool is applied, all parts that contain intersecting
mesh elements are highlighted in the tree, and the individual problematic mesh elements are
highlighted (shown in yellow) in the 3D view. The selection may be deleted, but typically each
intersection should be considered and repaired according to the intended mesh representation.

July 2011 FEKO User’s Manual

CREATING AND MESHING GEOMETRY IN CADFEKO 4-50

Figure 4-43: The Intersecting triangles button

July 2011 FEKO User’s Manual

IMPORTING AND EXPORTING 5-1

5 Importing and exporting

5.1 Importing and exporting geometry

There are a number of ways to import and export geometry in CADFEKO. CADFEKO can import
geometry from other CADFEKO models, Parasolid and other CAD formats. Importing and ex-
porting CADFEKO and Parasolid models do not require any special licences to be activated, but
translation to and from other formats does require extra licences.

5.1.1 Importing CADFEKO models

CADFEKO is able to import other CADFEKO models. The import allows the user to select whether
geometry, meshes, solution entities or optimisation settings should be included in the import
process or not. Click on the Application menu button and hover over the Import menu anchor.

From the Application menu select Import → CADFEKO model (*.cfx).

Figure 5-1: The import CADFEKO model dialog.

Variables, named points and media are always imported. An option is provided to merge variables
and media that are shared between the model being imported and the destination model (i.e.
variables and named points with the same name and value in both models). If there are naming
conflicts between names of imported entities and existing entities in the model (variables, named
points, geometry, meshes or solution entities), an error message will list the offending entities
and the import will be terminated. This problem can be overcome by adding an optional prefix
(in the Prefix field) that will be pre-pended to all imported entity names in order to ensure
uniqueness. When variables or media are merged, variables and media with the same name and
the same value result in a single instance of that variable or media after importing.

July 2011 FEKO User’s Manual

IMPORTING AND EXPORTING 5-2

It should be noted that the frequency, infinite planes and various other settings (e.g. meshing
settings in the Create meshing dialog) are not considered during the import process and will not
affect the destination model in any way.

5.1.2 Working with Parasolid

CADFEKO is based on the Parasolid solid modelling kernel and therefore models can be imported
and exported from/to the native Parasolid format without any translation.

Import of Parasolid models

Parasolid models can also be imported into CADFEKO. Click on the Application menu button and
hover over the Import menu anchor. From the Application menu, select Geometry → Parasolid.
This asks for a file name and then opens an Import Parasolid model dialog where the scale factor
to be applied during importing can be modified. This Scale by factor is the factor by which the
Parasolid model is scaled before it is imported. If the Parasolid model was created by exporting
from CADFEKO and the *.info file is available, the dialog will display the value in the file.
This value can then be changed. For example, if an antenna model was constructed in mm and
is to be imported into an automobile model constructed in m, a scale factor of 0.001 would
be appropriate (independent of any scale factor required by the definition of the model extents).
Note that importing a large model with a scale factor other than unity may be a very slow process.
Isolated vertices (acorns) in a Parasolid model are not the same as named points (see sec-
tion 4.2.2) in CADFEKO. These vertices are not imported, but CADFEKO does write the coor-
dinates to the message window during the import process. If required, the relevant named points
can then be created manually.

Export of Parasolid models

Click on the Application menu button and hover over the Export menu anchor. From the Appli-
cation menu select Geometry →Parasolid. This opens the Export Parasolid model dialog, where
the option of Text or Binary format can be set and the Parasolid version to export to can be cho-
sen. Once the export options have been chosen, a dialog prompting for the name of the file to
export to will be launched. CADFEKO exports the latest Parasolid version by default, but older
versions of Parasolid can also be selected. Only the final geometry is exported. Exporting and
importing the same model will result in the loss of the entire creation tree (this is similar to the
make primitive (see section 4.4.4) operation).
Parasolid models are inherently limited to a 1000x1000x1000 unit box centred at the origin.
CADFEKO introduces a scaling factor to make this more flexible (see section 4.1.2). The Scale
factor field on the Export Parasolid model dialog displays the factor by which the CADFEKO
model must be scaled during export to convert it to correct units required in the Parasolid model.
A scale factor of 0.1 implies that the dimensions of the saved Parasolid model are one tenth of
the native dimensions as set in CADFEKO. Typically, programs that import Parasolid models allow
specifying a factor by which the Parasolid model must be scaled during the import. In order to
maintain the correct units and scale, this factor should then be the inverse of the scale factor
used in the export of the model from CADFEKO.

July 2011 FEKO User’s Manual

IMPORTING AND EXPORTING 5-3

For large models (larger than 500 of the current CADFEKO units), the extents must be increased.
If, however, the model is smaller than 50 CADFEKO units, the extents should be decreased. In
general, changing the model extents is not recommended (unless the model is very small and
precision or geometric accuracy problems are encountered). Using the default extents results
in an unscaled Parasolid model, and it is not necessary to keep track of the scale factor during
model import/export.

5.1.3 General CAD formats

CADFEKO can also import other CAD formats using the Spatial InterOp translators. Currently
ACIS, CATIA V4, CATIA V5, AutoCAD, IGES, Parasolid, Pro/ENGINEER, STEP and Unigraphics
are supported import formats. The following are supported CAD export formats using the Spatial
InterOp translators: ACIS, CATIA V4, CATIA V5, IGES, Parasolid and STEP. AutoCAD DXF files
can also be imported (some advanced DXF options are not supported).
Note that all CAD import translator modules are licenced separately. If some import/export
menu items are disabled, these modules are not activated in the licence. If you wish to evaluate
or purchase any of these import/export modules, please contact your distributor.

Importing general CAD

From the Application menu, select Geometry. If an import menu item is selected, CADFEKO
requests the source file name and then opens the Import . . . model dialog. The Units group on
this dialog displays the units of the imported model and allows setting the Destination units. The
destination unit determines a scale factor and may be set different to the unit of the current model
to realise the required scaling effect. If CADFEKO cannot determine the unit of the imported
model it assumes a default value. It may be required to select a different unit to make the
imported model fit inside the Parasolid size limit discussed regarding the import of Parasolid
geometries.
Since all imported CAD models are converted to a Parasolid format during the import process,
all the comments for Parasolid imports also apply to imports from general CAD formats. Due to
the differences in the internal representations used by different CAD formats, format conversion
can have unexpected results. In particular, adjoining surfaces sometimes do not line up correctly
due to tolerance differences and models that use a numerical representations can cause faults
during scaling. Though CADFEKO does provide some basic diagnosis tools, no advanced CAD
fixing functionality is available natively in CADFEKO.
All model format conversions performed during a CAD import is logged in the file CADimport.log
in the directory %FEKO_USER_HOME%/logs. The information in this file is useful in cases where
the import conversion fails.

Importing AutoCAD DXF

AutoCAD DXF file can also be imported (no export). After selecting the file to import, a dialog
displays the contents of the file and various other options. It is possible to select only selective
components from the file by unchecking the check box of the unwanted geometry. The Advanced

July 2011 FEKO User’s Manual

IMPORTING AND EXPORTING 5-4

Figure 5-2: AutoCAD DXF import dialog (Contents and Advanced tab)

tab contains special options to extrude, to stitch faces together and the auto-merging of wires
during the import process.
Note that not all DXF entities are currently supported. The supported entities are:

• 3D face

• Arc

• Circle

• Ellipse

• Line

• Polyline and polyface mesh (3D)

Entities that are not supported are:

• Point

• Spline

• 3D solid

• Trace

• Hatch

• Text

• Dimensional annotations

July 2011 FEKO User’s Manual

IMPORTING AND EXPORTING 5-5

Exporting general CAD formats

Click on the Application menu and hover over the Export menu anchor. From the Application
menu select Geometry. If an export format is selected, the options and information available for
the particular format is displayed. Clicking on OK allows the user to select the file to export to.
Scaling is very important when translating between the different CAD file formats (ACIS, Catia,
IGES, Pro Engineer, STEP, Unigraphics, Parasolid) and this is often the source of many importing
errors. During an export a dialog displays the difference in the number of units displayed in
CADFEKO and the number of units that would be seen in the program where the file is imported.
The difference in the number of units is caused by the fact that CADFEKO does not perform
any scaling during the export (scaling could cause tolerance errors during the subsequent import
process). During an import, CADFEKO also displays the scaling that will be applied during the
import process - the scaling can be changed by changing the CADFEKO model unit or geometry
extents. If a model does not import as expected the scaling should be varied in an attempt to
import the geometry correctly.

5.1.4 Faults on geometry

It is important that the imported model is of a decent quality. If the imported model has overlap-
ping surfaces and/or gaps between supposedly connected surfaces, these will not be corrected
during the import process and the resulting CADFEKO model will also be of poor quality. Refining
such models may be a considerable task after importing. CADFEKO will try to import geometry
with faults and also indicate what the fault is. (A tool-tip with details regarding the fault will be
shown if the mouse pointer is held over the fault icon.) The user should consider each fault and
try to correct the model as far as possible. In order to locate specific faults in the geometry, the
entity of interest may be selected in the tree or in the details tree, and the zoom to selection tool
(see section 3.2.1) used. The Transparency mode under the geometry view options may also be
activated to allow visual location of the selected entity in the 3D view.
In certain cases it is possible to use the model without correcting a geometry fault, but this is
not advised. A list of possible faults and solutions are given in CADFEKO geometry faults (see
section 27).

5.2 Importing meshes

CADFEKO can import a variety of mesh formats. The import tools are accessed by clicking on the
Import menu anchor. From the Application menu select Mesh → . . .

5.2.1 Importing *.fek files

When importing *.fek files, only the mesh parts (wire segments, triangles, polygonal plates and
tetrahedra) are imported; information regarding the solution setup is completely ignored. As
CADFEKO specifies dielectric medium boundaries based on face labels, it appends an underscore
and a number to create unique labels if the *.fek file contains different dielectric boundaries
with the same label. In addition, since the imported labels are listed under a new mesh part, the
labels in the imported model contain an additional level. Medium information and segment radii
are retained during the import.

July 2011 FEKO User’s Manual

IMPORTING AND EXPORTING 5-6

5.2.2 Importing general mesh formats

CADFEKO imports other mesh formats by running PREFEKO and importing the resulting *.fek
file. Currently NEC, FEMAP neutral files (boundary surfaces bordered with line curves are im-
ported as polygonal plates), NASTRAN, meshed AutoCAD DXF files (only the LINE and POLY-
LINE structures which defines segments and triangles), STL, PATRAN, Ansys CBD, Concept files,
ABAQUS meshes, GiD meshes (for GiD meshes, hexahedra elements are ignored) and a spe-
cial ASCII format are supported. (For more detail on the ASCII format, see the IN card (see
section 13.22).
For all formats (except *.fek files), CADFEKO opens an additional Import mesh dialog. Here
the user can set a scale factor and control which elements to import. (For example, boundary
elements in FEMAP neutral files may be either part of the geometry — i.e. they should not be
imported — or intended as polygonal plates — i.e. they must be imported.) Quadrangles are
divided into triangles during the import process. Only Ansys CBD files support segment radius
information — for all other formats a default radius must be specified on the Import mesh dialog.
(The default radius is also used for Ansys segments where a radius is not defined in the file.)
Different radii can be set on the segments after importing. Finally, for meshed AutoCAD DXF
files, the LINE elements are divided in segments according to the value of the Segment length
field. If the LINE elements must not be subdivided, this value must be larger than the longest
line. Since these formats do not support specifying dielectric media, all segments, triangles and
polygonal plates are PEC structures in free space and all tetrahedra have the medium Unknown.
It is also possible to import meshes ‘manually’ by constructing a simple *.pre file that imports
the mesh in EDITFEKO, running PREFEKO and then importing the *.fek file in CADFEKO. This
approach provides more control over the import process, and gives access to some of the more
advanced import options, for example to only include certain labels.

5.3 Exporting meshes

CADFEKO is able to export NASTRAN mesh formats. It is exported directly from CADFEKO and
does not make use of PREFEKO.

5.4 Printing and image export

CADFEKO provides a tool to export the current 3D view to various image formats (*.jpg, *.bmp,
*.xpm, *.png and *.eps), as well as the direct printing of the current 3D view to a printer.
Image export may be accessed through the Image export option on the Tools tab ribbon. Printing
of images may be accessed through the Print option on the application menu.
The resolution of exported images may be specified and the exported image will be saved in the
chosen location. All printed and exported images will contain the current active 3D view, with
the current view-direction, zoom factor, colouring and visibility settings.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-1

6 Defining the solution configuration in CADFEKO

It is possible to completely define models and request solution results from within CADFEKO.
A standard model includes: a physical model; ports that define entry / exit points for energy;
excitations on the ports that define the behaviour of the energy; and a mesh that represents all
of these components. The creation of these components is discussed, along with instructions on
how to request results for later processing in POSTFEKO. It is also possible to run the FEKO solver
directly from CADFEKO and to edit CADFEKO models using EDITFEKO.

6.1 Using dielectric and lossy media

Media (representing different materials e.g. metals, dielectrics etc.) must be defined before they
can be used in the model definition.

Default media

The first three entries in the Media tree are Perfect electric conductor, Perfect magnetic conductor
(currently only used when setting magnetic ground planes) and Free space. The properties of the
Perfect electric conductor and Perfect magnetic conductor cannot be edited nor deleted, but the
properties of Free space may be edited. This provides for simulation where the free space media
is not a vacuum.

User-defined media

All user defined media/materials are listed under Media in the contents tree. Though each of the
different media types is listed separately, the medium names must be globally unique and contain
less than 43 characters. The coloured square next to each medium entry indicates which colour
is used to represent this medium in the 3D display (when the Colour by region medium, Colour
by face medium or Colour by coating medium colouring options are selected (see section 3.2.3))
and in the details tree (see section 3.3.2). The display colour for a medium can be changed
by selecting Change display colour from the medium’s pop-up menu. (The selected colours are
saved as part of the model in the *.cfx file and when the model is opened, the colours will
therefore be maintained, even if the session file is not available.)
All media (except the default media — Perfect electric conductor, Perfect magnetic conductor and
Free space) can be renamed in the tree or on the respective Properties dialogs. The name change
will be reflected in all instances where the dielectric is applied. Media may also be deleted,
provided they are not used by any other item.

6.1.1 Defining media

Dielectric media

Dielectric media can be applied to geometry regions (see section 6.1.2) and used as building
blocks in the definition of Layered dielectrics and planar substrates.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-2

Dielectric media are created by right clicking on Media in the tree and selecting Dielectric or by
selecting the Construct tab and clicking on the Media menu button.

Figure 6-1: The Dielectric face button

Changing the Name field at the bottom of the Create dielectric medium dialog has the same effect
as renaming the medium in the tree.
In cases where the geometry of a model is not available, dielectric properties may be set on mesh
faces. Dielectric media are set on the mesh in terms of the medium on the front side (normal -
displayed blue in the Default view colouring) and back side (away from the normal - displayed
brown in the Default view colouring) of the mesh. In this case, the onus rests on the user to
ensure that the dielectric definitions are consistent (e.g. the medium specified on the front sides
of two connected mesh triangles cannot differ, except if the normal of one of the elements has
been reversed.) This approach should only be followed for imported meshes with no geometry
representation.
A dielectric may be defined by specifying the definition method for the dielectric and magnetic
modelling.
Dielectric modelling of the dielectric is defined by means of the following dielectric definition
methods on the Dielectric modelling tab of the Create dielectric medium dialog:

• Frequency independent
The dielectric modelling of frequency independent dielectric media is defined in terms of
the Relative permittivity (ε r ), Relative permeability (µ r ), Magnetic loss tangent (tanδµ ),
and either the Dielectric loss tangent (tanδ) or Conductivity (σ). The loss tangent and
conductivity both specify dielectric losses, but have different frequency behaviour. The se-
lection depends on the specific problem. For example, low loss dielectric substrates are
typically specified in terms of the loss tangent, while human tissue (used in specific absorp-
tion rate studies) are specified in terms of conductivity. Expressions may be used in any of
the dialog fields.
The effective permittivity of the dielectric is given by
εe f f = ε0 ε r (1 − j tan δ) (6-1)
or
σ
εe f f = ε0 ε r − j (6-2)
ω
• Debye relaxation
The relaxation characteristics of gasses and fluids at microwave frequencies are described
by the Debye model. It has been derived for freely rotating spherical polar molecules in a
predominantly non-polar background. The method is defined in terms of the Relative static
permittivity (εs ), Relative high frequency permittivity (ε∞ ) and the Relaxation frequency
( f r ).

εs − ε∞
ε∗ = ε∞ + f
(6-3)
1+ jf
r

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-3

• Cole-Cole
The Cole-Cole model is similar to the Debye model, but uses one additional parameter
to describe the material. The Cole-Cole method is defined in terms of the Relative static
permittivity (εs ), Relative high frequency permittivity (ε∞ ), Relaxation frequency ( f r ) and
the Attenuation factor (α).

εs − ε∞
ε∗ = ε∞ + f 1−α
(6-4)
1+( j fr
)

• Havrillak-Negami
The Havrillak-Negami is a more general model and should be able to successfully model
liquids, solids and semi-solids.
The Havrillak-Negami method is defined in terms of the Relative static permittivity (εs ),
Relative high frequency permittivity (ε∞ ), Relaxation frequency ( f r ), Attenuation factor
(α) and the Phase factor (β).

εs − ε∞
ε∗ = ε∞ + h iβ (6-5)
f
1 + ( j f )1−α
r

• Djordevic-Sarkar
The Djordevic-Sarkar is particularly well suited as a broadband model for composite di-
electrics. It is defined in terms of the Variation of real permittivity (∆ε), Relative high
frequency permittivity (∆), Conductivity (σ), Lower limit of angular frequency (ω1 ) and
the Upper limit of angular frequency (ω2 ).
  ω + jω
∆ε ln ω2 + jω jσ
ε = ε∞ + 
∗
ω
 1
− (6-6)
log10 ω2 ln(10) ω ε0
1

• Specify points (linear interpolation)

The dielectric properties of the material can be defined at various frequency points. Values
for the dielectric properties are then linearly interpolated to obtain the dielectric proper-
ties at frequency points other than specified. Parameters required are Frequency, Relative
permittivity (ε r ) and either the Loss tangent (tan∆) or Conductivity (σ).

Magnetic modelling of the dielectric is defined by means of the following dielectric definition
methods on the Magnetic modelling tab of the Create dielectric medium dialog:

• Non magnetic
For non magnetic modelling, the relative permeability (µ r ) is set to 1.0 and the magnetic
loss tangent (tanδµ ) is set to 0.0.

• Frequency independent
The magnetic modelling of frequency independent dielectric media is defined in terms
of the Relative permeability (µ r ) and the Magnetic loss tangent (tanδµ ). The effective
permeability of the dielectric is given by
µe f f = µ0 µ r (1 − j tan δµ ) . (6-7)

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-4

• Specify points (linear interpolation)

The dielectric properties of the material can be defined at various frequency points. Values
for the dielectric properties are then linearly interpolated to obtain the dielectric proper-
ties at frequency points other than specified. Parameters required are Frequency, Relative
permeability (µ r ) and Magnetic loss tangent (tanδµ ).

Alternatively, a file may be imported containing the dielectric and magnetic properties of the
dielectric.

• Import from file

Note that the following keywords are used in the XML file format namely, freq, permittivity,
diel_loss_tangent, conductivity, permeability and mag_loss_tangent.
For demonstrative purposes, the keywords val_A, val_B and val_C will be used instead
in the demo XML file given below as the same format is also applicable when defining a
metallic medium or impedance sheet.
<?xml version="1.0" encoding="UTF-8"?>
<materialDB creator="FEKO" date="2010-07-01" version="1.0">
<material name="mediumA" val_B="7.0" val_C="9.0" >
<dataPoint freq="2.0" val_A="2.0" val_B="6.0" />
<dataPoint freq="3.0" val_A="3.0" />
<dataPoint freq="4.0" val_A="1.0" />
<dataPoint freq="5.0" val_B="5.0" />
<dataPoint freq="6.0" val_A="1.0" />
<dataPoint freq="8.0" val_B="6.0" />
<dataPoint freq="9.0" val_A="4.0" />
</material>
</materialDB>

In the following line, the static values for material with name mediumA, are defined.
<material name="mediumA" val_B="7.0" val_C="9.0" >

Next the frequency dependent data points are defined.

<dataPoint freq="2.0" val_A="2.0" val_B="6.0" />

The internal XML parser will then fill in the missing values in the frequency dependent
data points with static values (if they were defined). If only static data points are defined
and no frequency dependent materials data points are found, then one data point will be
generated with freq=“0.0” by the XML parser and filled with the static values.
Thus the above XML file will then be parsed as if it was specified by the user as:
<?xml version="1.0" encoding="UTF-8"?>
<materialDB creator="FEKO" date="2010-07-01" version="1.0">
<material name="mediumA" >
<dataPoint freq="2.0" val_A="2.0" val_B="6.0" val_C="9.0" />
<dataPoint freq="3.0" val_A="3.0" val_B="7.0" val_C="9.0" />
<dataPoint freq="4.0" val_A="1.0" val_B="7.0" val_C="9.0" />
<dataPoint freq="5.0" val_B="5.0" val_C="9.0" />
<dataPoint freq="6.0" val_A="1.0" val_B="7.0" val_C="9.0" />
<dataPoint freq="8.0" val_B="6.0" val_C="9.0" />
<dataPoint freq="9.0" val_A="4.0" val_B="7.0" val_C="9.0" />
</material>
</materialDB>

The Mass density is only used for specific absorption rate (SAR) calculations, but it must be
specified and must have a value larger than 0.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-5

Figure 6-2: A graphical illustration showing the result of the parsed XML file

Metallic media

Conducting (skin effect) losses are specified by setting the face media of the relevant face/s to
Metallic. (This cannot be done for faces bordering perfectly electric conducting regions.) Metallic
media are created by right clicking on Media in the tree and selecting Metallic medium, or by
selecting the Construct tab and clicking on the Media menu button.

Figure 6-3: The Metallic medium button

Note that it is not possible to apply a user defined metallic medium directly to a solid region
(except for the special case of the Perfect electric conductor medium). The medium of the internal
region should rather be set to free space (or dielectric) and a ‘thick’ metallic medium property
applied to the faces bounding the region. (‘Thick’ here implies that the property should be set
much thicker than the medium skin depth (as defined in Equation 6-8 in the section that deals
with parameters on faces (see section 6.1.2).)
Metallic media are defined in terms of the (i) Conductivity (σ), Relative permeability (µ r ) and the
Magnetic loss tangent (tan∆µ ), as frequency dependent data points or by importing the medium
from file. See the Dielectric media (see section 6.1.1) for the XML format. Keywords for metallic
media are freq, conductivity, permeability and mag_loss_tangent.
In this manual the phrase conducting media collectively refers to Metallic media as well as the
Perfect electric conductor.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-6

Media library

The media library consist of a list of predefined and user defined media. Predefined media may
be imported from the media library to be used in the CADFEKO model. Media may also be added
from the CADFEKO model to the Media library to be used in future CADFEKO models.
A medium is added from the Media library by selecting the Construct tab, clicking on the Media
menu button and selecting Media library from the menu. The Media library dialog may also be
launched by right-clicking on Media in the tree and selecting Media library from the popup menu.

Figure 6-4: The Media library button

The Media library has a filter functionality whereby a user can filter through the list of media for
a required medium. If more information regarding the medium is required, the Advanced view
mode may be used to display the medium properties of the selected medium.

Figure 6-5: The Media library dialog

Layered dielectrics

Surfaces that define the boundary between free space regions can be modelled as thin dielectric
sheets or as coated conductors. This is done by applying user defined Layered dielectric media
(which consist of an arbitrary number of dielectric layers of a specified thickness) to the surfaces.
Note that the Dielectric media constituting the various layers must be defined before the Layered
dielectric can be defined.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-7

Layered dielectric media are created by right clicking on Media in the tree and selecting Layered
structures → Layered dielectric, or by selecting the Construct tab and clicking on the Media menu
button.

Figure 6-6: The Layered dielectric button

The Create layered dielectric dialog contains a table with fields for the Thickness and Medium of
each layer. Only Dielectric media are available in the drop-down list for selection of the Media for
each layer. Clicking the Add button inserts a layer directly after the currently selected layer and
Remove deletes the selected layer from the list. Thus it is possible to insert new layers between
any two layers. Expressions may be used in the Thickness fields.
Coatings and thin dielectric sheets must be geometrically or electrically thin depending on the
selected solution. During EM validation of the model (see section 6.15), CADFEKO will give
warnings if the thickness limits are approached and errors if the thickness limits are exceeded.

Layered dielectric (anisotropic)

Layered dielectric (anisotropic) are created by right clicking on Media in the tree and select-
ing Layered structures → Layered dielectric (anisotropic) or by selecting the Construct tab and
clicking on the Media menu button.

Figure 6-7: The Layered dielectric (anisotropic) button

An anisotropic layered dielectric may be defined by specifying the Thickness, Principal direction
(deg), Material in principal direction and the Material in the orthogonal direction.
Note that the Dielectric media constituting the various anisotropic layers must be defined before
the Layered dielectric can be defined.

Impedance sheets

Impedance sheet structures can be used to apply a surface impedance to surfaces that define the
boundary between free space regions.
Impedance sheets are created by right clicking on Media in the tree and selecting Impedance
sheet or by selecting the Construct tab and clicking on the Media menu button.

Figure 6-8: The Impedance sheet button

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-8

These “media” define the real and imaginary parts of the Surface impedance (effectively the
ratio between tangential electric field on the surface and the electric surface current). In theory,
the same effect can be achieved by applying an appropriately defined Metallic medium to this
surface, but where only the surface impedance is known the properties of the equivalent metallic
medium need not be derived.
The following options are available for defining impedance sheets: Frequency independent, Spec-
ify points (linear interpolation) or Importing medium from file. See the Dielectric media (see
section 6.1.1) for the XML format. Keywords are freq, surf_imp_re and surf_imp_im.

Windscreen

Windscreens can be used to model multiple windscreen antennas on different windscreen di-
electric definitions. Windscreens are created by right clicking on Media in the tree and selecting
Windscreen layers or by selecting the Construct tab and clicking on the Media button.

Figure 6-9: The Windscreen medium button

When creating a windscreen antenna/analysis, the user needs to define the following three com-
ponents:

• Windscreen antenna.

• Windscreen reference plane.

• Windscreen dielectric/definition.

Note that these three components should share the same windscreen name.
When defining an active windscreen antenna element, the user needs to select the required
geometry in the tree. After right-clicking on the respective face/wire in the details tree, select
Properties. Under the Solution tab, select Windscreen under the Solve with special solution
method. Setting the Element type to Active element, will allow the user to specify the offset
(offset A) from the reference to the active element. Note that this offset is specifically required
because of the limitations of a finite mesh in combination with the curvature and rotation of the
model, compared to a smooth surface.
The windscreen analysis is only supported for MoM. Other techniques such as MLFMM, FEM, PO
and GO (ray launching) are not supported. Multiple windscreens with multiple glass definitions
are allowed in one model. The active windscreen antenna elements should always be defined
tangential w.r.t. to the dielectric windscreen reference plane (i.e. the different glass layers).
The windscreen reference may now be defined. Select the required geometry in the tree that
will be used as reference. After right-clicking on the respective face in the details tree, select
Properties. Under the Solution tab, select Windscreen under the Solve with special solution
method. Set the Element type to Reference element.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-9

Figure 6-10: The CADFEKO dialog for defining a windscreen antenna

Figure 6-11: The CADFEKO dialog for defining a windscreen reference

Finally, the dielectric properties of the glass layers can be defined. A dielectric medium can be
defined by right-clicking in the tree on Media and selecting Create dielectric medium. Next, a
layered dielectric is defined by right-clicking again on Media in the tree and selecting Create
layered dielectric. Enter a label and a thickness for the layered dielectric. The layered dielectric
can now be used in the definition of the windscreen.
The windscreen can be created by a number of ways:

1. Select the Construct tab and click on the Media button. Select the Windscreen medium
from the drop down menu.

2. In the tree, right-click on Media and select Windscreen layers from the popup menu

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-10

Figure 6-12: The CADFEKO dialog for defining a windscreen

On the Define windscreen dialog, enter the label of the layered dielectric in the Layer definition
field. Enter the distance from the Reference to the top of layer 1 in the Offset L field.

6.1.2 Setting material properties on geometry

By default all items are perfectly electric conducting (and will be solved using the MoM). Material
properties are set on regions, faces or edges / wires.
Regions are fully enclosed volumes that are listed in the details tree and may be selected in the
details tree or selected in the 3D view by means of the Auto selection mode. Some imported
models use surface parts to define the boundaries of solid regions. These models have no regions
and it is not possible to specify solid dielectrics. However, if the surface parts of such a model
are unioned to indicate connectivity, those parts that form closed surfaces will automatically be
identified as regions in the resultant union part. (Where the unioning of imported faces does not
results in closed region/s, the Stitch sheet parts tool may be used.)
Faces are the individual surface’s of the part and are also listed in the details tree when the part
is selected. Faces may be selected from either the details tree or by clicking in the 3D view when
face selection (see section 3.2.2) is active. The term “face” is used to distinguish it from “surface”,
which in CADFEKO refers to a 2D primitive such as a polygon.
Edges are also listed in the details tree when the part that contains them is selected. They include
the edges of faces as well as free curves (also called “wires”).
Where possible, medium properties are retained during all geometry operations. For example, if
a face is split in two by a subtract operation, both the resulting faces will inherit the properties
of the parent; or, if two overlapping faces are intersected, the resulting face will have all the
properties common to both parents. If it is not possible to merge the properties of the parents, the
items are marked suspect (see section 4.4) and a warning is provided indicating which properties
could not be resolved.

Dielectric volumes are indicated by applying a Dielectric media to the relevant geometry region.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-11

Conducting losses are accounted for by applying a user defined Metallic media to the relevant
conducting surfaces.

Coatings and thin dielectric sheets are specified by applying user defined Layered dielectric
media (constructed from individual pre-defined Dielectric media layers) to surfaces or wires
(free edges).

Impedance sheets (used to represent metal surfaces in cases where only the surface impedance
per unit area is known for a surface) can be modelled by applying an Impedance sheet
medium to the surface. (Although an impedance sheet is not strictly speaking a medium,
user defined Impedance sheets are also listed under Media in the tree, and are applied,
identified and displayed in a similar manner to the other media.)

Windscreens are indicated by applying Windscreen media to the relevant wires and faces.

Properties on regions

To set properties on regions, select one or more regions and select Properties from the pop-up
menu. This opens the Region properties dialog shown in Figure 6-13.

Figure 6-13: The Region properties dialog

On the Properties tab, the medium type of the selected region can be set. It may be Perfect
electric conductor, Free space or any pre-defined Dielectric.

Perfect electric conductor: Perfect electric conductor (PEC) is the default media for newly-created re-
gions. The local mesh size may be changed on the faces enclosing the region. It is possible
to delete faces that border a PEC region. Deleting a face that forms part of the boundary of
a region effectively removes the region.

Free space: Setting a region to Free space effectively converts it to a hollow shell (when setting a region
to Free space, by default the faces bounding the region are set to PEC). It is possible to delete
faces that border free space regions (if the media outside of the region is also free space).
Deleting a face that forms part of the boundary of a region effectively removes the region,
as the remaining face/s are no longer closed. Some face properties (e.g. a thin dielectric
sheet) can only be set on faces bordering free space regions.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-12

Dielectric: The Dielectric region medium is only available if the Media list contains dielectric media (see
section 6.1.1). (New dielectrics may be defined while the Region properties dialog is open
— the Dielectric medium will then immediately become available in the region properties
dialog.) Unlike label names (which must be unique) the same medium can be applied
to multiple regions. By default, the faces bounding a region that is denoted as Dielectric
assume the same dielectric property as the region. This may be modified to any conductive
medium (not a dielectric or layered dielectric medium) by setting the face properties for
each face (see section 6.1.2). The faces bounding a region that is denoted as Dielectric can
only be deleted under certain conditions. If the face defines a border between two different
media, it cannot be deleted.

Where geometry operations introduce intersection of existing regions, and the parent regions
have conflicting material properties, the resulting region/s will be marked suspect.
Initially the region properties dialog shows the current state of each property for the selected
item(s). If multiple regions are selected, fields which are not the same for all dialogs are left
blank. The properties represented by blank field are not modified when changes are applied.
This allows, for example, modifying the mesh size for a number of different dielectric regions
simultaneously without changing the media properties of the individual regions.
On the Solution tab of the Region properties dialog, the solution method for the selected region
may be set (shown in Figure 6-13).

MoM/MLFMM with surface equivalence principle (SEP) - default: The selected region will be solved with
MoM/MLFMM.

MoM/MLFMM with volume equivalence principle (VEP): The selected region will be solved by employ-
ing tetrahedra to accurately mesh arbitrarily shaped volumes with dielectric properties and
solved by means of the MoM/MLFMM.

Finite Element Method (FEM): The selected region will be solved by employing tetrahedra to accurately
mesh arbitrarily mesh shaped volumes with dielectric properties and solved by means of the
finite element method (FEM).

Uniform theory of diffraction (UTD) cylinder: The selected cylinder region will be solved with the uni-
form theory of diffraction (UTD).

Properties on faces

The Face properties dialog (shown in Figure 6-14) has two tabs: Properties and Solution. The
medium properties for the selected face/s are set on the Properties tab. Special solution options
for the chosen face can be set on the Solutions tab.
The Face medium options available in the Medium drop-down list depend on the media of the
two regions bordering the selected face/s. If there is free space on both the front and back sides of
the face/s, the options will be Perfect electric conductor2 , Lossy conducting surface, Impedance
sheet, Thin dielectric sheet and Default. (Some of these options are only available once the
corresponding media have been defined.)
2
Perfect electric conductor replaces the Metal surface option used on this dialog in CADFEKO 1.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-13

Figure 6-14: The Face properties dialog showing the face properties and solution settings options

The Default medium for faces that do not bound a region and for faces that bound a PEC or free
space region is the same as Perfect electric conductor. The Default medium for faces that bound
a dielectric region is the same as Dielectric boundary. If the medium selection is set to Default on
a face that bounds a region, it will take on the default properties according to the region medium
Type setting.
If a face is indicated as a Lossy conducting surface, a metallic Medium and a thickness must be
specified. Note that it is not possible to set general Metallic media directly on regions. A lossy
metallic region can be effectively modeled by setting the boundary faces of Free space region to
lossy surfaces that are much thicker than the skin depth (more than ×5) at the frequencies of
interest, where the skin depth is defined as

r
2
δskin = . (6-8)
ωµσ

The Impedance sheet (see section 6.1.1) and Thin dielectric sheet (see section 6.1.1) face types
will respectively list all defined Impedance sheets and Layered dielectric media in the Medium
options drop-down list.
The Impedance sheet may be used as an alternative method for defining conductive surfaces,
while the Thin dielectric sheet is used to model flat multilayer dielectric structures such as auto-
motive windows.3
The Face type selection is retained when a bordering region is changed. This may lead to invalid
settings. For example, if a Lossy conducting surface is set on a face bordered by free space,
this setting becomes invalid if either bordering region is changed to a Perfect electric conductor.
Face properties that become invalid due to changes in region media will cause a warning to be
reported in the message window and the face will become suspect (see section 4.4). Though the
media properties of such suspect faces are displayed as usual in the details tree, the Face type field
in the properties dialog enters a ternary state (even if all selected faces have the same setting).
Thus the region that caused the invalid setting can be reset without affecting the face settings
(which will then become valid again). The face will be marked suspect during this process and
must be set “not suspect” manually once the problem has been resolved.
In addition to the media type settings, Coatings may be applied to faces. Coatings are applied on
both sides of the face. Hence this option is only available when the face has free space on both
3
The kernel may refer to the SK card in error messages dealing with Impedance sheet and Thin dielectric sheet
specifications.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-14

sides. The Coating name must be selected from a list of Layered media. The coating is applied
such that layer 1 (according to the layered dielectric definition) is on the outside, as shown in
Figure 6-15.4

Figure 6-15: Coatings are applied to both sides of conducting surfaces.

On the Solution tab, the option to solve the selected face with a special solution method, is set.
The following options are available:

• None

• Physical optics (PO) - full ray-tracing

• Physical optics (PO) - always illuminated

• Physical optics (PO) - only illuminate from front

• Large element PO - full ray-tracing

• Large element PO - always illuminated

• Large element PO - only illuminated from front

• Geometrical optics (GO) - ray launching

• Uniform theory of diffraction (UTD)

• Windscreen

• Planar Green’s function aperture

Properties on edges

The Edge properties dialog (shown in Figure 6-16) provides the option to set a local wire radius
(overriding the default setting on the Create mesh dialog) for any selected edge in the model. If
the selected edge/s are wires (i.e. free edges that do not define the boundary of a face), Wire
segment radius, Wire core medium and Coating options are available in addition to the local
mesh size option.
Conducting losses on wires can be accounted for by selecting the pre-defined metal in the Medium
list drop-down list.
For a coating to be applied to a wire, the Coating name is selected from a list of the pre-defined
Layered dielectric media. Currently only single layer coatings can be applied to wires.

4
The kernel may refer to the CO card in error messages regarding coatings.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-15

Figure 6-16: The Edge properties dialog

On the Solution tab, the option to solve the selected edge with a special solution method is
available. The options available are:

• None

• Windscreen

6.1.3 Setting properties on mesh elements

Material properties should be set on the geometry or on a geometry mesh. Once the mesh action
is performed, a resulting simulation mesh will be generated and indicated by a blue mesh icon.
Any major changes to the geometry will automatically remove the simulation mesh, since it will
no longer accurately represent the geometry.
It is not possible to directly edit a simulation mesh. However, it is possible to unlink a simulation
mesh from its geometry and edit it as though it is a base geometry mesh. This type of mesh can
also be obtained by importing a model that is saved in a mesh format. It is possible to set media
properties on these meshes.
Select the required element labels (not the individual elements) in the details tree and access
the Properties from the right-click context menu. Properties can only be set on one type of mesh
element at a time (for example, if the selection contains mesh triangles and wire segments then
the properties dialog will not be available). According to the type of elements selected, the
relevant mesh properties dialog (as shown in Figure 6-17) will be launched.
For triangles, the medium is specified on each side. The media on either side of the mesh elements
are specified on the same dialog (rather than on a separate region dialog for faces that bound a
region). The Face medium and Coating properties may be set as for geometry faces.
For segment labels in the mesh, the surrounding medium can be set on the Mesh properties
dialog. The wire segment core medium and coating can be specified.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-16

Triangle element/s Wire segment/s

Tetrahedral element/s

Figure 6-17: The Mesh properties dialogs for triangle, wire segment and tetrahedral elements.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-17

For tetrahedra mesh elements, only the volume medium can be specified by referencing the label
of a user defined dielectric medium.
If any medium is changed in such a way that the Face medium or Coating setting on a mesh be-
comes invalid, that setting immediately reverts to a ternary state and has to be updated manually.

6.1.4 Displaying dielectric media, coatings and thin sheets

The display options may be accessed on the 3D view context tab, Display options contextual
tab. To select the required colouring option to colour the display according to the dielectric
regions/face media, click on the Colour button. The display of coatings in the 3D view may be
enabled/disabled by clicking on the Coatings button.

Figure 6-18: The (a) Colour button and (b) the Coatings button

In addition to the special 3D view colouring options, the face and coating parameters (media,
local mesh settings etc.) are indicated using icons in the tree (see section 3.3.2). Faces on
dielectric regions are indicated with an icon in front of the item in the tree.
Mesh segments are displayed with their actual radius in the individual 3D views. This radius may
include the coating radius. For disabling the display of the segment coating, click on the Coatings
button on the 3D view context tab, Display options contextual tab.

6.1.5 Displaying the model view, simulation mesh and overlay

6.2 Setting the frequency

Select the Source/Load tab and click on the Frequency button or double-click on Solution in the
tree to open the Solution frequency dialog (shown in Figure 6-20 and Figure 6-21). There are
four Frequency range options.

Figure 6-19: The Frequency button

Single frequency: Allows setting a specific solution frequency

Continuous (interpolated) range: All requested results are calculated using adaptive sampling in the
range Start frequency to End frequency. The sampling algorithm (see sec-
tion 21) uses finer sampling in areas where the results change rapidly to ensure
that all resonance effects are calculated accurately. Since all the requested re-
sults are interpolated, it is best to use this technique with as few result requests

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-18

Figure 6-20: The solution frequency dialog used to define the solution frequencies. The Single frequency
and Linearly spaced discrete points options are shown.

Figure 6-21: The solution frequency dialog showing the Continuous (interpolated) range and the Ad-
vanced tab.

as possible. The Maximum number of samples option limits the number of so-
lutions and hence the runtime, but the results may be inaccurate if not fully
converged. The Minimum frequency increment is used to limit how far FEKO
refines the frequency. This is mostly useful if there are small discontinuities in
the results. In general, the user should use the default settings for these two
options.
On the Advanced tab of the Solution frequency, the following Continuous fre-
quency options can be set: Specify maximum number of samples, Specify min-
imum frequency increment, Specify sampling for exported data files. The fre-

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-19

quency stepping can also be set as Linear or Logarithmically. An option is also

available to set the Convergence accuracy as the following:

High more samples, highly resonant structure

Normal default
Low fewer samples, smooth frequency response

A user may also elect to include certain quantities for adaptive frequency sam-
pling. Selecting specific quantities of interest will include it in the adaptive
sampling. Unselected quantities will be calculated at the discrete solution fre-
quency points.

Linearly spaced discrete points: The user can specify a fixed number of linearly spaced points between
the Start frequency and the End frequency. CADFEKO calculates and displays
the increment. This is typically used when the exact frequencies are known
where the solution is required.

Logarithmically spaced discrete points: The user can specify a fixed number of logarithmically spaced
points between the Start frequency and the End frequency. CADFEKO calcu-
lates and displays the multiplicative factor. This is typically used when the
exact frequencies are known where the solution is required.

6.3 Setting the total source power

The excitation of an antenna is normally specified as a complex voltage, but it may be useful to
specify the total radiated or source power instead. FEKO can therefore scale the result to yield
the desired source power level.
Select the Source/Load tab and click on the Power button or double-click Power under Excita-
tions in the tree to control the source power (the Power settings dialog is shown in Figure 6-23).

Figure 6-22: The Power button

Note that FEKO uses peak magnitude for all complex values — hence excitations (i.e. voltage
sources and current sources) must be specified with peak magnitude (as opposed to root mean
square values) if no power scaling is done. Note however that power settings are specified as
time-averaged values.

No power scaling: This is the default setting. If this option is selected, FEKO will calculate the results
using the specified source magnitudes. Note that plane wave excitations have infinite extent
and therefore infinite power. If a model contains plane wave excitations, the No power
scaling option must be selected.

Total source power: If Total source power is selected, FEKO will scale the results such that the total
source power (the sum of the power delivered by all the individual sources in a model with
multiple sources) is equal to the amount specified in the Source power field. No mismatch
is taken into account. This option can be used with any excitation except plane waves.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-20

Figure 6-23: The power settings dialog used to control the source power options

Incident power (transmission line model): If Incident power (transmission line model) is selected, all
structures are assumed to be fed using transmission lines with a complex characteristic
impedance Z0 . The Source power field specifies the sum of the incident power from all
these transmission lines. If there is a mismatch between the transmission line impedance
and the structure input impedance at the excitation point, some of the incident power will
be reflected back to the source. This is the mismatch loss. This option may be used for
models that contain only voltage source excitations. See illustration in the PW card (see
section 14.55)

FEKO always calculates the total source power for all solutions. For large models or models
with many sources, the calculation of mutual coupling — which is required for accurate source
power calculations — can be very time consuming. If Decouple all sources when calculating
power is checked, the mutual coupling is not considered when calculating the source power.
This is acceptable for sources which (in terms of the computation wavelength) are relatively far
from each other and from other structures in the model, or when accurate power values are not
required. (Gain and directivity extraction are based on source power, and are in general likely to
be inaccurate if the Decouple all sources when calculating power option is checked.)

6.4 General networks and ideal transmission lines

General networks (based on network parameter descriptions) and ideal transmission lines may
be included in the CADFEKO model. These network components may be interconnected and/or
connected to the model geometry and may also be excited or loaded directly.

6.4.1 Ideal transmission lines

Ideal transmission lines may be added to the CADFEKO model by selecting the Save/Load tab
and clicking on the TX line button or right-clicking on the Non-radiating networks branch on the
tree and clicking on the TX line.

Figure 6-24: The TX line button

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-21

The Add transmission line menu is shown in Figure 6-25. The option to automatically determine
the length of the transmission line based on the physical distance between the two points, is
available.
The parameters of the transmission line may be specified, and a label may be assigned. Note that
the transmission line length is specified in terms of the model unit (default metre).

Figure 6-25: The Add transmission line dialog

After a transmission line has been created, it will be shown under the Non-radiating networks
branch of the tree, and the transmission line properties may be edited by double-clicking on this
representation to open the Modify transmission line menu.

6.4.2 General networks

General non-radiating networks may be added to the CADFEKO model by clicking on the Source/Load
tab and clicking on the Source/Load tab and clicking on the Network button or right-clicking on
the Non-radiating networks branch and clicking on Network.

Figure 6-26: The Network button

The Add general network menu is shown in Figure 6-27. The network may be specified in terms
of S- Y- or Z-parameters imported from an external Touchstone file, a direct component-based
network imported from a SPICE *.cir file or manually specified in CADFEKO.
If a Touchstone or network file is used to specify the network parameters, the number of ports
must be indicated (this may be any number greater than or equal to 1) and the reference
impedances for the ports will be determined from the Touchstone file for the case of S-parameters.
If network parameters are specified manually in CADFEKO, a maximum of 4 ports may be used,
and the S-parameter reference impedances for each port should be specified (these reference
impedances must be real and need not be the same for all ports) when S-parameters are used.
Note that only the Touchstone format v1.0 is supported.
If a SPICE *.cir file is used to specify the network (e.g. RLC network), the following are
required: the number of ports and the label for the general network. Note that the number of
ports and the general network label in CADFEKO need to correspond to the number of ports and
subcircuit label in the *.cir file.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-22

Figure 6-27: The Add general network dialogs for Touchstone file and manual parameter definition.

The *.cir file should contain a description of the circuit defined in Berkeley SPICE3f5 syntax.
Please ensure that the correct syntax is used since FEKO only uses a subset of the Berkeley
SPICE3f5 syntax (see section 26). Also note that currently only linear circuits are supported
because FEKO is a frequency domain solver.
After a general network has been created, it will be shown under the Non-radiating networks
branch of the tree, and the definition of the network may be edited by double-clicking on this
representation to open the Modify general network menu.

6.4.3 Connecting, exciting and loading general networks

General non-radiating networks and Transmission lines that have been defined in a model may
be inter-connected or connected to geometry/mesh ports in the model. In order to specify the
interconnection of networks, transmission lines and ports, select the View tab and click on the
Schematic view button. The schematic tools may be accessed by clicking on the Schematic context
tab.
Connections between networks, transmission lines and ports can be created by clicking on the
Connector icon in the Connect non-radiating networks. A connection between two elements
can be created by clicking on the connector point (indicated by a white dot) and dragging the

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-23

Figure 6-28: The Schematic view button

connection until the mouse cursor is over the desired second connection point. When clicking
at the second, a connection between wires will be indicated by a black dot. Selected networks,
transmission lines, ports and connections will be indicated by a dotted outline. Connections
between elements can be deleted by selecting the respective element and pressing <Delete> or
by making use of the Delete icon.
Loads and excitations may be connected to general non-radiating network ports and transmis-
sion line ports in exactly the same way as they are connected to geometry and mesh ports (see
section 6.6).

6.5 Sources independent of the geometry

6.5.1 Plane waves

Select the Source/Load tab and click on the Plane wave button or right-click on Excitation in the
tree and select Add plane wave to specify a plane wave excitation.

Figure 6-29: The Plane wave button

All other sources (listed separately under Excitations) will also be considered active during the
computations for each plane wave incident direction. While defining or editing a plane wave
source, a preview of the incident and polarisation directions is shown in the 3D view. The preview
indicates the polarisation of the E-field (green arrow) and the direction of propagation of the
plane wave (blue arrow) for each of the defined incident directions. Once the field is created,
the E-field polarisation is depicted by a red arrow.
The plane wave source has two modes of operation.

Single incident wave: If Single incident wave is selected, a single plane wave is added to the existing
sources. Multiple Single incident wave excitations can be added to create specific field
distributions.

Loop over multiple directions: If Loop over multiple directions is selected, FEKO calculates a solution
for each specified direction of incidence. The incident direction is specified in a spherical
coordinate system, in terms of the angles θ and φ (in degrees). The user must specify
the Start angle, End angle and angle Increment for each angular coordinate. CADFEKO
calculates and displays the resulting number of incident directions. The actual end angle
is determined based on the start angle, the increment and the number of samples and may
not coincide exactly with the specified end angle. Note that with the Loop over multiple
directions a single plane wave source may also be defined (by setting the Start and End
angles the same). It is currently not possible to define multiple independent plane wave
excitation loops (more than a single incident direction) in a single CADFEKO solution setup.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-24

Figure 6-30: The incident plane wave excitation dialog

Polarisation angle: Polarisation angle specifies the angle η, in degrees, — measured in a right-handed
sense around the direction of propagation — from −θ̂ to the polarisation vector E ~ 0 (linear
polarisation) or major axis (elliptical polarisation). The angle fields in the dialog accept
point entry (see section 3.2.2) from the 3D view. For elliptical polarisation the Ellipticity
must be larger than 0 (linear polarisation) and smaller than or equal to 1 (circular polarisa-
tion).

The electric field strength of the incident field is given by

~ i (~r ) = E~ 0 + jv( E~0 × β̂0 ) · e− j β~0~r

” —
E (6-9)

~ 0 is the polarisation vector.

where v is the ellipticity, β̂0 is the direction of incidence and E
On the Workplane tab, a new workplane can be created. As a result, the plane wave is interpreted
in the chosen/specified workplane. When a new workplane was not set, the plane wave will be
influenced by the default workplane.

6.5.2 Electric and magnetic point dipoles

Incremental point sources may be added by selecting the Source/Load tab and click on the Elec-
tric dipole/Magnetic dipole button or by right-clicking on Excitation in the tree and selecting
Electric point source / Magnetic point source.
The dialogs for definition of electric or magnetic point sources are shown in Figure 6-32. The
Position and Orientation fields (the angles are in degrees) on these dialogs accept point entry
(see section 3.2.2).

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-25

(a) (b)
Figure 6-31: The (a) electric point dipole and (b) the magnetic point dipole buttons

Figure 6-32: The electric and magnetic point source dialogs

The electric point source

The Electric point source represents an elementary dipole element with the specified orientation,
magnitude and phase.

The magnetic point source

A magnetic point source can be either an Electric ring current (the magnitude is specified as the
product of the loop current and loop area) or a Magnetic line current (the magnitude is specified
as the product of the dipole length and the magnetic current). The electric ring current and
magnetic line current radiate the same near and far fields, though the radiated potentials differ.
The selection of which of these sources to use is dependant on the application.

6.5.3 Point sources with a specified pattern

To add a point source with a specified far field pattern, select the Source/Load tab and click on
the Radiation pattern or right-click on Excitation in the tree and select Radiation pattern point.
The Add radiation pattern point source dialog is shown in Figure 6-34.
The specified Magnitude scale factor and Phase offset is applied to each field point in the entire
pattern.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-26

Figure 6-33: The Radiation pattern button

Figure 6-34: The radiation pattern point source definition dialog

On the Workplane tab, a new workplane can be created. As a result, all the fields are interpreted
in the chosen/specified workplane. When a new workplane was not set, the Position field on the
Pattern tab will be influenced by the default workplane.
Radiation pattern data may be imported from an *.ffe file (created by FEKO) or from an ASCII
data file. The AR card can be viewed for more detail on the ASCII format (see section 14.19).
The user must ensure that the pattern information in the data file is applicable to the solution fre-
quency of the current model. Since far field patterns are typically frequency dependent, models
with radiation pattern sources will usually have a single solution frequency only. If the radiation
pattern was calculated using a frequency sweep in FEKO, the *.ffe file will contain multiple
patterns. The Start from point number field may then be used to indicate the position in the
file of the first line of the required pattern. For example, if the pattern was originally calculated
for 50 field directions in a frequency sweep, the pattern for the third frequency of the original
solution may be selected by setting the Start from point number field to 101.

Data file handling

The pattern information is not loaded into CADFEKO, only a reference to the Filename is stored.
The indicated file will therefore be re-processed each time PREFEKO is executed. This implies
that it is possible to overwrite the pattern data file and recompute the FEKO solution without

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-27

making any changes to the model (though the user needs to keep track of the changes to the file
and to ensure that the file can be located by FEKO during each computation).
The pattern data file is not archived by CADFEKO, since it may be in a different directory and may
be used by other models. If the CADFEKO model is saved to a different directory using the Save
as function, CADFEKO will first attempt to find the file in the same relative path and, failing this,
try to find it in the original, absolute path. CADFEKO will then (as part of the Save as operation)
update the file reference/s to the first valid location found.

6.5.4 Impressed spherical mode excitation

An Impressed spherical mode excitation can be defined based on pre-calculated spherical modes.
The excitation is added by selecting the Source/Load tab and clicking on the Spherical modes
button or by right-clicking on Excitation in the tree and selecting Spherical modes.

Figure 6-35: The Spherical modes button

The Add spherical mode source dialog (shown in Figure 6-36) allows the specification of the
source position, orientation and the spherical mode expansion to be excited. The source data
may be imported from an external TICRA *.sph file or specified manually.
When importing the source data from an external TICRA (*.sph) file, the following additional
options are available such as scaling the magnitude of the data file and offsetting the phase.
For manual specification, each mode must be defined separately per line of the table on the dialog
(also shown in Figure 6-36).
When specifying the spherical modes manually, individual modes may be added or removed using
the Add and Remove buttons. The options for manual specification of the spherical modes are
detailed as follows:

Propagation direction: This determines if the spherical waves propagate Inward (the model is illumi-
nated with modes propagating towards r = 0, i.e. spherical Hankel function of the first kind
(3) (1)
zN = hN ) or Outward (the modes radiate towards r = ∞, i.e. spherical Hankel function
(4) (2)
of the second kind zN = hN ). This option is only available when the modes are entered in
the *.pre file and not when the modes are imported from a TICRA file (*.sph), in which
case the outward propagating direction is used.

Index scheme: The Normal scheme uses the traditional smn index. If this option is selected, the user
can specify TE-mode (s = 1) or TM-mode (s = 2) and the indices M and N in the indices
columns. Here N is the mode index in radial direction and must be in the range 1, 2, . . . ∞
and M is the mode index in the azimuth direction ϕ. We do not distinguish between even
and odd modes (with cos(M ϕ) and sin(M ϕ) angular dependencies), but rather use the
angular dependency e j M ϕ . Thus the index M can also be negative, but it must be in the
range −N . . . N .
The Compressed scheme uses a compressed one-dimensional mode numbering scheme. The
J mode index is then specified in the index column. Here

J = 2 [N (N + 1) + M − 1] + s (6-10)

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-28

Figure 6-36: The Add spherical mode source dialog for excitations based on an external file

where s = 1 for TE-modes and s = 2 for TM-modes. This unified mode numbering scheme
allows the computation of an extended scattering matrix (with network and radiation ports).
The index J then represents a unique port number in the scattering matrix.

Mag. sqrt(W): Absolute value of the complex amplitude of this specific spherical mode p (due
p to the
applied normalisation of the spherical modes, the unit of this amplitude is W = VA).

Phase (deg.): The phase of the complex amplitude of this spherical mode in degrees.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-29

6.5.5 Impressed aperture excitation

Aperture excitations may be used to specify a planar, cylindrical or spherical aperture of measured
or calculated field values that are impressed as an excitation. The field values are converted into
an equivalent array of electric and magnetic dipoles in the solution.
Aperture excitation is added by selecting the Source/Load tab and clicking on the Aperture field
button or by right-clicking on the Excitation in the tree and selecting Aperture field to launch the
Add aperture excitation dialog (shown in Figure 6-38).

Figure 6-37: The Aperture field button

Figure 6-38: The Add aperture source dialog

In this dialog the source may be defined as follows:

Magnitude scale factor: The field magnitudes in the files will be multiplied by this factor. This is useful
when using data files where the field values are in different units (e.g. µV/m)

Phase of the aperture (deg): This phase (in degrees) will be added to the phase of the fields as defined
in the data files.

Aperture data type: The type of data to be used to generate the aperture excitation is specified here. An
aperture excitation may either be defined based on only impressed electric or magnetic field
distribution or a combination of both electric and magnetic fields.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-30

Source: The data for the E and H fields may be imported from *.efe and/or *.hfe files generated
by a previous FEKO solution (see section 6.12.5 and FEKO User Manual (section 14.31) for
information on generating these files) or a more general ASCII file. The ASCII file format is
discussed in FEKO User Manual (section 14.18). Note that the units for when reading data
from an ASCII file are V/m and A/m for the E-field and H-field respectively. The first line
number in the file from which data should be read may be specified in the Start reading from
line number field. This may be used, for example, when a single file contains information
for multiple aperture excitations, or when additional comment lines are inserted at the top
of an ASCII data file that should be ignored. If both electric and magnetic field data is read,
the starting line in the files will be the same.

Source Destination: The coordinate system in which the field data is specified, as well as the number
of field points in each coordinate direction is specified here. It is important to note that
data may only be specified in Spherical or Cylindrical coordinates when both electric and
magnetic fields are used. An additional Also sample along edges option is provided. This
should be used depending on how the physical location of the sample points relate to the
aperture defined in the Source Destination. When checked the outer sample points are
assumed to lie on the edges, when unchecked the sample point lie half an increment away
from the edges. When using multiple aperture excitations in a single model, sample points
should not lie on the edges of any two apertures that share a common side, otherwise two
elementary dipoles elements with the same location and polarisation will be included. If this
is the case the power calculation in FEKO will fail and incorrect results may be calculated.

Note that the position of the aperture excitation is by default at the global workplane. The
position may be adjusted by modifying the workplane on the Workplane tab.

6.5.6 Impressed current

An Impressed current may be defined as an excitation in a model. This is a special excitation

type, and it is subject to a number of limitations (see section 14.15). The impressed current
excitation is defined on the Add impressed current dialog (shown in Figure 6-40). This dialog is
launched by selecting the the Source/Load tab and clicking on the Impressed current button or
by right-clicking on the Excitation tab and selecting Impressed current.

Figure 6-39: The Impressed current button

The position of the start and end points of the current segment, as well as the magnitude and
phase (in degrees) of the current at the start and end points can be defined. The radius of the
impressed current must also be specified in the Radius field. This radius value may not be zero,
and the unit is determined by the model unit.
A Connect the end point to the closest triangle vertex option is provided. When checked, the
end position of the impressed current cannot be specified, but will be determined based on the
closest triangle mesh vertex available during the solution. This is useful when the impressed
current should terminate (and be connected to) the model geometry. There must be a mesh
vertex at the connection point, but it is not always possible to predict the exact location of the
mesh vertex to which the excitation should be connected. When using this option, care should

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-31

Figure 6-40: The Add impressed current dialog

be taken to confirm that the excitation does indeed connect at the required triangle vertex. (This
can be visually confirmed by running PREFEKO and observing the impressed current excitations
in POSTFEKO.)
Impressed current excitations are shown as red arrows in the 3D view. The preview arrow is
located between the start and end points, and has the specified radius of the impressed current.
When the Connect the end point to the closest triangle vertex option is used, the impressed
current is represented by a red sphere at the starting point of the impressed current. The radius
of the sphere is determined by the radius of the impressed current.

6.6 Setting ports, sources and loads on geometry

Ports and loads are mathematical representations for where energy can enter or leave a model.
A port can either serve as a source, or as a sink. Loads and excitations may be applied to a port
in a variety of ways to best model the equivalent real-world problem.
In CADFEKO, voltage sources and discrete loads are not applied directly to the model geometry
or mesh, but rather to ports which must be defined on the geometry or mesh before adding the
required source or load. Ports effectively define the locations where excitations and loads may
be attached to the geometry.

6.6.1 Ports

There are six types of ports that can be defined in CADFEKO: Wire ports, Edge ports, Microstrip
ports, Waveguide ports, FEM line ports and FEM modal ports. Each has a different application
and advantages or restrictions. The FEM ports, for example, may only be applied in FEM regions.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-32

Waveguide ports may be used with most solution methods, but only have certain predefined
shapes (i.e. rectangular, circular and coaxial). Using the correct port for a particular model can
help reduce resource requirements and provide more accurate results. Figure 6-41 shows the
ports that can be created.

Figure 6-41: List of ports that can be added. The list is obtained by right-clicking on the Ports entry in the
model tree.

All ports (irrespective of type) are listed under Ports in the tree. Each port contains sub-instances,
indicating the geometry- and/or mesh-linked manifestations of that port. Generally ports are
created on geometry items and such ports initially contain only a geometry instance. When the
geometry part to which the port is applied is meshed, a mesh instance is automatically created.
If the mesh part to which a port is connected is deleted, then the mesh instance of that port will
automatically be deleted if the port also has at a geometry instance. If the port does not have a
geometry instance, the port must first be removed before the mesh part can be deleted.
The tree representation for each port contains separate icon representations for the geometry
instance and/or each of the mesh instances of the port. (It is not possible for a port to have
multiple concurrent mesh instances - this was possible in previous versions of CADFEKO under
certain special circumstances. If a model created in a previous CADFEKO version which has
multiple mesh instances is opened in a newer CADFEKO version, separate mesh ports will be
automatically created for each of the mesh port instances and a message indicating the changes
provided.) It is generally not necessary to view the separate instances of a port themselves,
though if the port is suspect (see section 4.4) or contains a fault (see section 5.1.4), the suspect
or fault reason may only be displayed in the tool-tip for the individual instances.)
Ports may be created directly on meshes. This option is only available from the Mesh → Create
port menu and should only be used for imported meshes or in cases where the geometry is no
longer available. If a port is created directly on the mesh, it will only contain a mesh instance.
Ports created on meshes derived from CADFEKO geometry will be automatically removed if the
geometry is re-meshed, and the mesh is replaced.
If a port contains a geometry instance, selecting Properties on the pop-up menu of the main
port item opens the properties of the geometry instance. If the port contains only a single mesh
instance, this opens the properties of that instance. The properties of any specific instance may
be opened by expanding the port in the tree and selecting Properties on the pop-up menu of that
instance.
The Label field is used to uniquely identify each port. The label may be changed using the
Rename entry in the pop-up menu of the main port item or the Properties dialog of any geometry
or mesh instance of the port.
Some of the mesh refinement tools may remove elements on which ports are set. The affected
ports are then marked suspect and must be edited (and correctly re-assigned to geometry / mesh
items) to correct this setting.
The geometry and mesh instances associated with a port can be individually deleted. A port is
deleted when its last instance is deleted. If a port is applied to a mesh part (as described above)

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-33

or the geometry instance is deleted after meshing, that port contains only mesh instances. Any
mesh part that a port refers to cannot be remeshed unless the port is removed completely.
Note that only the mesh instances of ports are written to the *.pre and *.cfm files.
The convention is that if a positive voltage is applied to a port, the current flows through the port
from the negative side to the positive side.

Wire ports

Wire ports are created on wire edges, i.e. free edges that do not form a face boundary. Wire ports
may be created in a number of ways

1. Select the Source/Load tab and click on the Wire port button

2. Select the wire in the details tree and select Create port → Wire port from the pop-up menu

3. Right-click on the Ports tree entry and select Create port → Wire port from the pop-up
menu

Figure 6-42: The Wire port button

The Create wire port (geometry) dialog is shown in Figure 6-43. During the creation/editing of a
port, a preview of the current position of the port is indicated with a small sphere in the 3D view.

Figure 6-43: The Create wire port dialog for geometry instances

Edge: Determines which wire the port is located on. This can be changed by selecting an edge
in the 3D view or details tree. To select an edge in the 3D view, the edge selection mode
must be active. (The selection mode is changed to edge selection when the dialog is opened,
but the user may have changed it while the dialog is open.) If the dialog is opened from
the right-click pop-up menu of a single wire in the details tree, the selected edge label is
automatically entered into the Edge field.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-34

Place port on: When the wire is meshed, the mesh instance of the port can be placed on a single seg-
ment or on the vertex between two segments. This group allows specification of how the
port should be created. Vertex ports are mainly used where wires are connected to other
structures and the phase difference from the end point to the centre of the first segment
would have a significant influence on the input impedance.

Location on wire: This group is used to specify where on the selected wire the port is to be located. If
Segment points are specified at the end points, the first (specially shortened) segment on
this end is used. With Other selected, the user may specify an arbitrary position along the
wire in terms of the position as a percentage of the total wire length. 0% is interpreted as
placement at the start point and 100% at the end point. If the wire is modified after the port
has been applied in such a way that there is still only one wire with this label, the port will
maintain the same relative position along the wire. Thus if the port was one third from the
end of a wire and the wire is shortened to work at a different frequency, the port remains
one third from the end of the shorter wire. The absolute position of a port may be fixed
by entering a named point or a ‘pt’ expression (see section 4.2.2) in the % field. The port
is then located at the projection of this point onto the wire. If the wire is modified after
application of the port, the point will remain as close as possible to this absolute position.
This field accepts standard point entry (see section 3.2.2)

Reverse polarity: The option to reverse polarity of the port can be set by checking Reverse polarity.

If a model is modified such that the wire label to which a port is applied no longer exists, or if the
wire is divided into multiple parts during a geometry operation, the port may be marked suspect.
The user must then edit it and select the appropriate single edge to which the port should apply.
If the wire is split into multiple edges in the case where the port position has fixed using a named
point, the port will be located on the edge where the projection is closest to this point where
posible.
Vertex ports are now allowed to be set on the end of wires that are connected to infinite ground
planes and UTD plates.
Wire ports can be set directly on the mesh by selecting the Source/Load tab and clicking on the
Wire port button. Note that ports should generally be applied to the geometry, mesh instances
should only be directly defined on imported and unlinked meshes.
The Create wire port (mesh) dialog (shown in Figure 6-44) is similar to the Create wire port
dialog for geometry. The Locate port on group allows direct Segment / Vertex selection. When the
Segment or Vertex field is active (according to the Locate port on selection), the user can select a
segment/vertex by clicking in the 3D view. (The correct selection mode (see section 3.2.2) must
be active.) The Segment/Vertex field reflects the wire label to which the port will be applied,
while the start and end point coordinates of the segment or the coordinates of the vertex are
displayed in the read-only fields below this.
In the 3D display the segment on the positive side of the port is indicated with a red cylinder and
the one on the negative side with a blue cylinder. Ports which are set on vertices (as opposed
to segments) are identified by an additional small sphere at the vertex. If a vertex port is at a
connection to a surface, only the segment display is coloured.

Edge ports

Edge ports are created along an edge defining the boundary between two sets of faces. All the
faces referenced in the port definition must belong to the same part for the port definition to be

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-35

Figure 6-44: The Create wire port (mesh) dialog for mesh instances

valid.
Select the Source/Load tab and click on the Edge port button or by right-clicking on Ports in the
tree and selecting Edge port.

Figure 6-45: The Edge port button

The Create edge port (geometry) dialog — shown in Figure 6-46 — contains two lists of labels.
The Positive faces list defines the set of faces on the positive side of the port, while the Negative
faces list defines the set of faces on the negative side.
When the dialog is launched, the Positive faces list is populated based on the current selection.
If the current selection contains a face selection, then the list is automatically populated with all
of the selected face labels. If the selection contains edges, then the labels of all of the associated
faces are also listed. If the selection contains only parts, then no faces are added to the initial list.
(Note that this initial list of faces may contain labels of faces from different parts, and is simply
intended as a starting point for the port definition. Some labels will need to be moved to the
Negative faces list and other labels may need to be added or entirely removed to specify a valid
port definition).
Faces that use the UTD solution method may also be used in the definition of an edge port, but
one side of the port must consist only of one valid UTD face, and all faces on the other side of
the defined port must be standard MoM faces.
Where an infinite ground plane is present in the model, any edges that lie in the plane may be
excited with respect to that ground plane. The Connect to infinite ground checkboxes on the
dialog allow the positive or negative side of the port to connect to the infinite ground.

Add/Remove: The user may add and remove face labels from each of the lists as required.

Move to negative/positive face: Faces can be switched between the lists by clicking the button between
the lists (double-clicking on an entry will also shift that entry to the other list)

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-36

Figure 6-46: The Create edge port dialog for geometry instances.

Connect to infinite ground: When this option is selected, the infinite ground will form the positive/neg-
ative side of this edge excitation (according to the selection). Only one side of the excitation
may be connected to ground at a time, and all faces in the list on the other side of the port
must have edges that lie in the plane of the infinite plane.

Only one entry in either of the lists of face labels may be selected at a time. The background of
the entry that has the selection focus is coloured yellow. While one of the list entries has focus,
clicking on a face in the 3D view or on the details tree representation of a face will enter the
face’s label into the selected field. In this way, the relevant face labels may be added to the lists
to complete the port definition.
During creation/editing of an edge port, the boundary between the two sets of faces (the actual
location of the port) is highlighted with a grey cylinder in the 3D view. The edge is marked on
the side of the positive faces with a second, red cylinder while the negative faces are marked with
a third, blue cylinder.
Note that the edge on which a port is defined need not be straight — it may even close on itself,
for example to create a port at the centre of a thick dipole modelled using a cylinder meshed into
triangle elements rather than wire segments.
For the edge port, the mesh instance creation dialog (launched from Geometry → Create port
→ Edge port on the main menu) and the 3D display are exactly the same as for the geometry
instance.

Microstrip ports

Microstrip ports are used to represent feed lines on microstrip structures. These ports are speci-
fied on an edge or set of edges that form a continuous, straight, horizontal (i.e. lie in a constant

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-37

z plane in the global coordinates) edge bordering only one face. In addition, the model must
contain a planar dielectric substrate with a conducting ground plane at the bottom.
Select the Source/Load tab and click on the Microstrip port button.

Figure 6-47: The Microstrip port button

The Create microstrip port (geometry) dialog (shown in Figure 6-48) contains a single list of
edges (similar to the face lists of the Edge port discussed above).

Figure 6-48: The Create microstrip port dialog for geometry instances.

This port is applied by selecting the Source/Load tab and clicking on Microstrip port or by se-
lecting a set of edges and selecting Create port → Microstrip port from the pop-up menu in the
details tree. The port is previewed and displayed in the 3D view in a similar fashion to the edge
port. The orientation or polarity of the microstrip port is shown in the 3D view port preview. If
the excited edge is highlighted blue, the positive side of the port is connected to the edge and
the negative side to the ground plane. If the edge is highlighted red, then the negative side of
the port is connected to the edge. The orientation of the port can be reversed by checking or
un-checking the Reverse polarity checkbox.
A microstrip port may be created directly on the mesh. The mesh microstrip port dialog is
accessed by selecting the Source/Load tab and clicking on the Microstrip button. On the mesh
instance dialog (shown in Figure 6-49) the edge to which the port is to be applied is specified in
terms of its start and end points. These must be mesh vertices and must be selected by clicking on
the relevant vertices in the 3D view while the Start vertex and End vertex fields are active. (Mesh
vertices must be displayed.) The vertex fields then display the label of the elements associated
with the appropriate vertices. The display is the same as for the geometry instance.

Waveguide ports

Waveguide ports are used to define the planes of excitation for waveguide structures. Three basic
waveguide cross-sections are supported namely:

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-38

Figure 6-49: The Create microstrip port dialog for mesh instances

1. Rectangular waveguide

2. Circular waveguide

3. Coaxial waveguide

These ports are specified on a single face with the correct shape. In order to apply a port to a
face, the face must be flat, must not contain any internal edges, must not form the boundary
of a PEC or dielectric region, and must not have any special material properties (e.g. Dielectric
coating) or be solved using special solution methods (e.g. UTD).
Waveguide ports may be created by selecting the Source/Load tab and clicking on the Waveguide
port button.

Figure 6-50: The Waveguide port button

The Create waveguide port dialog (shown in Figure 6-51) allows the selection of a single face to
which the port is to be applied. When a face is selected, a propagation direction and reference
direction for the port are automatically chosen and displayed in the port preview. If these are not
as desired, they can be changed by selection the Propagation direction opposite to normal and
by setting the relevant rotation angle in the Rotate reference direction list respectively.
On the Advanced tab of the Create waveguide port dialog (shown in Figure 6-52) certain special
settings may be chosen. By checking the Manually set the number of modes considered option,
the maximum mode indices can be set. If this is not selected, then FEKO will automatically
determine the maximum number of modes to consider. By selecting Manually set the reference
vector the automatically chosen reference vector for the port can be overridden. In this case,
the reference vector is specified from the centre of the port face in the direction of the specified
vector.
Waveguide ports may also be applied directly to mesh faces. In order to do this, select the
Source/Load tab and click on the Waveguide port button. Specification of a waveguide port on
the mesh is very similar to the specification of a waveguide port on geometry in that CADFEKO
automatically calculates the relevant port shape based on the shape of the selected mesh face
(the mesh must represent one of the supported waveguide shapes sufficiently or CADFEKO will
give an error). For direct application of a waveguide port to the mesh, the reference direction of
the port should be manually specified in terms of a vector from the centre of the excitation face.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-39

Figure 6-51: The Create waveguide port dialog for geometry instances.

Figure 6-52: The Advanced tab of the Create waveguide port dialog for geometry instances.

The waveguide port differs from other geometry-linked ports in CADFEKO in that when no ex-
citation is applied to the port it will still be considered in the FEKO solution. A waveguide port
without any waveguide excitation(s) (see section 6.6.2) applied will be considered as an ideal
waveguide termination in all parts of the solution (i.e. all energy that propagates into the wave-
guide port will be absorbed and not reflected).

FEM line ports

FEM line ports are used to define the location of impressed current source excitations and loads
in the FEM region. A FEM line port may be added by selecting the Source/Load tab and clicking
on the FEM line port button or by right-clicking on the Ports branch of the tree and selecting
Create port → FEM line port.
The Create FEM line port (geometry) dialog is shown in Figure 6-54.
The port position may be specified in one of two ways — based on the location of a free edge (or
a connected set of free edges that form a continuous straight line) inside of a dielectric region in

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-40

Figure 6-53: The FEM line port button

Figure 6-54: The Create FEM line port (geometry) dialog for the definition of a FEM line port in the FEM
region

the model, or based on coordinates defining the beginning and end points of the FEM line port
(in global coordinates).
If the port location is specified based on an edge (or set of connected edges) the orientation of the
port is determined automatically. The orientation can be visually determined from the 3D view
preview of the port, and may be reversed if required by checking the Reverse the port orientation
option. If the geometry on which the port is defined is modified (translated or scaled etc.) the
FEM line port location and size will be adjusted accordingly. If the FEM line port definition is
based on coordinates, the user must ensure that the coordinates remain correct.
During meshing, all edges that are associated with a FEM line port definition will be excluded
in the mesh, but are used to determine the start and end points of the FEM line port. The FEM
line port has a mesh instance which may be directly defined or edited (after it is generated from
a geometry port during meshing). The mesh instance of the FEM line port is defined based on
either the start and end coordinates, or by attaching the start and end point of the FEM line port
to specific mesh vertices within (or on the boundary of) a tetrahedral FEM mesh. Mesh instances
of the FEM line port that are generated based on geometry ports during meshing will be defined
based on mesh vertices where possible, or based on the coordinate values directly.
The Create FEM line port (mesh) dialog is shown in Figure 6-55.
Where the mesh instance of a port is defined based on mesh vertices it will be maintained through
operations on the mesh (scaling, translation etc.). Where the mesh instance of the FEM line
port is based on coordinate points, the user must ensure that the port definition is as intended.
(Note — during meshing, mesh instances of FEM line ports that are based on coordinate point
definitions are generated with the label ’user defined’ and do not inherit their name from the
geometry instance of the port. This indicates that CADFEKO will not maintain this port position

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-41

Figure 6-55: The Create FEM line port (mesh) dialog for the definition of a FEM line port in a tetrahedral
mesh region

through mesh manipulations.)

Loads (see section 6.6.3) and current sources (see section 6.6.2) may be applied to a FEM line
port and FEM line ports may be included in S-parameter calculations (see section 6.12.2).
If a model created in a version of CADFEKO released before the Suite 5.4 FEKO release, any FEM
current source excitations in the model will automatically be converted into an equivalent FEM
line port with an applied current source excitation (see section 6.6.2). This should not influence
the solution, but the line port will have to be maintained by the user through any changes made
to the model.

FEM modal ports

FEM modal ports are used to apply a modal port boundary condition on the boundary of a finite
element (FEM) region. A FEM modal port essentially represents an infinitely long guided wave
structure (transmission line), connected to a dielectric volume modelled with FEM. The FEM
modal port may be excited with the fundamental mode of the associated guided wave structure,
or it can act as a passive port. S-parameters can be computed between the fundamental mode of
the FEM modal port and other excitations in the model. In order to apply a FEM modal port to a
face, the face must be flat and on the boundary of a FEM region.
A FEM modal port may be added by selecting the Source/Load tab and clicking on the FEM modal
port button or by right-clicking on the Ports branch of the tree and selecting Create port → FEM
modal port.

Figure 6-56: The FEM modal port button

The Create FEM modal port dialog (shown in Figure 6-57) allows the specification of the port
either as a list of faces or as points.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-42

Figure 6-57: The Create FEM modal port for geometry instances.

6.6.2 Sources

When adding an excitation, a new item will appear under the Calculation part of the tree to
indicate that source data will be calculated by simply adding an excitation. Note that when only
the reflection coefficient or input impedance is required, it is not required to add an S-parameter
request.

Voltage sources

The voltage source excitation can be applied to any wire, edge, line, network or transmission-line
port to define the excitation of that port. Any port of these types to which a voltage source (or
load) is not applied will have no effect on the solution — though it may be used in the definition
of a port for S-parameter calculations.
In order to define a voltage excitation, select the Source/Load tab and click on the Voltage source
button or right-click on Excitation in the tree and select Voltage source.

Figure 6-58: The Voltage source button

A voltage source may also be added by right-clicking on a defined port in the Ports tree and

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-43

selecting Add voltage source (in this case, the port label is automatically selected in the Port field
on the voltage source dialog — the port selection may, however, be changed freely).
The voltage source creation dialog is shown in Figure 6-59. The dialog contains a Port field which
lists all the ports in the model to which a voltage source may be applied. The list of valid ports
will include all defined ports, and when the excitation is applied, it will be reflected in all of the
mesh and geometry instances of that port.

Figure 6-59: The dialog for the application of a voltage source to a port.

The input fields for the voltage magnitude and phase define the excitation to be applied. The
specified voltage gives the potential difference between the positive side of the port relative to
the negative side. A positive voltage will result in positive current flowing out of the positive side
and into the negative side of the port.

Current sources

The current source may be applied to a line port (see section 6.6.1) in a dielectric region to be
solved with the FEM in order to realise an impressed current excitation (this is equivalent to the
FEM current source excitation available in FEKO versions before the Suite 5.4 release).
Add a current source by selecting the Source load tab and clicking on the Current source button.

Figure 6-60: The Current source button

The Add current source dialog is shown in Figure 6-61. The excitation is specified based on a
constant current (Magnitude and Phase) that is applied in a filament along the line defined by the
line port to which it is attached. The line port should therefore be short relative to a wavelength
in the medium in which it is located.

Figure 6-61: The dialog for the application of a current source to a line port.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-44

An intrinsic limitation of the impressed current source is that no radius is taken into account,
therefore the field is singular in the vicinity of the filament. This affects the accuracy of the
computed input impedance of the source.

Waveguide excitations

Specific waveguide mode excitations can be applied to waveguide ports (see section 6.6.1)
that have been defined in the model. A waveguide mode excitation is added by selecting the
Source/Load tab and clicking on the Waveguide excitation button or by right-clicking on Excita-
tions in the tree and selecting Waveguide excitation.

Figure 6-62: The Waveguide excitation button

The Add waveguide excitation dialog is shown in Figure 6-63.

Figure 6-63: The dialog for the application of a waveguide excitation to a waveguide port (options for
manual and fundamental mode selections shown).

The Excite fundamental mode only option should be selected if the fundamental propagating
mode at the solution frequency must be excited (the FEKO kernel will compute this mode based
on the waveguide size and cross-section during the solution). The Magnitude and Phase (in
degrees) of the excitation, as well as the physical Rotation of the mode (also in degrees) can
be specified. The rotation of an excited mode is relative to the reference direction of the port
to which the mode is applied and for rectangular ports, the rotation must be a multiple of 90
degrees.
If a specific mode or a specific set of modes are to be excited, the Manually specify modes option
should be selected. In this case, the type and indices (where applicable) as well as the magni-
tude, phase and physical rotation of the mode(s) must be entered in the table. Modes may be
added or removed from the list using the Add and Remove buttons. It is important to note that
if multiple modes are applied to a single waveguide port (as part of one waveguide excitation
definition or in separate definitions) the modes will all be applied simultaneously during the
solution phase. When using waveguide ports as part of an S-parameter calculation, the wave-
guide excitations that are applied to the port will not influence the S-parameter calculation (see

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-45

section 6.12.2). During the S-parameter calculations, only the mode(s) specified as part of the
S-parameter request will be considered.

FEM modal excitations

The FEM modal excitation can be applied to FEM modal ports (see section 6.6.1) that have been
defined in the model. A FEM modal excitation is added by selecting the Source/Load tab and
clicking on the FEM modal excitation or by right-clicking on Excitations in the tree and selecting
FEM modal excitation.

Figure 6-64: The FEM modal excitation button

The Add FEM modal excitation dialog is shown in Figure 6-65.

Figure 6-65: The dialog for the application of a FEM modal excitation to a FEM modal port.

If the FEM modal excitation is defined, the modal port will be excited with the fundamental
mode of the associated long guided wave structure of the FEM modal port. When no excitation
is defined, the modal port will act as a passive port (sink) for fields incident on the port.

6.6.3 Loads

An impedance load may be applied to any wire, edge, line, network or transmission-line port.
The Create load dialog (shown in Figure 6-67) is launched by selecting the Source/Load tab and
clicking on the Add load button or by double clicking on Loads in the tree, or by right-clicking on
a specific port and selecting Add load (if this last method is used, then the port is automatically
selected in the ports list on the Create load dialog — though the selection may be changed freely).

Figure 6-66: The Add load button

There are three types of loads that may be specified.

Complex impedance: A frequency-independent load consisting of a constant real and imaginary part.
This load type can be applied to wire, edge, stripline, network or transmission-line ports.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-46

Figure 6-67: The impedance load dialog showing the options for the 3 types of loads that may be applied
to a port.

Series circuit: A frequency dependent load consisting of a series-connected resistor (R), capacitor (C)
and inductor (L). This load can only be applied to wire ports. The load impedance is given
by
1
Zs = R + jωL + . (6-11)
jωC

Parallel circuit: A frequency dependent load consisting of a resistor (R), capacitor (C) and inductor (L)
connected in parallel. This load can only be applied to wire ports. The load impedance is
given by
1
Zp = 1 1
(6-12)
R
+ jωL
+ jωC

where the resistance or inductance is taken as infinite when set to 0 (i.e. it does not con-
tribute to the impedance).

If multiple loads, or loads and sources are applied to the same port, they are placed in series.
This also applies to the Parallel circuit load — only the three elements of the load are connected
in parallel.

6.7 Infinite planes — Planar Green’s functions and ground planes

FEKO can include the effect of an infinite ground plane and/or planar substrates.
Select the Construct tab and click on the Infinite structures button and select Infinite plane from
the dropdown menu or double-click Infinite planes under Solution in the tree to open the Infinite
planes / ground options dialog (shown in Figure 6-69).

Figure 6-68: The Infinite plane button

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-47

No ground (homogeneous free space medium): This is the default for any new model. When this option
is selected, the model is solved in an homogeneous environment filled with the Free space
medium. The properties of the Free space medium may be edited in the media tree as
required.

Perfect electric (PEC) ground plane at z=0: This provides for the inclusion of an infinite ground plane
that coincides with the z = 0 plane (in the global coordinate system). Perfect electric con-
ductor is used as the ground medium. In the reflection coefficient approximation, a reflected
component is added to each field. This technique is a lot faster than the exact Sommerfeld
integral method, but is also less accurate. For real grounds all conducting structures must
be more than a tenth of a wavelength above the ground. (For perfect ground, structures
may be connected to the ground, but not below it.)

Perfect magnetic (PMC) ground plane z=0: This provides for the inclusion of an infinite ground plane
that coincides with the z = 0 plane (in the global coordinate system). Perfect magnetic con-
ductor is used as the ground medium. In the reflection coefficient approximation, a reflected
component is added to each field. This technique is a lot faster than the exact Sommerfeld
integral method, but is also less accurate. For real grounds all conducting structures must
be more than a tenth of a wavelength above the ground. (For perfect ground, structures
may be connected to the ground, but not below it.)

Homogeneous half space in region z<0 (reflection coefficient approximation): This provides for the inclu-
sion of an infinite ground plane that coincides with the z = 0 plane (in the global coordinate
system). The medium is selected from the list in the Ground medium group. This list con-
tains all of the user-defined Dielectric media. In the reflection coefficient approximation,
a reflected component is added to each field. This technique is a lot faster than the exact
Sommerfeld integral method, but is also less accurate. For real grounds all conducting struc-
tures must be more than a tenth of a wavelength above the ground. (For perfect ground,
structures may be connected to the ground, but not below it.)

Homogeneous half space in region z<0 (exact Sommerfield integrals): This provides for the inclusion of
an infinite ground plane that coincides with the z = 0 plane (in the global coordinate sys-
tem). The ground medium is selected from the list in the Ground medium group. This list
only contains all of the user-defined Dielectric media (Perfect conductors are not supported
in the Sommerfeld ground). The Sommerfeld integral solves the exact boundary condition
using the appropriate Green’s function. This method supports conducting structures inside
the ground, but no mesh element that crosses the surface of the ground may be connected
to an element below the ground. For example, partly buried wires are allowed as long as
there is a vertex at the position where the wire passes through the surface of the ground.).

Planar multilayer substrate: This adds an infinite planar multilayer substrate to the model. The user
can define the Medium and Thickness of each layer in the planar structure. A ground plane
may be specified at the bottom of each layer by selecting Bottom from the Ground plane
dropdown list. If a groundplane is enabled at the bottom of a layer, it will be indicated by
an orange line underneath Media preview. The model is then solved using planar Green’s
functions to account for the layered structure. The Medium lists contain all the user defined
Dielectric media. Layers can be added and removed with the Add and Remove buttons.
Conducting planes can be added on top of layer 1 or below the bottom layer by checking
the appropriate fields in the Conducting ground plane group. Note that Layer 0 is infinitely
thick and it is ignored if there is a conducting plane at the top of the layered structure. The
thickness of the bottom layer is only relevant if there is a conducting plane at the bottom.
Otherwise this layer extends towards infinity. The planar structure is always orthogonal
to the global z-axis, the vertical position can be modified by specifying the Z-value at the

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-48

top of layer 1. Planar substrates may be used without ground planes in complex (layered)
models (for example models of real earth); with a ground plane at the bottom (for example
in microstrip applications) and with both conducting top and bottom planes (for example
in stripline applications).

Figure 6-69: The infinite planes dialog for the application of ground planes and planar substrates

The infinite ground plane is displayed in the 3D model view.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-49

6.8 Symmetry

CADFEKO allows the specification of planes of symmetry that may be used to accelerate and
reduce the memory requirements for the solution of a problem.
To add symmetry to a model select the Solve/Run tab and click on the Symmetry button.

Figure 6-70: The Symmetry button

The Symmetry definition dialog is given in Figure 6-71.

Figure 6-71: The Symmetry definition dialog

During meshing CADFEKO will validate that the geometry to be meshed does indeed adhere to
the specified model symmetry (both geometric symmetry as well as symmetry of excitation and
loads where magnetic/electric symmetry is used). If the model is found not to adhere to the
specified model symmetry then CADFEKO will abort the meshing process, and provide a list of
the parts of the model that break the symmetry.
When writing the mesh to the *.cfm file (during save or exporting from the File → Export
→ CADFEKO Mesh dialog) the symmetry of the mesh is verified. If the mesh does not adhere
to the symmetry (both geometric and excitation/loads), then the *.cfm file is written, but the
symmetry is excluded. In this case a warning is given with the reasons why the symmetry cannot
be applied.
It is possible that the geometry and/or mesh of a model appear to be symmetrical but the CAD-
FEKO symmetry tests indicate that this is not so. In this case it may be more appropriate to
delete half of the model, and use the mirror operation to ensure a symmetrical model rather than
trying to resolve each asymmetry. This is particularly relevant when working with imported CAD
geometry or imported meshes where there may be slight differences with respect to tolerances.
CADFEKO indicates the current symmetry of the model in the 3D view (as shown in Figure 6-72
for three planes of symmetry). This preview is always shown while the Symmetry definition
dialog is open, but may be deactivated by toggling the Show symmetry planes icon in the relevant
3D view. The preview is coloured according to the symmetry type (green for geometric symmetry,
orange for electric symmetry and grey for magnetic symmetry).
When using magnetic or electric symmetry planes, all loads, excitations, media properties etc.
must be defined such that the model properties are identically symmetrical. For geometric planes

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-50

Figure 6-72: Preview of symmetry planes in the 3D view

of symmetry, the geometry must be symmetric though symmetrical loading, excitation and media
properties are not required. It is important to note that geometry-linked ports (wire ports, wave-
guide ports, edge ports, line ports) may affect meshing and must therefore always have sym-
metrical counterparts around any defined symmetry plane, irrespective of the symmetry plane
type.

6.9 Periodic boundary conditions

Periodic boundary conditions may be specified in CADFEKO. Like symmetry, periodic boundary
conditions are considered a model setting and may be accessed by selecting the Infinite structures
tab and selecting Periodic boundaries from the dropdown menu.

Figure 6-73: The Periodic boundaries button

The dialogs for 1D and 2D periodic boundary conditions are shown in Figure 6-74.
The definition of the unit-cell of the periodic boundary condition solution is based on vectors. For
the 1D case, the start– and end–point of a single vector is required. Periodicity is then defined
based on two planes passing through these start and end points, and normal to the vector formed
between them. The vector used to define 1D periodicity can have any orientation, but must
have a non-zero length. For the 2D case, two vectors are required. These vectors form the two
boundaries of the unit cell which is infinite in the direction normal to the plane on which both
vectors lie (see Figure 6-75). The vectors that define the unit-cell for 2D periodicity must have
non-zero length, and cannot be oriented in the same direction.
CADFEKO allows the specification of the phase shift to be applied in the direction of each of the
vectors defining the unit-cell. The specified values for the phase-shift are only used if a plane-
wave excitation is not present. (If phases are specified and a plane wave excitation is present in
the solution, the FEKO kernel will return an error during solution.)

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-51

Figure 6-74: The Periodic boundary condition dialog (1D and 2D)

Figure 6-75: The Periodic boundary condition for 2D periodicity

For array modelling using periodic boundary conditions, the beam (squint) angle can be specified
by defining the theta and phi angle. The phase along the periodic lattice vectors will then be
computed automatically to ensure the specified beam direction.

6.10 Numerical Green’s function (NGF)

CADFEKO allows the numerical Green’s function (NGF) to be applied to a MoM problem consist-
ing of a dominant, fixed static part and a smaller dynamic part. A user may elect to save the static
interaction matrix to a *.ngf file. The CPU time may then be reduced by reusing the *.ngf file
to obtain solutions for various dynamic domains

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-52

The Numerical Green’s function may be activated by selecting the Solve/Run tab and clicking on
the NGF button.

Figure 6-76: The Numerical Green’s function button

Note that domain decomposition is recommended for MoM models consisting of a large static
part and a smaller dynamic part. The static part must be large in relation to the dynamic part to
obtain a reduction in CPU time.
The static parts of the MoM model is defined on the numerical Green’s function settings dialog
(see Figure 6-77) on the Solve/Run tab.

Figure 6-77: The numerical Green’s function settings dialog

6.11 Solution control and special solver settings

CADFEKO provides a number of options for the different solution techniques available in the
FEKO Suite. These options are advanced and the implications of each setting should be well
understood before use. The inappropriate application of these settings may result in poor result
accuracy or inefficient calculations. For most cases, the default solver settings are appropriate
and no changes are necessary before solving the model.
The solver settings may be accessed by right-clicking on Solution in the tree and selecting Solver
settings or by selecting the Solve/Run tab and clicking on the Solver settings button.

Figure 6-78: The Solver settings button

The General tab

On the General tab, setting Data storage precision to Single precision forces FEKO to store certain
memory-critical arrays in single precision. Single precision is recommended unless the FEKO

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-53

Figure 6-79: The Solver settings dialog.

kernel gives warnings and suggests that Double precision be used (this might happen for instance
at very low frequencies where increased accuracy is required).
For low frequency modelling, the low frequency stabilisation for MoM may be activated in the
Low frequency modelling group.
For large models, a significant amount of time can be saved if the solution coefficients are saved
during the solution phase. Note that for smaller models (where the run time is short), storage
of the solution coefficients is typically not required. Storage of the solution coefficients creates
large files that need to be maintained by the user.
The reading/creation of the matrix elements (*.mat), LU-decomposition (*.lud) and solution
vector of the linear equations (*.str) may be enabled in the Output files group.

The MLFMM / ACA tab

The MLFMM may be activated on the MLFMM / ACA tab by selecting Solve model with the fast
multipole method (MLFMM). For most cases the default settings are appropriate. The MLFMM
is typically used to solve electrically large, high frequency problems faster and with less memory
resources than the method of moments (MoM). If monostatic RCS is requested, solution time can
be decreased if the angular increment is less or equal to 5◦ . Note that the MLFMM can currently
not be used simultaneously with a multi-layer Green’s function in the same model. MLFMM may
be used with the hybrid FEM/MoM.
The parameters on the MLFMM / ACA tab are only enabled if solution using the MLFMM is
activated. The MLFMM is based on a hierarchical tree-based grouping algorithm, and FEKO

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-54

automatically determines the ideal number of levels for each model. If the solution does not con-
verge, advanced users may try to obtain convergence by modifying the Box size in wavelengths
option. A starting point of 0.23 is recommended and values should not be much less than this.
(The value specifies the box size in fractions of the wavelength.) Alternatively, the user can select
Advanced solver settings and set the number of iterations or the stopping criteria for the iterative
process or even select a different preconditioner. Any text field on this tab may be left empty.
It should be emphasised again that setting these parameters can have significant consequences,
both in accuracy and solution time — users without a working knowledge of the MLFMM or
iterative solvers should preferably only use the default settings. The behaviour of the MLFMM
can also be influenced by selecting the combined field integral equation, see section 6.11.1.
The adaptive cross approximation (ACA) may be activated on the MLFMM / ACA tab by selecting
the Solve model with the adaptive cross-approximation. The ACA is a fast method similar to the
MLFMM but is also applicable to low frequency problems or when using a Green’s function in
the model. If the solution does not converge, advanced users may try to obtain convergence by
selecting the Advanced solver settings and setting the number of iterations, stopping criteria for
the iterative process or even selecting a different preconditioner.
The advantages of the ACA is that it is done on the matrix level and has no low frequency
breakdown like MLFMM. It is not restricted to free-space Greens function like MLFMM and each
matrix-vector multiplication is very fast.

The FEM tab

The FEM tab is used to control the FEM (finite element method). The FEM is used if the model
contains tetrahedral mesh elements (see section 4.8). The following conductor properties of
metallic surfaces inside a FEM region are supported: skin effect, thin dielectric sheet, impedance
sheet and electrically thin surface coating. If the FEM and MoM are de-coupled, the imperfect
conductor surface may be on the FEM boundary. If Decouple from MoM (use FEM absorbing
condition) solutions is checked, the FEM region (the tetrahedral elements and any conducting
surfaces on their boundaries) will not influence the MoM solution. This decoupling may save
significant run time, but is only valid if the FEM and MoM regions are electrically far enough
apart.
Switching to First order elements reduces the required memory and run time as well as the
accuracy. To get the same accuracy using First order elements, the mesh size must be reduced
— normally such that the solution actually requires more memory and run time than the larger
mesh with second order elements. Thus switching to First order elements is only recommended
if the mesh is already very fine, for example to account for highly inhomogeneous media or very
complex geometry.

The High frequency tab

The High frequency tab provides settings for the high-frequency asymptotic methods available
in FEKO (Physical optics, the UTD (uniform theory of diffraction) and Geometrical optics (ray
launching)).
Any changes to the high frequency settings should be based on the physical considerations of
each model.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-55

Solution using PO, UTD or GO is activated for each face (see section 6.11.1). If Decouple PO and
MoM solutions or Decouple UTD/GO and MoM solutions are checked, the PO or UTD/GO struc-
tures are not taken into account when calculating the MoM currents. In this case, for example,
the input impedance of a horn fed parabolic reflector, where the horn is solved using the MoM
and the reflector using PO, will be the same as that of the MoM horn in free space. In general,
where the MoM and PO/UTD/GO regions are electrically far apart, and far field quantities are of
interest, decoupling the solutions may save a significant amount of memory and run time without
significantly affecting the results.
During calculations using the PO formulation, a large amount of the time is spent in determining
which surfaces are illuminated from each source. This information may be saved by activating
the Store / re-use shadowing information to speed up subsequent runs (re-use is only possible
if the physical geometry of the model is unchanged). Storage of the shadowing information,
however, does lead to huge *.sha files on disk.
The option Use symmetry in ray-tracing (when possible), is set by default for PO. If Export ray
file for post-processing is checked, FEKO outputs the rays used during the UTD or GO solution
process to a *.ray file. The ray-tracing information is also stored in the *.bof file and the rays
may be viewed in POSTFEKO. It should be noted that *.ray files can be very large, especially
when using a MoM / UTD solution that has not been decoupled and there are more than a few
hundred mesh elements in the MoM region.
The Max. no. of ray interactions field limits the number of reflections/transmition interactions
taken into account for each ray that is tracked during the solution process. For example, a
ray path that includes a double diffraction and one reflection has three interactions, or a path
that includes a diffraction and three reflections has four interactions. The options in the Select
ray contributions group allow selection of the special UTD ray interactions that are to be taken
into account. (Note that some combinations go hand in hand. For example, the contributions
of edge and corner diffractions in the UTD solution are usually of the same order and they
should be switched on or off together.) Selecting more contributions or increasing the number of
ray interactions will increase both accuracy and computation time. The chosen settings should
therefore be a compromise between accuracy and solution time.
For geometrical optics (ray launching), special ray launching settings may be accessed in the
Dielectric GO ray launching settings section. When Specify angular increment is selected, the θ
and φ angular resolution for ray launching and the increments for the parallel ray front in the U
and V direction may be manually specified. Though the runtime for a problem involving GO may
be decreased by using this option, it may influence the accuracy of the solution. Manual specifi-
cation of the angular increment should only be used in specific conditions after the implications
have been carefully considered.

The Preconditioner tab

Advanced users may elect to use the Direct sparse solver or the Iterative solver for FEM models
which do not give a solution with the default settings. For the iterative solver it is possible to set
the maximum number of iterations, termination criterion and a Preconditioner. The Stabilisation
factor only applies to the incomplete (Multilevel ILU / diagonal decomposition) preconditioner
and may improve convergence in critical cases. The value should be between 0 and 1. Note that
any field on the Preconditioner tab may be left empty.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-56

6.11.1 Setting solution parameters on faces

By default, FEKO will employ method of moments based (MoM) solution methods for all surfaces
(dielectric and conductive) in a model. Other solution methods may, however, be set individually
on each face in the model. For example, an area of the model may be solved with PO. The
solution method is specified on the Solution tab of the Face properties dialog for the relevant
geometry faces or on the Mesh properties dialog for the relevant triangle labels. (Properties set
on the geometry will be transferred to the mesh during meshing.)
The different solution methods available are mutually exclusive, and the choice between the
special solution methods is reflected in the Solve with special solution method drop-down list.
The available options are:

None: No special solution method is used and the face will be solved using the default, method
of moments solutions.

Physical optics (PO): The high frequency physical optics method is used when calculating the
currents on the face.

full ray-tracing With this method, a significant fraction of the solution time is spent on ray
tracing (determining which elements are visible from each non-PO radiating element).
The advanced PO options (see section 6.11) allow manual optimisation of ray tracing.
always illuminated This option is applicable when the model is such that the entire sur-
face will always be illuminated (such as for a basic reflector antenna analysis). Shad-
owing is not analysed, and non-PO or large element PO surfaces are considered trans-
parent for this purpose.
only illuminate from front This option is applicable to all closed bodies, but the normal
vector must point outward (away from the closed body) from each face. Smaller mesh
elements are required close to the shadow boundaries.

Large element PO: The high frequency large element physical optics method is used for large
smooth areas when calculating near and far fields. This method is only to be used when
there are no discontinuities in the incident fields (e.g. close to a point source). If the large
element PO is used together with the MoM method, the MoM and PO regions are to be
de-coupled. Note that the same options as given above for PO (full ray-tracing, always
illuminated etc.), are also valid for large element PO.

Geometrical optics (ray launching): The geometrical optics (ray launching) method is used.
This method is applicable to the simulation of thin dielectric or metallic objects, such as
radomes and dielectric lenses. The use of ray-launching GO is also supported for thin
dielectric sheets, coatings and anisotropic media. Note that the use of plane wave as an
excitation is also supported (required for RCS requests).

Uniform theory of diffraction (UTD): The UTD will be used for this face. This method is appli-
cable to flat electrically large PEC faces.

Windscreen: The windscreen will be used for this face. This method is applicable to the simula-
tion of multiple windscreen antennas on various glass definitions.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-57

Planar Greens function aperture: The planar Green’s function aperture modelling will be used
for this face. This method is applicable to the simulation of a slot/aperture in an infinite
electrical conductor.

When the properties are viewed for a selection that contains multiple faces, if any of these options
are different for the faces in the selection, that field is opened in ternary state. The Integral
equation field allows selection of the formulation to be used during the solution phase. By
default the Integral equation is set to Electric field (EFIE). This is the most general formulation
and can be applied to both open and closed bodies. The Combined field (CFIE) setting can only
be used on closed objects, and then it is generally more stable. Note that the triangle normal
should point away from the region with zero field. Where applicable, the CFIE is particularly
useful when used with the MLFMM as it generally requires less memory and convergence can be
much faster than the EFIE based solution. It is possible to use the Combined field setting on a
subset of the model, but the surface must be a closed PEC surface and all the normals must point
outward (away from the zero field region).

6.11.2 Viewing special solution parameters on faces

Keeping track of different solution parameters that have been set on different regions of the
geometry can become difficult. CADFEKO provides a View by solution tool, which allows high-
lighting of all the geometry with the same solution settings in the 3D view, while the display
of all other geometry is semi-transparent. This tool allows the user to quickly check that the
correct solution parameters have been set before solution. The View by solution tool (shown in
Figure 6-81) may be enabled by selecting the Solve/Run and clicking on the View by solution
button.

Figure 6-80: The View by solution button

6.12 Requesting results

Before running a simulation, the required results must be requested. The CADFEKO model con-
tains a list of requested calculations. These are listed in the tree under Requests. The same set of
results are calculated for all frequencies with the specified loads and excitations. The following
operations may be performed on individual requests:

• Requests that are represented in the 3D view may be hidden/shown.

• Requests that are represented in the 3D view may be included/excluded without deleting
them by selecting Include/exclude from the popup menu. Excluded solution requests are
indicated by the addition of a small red cross to the icon and will not be included in the
FEKO solution (excluded solutions cannot be used during optimisation).

• Requests may be copied (to create similar requests).

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-58

Figure 6-81: The View by solution parameters dialog.

• Requests may be renamed. (The names or labels of some requests are referenced in POST-
FEKO. OPTFEKO also uses the labels of solution requests to uniquely identify the results
applicable to optimisation).

• Requests may be deleted from the pop-up menu of the tree item, or using the <Del> key.

• Requests may be edited by double-clicking on them or selecting Properties from the pop-up
menu.

All results may be requested from the the pop-up menu on Requests or by double clicking the
appropriate header if other requests of this type have already been defined and are present in the
model.

6.12.1 Currents

The currents that should be provided in the output of the simulation (and therefore be available
for analysis or visualisation) are specified in the Request dialog (shown in Figure 6-83). This
dialog is launched by selecting the Request tab and clicking on the Currents button.

Figure 6-82: The Currents button

All currents: All currents will be computed and stored in the simulation output.

Only currents on specified labels: The currents on the segments and triangles with the specified labels
are computed.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-59

Figure 6-83: The Request currents dialog

Only segment currents: The currents on segments are computed.

Only triangle currents: The currents on triangles are computed.

Export currents to ASCII file (*.os/*.ol): The currents and charges that are stored in the solution output
are also written to a *.os and *.ol files respectively which can be read by other post-
processors.

Export currents to *.out file: In addition to the inclusion of the currents in the standard solution output
(used by POSTFEKO), the currents are written to the text *.out file.

Though currents must be included in the output to enable the display of current distributions,
in large models the unnecessary storage of currents can lead to large output files. When using
adaptive sampling (see section 6.2) the unnecessary storage of currents may also increase the
number of frequencies required for solution convergence. In general, only the currents that are
needed should therefore be stored.

6.12.2 S-parameters

The calculation of S-parameters between an arbitrary number of ports may be requested by

adding an S-parameter solution request. A S-parameter request may be added by selecting the
Request tab and clicking on the S-parameter button.

Figure 6-84: The S-parameters button

The Request S-parameters dialog (shown in Figure 6-85) contains a table where the user can
specify the different ports to be considered in an S-parameter extraction. For all ports that are
not waveguide ports, the reference Impedance can be specified (if no reference impedance value
is given, a default reference impedance of 50 Ohm is used). For waveguide ports, type (TE/T-
M/TEM), indices and rotation of the mode that is to be considered can be specified. The Add
and Remove buttons can be used to add or remove ports to the list. If the Active field of any port
in the table is checked, the port is used as a source. Otherwise it is only a receiving port. For

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-60

Figure 6-85: The Request S-parameters dialog.

example, if the list contains two ports and only Port 1 is active, FEKO will calculate S11 and S21
but not S12 and S22 . If Port 2 is also active, S12 and S22 will also be calculated.
The port numbers in an S-parameter solution request are indexed based on the order of appear-
ance in the port list on the S-parameter request dialog, and not on the label of the selected
port.
If Export S-parameters to Touchstone file (*.snp) is checked, the calculated S-parameters included
in the solution output are written to an additional *.snp file. When exporting S-parameters to a
Touchstone file it is important to note that FEKO does not normalise the S-parameter values to a
global reference impedance, but rather provides the values referenced to the impedance specified
on each port. Some tools that use the Touchstone format, however, assume that all values are
referenced to a common impedance. This may result in misinterpretation of the S-parameter
results provided by FEKO. When exporting S-parameters for use in a tool that supports only a
single reference impedance, the reference impedance provided at each port in the S-parameter
solution request should be set the same to ensure the correct interpretation.
During the calculation of S-parameters, FEKO adds loads of the specified reference impedances
at all the port locations. These loads normally remain in place after the S-parameter calcula-
tion. If Restore loads after calculation is checked the S-parameter loads are removed once the
S-parameter calculation is complete. However, restoring the loads require repetition of a full
matrix computation and LU decomposition step for the MoM. This is typically the most time
consuming step in the analysis.

6.12.3 Transmission/reflection coefficients

The calculation of transmission/reflection coefficients for a plane wave interacting with a planar
structure may be requested by selecting the Request tab and clicking on the Transmission/reflec-
tion button.

Figure 6-86: The Transmission/reflection button

The Request transmission/reflection coefficients dialog is shown in Figure 6-87.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-61

Figure 6-87: The Request transmission/reflection coefficients dialog

Plane position for phase reference: The position of the plane wave in Cartesian coordinates are defined
here.

Figure 6-88: A plane wave interacting with a planar structure

The transmission coefficient is defined as

Et
τ= (6-13)
Ei

The reflection coefficient is given as

Er
ρ= (6-14)
Ei

Note that only a single plane wave is allowed (i.e. no other sources are allowed in the model).
The model must either contain a:

• Planar multilayer substrate without any other geometry/mesh in the model or

• 2D periodic boundary condition (PBC)

6.12.4 Far fields

The calculation of far fields is specified by adding a far field solution request. Select the Request
tab and click on the Far fields button.
The Request far fields dialog is shown in Figure 6-90.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-62

Figure 6-89: The Far field button

Figure 6-90: The Request far fields dialog

The Position tab

On the Position tab, a selection can be made between two different far field calculation types.

Calculate fields as specified: This option provides for the extraction of general far field patterns. The
Start and End angles and the Increment for each angular axis can be specified. (All these
fields do accept point entry from the 3D view, though it may be counter-intuitive to enter 3D
angles from a planar surface.) CADFEKO calculates and displays the number of field points.
The actual end angle depends on the start angle, the number of points and the increment
and may not coincide exactly with the specified end angle value. In models that contain
plane wave sources, the direct contribution of these sources is always ignored. The preview
shows a spherical “surface” of the specified range, with markers on the surface, indicating
the actual calculation directions. Buttons are provided on the dialog that can be used to
quickly define some commonly used pattern requests.

Calculate fields in plane wave incident direction: This option is used in conjunction with a plane wave
excitation (see section 6.5.1), mostly to calculate monostatic radar cross section (RCS).
If this option is selected, the model must contain a plane wave excitation (with single or
multiple directions of incidence). If no plane wave excitations are defined, an error will be
reported during the solution. No additional parameters are required in the far field request,
as scattered fields are only calculated in the direction that the incident wave is coming from.
This option must be selected if an RCS optimisation search is based on this far field request.

When extracting large numbers of far field samples for both axes, the impact on the solution time
may be significant.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-63

When the number of requested far field points is larger than 1 for both angular axes, FEKO
computes (in addition to the far field values themselves) the integral of the Poynting vector (i.e.
the total radiated power) through the spherical segment defined by the start and actual end
angles in each angular direction.
A full φ-cut runs from 0◦ to 360◦ and a full θ -cut runs from 0◦ to 180◦ . In order to account for a
potential ambiguity when specifying such a closed surface representing a full pattern in spherical
coordinates, FEKO also computes the radiated power over a second sector that extends by half
an angular increment beyond the specified start and end points in both angular directions.
For the case where the φ and θ start and end angles exactly overlap, the power computation
in the first-sector correctly represents the radiated power in all directions. If, however, the first
and last angular directions do not exactly overlap, but are separated by a single increment angle
value, the power computed over the second sector represents the actual power over the full
pattern. (Both radiated power values — with and without the overlap-correction — are computed
and stored in the text *.out file below the far-field information. The actual ranges represented
by each of the power values is also indicated. The value that represents the intended angular
ranges should be taken — for example, when considering a full pattern, the value for which the
φ range is exactly 360◦ and the θ range is exactly 180◦ should be used.)

The Advanced tab

Directivity/Gain: A selection may be made between the calculation of directivity or gain (when not cal-
culating RCS). Note that both POSTFEKO and OPTFEKO can access both gain and directivity
results independent of this setting — this setting controls only what is written to the *.out
file.

Export fields to ASCII file (*.ffe): If this option is checked, the computed far fields are exported to an
*.ffe file. This file can be used by other post-processors or as a source pattern for the
radiation pattern point source (see section 6.5.3) or a receiving antenna (see section 6.12.8)
element.

Export fields to *.out file: If this option is checked, the fields are included in the text output *.out file.
This file is not used by POSTFEKO or OPTFEKO and is only necessary if the user would like
to view or extract the relevant data from the text output directly.

Calculate only the scattered part of the field: This option indicates that the radiated contribution from
impressed sources (such as electric and magnetic point sources) should (in addition to the
contribution from plane wave sources) be ignored, yielding only the scattered fields. Nor-
mally this option should remain unchecked.

Only determine radiated far field power by integration: If this option is checked, the far fields — and the
total radiated power — are calculated, but the field values are not written to the *.bof or
*.out output files (fields are still written to the *.ffe file if this has been requested). This
option should only be used if the individual field values are not required and the output files
would otherwise become too large. If this option is selected, the far field values will not be
accessible for viewing in POSTFEKO or for optimisation in OPTFEKO

Calculate spherical expansion mode coefficients: This can be checked to calculate these coefficients. The
coefficients can be exported to a TICRA *.sph file. Note that the gain in GRASP will only
be correct if the radiated power in FEKO has been set to 4π Watts. For more detail see
the description of spherical modes in the FF card (see section 14.37). Spherical mode

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-64

coefficients can also be imported from a *.sph file in the spherical mode excitation (see
section 6.5.4).

Calculate far field for an array of elements: If this option is checked, the far field can be calculated for
an array of elements. Optional is to set the Number of elements along vector 1 and Number
of elements along vector 2.

6.12.5 Near fields

The calculation of near fields is requested by adding a near field solution request. Select the
Request tab and click on the Near fields button.

Figure 6-91: The Near field button

The Request near fields dialog is shown in Figure 6-92.

Figure 6-92: The Position and Advanced tabs of the Near field request dialog

The Position tab

On the Position tab, the Start, Increment and approximate End positions for each axis as well as
the type of Coordinate system , must be specified.

Specify increments: A user specifies the Start, End and Increment. The Number of field points are then
automatically calculated.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-65

Specify number of points: A user specifies the Start, End and Number of field points. The Increment is
then automatically calculated.

The actual end position depends on the start position, the number of samples and the increment
and may not coincide exactly with the specified end position. The fields and units of the fields
are interpreted according the coordinate system selection on the coordinates tab (distances in
terms of the model unit and angles in degrees). CADFEKO dynamically calculates the number of
sample points in each direction and displays a preview of in the 3D view. This preview shows
the specified range as surfaces enclosing the specified volume, with markers at the actual sample
positions. For finely sampled near field requests, where there are too many markers, the markers
may not be displayed. (Note that using even moderate numbers in all three directions can quickly
result in a large number of samples.)

The Advanced tab

On the Advanced tab, a selection may be made between the calculation of fields or potentials and
special storage options may be set.

Field: The actual near field components are calculated and stored in the solution output. The
electric and/or magnetic field components may be included.

Potential: This option may be used for advanced field analysis. Only one potential type may be in-
cluded in a single near field request. The vector potential, or the gradient of the scalar
potential for either the electric or magnetic field may be selected. The data from near field
requests in which potentials are calculated cannot be viewed or processed in POSTFEKO.

The following options are common to both the Field and Potential near field request types.

Export fields to ASCII file (*.efe, *.hfe): Calculated electric fields/potentials are written to a *.efe file
and magnetic fields/potentials to a *.hfe file. These are ASCII files that may be used by
other post-processors.

Export fields to *.out file: If this option is checked, the fields are included in the human readable *.out
file. (The *.out file is not required for viewing in POSTFEKO or for optimisation in OPT-
FEKO.)

Export fields to SEMCAD *.dat file: If this option is checked, the fields are written to a *.dat file.

Calculate only the scattered part of the field: If this option is checked, the radiated contribution from im-
pressed sources (such as electric and magnetic point sources) is ignored, yielding only the
scattered fields. Normally this option should remain unchecked.

6.12.6 SAR

Specific absorption rate (SAR) studies typically require the average absorption over a volume
(the Volume-average SAR) or the maximum absorption in a 1 g or 10 g cube anywhere in a given
volume (the Spatial-peak SAR). The calculation of these quantities can be requested by adding a
SAR solution request. Select the Request tab and click on the SAR tab.
The Request SAR dialog is shown in Figure 6-94.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-66

Figure 6-93: The SAR button

Figure 6-94: The SAR request dialog

Select calculation: The specific SAR related quantity that is required is defined here.

Specify the search region: The region (or volume) is which the SAR should be calculated is defined here.
The available search regions are:

• Entire model : All dielectric regions in the model will be considered, and a single
average or peak SAR value will be computed.
• By medium : Only the set of regions with specific media types will be considered.
All, or only a specific medium may be selected. For All media average or peak SAR is
calculated separately for each medium, and thus it is not the same as selecting Entire
model.
• In a planar substrate : A specific layer (layer 0 is the upper free space region, layer 1
the uppermost dielectric layer, etc.) or All layers may be included. This option cannot
be used with volume averaging.
• At a specified position : The 1 g or 10 g cube SAR value may be calculated at a specific
position. This option cannot be used with volume averaging.

6.12.7 Cable harness

The Cable harness request may be used to calculate the coupling into a cable near a large con-
ducting surface (see section 4.6.4).
The results of a cable analysis calculation — the voltages, currents, impedance and power be-
tween the inner conductor and the shield — can be displayed in POSTFEKO. Note that even
though FEKO calculates the coupling into the cables, the cable geometry is not taken into ac-
count in any other calculations. The presence of the cable therefore does not influence any other
result — not even other cable analysis calculations.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-67

6.12.8 Ideal receiving antenna

The Ideal receiving antenna provides a tool by which the power that would be received by an
ideal antenna with a specified pattern can be calculated.
Select theRequest tab and click on the RX antenna button.

Figure 6-95: The RX antenna button

The Request ideal receiving antenna dialog is shown in Figure 6-96.

Figure 6-96: The ideal receiving antenna request dialog

The receiving antenna is reciprocal to the point source with a specified pattern (see section 6.5.3)
and the parameters on the dialog are identical.
Note that the following assumptions must be kept in mind when using the ideal receiving an-
tenna:

• The antenna is considered to be matched (i.e. no mismatch loss is taken into account)

• The antenna is assumed to exist only at one point

• The antenna and model are assumed to have no impact on each other during the solution
phase (no coupling is taken into account)

On the Advanced tab the option is available to include only the scattered part of the field.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-68

6.13 Error estimation

Error estimation is an a-posteriori error indicator which gives feedback on the mesh quality.
The mesh quality is determined by testing the solution against an unconstrained physical test.
Requesting of error estimates will enable the user to find areas which might require local mesh
refinement, (see section 4.8.4).
Select the Request tab and click on the Error estimation group.

Figure 6-97: The Error estimation button

A user may elect to request error estimates only on All mesh elements, Only error estimates on
triangles, Only error estimates on segments, Only error estimates on tetrahedra or Only error
estimates on specified labels.
The error estimate dialog is shown in Figure 6-98.

Figure 6-98: The Request error estimation dialog

Export error estimates to *.out file: In addition to the inclusion of the currents in the standard solution
output (used by POSTFEKO), the error estimates are written to the text *.out file.

Adaptive mesh refinement may be applied by reading the error estimates from an earlier solution.
Point refinement rules will be created in the areas where the errors are estimated to be the
highest.
An Adaptive refinement may be added by selecting the Mesh tab and clicking on the Adaptive
refinement button.

Figure 6-99: The Adaptive refinement button

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-69

6.14 Mesh manipulation

6.15 CEM validation of the CADFEKO model

CADFEKO provides a tool that performs a number of basic consistency checks on the model.
These checks are largely related to electromagnetic modelling rules and limitations. As meshes
may consist of millions of elements, CADFEKO does not perform continuous consistency valida-
tion of all of the mesh parameters, but requires that the user launch the validation tool manually
to check the model. EM validation of the model is generally advisable and all indicated prob-
lems should be addressed before the FEKO solver is launched. Considerable amounts of time and
frustration can be spared if modelling errors are identified as early as possible.
Select the Solve/Run tab and click on the CEM validate button.

Figure 6-100: The CEM validate button

As the electromagnetic validation for the various components are completed, the details are
displayed on the dialog. The following components are checked: Frequency, Geometry, Mesh
and Solution. As each component is validated, an icon is displayed giving the user information
regarding the status of the completed EM validation. When selecting a received warning or error,
a message detailing the warning/error is given in the Error/Warning details of the selected item
window.

Figure 6-101: The Computational electromagnetic validation dialog

6.16 Running the FEKO solver components

The CADFEKO model is saved as a *.cfx file. In order to solve a model built in CADFEKO,
two input files are required by the FEKO pre-processor. These are the *.cfm file (which con-
tains the mesh information) and the *.pre file (which refers to the *.cfm file and contains

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-70

information regarding the solution process and settings). When saving the model, CADFEKO
automatically exports the current mesh to the *.cfm file (if solution parameters are enabled (see
section 6.18.1), the *.pre file is also updated). CADFEKO also exports the current mesh to the
*.cfm (and *.pre) file each time a component is started from the Run menu.
If CADFEKO encounters any invalid settings, it writes error states to the *.pre file so that the user
will get the appropriate error message, even when running PREFEKO from another component
of the FEKO Suite.
Typically, once the model is set up in CADFEKO, PREFEKO is executed to create the FEKO input
file (*.fek) from the *.cfm and *.pre files. The geometry, excitations and solution requests
represented in this file can then be viewed and visually validated in POSTFEKO before running
FEKO or OPTFEKO.
Once the FEKO solution is completed, the results (stored in the *.bof file) can be viewed in
POSTFEKO, or in the text output file (*.out) .

6.16.1 Component parameters

The Component parameters item is situated on the Solve/Run tab. Click on the dialog launcher
button to display the Component launch options dialog. This dialog provides for the specification
of command line parameters for the various FEKO components. Note this dialog is the same for
CADFEKO, POSTFEKO and EDITFEKO.

Figure 6-102: The Dialog launcher button

The PREFEKO tab

If Treat errors as non-fatal. . . is checked, PREFEKO will continue running after encountering an
error. This is used to see all geometry modelling errors identified during the PREFEKO geometry
validation at once. (PREFEKO can also generate additional debug output, but these options
should only be activated to debug specific issues.)
Shown in Figure 6-103 is the PREFEKO tab on the Component parameters dialog. The follow-
ing options are available: Export of variables (names, values, comments), Debug options and
Advanced.
The Advanced field allows manually typing the required command line options as would be done
after the file name in a command shell. More detail can be found in the PREFEKO chapter (see
section 17).

The FEKO tab

If Only check the geometry on the FEKO tab is checked, FEKO will perform all the geometry
checks and exit before any computations commence. This option is usually used to check model
validity before solving the model on a separate machine.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-71

Figure 6-103: The PREFEKO tab on the Component parameters dialog

The user can also set priority of the FEKO run on this tab. If the priority is set to Low the run
will take slightly longer, but the computer will still be responsive for other work. (Note that for
parallel runs, all machines in the cluster operate at the speed of the slowest PC, so starting other
CPU-intensive jobs on one of the PCs in a cluster is generally not recommended.)
Users may elect to utilise their NVIDIA GPUs based on the Compute Unified Architecture (CUDA)
by checking the Use GPU (graphical processing) for NVIDIA CUDA devices checkbox.
GPU acceleration is supported for Windows and Linux platforms. Always ensure that the latest
NIVIDIA drivers are installed for the graphics card. More details regarding GPU acceleration and
CUDA and a list of supported graphics cards is available at http://www.feko.info/GPU.
Note that enabling GPU acceleration is only applicable if a compatible NVIDIA GPU device is
found. If no compatible NVIDIA GPU device is found, a warning will be issued during the com-
putation run and the setting is reverted to the default state of no GPU acceleration.
The Remote host information is used for remote execution. The Remote execution is enabled on
the Solve/Run tab. Additional information is also available in Section 18.2.1.
The Remote execution group deals with the parameters defining the settings used when launching
FEKO on another machine remotely. First one specifies which machine to use in the Remote
host (hostname or IP address) field and second the Remote execution method for the launching
mechanism can be chosen. Here two options are available:

• ssh/rsh: This is the method using a remote shell (either RSH or SSH or similar) for launch-
ing the process. For copying of the files SCP (or similar) will be used. Thus the remote
machine must be able to serve such connection attempts (i.e. a SSH daemon must be setup
and running with public key authentication). This method can be used between different
platforms.

• MPI: This is only supported between Windows machines (i.e. both machines must run a
Windows operating system), since this method uses native windows file copy methods and
a shared network folder on the remote machine for transferring the model files and results.
The launching on the remote machine is then done by the MPI daemon which is installed
already during the FEKO setup for parallel launching. Authetnication is done by Windows
internal mechanims, so the remote machine must be able to authenticate the current user
either against a domain or its local user database to grant access.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-72

Figure 6-104: The FEKO tab on the Component parameters dialog

Additional information is also available in Section 18.2.3.

The options under Parallel execution are used for the parallel version of FEKO when Run →
Parallel FEKO execution is selected on the Run menu.
The option Specify number of parallel processes along with the associated edit field specifies the
total number of parallel processes to be launched. If this is not active, then all entries in the
machines list (see below for the Configure button) with their specified number of processes will
be used for launching. If this option is active then only the given number of processes will be
started using the list from the beginning until the number is reached, regardsless if there were
more entries available.
If the Use shared memory/OpenMP threading for multiple CPU nodes is selected, the hybrid
MPI/OpenMP parallelisation is uitilised. It combines the benefits of OpenMP (implementation
of multi-threading) and MPI (communications protocol used to program parallel computers).
Processes are distributed over multiple machines/nodes by means of MPI. Multi-threading is
then used on each node for subsequent parallelisation and memory saving.
Clicking Configure in this group allows specification of the machines to be used for the parallel
solution and the number of processes to run on each. The order of the entries is of importance
as this order is used upon starting the processes (i.e. the processes are started on the machines
from top to bottom).
The options Full CPU report with runtimes for individual processes, Output MFLOPS rate of each
process (with network communication time) and Network latency and bandwith enable addi-
tional diagnostic tests and output and should be disabled for normal FEKO runs to not degrade
performance.
In the Parallel authentication method group one selects the mechanism to be used for authenti-
cating the parallel processes on the individual machines. Options are:

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-73

• Use encrypted credentials in registry (Windows only): This option uses previously stored
encrypted username and password from the Windows registry. The user has to save these
credentials previous to starting any parallel computation using the tool Update parallel
credentials (MPIregister) provided with the FEKO installation (START → All Programs →
FEKO → Suite x.y). This is a per-user setting and has to be updated on each change of the
user’s password. If using remote-parallel launching, this must also be done one the remote
host where then parallel FEKO will be started from.

• Use SSPI (Active Directory) integration (Windows only, requires domain): If selecting
this option all the machines must be member of a Windows (Active Directory) domain
and also the user accounts must be domain accounts. Then the authentication will be
carried out using internal functions of Windoes without having to encrypt anything into
the registry. Note that maybe additional (one time) configuration settings have to be
done (by the domain administrator) to prepare the Windows domain for this kind of au-
thentication requests (see the Intel MPI (User Authorization — Active Directory Setup)
and/or MPICH2 (Runtime Environment — Security) documentation shipped in the direc-
tory mpi\<mpi-version>\doc of the FEKO installation directory for details).

• Local run only (no authentication required): This option signals to FEKO that all processes
run on the same local machine and thus no authentication is required at all. This is most
commonly used if running parallel FEKO only locally on a multicore or multiprocessor
machine.

• Default (rsh/ssh for UNIX, registry for Windows): Selecting this option will always use
the default authentication method for the target operting system. For UNIX systems the
public key authentication of rsh/ssh will be used and for Windows the registry method is
considered the default.

Additional information is also available in Section 18.2.2.

The Advanced field allows manually typing the command line options as would be done after the
file name when launching FEKO from a command shell. The installation of remote launching as
well as the parallel version is covered in the FEKO Installation Guide. More detail can be found
in the FEKO section (see section 18).

The Utilities tab

On the Utilities tab, the user can select if the temporary files created during OPTFEKO, TIMEFEKO
and ADAPTFEKO runs should be deleted or kept (in the case of OPTFEKO, the files relating to the
optimum model and solution are not considered temporary files and are not deleted), shown in
Figure 6-105, TIMEFEKO and OPTFEKO.
The following options can be set on the Utilities tab: ADAPTFEKO, TIMEFEKO and OPTFEKO.
The Restart analysis number option for ADAPTFEKO can be used if the run was discontinued
(and the temporary files were not deleted). The solution can be restarted at the number of the
first unfinished model.
The Restart from solver run option for OPTFEKO can be used if the optimisation process was
terminated or interrupted and no temporary files were deleted. The optimisation can be restarted

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-74

Figure 6-105: The Utilities tab on the Component parameters dialog

at the number of the last completed optimisation iteration. No changes whatsoever may be made
to the model before restarting the optimisation process.
The Farming out kernel runs section provides a Configure button that launches the Machines
configuration dialog. This can be used to set up the machines and processors that should be used
for an optimisation run where farming out of the kernel solutions (see section 20.4) is used.

The Environment tab

The Environment variables field may be used to enter environment variables which can control
the FEKO solution.

6.17 Additional solution configurations

The Remote FEKO execution, Parallel FEKO execution and Farm out by OPTFEKO can be toggled
on or off by selecting the respective configuration on the Solve/Run tab.

Figure 6-106: The (a) Remote button, (b) the Parallel button (c) the Farm out button

If activated, a check mark will be displayed in front of the configuration method. When the
combination of Parallel FEKO execution and Farm out by OPTFEKO are both selected, farming
will be used during optimisation whilst using parallel FEKO runs.

6.18 Working with CADFEKO models in EDITFEKO

CADFEKO can control the entire solution configuration and it should generally not be necessary
to work with the *.pre file directly. Some advanced features, however, are only available in the
EDITFEKO interface and more flexible solution configurations can be realised by direct editing of
the *.pre file.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-75

Users may prefer to continue using EDITFEKO for existing models, particularly those with com-
plex control settings. Even though CADFEKO can create new *.pre files, it cannot import set-
tings or card-based geometry from an existing *.pre file. If a *.pre file created by CADFEKO
is modified outside of CADFEKO, then CADFEKO will relinquish control of the *.pre file and it
will have to be maintained outside of CADFEKO from that point onwards.
The rest of this section assumes that the user is already familiar with EDITFEKO (see section 11).

6.18.1 Disabling the CADFEKO solution configuration

As mentioned above, CADFEKO supports old models (or cases where the user wants to edit
the *.pre file to create complex solutions) by working in a geometry only type mode. This is
activated by activated by selecting the Solve/Run tab and click on the Enable solution button.

Figure 6-107: The Enable solution button

Normally this item is disabled and all solution related operations are enabled. If a model is
loaded or saved in CADFEKO, it checks if the *.pre file has been edited outside CADFEKO. If this
is the case, the user can elect to ignore the changes made to the *.pre file or disable the solution
control in CADFEKO.
If the CADFEKO solution is disabled, the Enable solution menu will be enabled and unchecked.
All solution settings are unavailable and any existing settings are ignored when saving the model.
(They are maintained and can be re-instated by enabling the solution again.) Dielectric media
can be created, but their material parameters cannot be set.
Selecting Enable solution will re-enable all the solution settings and rename the existing *.pre
file to name_orig_x.pre (x is a number chosen to ensure that no existing file is overwritten).
After re-enabling the CADFEKO solution, all settings made exclusively in the *.pre file will be
ignored, and only solution components and settings defined in CADFEKO will be used in the
solution.

6.18.2 Setting units

If the solution configuration is disabled, FEKO expects all input to be in metres unless the SF
card is used in the geometry section of the *.pre file to specify the units. For example, if the
model was constructed in millimetres, an SF card with a 0.001 scale factor should be added to
the *.pre file.

6.18.3 Referencing elements

In EDITFEKO, properties can be set on specific elements using their full labels (see section 4.4.4).
Segments have the label of the edge (typically called Wire. . .), triangles that of the face (typically
Face. . .) and tetrahedra that of the dielectric region (typically Region. . .). Of course, these names
can be modified on the geometry or the mesh elements. See section 4.10.2 for more details.

July 2011 FEKO User’s Manual

DEFINING THE SOLUTION CONFIGURATION IN CADFEKO 6-76

Since setting sources or loads on wire segments require unique labels, CADFEKO exports the port
segments with unique labels. These labels are created by appending the port name to the wire
label. For example, if Port1 is located on the centre segment of Line1.Wire1, this segment will
be written with the label Line1.Wire1.Port1 while the remaining segments will have the label
Line1.Wire1. For vertex ports, the associated segment is the shorter segment connected to the
vertex — POSTFEKO can be used to check which segment was relabelled.
Since every region / face / edge has a unique label, a typical model may contain a large number
of labels. Setting properties such as losses then requires a large number of cards. To simplify
this, PREFEKO supports renaming multiple labels using the wild cards * and ? (where * expands
to any string and ? to any character). For example, renaming Wheel?.Face* to Wheel will
rename the faces Wheel1.Face1 and Wheel2.Face17 (and all others matching this pattern, but
not Wheel10.Face1 or Wheel1.Wire1) to Wheel. They can then be referenced with a single name.
After renaming these elements, the original names no longer exist, i.e. referencing items using
those names will not find any elements. Care should, however, be taken not to rename labels —
such as ports — which need to be unique.

6.18.4 Using variables and named points in EDITFEKO

When exporting the *.cfm file, CADFEKO includes all variables and named points with their
current values (i.e. all expressions used in the definitions of variables and named points are
evaluated and reduced to numerical values before inclusion in the *.cfm file). If requested in
the IN card settings, these variables and named points are then imported by PREFEKO and can
be referenced in the *.pre file — at any point after the IN card.
Note that all expressions used in the solution configuration in CADFEKO are evaluated and re-
duced to static numerical values before writing the *.pre file — hence changing these variables
by redefining them after the IN card in EDITFEKO will not have any effect on expression-based
solution configurations defined in CADFEKO.

6.18.5 Setting dielectric parameters

If the solution is disabled in CADFEKO, dielectric media may still be defined and applied to
regions or mesh elements in CADFEKO.
The parameters of each additional dielectric must, however, be specified by the manual addition
of a DI card — referencing the name of the dielectric specified in CADFEKO — to the control
section of the *.pre file. As dielectric names are not hierarchical (like the label names) and the
same dielectric can be applied to multiple, entirely separate regions.

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-1

7 Setting up optimisation in CADFEKO

CADFEKO provides a facility to allow the set up of an optimisation. This facility replaces the
EDITFEKO-based optimisation approaches that are available in previous FEKO releases. The
implementation is such that (though optimisation settings will have to be manually recreated in
CADFEKO) all existing FEKO models can be used as part of an optimisation process.
In this section, setting up of optimisation and various optimisation-related options in the CADE-
FEKO interface are considered. For further information regarding the optimisation work-flow, op-
timisation algorithms and the interpretation of the various optimisation-related settings, please
refer to the description of OPTFEKO (see section 20). An example of the usage of the optimiser
may be found in the Getting started manual.
It is important to note that adaptive or continuously sampled results (generated using ADAPT-
FEKO) cannot be used in an optimisation. Only results generated when using single or discretely
sampled frequency settings may be used.

7.1 Defining optimisation searches

The first step in setting up an optimisation is the definition of an optimisation search. A search
can be created in multiple ways:

1. Select the Request tab and click on the Add search button

2. Right-click on the Optimisation branch in the tree and select Add search.

Figure 7-1: The Add search button

When a new optimisation search is created, the Add optimisation search dialog will appear. In this
dialog, the optimisation method and method-specific settings can be chosen for the optimisation
search before the search is created. Once a search has been created, it is represented under the
Optimisation branch of the model tree. In order to change the optimisation search settings, the
Modify optimisation search dialog (as shown in Figure 7-2) may be opened by double-clicking on
the relevant search in the Optimisation tree, or by right-clicking on the search in the Optimisation
tree, and selecting Properties at any time after creation.

Multiple searches

Multiple searches may be defined in one model and are represented as individual branches in the
Optimisation part of the model tree. Only one optimisation search may be activated at a time (if
there is only one search defined in the model, then this search will always be active). The settings
and options chosen in each search are completely independent, and only the settings specified in
the active search will be saved to the *.opt and *.pfg files for use during an optimisation run.

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-2

Figure 7-2: The optimisation search settings dialog

Figure 7-3: The Activate button

In order to activate a specific search, right-click on the relevant search in the tree and select
Activate optimisation search or select the Request tab and click on the Activate button.
The active search is indicated by the addition of a green check-mark to the icon representation
of the search in the tree.

7.1.1 Setting the optimisation method and stopping criteria

There are 5 methods that can be chosen for a specific optimisation search. These are:

Automatic: A method will be automatically chosen by the optimiser.

Simplex (Nelder-Mead): A gradient-based or ‘hill-climbing’ method (section 20.2.1).

Particle swarm optimisation (PSO): A swarm-based global search method (section 20.2.2).

Genetic algotithm (GA): An evolutionary global search method (section 20.2.3).

Grid search: This method searches over a pre-defined grid of parameter sets (section 20.2.4).

For the Automatic, Simplex, PSO and GA methods, two options that will influence the stopping
criteria of the method may be chosen. The first option is Specify the maximum number of solver
runs. If this option is activated, then the optimisation process will be terminated once the FEKO
solver has been launched the specified number of times during the optimisation process (note
that for the PSO and GA methods, the optimisation may terminate before the indicated number
of solver runs when a full swarm or generation cannot be generated within the remaining num-
ber of allocated runs). The second termination option is the Optimisation convergence accuracy
(standard deviation) option. This option allows the level of accuracy required for the optimi-
sation process to be adjusted. There are three options, High, Normal and Low. Each of these
options modifies the conditions under which the search algorithm will converge, and the effect

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-3

is dependant on the method chosen (for more information regarding the effects of this option on
each method see section 20.2). The default accuracy for all methods is Normal.
For the PSO and GA optimisation methods, the seed value can be set for random generation (the
Advanced tab of the Modify optimisation search, shown in Figure 7-2).
The following options are available:
Default: The seed value is set equal to a fixed default.

Generate random seed: The seed value is set equal to a random integer number.

Specify seed value: The seed value can be entered as any number greater or equal to 1.

When an optimisation process is prematurely terminated due to the limitation placed on the num-
ber of solver runs, the optimum solution found up to that point, as well as all other information
regarding the optimisation process will be available.
For the special case of the Grid search method, the termination criteria available for other meth-
ods are not relevant. If this method is chosen a single option is provided. The Default number of
points setting specifies the default number of grid points that are to be used for each optimisa-
tion parameter in the predefined grid. This value will be used for each parameter for which the
Grid points field in the Optimisation parameters dialog for that search is left empty (parameter
settings are described in section 7.1.2).

7.1.2 Defining optimisation parameters

Parameters define the boundaries on variables of the model that can modified during an optimi-
sation search. Select the Request tab and click on the Parameters button.

Figure 7-4: The Parameters button

The parameters are local to each optimisation search and are defined in the Optimisation param-
eters dialog (shown in Figure 7-5). This dialog is accessed by double-clicking on the Parameters
entry in the tree under the relevant optimisation search. A valid optimisation search must have
at least one parameter defined. (It is important to note that all settings, including the parameter
selections and settings are local to each search.)
Using the Add and Remove buttons, parameter definition entries may be added or removed from
the parameter list.

Parameter names

Each parameter name is chosen from a drop-down list containing all of the Variables defined in
the CADFEKO model. Once selected, the parameter inherits the name of the variable and if the
name of the variable is changed, the parameter name will be updated accordingly. (The definition
of optimisation parameters based on variables that are only referred to in EDITFEKO card-based
geometry and/or solution settings that have been manually added to the pre-file is possible. An
identically named CADFEKO variable must, however, first be added to the CADFEKO model —
information regarding the creation of variables in CADFEKO may be found in section 4.2.1.)

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-4

Figure 7-5: The Optimisation parameters dialog

Parameter boundaries

For each parameter a Min value and Max value must be specified. These may contain any valid
CADFEKO expression given in Table 4-1 and may be a function of any other variables defined
in the CADFEKO model provided that the expression does not imply a dependency between two
optimisation parameters in the same search. Any dependency between parameters will result in
an error when the settings are applied. The offending expression field/s will be highlighted and
details regarding the dependency provided in the message window (see section 3.1). The evalu-
ated maximum value expression must be larger than the evaluated expression for the minimum
value.
In addition to the required minimum and maximum values for each parameter, a Start value and
Grid points value may be set for each parameter. The Grid points value specifies the number of
points to be used for that parameter when the Grid search method is used. If no value is chosen
here, then the default value specified in the Grid search method settings (see section 7.1.1) will
be used. The Start value specifies the first point that the optimiser should use for that parameter.
This is particularly useful if some information is available regarding the approximate location
of the optimum value for that parameter and the search can be started close to the optimal
value. Particularly when the Randomised search methods (PSO (see section 20.2.2) and GA (see
section 20.2.3)) are used, the starting value will have an effect on the optimisation process. If
no start value is chosen for a given parameter, the value in the middle of the specified range will
be chosen as the start value for that parameter. The specified starting point must lie in the range
between the minimum and maximum values.

Constraints between optimisation parameters

The Optimisation parameters dialog provides a Constraints tab (shown in Figure 7-6). Con-
straints may be set on any optimisation parameter, defining dynamic boundaries for that param-
eter during the optimisation process.

Parameter and constraint deactivation

For each parameter in the parameter list or constraint in the parameter constraints list, there is a
Use check-box that can be used to specify the inclusion of each specific parameter or constraint
in the optimisation search process. If the Use check-box for a specific parameter or constraint is

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-5

Figure 7-6: Setting Constraints on the Optimisation parameters dialog

un-checked, then that parameter or constraint will not be included in the *.opt or *.pfg files
and will not influence the optimisation search. If a parameter is deactivated, the value of the
variable as specified in the CADFEKO variables list will be used as if it were not defined as an
optimisation parameter. (It is important to note that all parameter and constraint settings are
local to each search. Deactivating a specific parameter or constraint in the parameter settings of
one search will therefore not deactivate that parameter or constraint in any other search.)

7.1.3 Defining optimisation masks

An optimisation mask may be defined by selecting the Request tab and clicking on the Add mask
button or by double clicking on Masks in the tree.

Figure 7-7: The Add mask button

The Add optimisation mask dialog (shown in Figure 7-8) allows the definition of the mask either
by importing values from an external text file (tab-, comma- or space-delimited) or by manually
entering the x and y coordinates of the mask points. No linkage is maintained to an external file
from which mask coordinates may have been imported, and if the file changes, the points must
be re-imported. When importing points from an external file, any existing coordinate definitions
in the mask are over-written by the imported data. The import therefore always starts from the
first coordinate point of the mask.
A graphical representation of the mask is shown. This is useful for validating that the mask data
is in fact as intended, particularly when working with large numbers of data points imported
from an external file.
The mask can be given a label that is used when referencing the mask in the optimisation objec-
tive of the optimisation goal(s). Note that masks are not specific to an optimisation search and
may therefore be referenced from multiple optimisation searches without having to redefine the
mask within each search.

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-6

Figure 7-8: The Add optimisation mask dialog

How masks are used for optimisation

An optimisation mask a set of values that form a continuous line (or trace). The mask is used to
determine the goal value for the FEKO simulations. All the points that are calculated by FEKO
that satisfy the criteria for the goal type and name will be added to a long array of values and
then compared to the mask. Three examples will be used to illustrate how mask are used during
optimisation.
Example 1: If a user wants to optimise his far field pattern at a single frequency using a mask, the
user needs to create the mask with the required shape and also create the far field request that
should be compared to the mask. The first point in the far field calculation will map to the first
point in the mask and the last point in the far field calculation will map to the last point in the
mask. All other points of the far field will be compared to points in the mask (linear interpolation
is used to ensure a continuous mask).
Example 2: If a user want to optimise the antenna gain in a predefined direction to have a gain
profile that can be described by a mask, the user needs to create the single point far field request
and the mask. The gain at the first frequency will now map to the first point in the mask and the
last frequency will map to the last point in the mask. The gain at the frequency values within the
range will be compared to values in the mask using linear interpolation.
Example 3: Using a combination of the two examples above the user can create a complex
optimisation requiring a predefined gain pattern that changes as a function frequency. The user
needs to create the multi-point far field request and create a mask such that the first point of the
mask maps to the first point in the far field request at the first frequency and the last point in the
mask maps to the last point of the far field request of the last frequency. Suppose the required
far field pattern does not change over frequency, then the mask will have the same required far
field pattern repeated several times (once for every frequency).

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-7

It is clear that very complex optimisations are possible using masks. Care should be taken to
create the correct mask for the optimisation since the optimisation will not fail due to an incorrect
mask, but the user will not get the optimum result that was expected.

7.1.4 Setting up optimisation Goals

Within each optimisation search, Goals must be defined that specify the desired state that the
optimisation process should attempt to achieve by varying the specified parameters within the
bounded region defined by the parameter definitions (see section 7.1.2). Though multiple Goals
may be defined, a valid optimisation search must contain at least one Goal definition. Optimisa-
tion Goals are local to each search, and are created under the Goals branch of any optimisation
search.
Optimisation with FEKO makes provision for multiple Goals in each search. These Goals can be
related in various ways to each other in order to define the required result for the optimisation
search. The definition of the set of Goals for an optimisation is very flexible, and there are many
options available.
A Goal may be created in one of 2 ways:

1. Right-click on the Goals branch in the relevant search in the Optimisation tree and select
the Goal type required

2. Select the Request tab and click on the Add goal function button.

Structure of an optimisation Goal

All optimisation Goals (irrespective of type) have the same basic structure. In this section the
generic concepts and features of all Goals are considered.
Goals are divided into four basic parts.

Focus: The part of the FEKO solution that is to be considered for optimisation. The Focus is
based on a quantity that is computed directly by the FEKO solver, and is uniquely identified
based on the Request label.

Processing steps: A number of conversion steps or mathematical operations that are to be car-
ried out on the Focus before the Goal is evaluated. Some processing steps are specific to the
Focus and Goal type, while other processing steps are generic to all Focus and Goal types.
The number, order and type of processing steps can be freely chosen by the user to provide
flexibility in the Goal definition.

Objective: The Objective describes a state that the optimisation process should attempt to achieve.
The Objective is predefined and assumes the same unit as the Focus.

Operator: The Operator indicates the desired relationship between the Focus and the Objective.

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-8

The different Goal types are differentiated only by the Focus (and the focus-specific options). The
Operator and Objective options as well as many of the processing steps that may be chosen are
identical for each Goal type.
Before considering the variations that are available for each Goal type we will consider the op-
tions and concepts that are common to the different Goals.

The Request label

The Request label for each Goal uniquely identifies the simulation results that are to be consid-
ered during the evaluation of the Goal. Every solution request and excitation that is included in
the FEKO model can be assigned a label.
In the CADFEKO Solution tree, the name (label) of each calculation request and excitation that
has been defined is displayed and can be changed by the user. When defining a Goal, the relevant
label may be typed into the field. In addition to the text-entry option, the labels of the calculations
or excitations defined in CADFEKO that may be referenced in a Goal of that specific type are
included in the Request type drop-down list. If the Request label is selected from this list, then
CADFEKO will maintain the reference even if the name of the calculation or excitation is changed
in the Solution tree.
The labels of calculations and excitations defined directly in the *.pre file (and not in CADFEKO)
may also be typed into the Request label field. The label of a calculation or excitation defined
in EDITFEKO is indicated by a comment field (see section 13.1) added at the end of the line in
which the card is defined (see section 11.4.1). No validation of the correct solution type or label
validity is performed when a label is typed into the Request label field, and any changes must be
manually maintained.
It is important to note that all data in the output that is of the correct type and with the indicated
label will be included in the Goal evaluation, irrespective of frequency. When defining calcula-
tions and excitations in CADFEKO, it is not possible to have two entities of the same type with the
same label. In EDITFEKO, however, there is no limitation in this regard. When solving the model
at multiple frequencies, the data at all frequencies will be considered in the goal evaluation.
After the first optimisation iteration (during the goal evaluation phase), if no data of the correct
type with the correct label is found in the FEKO output (or if the settings in the data request are
such that the data relevant to the goal setup was not calculated) an error will be returned.

Focus processing options

The following processing steps are common to all Goals.

No processing: Where the Focus is non-complex, no processing steps are required. In order to
consider the Focus directly, the No processing option is provided.
x→x

Real/Imaginary/Magnitude/Phase: Allows selection of a specific component of a complex Focus type

— for an array, the complex component of each array element is taken, deliv-
ering a non-complex array.
x → Re(x)/I m(x)/Phase(x)/M a g(x)

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-9

Unwrap: Unwraps a phase component — for a phase array, the whole array is considered
in the unwrap process. This operator can only be applied directly after phase.
(x → unw r ap(x))

Absolute value: Takes the absolute value — for an array, the absolute value of each element is
taken.
x → |x|

Average/Minimum/Maximum: Finds the average, minimum or maximum value of an array — this has
no effect on a single value.
x → ave(x)/min(x)/ma x(x)

Normalise: Normalises to the largest value in an array — for a single value, ‘1’ will be
returned.
x
x→ ma x(x)

Log: Takes the base-10 logarithm — for an array, the base-10 logarithm of each
element of the array is taken. This operator is only available for non-complex
values or arrays.
x → l o g10 (x)

Offset: Adds a specified non-complex value — for an array, the value is added to each
element of the array. This operator is only available for non-complex values or
arrays.
x → x +n

Scale: Multiplies by a specified scale factor — for an array, each element of the array
is multiplied by the scaling factor.
x → nx

Exponent: Applies an exponent — for an array of values, the exponent of each value in
the array is taken.
x → xn

Undefined: When a processing step is modified so that a previously chosen processing

step becomes invalid, the processing step that is no longer valid reverts to an
Undefined state. All Undefined steps must be deleted or redefined before the
changes to the goal can be applied.

The Objective

The optimisation Objective may be defined in each Goal as a Single value or as a mask.

The Single value Objective This Objective is defined in the Value text-box. The error is evalu-
ated by comparing this value to the processed Focus value according to the defined Oper-
ator. Where the Focus remains an array after the processing steps have been applied, the
Objective value will be compared to each of the array values separately, and the cumulative
error will be extracted according to the Operator type by a summation of all of the errors.

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-10

The Mask Objective A 2D mask may be pre-defined (see section 7.1.3) and used as the objective
of an optimisation goal. This allows for the comparison of an array of calculated data with
a predefined array in the evaluation of the fitness of the optimisation step. This type of
objective would typically be used when a quantity that varies with position, observation
angle or frequency within one simulation result is to be optimised. The mask array need
not be the same length as the computed data array to which it will be compared as the
optimiser will use a piece-wise linear fitting on the mask array to determine the values for
comparison at the correct points (points as calculated according to the solution set up).

The Operator

There are five Operator types that are common to all Goals.

1. Equal - Indicates that the processed Focus should be equal to the Objective.
N
X
er r or = Focus(n) − Ob jec t ive (7-1)
n=1

2. Greater than - Indicates that the processed Focus should be greater than the Objective.

N 
X Focus(n) − Ob jec t ive for Focus < Ob jec t ive
er r or = (7-2)

n=1 0 for Focus ≥ Ob jec t ive

3. Less than - Indicates that the processed Focus should be less than the Objective.

N 
X Focus(n) − Ob jec t ive for Focus > Ob jec t ive
er r or = (7-3)

n=1 0 for Focus ≤ Ob jec t ive

4. Maximise - Indicates that the processed Focus should be maximized (no Objective is re-
quired for this Operator).

5. Minimise - Indicates that the processed Focus should be minimized (no Objective is required
for this Operator).

When a goal is evaluated, a single value error representation of the goal is extracted according
to the Operator type. When the Focus remains an array after the processing steps, an error is
evaluated at each point in the array, and the cumulative error is taken. For the comparative
Operator types (Equal, Greater than and Less than), where the relationship between the Focus
and Objective satisfies the operator, the contribution to the error representation will be zero.

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-11

Figure 7-9: The Impedance goal button

7.2 Optimisation Goal types

7.2.1 The Impedance Goal

The Impedance Goal provides for optimisation based on factors relating to the impedance and
admittance at any voltage or current source that is solved as part of the FEKO model. The Focus
is identified based on the label of a voltage source or current source in CADFEKO or of a card-
defined excitation of the type A1, A2, A3, A4, AF or AN in EDITFEKO.

Figure 7-10: The menu for the Impedance optimisation Goal

Focus type

Input impedance/Input admittance: Both of these are complex quantities that represent the load char-
acteristics (based on the currents and voltages at the source points). As these Focus types
always consists of complex values, the Focus processing options require that there be at least
one general processing step indicating the selection of one of the complex components.

Reflection coefficient (S11): The reflection coefficient is computed with respect to the indicated refer-
ence impedance. (Note that for the impedance goal, the reflection coefficient is computed
directly from the observed input impedance. This value is then in effect the ’active’ reflection
coefficient (Γ) and may differ from the S11 computed during an S-parameter calculation in
a multiport model.)

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-12

Transmission coefficient: The transmission coefficient (1 − Γ) is considered with respect to the indicated
reference impedance.

VSWR: The Voltage Standing Wave Ratio ([1 + |Γ|]/[1 − |Γ|] for the observed input impedance is
considered with respect to the indicated reference impedance.

Return losses: The return loss (−20l o g|Γ|) for the observed input impedance is considered with respect
to the indicated reference impedance.

Current: The current flowing through the segment on which the selected voltage source is located.
(Note that in order to use this optimisation goal, a port with an excitation (or an A1 card
(see section 14.4)) must be applied to the segment of interest. The excitation should be set
to zero magnitude, and referenced in the voltage source label field.)

Reference impedance

The impedance to be used during the calculation of the relevant Focus types can be indicated
here. The impedance must be a single non-complex value and is local to each impedance goal
(i.e. different reference impedances may be used for different impedance goals in the same
optimisation search).

7.2.2 The Near field Goal

Figure 7-11: The Near goal button

The Near field Goal provides for optimisation based on factors relating to all near fields computed
during of the FEKO solution. The Focus is identified based on the label of a Near fields request in
CADFEKO or of an NF-card in EDITFEKO.
The Near field Goal dialog is shown in Figure 7-12. There are three Near Field Goal-specific
options available in the choice of the Focus that are described below.

Field component

Electrical/Magnetic: The Electrical or Magnetic part of the near field must be chosen. If the data com-
puted based on the referenced label does not include the chosen part of the near field, then
an error will be returned during the evaluation of the Goal in the first optimisation iteration.

Coordinate system

The coordinate system in which the directional component of the near field is required must
be selected. The available coordinate systems are Cartesian, Cylindrical(X)/(Y)/(Z), Spherical
and Conical. This coordinate system selection defines the options available in the Directional
component selection list.

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-13

Figure 7-12: The menu for the Near field optimisation Goal

The coordinate system chosen here differs from the coordinate system chosen as part of the near
field computation request, in that the coordinate system choice in the near field Goal is related
to the near field component of interest, while the coordinate system chosen in the near field
request dialog is related to the positioning of the sample points for near field calculation. This
distinction makes it possible consider the near field component in any direction independently of
the physical placement of the near field sampling points.

Directional component

The options available in the Directional component list depends on the choice of Coordinate
system, but are independent of the near field request sampling point positions.
Radial or x/y/z/phi/theta-directed: In the chosen Coordinate system, the field in any of the 3 coordinate
directions may be requested. Each individual component of the electric or magnetic near
field is a complex quantity, and the selection of a specific field component requires that there
be at least one general processing step which indicates the selection of one of the complex
components.

Combined: In addition to the individual components in the coordinate directions, the Combined near
field value may be requested. This value is computed by combining all 3 directional compo-
nents of the field at each point as follows (shown for Cartesian components).
Æ
Fcombined = |F x |2 + |F y |2 + |Fz |2 (7-4)
The choice of Coordinate system has no effect on the value of the Combined component.
The combined field is always a non-complex value (or an array of non-complex values) and
it is therefore not required that any further processing be performed.

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-14

7.2.3 The Far field Goal

Figure 7-13: The Far field goal button

The Far field Goal makes provision for optimisation based on factors relating to all far field
quantities computed as part of the FEKO solution. The Focus result is identified based on the
request label of a Far fields request (see section 6.12.4) in CADFEKO or an FF-card request (see
section 14.37).
Fields that are requested in invalid directions (for example fields requested below an infinite
ground plane) are ignored during the Goal evaluation. If no valid far field results with the
correct request label are found in the solver output, an error will be generated during the Goal
evaluation phase of the first optimisation iteration.

Figure 7-14: The Far field optimisation Goal definition dialog

The Far field Goal dialog is shown in Figure 7-14. Goal-specific options for the Far Field Focus
choice are described below.

Focus type

E-field: The E-field Focus type considers the radiated fields associated with specified far field solution
request directly. The fields are considered according to the settings of the far field request
(for example if only the scattered fields from a single object are requested, then only these
will be taken into account in the Goal evaluation).

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-15

Directivity and Gain: With this Focus type, only the gain/directivity of the model is considered. This
option can only be based on a far field request where the Calculate fields as specified option
is chosen and is independent of the Directivity/Gain selection in the far field request.

Radar cross section (RCS): This Focus type is only valid for far field solutions that have been computed
with an exclusively plane-wave excitation. The RCS Focus type delivers non-complex val-
ues (or an array of non-complex values) representing the derived RCS (see section 19-8)
according to the options set in the far field calculation request. If no valid RCS information
is found in the computation output, an error will be generated during the Goal evaluation.

Polarisation

The Polarisation option allow specification of the component of the far field to be considered in
the Goal.

Horizontal (Phi) / Vertical (Theta): These options allow specific selection of the θ - and φ-directed com-
ponents of the far field. For the Gain and Directivity Focus types, only the component of
the field with the selected polarisation is used in the calculation of the required quantity,
delivering a non-complex value (or array of non-complex values).

LHC / RHC: These options allow specific selection of the left-hand-circular and right-hand-circular com-
ponents of the far field (as defined in Equations 19-4 and 19-5). For the Gain and Directivity
Focus types, only the component of the field with the selected polarisation is used in the cal-
culation of the required quantity, delivering a non-complex value (or array of non-complex
values).

S / Z: These options allow specific selection of the S- or Z-polarised components of the far field
(as defined in Equations 19-2 and 19-3). For the Gain and Directivity Focus types, only
the component of the field with the selected polarisation is used in the calculation of the
required quantity, delivering a non-complex value (or array of non-complex values).

Axial ratio: The Axial ratio option is only available for the E-field Focus type. This provides the ratio
between the magnitudes of the θ - and φ-directed field components (as described in sec-
tion 19.6 and specifically Equation 19-22). For the purposes of optimisation, an additional
sign is added to the Axial ratio value considered by the optimiser. The sign indicates the
handedness of the radiated field, with a negative sign implying left-handedness, and a posi-
tive sign implying right-handedness. This makes provision for the inclusion of the required
handedness directly in the Axial ratio optimisation.

Total: For the E-field Focus type, the Total option provides a magnitude combination of the θ - and
φ-components of the far field. The total field is calculated as:
Æ
Etotal = |Eθ |2 + |Eφ |2 (7-5)

This value is representative of the power in the far field.

For the Gain and Directivity Focus types, the polarisation-independent quantities are con-
sidered. This is the only Polarisation option for RCS.

7.2.4 The S-parameter Goal

The S-parameter Goal dialog is shown in Figure 7-16. The S-parameter-specific options for the
Focus choice are described below.

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-16

Figure 7-15: The S-matrix goal button

Figure 7-16: The dialog for the S-parameter optimisation Goal

Quantity

Coupling coefficient (Smn ): Only the coupling between different ports will be considered in the optimi-
sation (all S-parameter values where the port indices are not equal). It is important to note
that if Snm and Smn are computed in an S-parameter solution, then both of these values
will be considered in the Goal evaluation. If the consideration of only the coupling in one
direction is required, the relevant port should be deactivated in the S-parameter calculation
request.

Reflection coefficient (Snn ): Only the reflection at the port/s will be considered in the optimisation. The
reflection coefficient at all ports that are active for the S-parameter computation will be
considered.

Return loss: The return loss at the port/s will be considered in the optimisation. Return loss is calculated
from the reflection coefficient at each active port as:

RL = −20 log |Snn | (7-6)

Transmission coefficient: The transmission coefficient at the port/s will be considered in the optimisa-
tion. The transmission coefficient is calculated from:

γ = 1 − |Snn | (7-7)

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-17

VSWR: The voltage standing wave ratio at the port/s will be considered in the optimisation. Return
loss is calculated from the reflection coefficient at each active port as:

1 + |Snn |
VSWR = (7-8)
1 − |Snn |

Port selection

Specify input port number (n): By default, all of the active ports will be considered during the Goal
evaluation. When activated, this option allows the selection of a single port to be used as
the input port. For example, if all of the S-parameters in a 3-port device are computed and
the Focus quantity is chosen as Coupling coefficient (Smn ), then if the input port is specified
as 2, only the values of S12 and S32 will be considered during the Goal evaluation.

Specify output port number (m): In a similar manner to the input port selection option, this option
(when selected) allows the selection of a single port to be used as the output port. For
example, if all of the S-parameters in a 3-port are computed and the Focus quantity is cho-
sen as Coupling coefficient (Smn ), then by specifying the output port as 2, only the values of
S21 and S23 will be considered during the Goal evaluation.

7.2.5 The SAR Goal

Figure 7-17: The SAR goal button

The SAR Goal dialog is shown in Figure 7-18. The SAR Focus delivers a non-complex value (or
an array of non-complex values) based on the FEKO solution. Besides the general optimisation
goal options (see section 7.1.4), there are currently no SAR goal-specific options.

7.2.6 The Power Goal

The Power goal allows the user to optimise the total antenna efficiency, total power and power
loss. The power goal does not accept any request name since the operates on the total power in
the model. Besides the general optimisation Goal options (see section 7.1.4), there are currently
no Power goal-specific options.

7.2.7 The Receiving antenna Goal

The receiving antenna goal allows the user to optimise receiving antenna efficiency, received
power or power loss. Besides the general optimisation Goal options (see section 7.1.4), there are
currently no Power goal-specific options.

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-18

Figure 7-18: The dialog for the SAR optimisation Goal

Figure 7-19: The Power goal button

Figure 7-20: The RX antenna goal button

Figure 7-21: The Transmission/reflection goal button

7.2.8 The Transmission/reflection Goal

The Transmission/reflection Goal makes provision for optimisation based on factors relating to all
transmission/reflection coefficients quantities computed as part of the FEKO solution. The Focus
result is identified based on the request label of a Transmission/Reflection coefficient request (see
section 6.12.3) in CADFEKO or a TR-card request (see section 14.62) in EDITFEKO.
The Transmission/reflection Goal dialog is shown in Figure 7-22. Goal-specific options for the
Transmission/reflection Focus choice are described below.

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-19

Figure 7-22: The Transmission/Reflection optimisation Goal definition dialog

Focus type

Transmission: The transmission coefficient is calculated as

Et
τ= (7-9)
Ei

Reflection: The reflection coefficient is calculated as

Er
ρ= (7-10)
Ei

Polarisation

The co-polarisation and cross-polarisation is available for optimisation.

7.3 The global Goal: combining and weighting of multiple Goals

When multiple goals are included in a search, in addition to the individual goal definitions, the
way in which the goals should be combined during the evaluation of the global goal must be
defined. CADFEKO provides two mechanisms by which the evaluation of the global goal can be
controlled.

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-20

7.3.1 Goal weighting

The first mechanism is a weighting that can be defined for each goal on the Goal definition dialog
(see section 7.1.4). This weighting is used to modify the contribution of the goal’s error to the
global error during the fitness evaluation. The global error in each level of the tree is computed
by taking the evaluated error of each goal, multiplying it by the indicated weighting factor, and
then summing all of the resultant weighted errors in each branch-level of the tree.
The weighting of each goal is shown in brackets next to the tree representation of the goal.

7.3.2 Goal combination tools

A goal combination tool is provided, which allows the extraction of a single error value from a
set of goals. The extraction type can be chosen as Maximum, Minimum or Average. When a set
of goals are combined using this tool, only the minimum, maximum or average value of all of the
errors of all of the goals in the set is taken.
In order to combine goals using a combination tool, goals in one search in the same tree level
should be selected. The Combine goals tool can then be selected from the Optimisation toolbar or
menu. The Combine goals dialog (shown in Figure 7-23) is launched in which the combination
method can be chosen. Goals may be added to an existing combination by right-clicking on the
combination in the tree, and selecting the type of goal to add. Goals can be removed from the
combination by deleting them. If all goals in a combination are deleted, then the combination is
automatically deleted as well. Goals can be copied out of a combination to the root of the goals
tree by right-clicking on the goal and selecting Copy. The original goal can then be deleted.

Figure 7-23: The goal combination tool dialog

The Average, Minimum and Maximum options define how the evaluated errors of the goals in the
combination should be reduced to one error value. For example, if Average is chosen, then the
average error of the goals in the combination will be returned, while if Maximum is chosen, the
maximum error will be returned. Each combination can be assigned a weighting which indicates
how the error should be combined with other goals and combinations in the same level of the
tree during the global fitness evaluation. The combination tools may be nested to as many levels
as required.

7.4 Specifying special optimisation solver settings

Special options relating to OPTFEKO can be set by selecting the Solve/Run tab and clicking on
the dialog launcher button, the Utilities tab under OPTFEKO.

July 2011 FEKO User’s Manual

SETTING UP OPTIMISATION IN CADFEKO 7-21

Restart from solver run n: If this option is selected, then OPTFEKO will attempt to restart the optimisa-
tion process from the indicated iteration number (n). The optimisation can only be restarted
if the temporary files have been kept during a previous optimisation run. If solution files are
missing for a specific optimisation iteration, OPTFEKO will run the FEKO solver to recreate
the missing files. If any changes have been made to the model, solution setup or optimisa-
tion setup, OPTFEKO will ignore all existing results, and re-compute all results as required.

Delete all files (except optimum): If this option is selected (default), then all of the temporary files will
be deleted during the optimisation process. When the optimisation process is completed
(or if the optimisation process is interrupted), the original model, as well as the optimum
will be available along with all related simulation results. The optimum model and results
are indicated by the addition of _optimumat the end of the file name/s. If this option is
unchecked then no model or results files will be deleted during the optimisation process.

Farming out of optimisation solutions: This section allows the specification of the distributed computing
system when farming out of the solutions during an optimisation is required. The Configure
button launches a dialog in which the machines in the cluster as well as the number of
processes that may be launched on each of the machines may be specified (this is identical
to cluster configuration for parallel launching (see section 18.2.2)). Details regarding the
use of farming during optimisation may be found in section 20.4.

July 2011 FEKO User’s Manual

Part III

Working with POSTFEKO

INTRODUCTION TO POSTFEKO 8-1

8 Introduction to POSTFEKO

POSTFEKO is used mainly for two purposes: to validate meshed geometry and to analyse results.
Validation of mesh geometry is done so that users can confirm that their models are correct before
starting a simulation. This is particularly useful when models are created using EDITFEKO, but
is just as relevant for CADFEKO modelling. Analysis of results is the other primary function of
POSTFEKO. Once a model has been simulated, POSTFEKO can be used to display and review the
results. A variety of tools are available to help visualise data in a constructive manner.
Multiple models, with their geometry (in *.fek files) and results (in *.bof files), can be dis-
played in a single POSTFEKO project session. The displays are automatically updated each time
the model and/or results are updated.

Figure 8-1: The main POSTFEKO display. Here the geometry and 3D results of a single model is displayed.

8.1 Environment

The POSTFEKO environment (shown in Figure 8-1) consists of a ribbon toolbar at the top, a
viewing area in the middle, a project browser on the left and a control palette on the right-hand
side. The ribbon is used to control views, their contents, and tools that are used for interpreting
data. The viewing area can contain multiple 2D and 3D views, each with its own context-sensitive

July 2011 FEKO User’s Manual

INTRODUCTION TO POSTFEKO 8-2

ribbon tabs. The project browser (hidden by default, but can be shown by View → Window →
Project) provides additional information about the project and advanced details about models
and their contents. The result palette on the right-hand side allows the user to select the compo-
nents of a result that should be displayed. Results shown in views can be tuned to provide the
desired information by using this palette.
The ribbon is the main mechanism for manipulating views. Data is added to either 2D or 3D views
by selecting an enabled result from the ribbon menu. All system controls are contained here and
grouped for convenience. The different view types that may be created are 3D views, Cartesian
graphs, Smith charts and polar graphs. When any of these view types are in focus, the ribbon will
adapt to provide options that are relevant to that view type and the contextual tabs belonging
to the selected view are marked by a coloured bar above them. Whenever there are more items
available than shown in the ribbon group, there is a small indicator at the right bottom of the
ribbon group (besides the group title) which then opens a dialog window containing extended
controls.

Figure 8-2: Example of the FEKO ribbon.

Views are tabbed in the viewing area by default, but may be rearranged by altering the window
settings on the View tab. The context sensitive ribbon tabs give the user control over axis settings,
titles, captions and a variety of other view specific settings.

Figure 8-3: The POSTFEKO viewing area

July 2011 FEKO User’s Manual

INTRODUCTION TO POSTFEKO 8-3

On the right-hand side, the result palette is shown (see Figure 8-4). When a result is added to
a view, it is populated in a list at the top of the panel. The currently selected solution is also
displayed here. For a selected result, the slicing options are provided where the user can supply
more specific information about the type of data that should be displayed. Only properties which
are relevant for that result are populated in the palettes. The palette will then look different for
every type of result that can be displayed.
The Project browser panel may be displayed to provide more control over the project session (see
Figure 8-4). This project browser is not displayed by default. It provides additional information
in a tree format about the current project, the models in the project and specific components of
a model. The Project Browser panel lists all models that are loaded in the current project, as
well as all stored or imported data. The Model Browser shows all properties of a given model,
including details regarding all solution configurations, optimisation details and combined data.
The Details Browser shows in-depth detail of any component selected in the model browser.
It shows all relevant details regarding the entity settings (Note that no information about the
calculated results are given here).

Figure 8-4: The POSTFEKO project browser (left) and quantity palette (right).

8.2 Start page

When a blank instance of POSTFEKO is opened (i.e. no models are loaded), the start page
is displayed. It is also displayed when a project is cleared by pressing the New project but-
ton. The start page provides quick access to recent files, documentation and the FEKO website
(http://www.feko.info). Additional help can also be found by pressing the Help button.

July 2011 FEKO User’s Manual

INTRODUCTION TO POSTFEKO 8-4

Figure 8-5: The start page.

8.3 Application menu

The application menu is accessed by clicking on the green application button in the top-left
corner of POSTFEKO. It provides control over managing project sessions by creating new projects,
opening existing projects, saving projects and adding models to a project session. Rendering
options and updates settings can also be reached here, along with information about the current
POSTFEKO version and third party libraries. The Recent files list provides quick access to files
that have been used recently. Figure 8-6 depicts the application menu.

Figure 8-6: Application menu

The Rendering options are the same as those described in the section about rendering options
(see section 3.1.5). These settings help configure how views are drawn and whether the main

July 2011 FEKO User’s Manual

INTRODUCTION TO POSTFEKO 8-5

calculations should be done in software or on the system’s graphics card. Checking for updates
ensures that the most recent version of FEKO is installed on a system. More information on
configuring the updater can be found in the section about the FEKO updater (see section 23). All
default settings for creating new views, adding data to graphs and other display settings can be
modified under the default Preferences settings.

8.4 Ribbon tabs: tools and controls

The ribbon is divided into two classes, namely “default tabs” and “context-sensitive tabs”. Default
tabs contain controls and tools that are applicable to all view types and provide a central point
for controlling the application. For each view type, there also exist the context-sensitive tabs that
contain all controls and tools that are only applicable to that view. In general, the way in which
a user would like to interact with a 3D view is very different to how they would like to interact
with a graph. Context-sensitive menus make it easier to find those tools for a selected view.

8.4.1 Default tabs

The default tabs include:

• Home: Groups the most common actions together for quick access. The Home tab is a good
place to start for most basic operations, including: managing files and projects, adding new
views, adding results to views and running suite applications.

• Reporting: Tools for exporting data, images and animations are provided. The generation
of MS Word and MS PowerPoint presentations can also be managed from the Reporting
tab.

• View: All camera position settings can be altered here. Windows can be tiled, cascaded,
closed, etc. The visibility of the project browser and result palette can also be set. The
viewing angle can be set to focus on any origin, distance and angles. Figure 8-7 shows the
Transform view dialog that can be used to manually edit these properties.

Figure 8-7: View transformation dialog.

July 2011 FEKO User’s Manual

INTRODUCTION TO POSTFEKO 8-6

8.4.2 Context-sensitive tabs

Context-sensitive tabs depend on the view type and are unique for each 3D view, Cartesian
graph, polar graph and Smith chart. Depending on which view type is selected, a different
set of operations is made available. This means that only those actions that are relevant to the
current selection are available.

3D View context-sensitive tabs

The tabs that will appear when a 3D view is active are:

• Display: Affects the visualisation settings for a model. The visualisation of model entities
can be set, including infinite planes, legends, symmetry, periodic boundary conditions, etc.
Tools such as greyscale and cutplanes are also available. Views can also be duplicated or
exported from this tab.

• Mesh: All settings pertaining to the mesh visualisation can be set here. This includes colour-
ing and opacity options, as well as visibility settings for each mesh element type. Tools such
as distance measurement, searching for mesh elements, highlighting certain solution meth-
ods or checking mesh connectivity are also provided here.

• Result: Settings regarding a selected result are provided. This includes visibility settings
such as opacity, surface/grid visibility, colouring options, extrusion, etc. The requested
points can also be displayed for a result, arrows can be displayed and all ray display settings
are available. It is also possible to add new results to the view from the Result tab, duplicate
the selected result or to store a local copy of the result in the current project.

• Animate: Provides access to animation settings, including the animation variable (phase,
frequency, camera angle, etc.), rendering options and export settings.

2D View context-sensitive tabs

All 2D graphs (i.e. Cartesian graphs, polar graphs and Smith charts) have a Display and Trace
tab. The tabs differ slightly between the different graph types, but their main function stay the
same.

• Display: The Display tab allows the user to: duplicate views or generate equivalent copies
of a graph in a different format, edit chart text (e.g. titles or axis labels), set greyscale, edit
legend settings and modify axis and grid settings. In general, all settings pertaining to the
view can be set in this tab.

• Trace: Duplications, new math traces and local copies can be managed here. Also, the ren-
dering settings for the traces (i.e. the drawing order and sampling settings), independent
axis scaling and units can be changed.

• Measure: Cursors can be managed here that display the value of the traces on a graph.
Annotations can also be added to a trace from here, which update with the results and
can give additional information about a trace (such as beamwidth, bandwidth, maxima,
minima, etc.)

July 2011 FEKO User’s Manual

INTRODUCTION TO POSTFEKO 8-7

• Format: Provides access to font settings, colouring options (e.g. legend box or graph back-
ground colours), line settings, marker settings and legend drop shadow settings.

8.5 System operations

Along with the tools that manipulate displayed data, POSTFEKO also gives a set of tools that
work on a system level. These tools include saving and loading projects, animating results or
models in a 3D view, importing and exporting data (e.g. from measurements, calculations or
other simulations) and to store data for later use.

8.5.1 Application Button and quick access toolbar

The green application button is the collection of some most used system operations for easier
access and also contains recently used files, the options for printing, checking for updates, setting
rendering options and showing help and version information as well as closing POSTFEKO. A
detailed description follows in Section 8.3.
The quick access toolbar is located just above the application button and contains some opera-
tions such as (New project, Save, Undo and Redo) for one-click access without having to switch
the ribbon context.

8.5.2 Save and load project sessions

Saving can be done by clicking on the application button and choosing either Save or Save as,
depending on whether the file-name should be specified. The resulting project file will have a
*.pfs extension and will store all settings and references to result files that were present at the
time of save. There is also a quick save short-cut in the top-left corner above the application
button.
To “load” could either mean loading a saved project, or to load a model into an existing project.
The application button will provide both a Load model button as well as a list of recent files
that were accessed by the user. The Open project button on the Home tab will restore a saved
POSTFEKO project. Both models and projects can be accessed on the start page as well.

8.5.3 Preferences

In many cases, especially when it comes to styling, it is desired to always change a setting to
something other than the default. A Preferences tool is provided that allows custom settings to
be stored. This tool can be found under Application menu → Settings → Preferences. Figure 8-8
shows the dialog. A variety of options can be set, from default display settings to data sampling
and presentation settings.

July 2011 FEKO User’s Manual

INTRODUCTION TO POSTFEKO 8-8

Figure 8-8: Default preferences dialog.

8.5.4 Animation

In 3D views, it is often useful to see how properties change over time. It is possible to animate a
view or its results according to:

• Phase

• Frequency

• Phi rotate

• Theta rotate

• Theta & Phi rotate

For phase and frequency animations, results must be displayed that can be animated. For exam-
ple, currents or near fields may be animated over phase and any result that was simulated over a
continuous or discrete frequency range can be animated over frequency. Any 3D model may be
animated over the theta or phi angles, since this is purely a camera view animation and does not
require any results. Figure 8-9 shows the dialog where these settings are changed.

Figure 8-9: Animation settings dialog.

July 2011 FEKO User’s Manual

INTRODUCTION TO POSTFEKO 8-9

Animating by instantaneous phase gives insight as to how results change over time. The display
shows the magnitude of a result at a given phase and increments this phase at a user-defined
rate. Similarly, when multiple frequencies were simulated, the results at a certain frequency can
be shown. In this case the frequency is incremented and the display of the results are varied
accordingly.

8.5.5 Import and export features

Importing and exporting of data into POSTFEKO is also possible. Imports can be performed on
most text-based data files, as well as data that is generated by FEKO, but not included in a model’s
results file. Calculated data can also be exported for external processing or for use in a different
project (where the full set of results is not required).

Figure 8-10: Dialog for specifying column details during an import.

Importing

Data that can be imported include

• POSTFEKO graph files (*.pfg files)

• FEKO field results files: electric fields (*.efe); magnetic fields (*.hfe) and far fields
(*.ffe)

• Any ASCII based data file. For these files, the formatting of the data must be specified by
the user in order for the import to be successful.
Data that was imported can be added to a 2D graph in the same manner as any result.
There is an Imports button under the Add results group where imports can be accessed.
The project browser contains an entry for each import under Stored data, where imports
can be deleted from the project if they are no longer needed.

July 2011 FEKO User’s Manual

INTRODUCTION TO POSTFEKO 8-10

POSTFEKO will manage imports of familiar types by itself, without the need for the user to
specify any further details. However, for the custom import of raw data, an import template
must be defined. It is necessary to define how data columns are separated (i.e. with tabs,
spaces, commas, etc.), which lines to read and whether column titles are present. The
preview section will show where columns are separated and what their titles are.
Once the format is defined, the user must specify what type of data each column contains.
The label can be specified, along with any scaling considerations (e.g. if the data is in dB
instead of linear, or MHz instead of Hz, etc.). Figure 8-10 shows the import dialog where
these properties are defined. Each column can be specified by what type of structure it has,
along with more detail regarding the specific contents of the data:

– Axis (scalar): If the column will be used as an independent axis on a 2D graph, this
option should be chosen. Quantity options that are available include:
∗ Frequency, position, radius, angle or a user defined quantity.
– Scalar: Any scalar result type may be used. Quantity options that are available in-
clude:
∗ Far field, near field, voltage, current, power, specific absorption rate, impedance/ad-
mittance, scattering parameters, axial ratios, gain/directivity, radar cross section,
voltage standing wave ratio, reflection coefficient, user defined quantities and
several other typical data types.
– Complex pair (Real + Imaginary): If two adjacent columns contain the real and imag-
inary components of a complex number, this option should be chosen.
– Complex pair (Magnitude + Phase): If two adjacent columns contain the magnitude
and phase (in degrees) of a complex number, this option should be chosen. Complex
pairs can be classified as any of the above quantities that can be complex, such as:
∗ Far field, near field, voltage, current, impedance/admittance, scattering parame-
ters, reflection coefficient, or a user defined quantity.

Figure 8-11: Export dialogs for animation (left) and image (right) exports.

Exporting

The export operation can refer to

• Exporting an image of the current view (2D or 3D)

• Exporting an animation in a movie file format (only 3D)

July 2011 FEKO User’s Manual

INTRODUCTION TO POSTFEKO 8-11

• Exporting data in a text format (only 2D)

When exporting an image of the current view, a visual copy of whatever the view contains will
be stored. The user must specify the file location, file name, image format and image resolution.
The animation export operation works in a similar fashion to the image export. For animations
the file location, file name, file format, export quality, video resolution and framerate needs to be
specified.
For exporting data on 2D graphs, only the file location and file name is required. If, however,
continuous data is being exported, the number of sample points must also be specified.

8.5.6 Storing data and using stored data

It is possible to store a copy of most calculated results in a project for later reference. A stored
result will be unaffected by changes to the model, making it a useful reference point against
future data. Stored data can be accessed in the same way as data from a file. When adding a
result from the ribbon, a new entry for Stored data is created containing the stored result.

Figure 8-12: Accessing stored data from the ribbon.

In general, all results that can be plotted on a 2D graph can be stored. Results that cannot be
stored include: cable paths, error estimates, imported data, rays, already stored data, currents
and charges.

8.5.7 Math scripting using Lua in POSTFEKO

Scripting provides a method of creating custom results within the POSTFEKO environment. Any
supported result type can be created using a script, imported from a file or modified from existing
session results. These based on the Lua scripting language; a scripting language commonly
embedded in applications. The Lua language is simple and extendible, making it well-suited to
generating scripts for a wide variety of applications.

Figure 8-13: Using the ribbon to create or modify scripts.

All scripts are stored in the POSTFEKO session and can be modified or re-evaluated using the
editor. See the section on “Using math scripts and Lua” (see section 9.2.9) for more information

July 2011 FEKO User’s Manual

INTRODUCTION TO POSTFEKO 8-12

on how the editor is to be used. Data generated by a script can be treated the same as any of
the internal types, namely: near fields, far fields, sources, loads, networks, s-parameters, power,
transmission/reflection coefficients or an arbitrarily defined format.

July 2011 FEKO User’s Manual

USING POSTFEKO 9-1

9 Using POSTFEKO

The POSTFEKO application can be launched from different suite components and in different
ways, including a console command prompt, desktop icons, or from other suite applications.
How this is done will be discussed, as well as what to do once the application starts up.

9.1 Launching POSTFEKO

POSTFEKO can be launched (i) from the command line in a console environment, (ii) by double
clicking on the POSTFEKO icon, or (iii) by launching POSTFEKO from other suite components
such as CADFEKO or EDITFEKO. If the application icon is used to launch POSTFEKO, then no
models will be loaded and the start page will be shown. Launching POSTFEKO from other suite
applications will automatically load the model into a new project.
The command line method give users a choice as to how they would like to launch POSTFEKO. If
a model (or set of models) is specified, then it will be added to a new project; otherwise a blank
project will be given. Command line arguments can be used to specify additional parameters
when launching POSTFEKO. Arguments that may be used are listed in Table 9-1:

Table 9-1: Command line arguments for launching POSTFEKO

Argument Description
−−version Displays the current version of POSTFEKO.
*.pfs, *.fek, *.bof
*.pre, *.pfg, *.wfg Files with these extensions may be loaded into POSTFEKO.

Launching POSTFEKO from a console with no command line arguments is the same as launching
it from the desktop. No models will be loaded and a blank project will be presented. A list
of arguments can be given and file formats with the extensions *.pfs, *.fek, *.bof, *.pre,
*.pfg, *.wfg and *.out. Note that only one *.pfs file may be listed at a time and that the
*.out file argument will only show a message indicating that *.out files cannot be loaded
directly.

9.2 Typical user scenarios

Common user scenarios are presented to illustrate how to perform common operations. Scenarios
include operations such as opening models, clearing projects, populating views, modifying the
displayed data and using a range of POSTFEKO tools to interpret and verify the model and its
results.

9.2.1 Managing projects/models

POSTFEKO can manage multiple models simultaneously in a given project. Models can be added
to a project by clicking on the Add model button, using the start page (See section 8.2.), or by
opening POSTFEKO with a model file (e.g. when launching POSTFEKO from CADFEKO).

July 2011 FEKO User’s Manual

USING POSTFEKO 9-2

To add files to a project, click on the Add model button. It is available in the Home
→ File group, as well as on the application menu. Models can also be added to a
project from the start page from the recent files list.

Whether a single model or a complex project was created, the project can be saved
for later use. Clicking on the Save or Save as buttons allow the user to specify a
location and name for the current project. Projects that have been saved can later
be opened again by using the start page or by clicking on the Open project button.

The Open project button provides access to any saved project. Such projects have a
*.pfs extension and must be explicitly generated by the user. Project sessions can
also be opened from the start page (if they have recently been created/opened) or
from the application menu.
See also the options for launching POSTFEKO with multiple models or a project using the com-
mand line as described in section 9.1.

9.2.2 Add results to a view

Results can be added to both 2D or 3D views. POSTFEKO will only enable the buttons for those
results that are present in the current model or project. Clicking on a result button will provide a
list of all results of that type that may be added to the current view. Note that a 3D view is always
associated with a specific configuration for a single model. For graphs, however, any valid data
for all the loaded models may be added.
Far fields: Adds far field results to a valid view.

Near fields: Adds near field results to a valid view.

Error estimates: Displays error estimates in a 3D view.

Currents: Adds currents to a valid view.

Rays: Adds rays to a 3D view.

Sources: Adds sources to a valid view.

Loads/Networks: Adds load/network data to a valid view.

S-Parameters: Adds scattering parameters (s-parameters) to a valid view.

Power: Adds power data to a valid view.

July 2011 FEKO User’s Manual

USING POSTFEKO 9-3

Imports and Scripts: Adds imported data or data generated by a script to any view
on which the data is valid.

Transmission/reflection: Adds transmission/reflection coefficients to a valid view.

Optimisation: Adds optimisation data to a valid view.

Receiving antenna: Adds receiving antenna data to a valid view.

Standard absorption rate (SAR): Adds Standard Absorption Rate (SAR) data to a
valid view.

3D views are able to display: far fields, near fields, error estimates, currents and rays. In cases
where the SAR is calculated at a known location, the SAR may also be displayed.
Cartesian graphs may contain all data, except for: error estimates and currents on triangles. Polar
graphs may contain data where an angle defines the independent axis. That is, any result that
varies according to Theta (θ ) or Phi (φ). Near and far fields are the only results that meet this
criteria. Smith charts may contain complex source data, such as impedance or s-parameters. For
all 2D graphs, it is possible to import data that meet the criteria for being added to a view. See
Table 9-2 for a summary of the valid results on each graph type.

Table 9-2: Summary of which results may be plotted on various graphs.

Result type Cartesian Smith Polar

Wire segment data (Charges/Currents/Error Estimates) X
Far Fields X X
Impedances X X
Loads X
Near Fields X X
Networks X
Power X
Sources X X
S-Parameters X X
Transmission/reflection coefficients X
Optimisation results X

9.2.3 Convert graph types

It is possible to convert from one graph to another if the data is compatible with both. Table 9-2
summarises which result types my be plotted on which graphs. Note that there are some restric-
tions:

• Charges, currents and error estimates on mesh triangles cannot be plotted on 2D graphs

July 2011 FEKO User’s Manual

USING POSTFEKO 9-4

(a) (b) (c) (d)

Figure 9-1: The (a) Duplicate view (b) Cartesian copy (c) Polar copy and (d) Smith copy buttons

• Only data that varies over angle can be plotted on a polar graph

In the Display →Duplicate group, each graph provides the option to duplicate itself (Figure 9-1
a), or to create a different type of graph (Figure 9-1 b, c, d). For instance, a Cartesian graph
will provide the options to Duplicate view (Figure 9-1 a), create a Polar copy (Figure 9-1 c) or to
create a Smith copy (Figure 9-1 d) that contains the same data. If the traces on the source graph
are incompatible with the destination graph, an error will be given with a list of the traces that
do not adhere to the requirements.

9.2.4 Using 2D graphs (Cartesian, Smith and polar)

There are three graph types available in POSTFEKO, namely the Cartesian graph, the polar graph
or the Smith chart. Mostly, the same display options are available. However, for each graph,
there may be some options that are specific to that graph. To see a list of what types of data may
be displayed on which graphs, see Table 9-2. The Trace tab is common to all graphs and provides
control over the styling of the plotted data.

The Display tab

The Duplicate group The Duplicate group provides the options to make a copy of the current
graph (Figure 9-1 a). Simply duplicating the view will provide an identical copy of the current
view, complete with all settings (excluding cursors). The functionality is also provided to derive
a graph of a different type from the current graph (Figure 9-1 b, c, d). Section 9.2.3 contains a
more detailed explanation of how to convert between different graph types.

The Display group Any text (other than the text for legend entries) can be set in the Display
group with the Chart text button. Other options provided by the Display group include setting
the view to Greyscale or to add a Minor grid (i.e. additional minor grid lines will be added
between the existing grid lines). More information on changing the text on a graph can be found
in Section 9.2.8.

The Legend group Legends can be added to either predefined positions, or by manually se-
lecting a location using the tools provided in the Legend group. The legend entry for a selected
trace can also be modified with the Trace text button. Section 9.2.6 further discusses the use of
legends for both 2D and 3D views.

July 2011 FEKO User’s Manual

USING POSTFEKO 9-5

(a) (b) (c) (d) (e)

(f) (g) (h) (i) (j)

Figure 9-2: Axes group buttons

The Axes group The Axes group is available for Cartesian and polar graphs. Functions that are
available here include setting the data ranges (Figure 9-2 a; see also Section 9.2.8), normalising
the graph (Figure 9-2 b, c, d), changing an axis scaling to a logarithmic scale (Figure 9-2 e, f, g) or
(in the case of polar graphs) changing the orientation of the graph (Figure 9-2 h, i). To normalise
the data on a graph, select the Normalise (Figure 9-2 b) button. One of two normalisation modes
can then be utilised:

• Normalise to maximum of all traces (Figure 9-2 c): This setting will look at all of the traces
that are visible on a graph to determine the maximum value between them. All traces will
then be normalised to that value, meaning that all of the traces will maintain their relative
scaling to one another.

• Normalise to maximum of individual traces (Figure 9-2 d): This setting will look at each
trace in isolation. The maximum value of that trace will be used for normalisation. The
effect is that every trace will have an absolute maximum of 1.

Polar graphs are divided into 360◦ steps. The position of the 0◦ position can be set by using the
Orientation (h) options. The orientation can be set to North, South, East or West. In addition,
the direction in which the angle grows can be set to either a Clockwise (Figure 9-2 i) or Counter-
clockwise (Figure 9-2 j) direction. The direction of growth is only enabled if the orientation has
been set manually (Figure 9-2 h).

The Grid group The Smith chart has a unique group that allows the user to specify the type of
Smith chart to use. The choice between an admittance and impedance Smith chart is provided.

The Trace tab

The Trace tab mainly controls the styling of the plotted data. Settings pertaining to the sampling
of continuous data are also available.

The Manage group The Manage group provides the options to Duplicate the trace, to create a
New math trace or to Store a copy of the currently selected data in the project session. Duplicating
the view will result in another trace that contains the exact same settings as the current one.
Math traces can be used to perform calculations on data sets or to provide purely mathematical
reference curves. Storing a local copy of a data set saves the current state of the displayed
data. This data can then be compared to later runs of the same model (or to different models)
independently of how the models are changed.

July 2011 FEKO User’s Manual

USING POSTFEKO 9-6

(a) (b) (c)

Figure 9-3: Rendering options for traces

The Rendering group Rendering options affect how the traces are drawn on the graphs. For
continuous data, the sampling can be specified (Figure 9-3 a). The default rendering is automat-
ically determined based on the sampled data. However, when the sampling is set, the amount
of points can be reduced or increased to suit requirements. Raising (Figure 9-3 b) or lowering
(Figure 9-3 c) a trace is purely a display feature. The top traces will be visible above lower traces.

The Units group For Cartesian and polar graphs, there is a Units group that provides control
over the independent axis, as well as the displayed units for the traces that have been added to
the graphs. The units which best fit the displayed data are automatically determined based on
the range of the data. This Auto mode can be overwritten by manually specifying the desired
display units.

Figure 9-4: Transform the independent axis of a graph

Figure 9-4 shows the independent axis transformation dialog. The independent axis
can be stretched or shrunk using the Scale unit and a constant offset can be given
using the Offset unit.

The Measure tab

The measure tab helps with the reading off and interpretation of plotted results. Two classes of
measurements are provided, namely “annotations” and “cursors”. Annotations can be added to a
trace and highlight values of interest. The annotation will update along with the data and always
display the value according to its definition. Cursors are more dynamic in the sense that they
can be dragged around until they are placed at the desired positions. Cursors have the benefit
of being able to read data off of several traces simultaneously, but suffer from the limitation that
they cannot update along with the results.

(a) (b) (c)

(d) (e) (f)

Figure 9-5: Source annotations group buttons

July 2011 FEKO User’s Manual

USING POSTFEKO 9-7

The Source annotations group Source annotations are used primarily for indicating band-
widths. Due to the varying definitions of “bandwidth” between industries and applications,
several definitions of both transmission (Figure 9-5 (a)–(c)) and reflection (Figure 9-5 (d)–(f))
bandwidths are provided. The corresponding values for linear traces are supported. Reflection
bandwidths are typically used for applications such as antenna problems, where transmission
bandwidths are used more for filters or other multi-port problems.
Note that the −3 dB bandwidth refers to the half power bandwidth of an appropriate result,
which is closer to −3.01029995663981 dB.

(a) (b) (c) (d)

Figure 9-6: Far field annotations group buttons

The Far field annotations group Figure 9-6 show all annotations that are specific to far field
results. These focus on the pattern data and provide various definitions of beamwidth and the
sidelobe level. The half power beamwidth (Figure 9-6 (a)), first null beamwidth (Figure 9-6 (b))
and null to null beamwidths (Figure 9-6 (c)) are provided. The sidelobe level (Figure 9-6 (d)) is
defined as the ratio between the maximum beam strength divided by the second largest radiated
beam strength.

(a) (b) (c) (d)

(e) (f) (g) (h)

Figure 9-7: Custom annotations group buttons

The Custom annotations group The annotations defined so far refer to information that can
be extracted from a trace with a very specific type of data. More generic definitions can be found
under the Custom annotations. Three definition methods exist for defining annotations. All of
the previous definitions are defined in terms of one of these methods:

Single point A single point of interest is isolated. Points that are typically of interest include
maxima, minima or the value at a specific point. Figures 9-7 (a)–(f) are single point values.
Selecting the Other definition method will show the dialog in Figure 9-8 (left).

Delta Two points are defined and the difference between the points is used. Values such as
sidelobe level are defined using this method. Choosing to create a Delta definition will give
the dialog in Figure 9-8 (centre).

Derived width A single point of interest is isolated. Two surrounding points are then derived
from this value. Values such as bandwidth are defined using this method. Choosing to
create a Derived width definition will give the dialog in Figure 9-8 (right).

July 2011 FEKO User’s Manual

USING POSTFEKO 9-8

Figure 9-8: Defining a custom annotation using the three definition dialogs.

(a) (b)

(c) (d) (e) (f) (g) (h)

Figure 9-9: Measurement group buttons

The Measurement group Cursors are available to help read off information from a graph. The
Cursors (Figure 9-9 a) button enables the cursors. Each graph type has a different type of cursor.
Cartesian and polar graphs have two cursors. By enabling the Cursor table (Figure 9-9 b), a
summary of the information is presented. This table contains both the data at the displayed
points for all traces, but also the difference between the two points for all traces. Note that the
Smith chart only has one cursor per trace.
If a cursor is moved outside of the visible region of a graph, a handle will appear in the corner of
the graph so that the cursor can still be moved.
Several predefined positions for cursors are defined per trace. Using any of these buttons will
make the currently selected cursor jump to the indicated location. The selected cursor is indicated
by a solid line. The buttons in Figure 9-9 (c)–(e) will respectively jump to the global maximum,
the next local maximum to the left and the next maximum to the right. Similarly, the buttons in
Figure 9-9 (f)–(h) will jump to the corresponding minima.

The Format tab

The Font, Colour and Effects groups Settings that relate to the styling of a graph are provided
on the Format tab. Fill colours, text colours, fonts, shading, text styling and general tools that
pertain to the appearance are provided.

The Line and Marker groups Access to the formatting tools for the line and markers is also
provided. For both lines and markers, the style, colour and weighting can be set. In the case
of markers (Figure 9-11 a), there are three methods by which they can be drawn: markers can

July 2011 FEKO User’s Manual

USING POSTFEKO 9-9

Near Field
80
Nearfield E-Field [V/m] 70 B1
60
50
40
30 A1

20
Trace A B B-A
10 1 36.83 71.78 34.95
Horizontal axis 7.11 7.629 0.519
0
6.8 7.0 7.2 7.4 7.6 7.8 8.0 8.2
Frequency [GHz]

Figure 9-10: Cursors provide a means of quickly reading off data from a graph.

(a) (b) (c) (d)

Figure 9-11: Marker placement settings

be drawn on the calculated frequencies (Figure 9-11 b), sparsely spaced markers (Figure 9-11 c)
can be drawn at constant intervals and densely spaced markers (Figure 9-11 d) can be drawn at
constant intervals. The latter two options are trace decorations that will always be visible in a
view, irrespective of the zoom level.

9.2.5 Using 3D views

The 3D view is a powerful tool for getting a feel for how the electromagnetic properties of an
environment behave. By setting the display properties of the meshed geometry, the theoretical
model properties and the results, information can be presented in such a way as to aid under-
standing. For each category, a tab exists that groups similar features together and make finding
features more intuitive.

The Display tab

The Display tab is used to change visibility settings of most of the model entities. Tools such as
legends, cutplanes and axes are also available to help visualise the model and calculated results
to help derive useful information. The Display tab is the main controlling tab for the 3D view.
This means that the duplication and export of the views are also controlled from here.

The Display group

July 2011 FEKO User’s Manual

USING POSTFEKO 9-10

(a) (b) (c) (d) (e) (f)

Figure 9-12: The (a) Cutplanes (b) Greyscale (c) Boundary and (d–f) Annotation buttons

Figure 9-12 shows the Display group, which contains buttons that enable greyscale (b) and the
bounding box (c) visibility. The bounding box visibility applies to the geometry, but ignores the
displayed results and non-mesh geometry. Cutplanes can be added by clicking on the Cutplanes
button (a). Figure 9-13 shows the dialog that is presented.
Annotations in the 3D view can be shown or hidden by clicking on Annotations button (Fig-
ure 9-12 (d)). To select how the annotations must be shown, two options are available. The first
is to highlight the selected element and to show an annotation (Highlight and annotate elements
as shown in Figure 9-12 (e)). The second is to only highlight the element (Highlight elements as
shown in Figure 9-12 (f)).

Figure 9-13: The (left) Plane definition and (right) Visibility filter tabs

On the plane definition tab, the location of the cut plane can be set by defining a flat plane. The
Flip button alternates the normal direction of the plane, which in turn determines which side
of the plane will be hidden. Multiple planes may exist and a given plane can be removed or
deactivated on the tab for that plane. A global Visibility filter (shown on the right in Figure 9-13)
provides control over which entities should be affected by the cutplane. By default, everything
that can be visualised and is in the model will be affected. To change this, move the desired
components over to the Do not cut entities list on the right and click Apply. Note that the
visibility filter is shared between all cutplanes.

(a) (b)
Figure 9-14: The (a) Add legend to a position and (b) Individual range buttons

The Legends group Figure 9-14 shows some of the buttons that are typically used to manipu-
late a legend. Legends can be added to any of the four corners in a 3D view. Each of the legend

July 2011 FEKO User’s Manual

USING POSTFEKO 9-11

buttons will have a list of entities that it can bind to, depending on what is currently being dis-
played. The ranges of the legends (and in effect the colouring of the results) can be set using the
Individual range settings as well as the advanced scaling settings found in the advanced dialog.
Legends are discussed in more detail in section 9.2.6.

(a) (b) (c) (d) (e) (f) (g)

Figure 9-15: Display entities group buttons

The Entities group Each of the buttons in Figure 9-15 shows or hides all components of that
type in the 3D view. By default, all entities except for named points are visible. Clicking on any
enabled button will toggle that entity’s visibility. The buttons in Figure 9-15 are: Sources (a),
Loads (b), Cables (cable paths) (c), Points (named points) (d), Networks (general networks) (e),
TX line (transmission lines) (f) and RX antenna (ideal receiving antennas) (g).

Figure 9-16: Advanced source and load display settings

The buttons in Figure 9-15 only provide a show/hide functionality. For more options regarding
the visibility of sources and loads, the advanced dialog shown in Figure 9-16 can be used. For
both sources and loads, a Show and Hide column is presented which contains all sources/loads
present in the model. By placing entries in the Show column, sources/loads of that type will
be shown only if the visibility for the entity type is enabled. Additional options are provided
for source visualisation. Here, sources can be coloured and/or scaled by magnitude. This is
often used in conjunction with aperture sources, electric dipoles, magnetic dipoles and impressed
currents.

(a) (b) (c)

Figure 9-17: Display of (a) Symmetry, (b) PBC (Periodic Boundary Conditions) and (c) Infinite planes.

The Method display and Infinite planes groups Figure 9-17 shows the main display buttons
for enabling the visualisation of symmetry planes, periodic boundary conditions (PBC) and infi-
nite planes. As with the entity visualisation, these buttons merely enable/disable the visibility of

July 2011 FEKO User’s Manual

USING POSTFEKO 9-12

the method. For infinite planes, the opacity can also be set and the infinite layers may be shrunk
to help improve rendering.

The Axes, Export and Duplicate groups are simple in function. Exports are discussed in more
detail in section 8.5.5. The Duplicate view button does exactly what it says, in that an exact
replica of the current view is created. The axis visualisation provides two main axis settings:
show/hide the global axis (at the origin of the model) and show/hide the mini axis (in the
bottom-left corner of the view). Axis tick marks can be shown on the global axis if it is displayed.

The Mesh tab

The Mesh tab also contains display settings, but since there are so many that apply specifically to
the mesh, they are grouped together for convenience. The groups include settings for Rendering,
Opacity and mesh Visibility, along with a set of tools to help inspect the model.

The Rendering group affects how the mesh is drawn in the 3D view. Features in this group
are displayed in Figure 9-18. The “colour by” options display the mesh using different colours.
Available options include colouring by: Element face media (Figure 9-19 a), Model outline (Fig-
ure 9-19 b), Element region media (Figure 9-19 c), Element label (Figure 9-19 d), Element nor-
mal (Figure 9-19 e) and Element type (Figure 9-19 f). The main “colour by” button (Figure 9-18
a) will then change and show the according icon that is selected for colouring.

(a) (b) (c) (d) (e) (f)

Figure 9-18: Mesh rendering options

(a) (b) (c) (d) (e)

Figure 9-19: Mesh colouring options

Element normals can also be displayed using the Normals button (Figure 9-18 e), which displays
a line indicating the normal direction. For segments, the radii can be artificially enlarged using
the Segment radius button (Figure 9-18 b), which multiplies the defined radius by a factor be-
tween 1–10 times the original radius. Figure 9-20 shows the dialog for displaying the principal
direction for anisotropic layers (Figure 9-18 d). Anisotropic layers are applied on a label. For
each geometry part that is covered with an anisotropic dielectric layer, the option is made avail-
able to show the principal direction. Since multiple layers can be present, it is also necessary to
specify which layer to show.

The Opacity group Figure 9-21 shows the opacity group that pertains to the opacity of mesh
elements. Since windscreen and aperture triangles have an inherent opacity setting, they can be
customised individually. Mesh opacity (a) applies to all visualised geometry and the effect on
aperture (c) and windscreen triangles (b) are compounded. For a setting of 0%, the element will
be purely transparent and will not be visualised. At 100%, no transparency will be applied.

July 2011 FEKO User’s Manual

USING POSTFEKO 9-13

Figure 9-20: Anisotropic layer visualisation settings

(a) (b) (c)

Figure 9-21: Opacity settings

(a) (b) (c) (d) (e)

(f) (g) (h) (i) (j)

Figure 9-22: Mesh visibility settings

The Visibility group For most mesh elements, it is possible to individually set the visibility for
the faces, edges and vertices of the triangles. Elments include: wire segments (Figure 9-22
a), metallic triangles (Figure 9-22 b), dielectric triangles (Figure 9-22 c), aperture triangles
(Figure 9-22 d), windscreen triangles (Figure 9-22 e), tetrahedra (Figure 9-22 f), cuboids (Fig-
ure 9-22 g), UTD polygons (Figure 9-22 h) and UTD cylinders (Figure 9-22 i). For the some spe-
cial cases, different settings are given. These cases are wire segments (surfaces, lines, vertices),
tetrahedra (faces, edges, vertices, volume) and UTC cylinders (faces, edges). The visibility filter

Figure 9-23: Mesh visibility filter

(Figure 9-23) provides some additional control over the visibility of mesh elements. The visibility
filter is accessed through the associated button (Figure 9-22 j) and provides the functionality to
filter out mesh regions with a specific label, or to filter out specific media.

July 2011 FEKO User’s Manual

USING POSTFEKO 9-14

The Tools group

(a) (b) (c) (d) (e)

Figure 9-24: Mesh tools

The mesh tools group provides a set of tools that can be used to help identify specific parts of
a geometry or to ascertain certain properties about the geometry. The Measure distance feature
(Figure 9-24 a) places the 3D view into a mode that will measure the distance between any two
selected points. To use the distance measurement mode, use the following procedure:

• Click on the Measure distance button to open the distance measurement tool.

• Hold down <Ctrl>+<Shift>

• Click to remember the first point.

• Repeat this process for the second point. The distance will be displayed in the tool.

The Measure angle tool works in much the same way, except that three points are required.

Figure 9-25: Find mesh elements

The Find elements tool (Figure 9-24 c) is used to identify mesh elements by their internal ID’s.
This is often used when kernel errors or warnings are given. An element type must be specified
before entering the element ID in the dialog window (see Figure 9-25). Multiple ID’s may be
found at the same time by using a comma separated list. The 3D view will add a preview
annotation to the specified elements. When the Add annotation(s) button is pressed, a permanent
annotation is linked to that element(s), otherwise annotations will disappear once the dialog is
closed or the selection is cleared.
Mesh connectivity (Figure 9-24 d) is a tool that indicates the free edges of a geometry by drawing
a red line on the edge. In other words, mesh elements that do not connect to other elements are
shown. Such elements typically appear at the edges of plates or at the ends of wires. When com-
plicated or imported geometries are used, the mesh connectivity can show where the geometry
does not line up correctly for an enclosed mesh. The mesh connectivity tool is a mode that can
be either enabled or disabled on the ribbon.
Figure 9-26 shows the options that are available for mesh highlighting, if any is desired. The se-
lected solution type will be highlighted in the 3D view by displaying a yellow grid over elements
that are solved with that method. Solution types include: lossy metals (b), coatings (c), CFIE/M-
FIE (d), EFIE (e), impedance sheets (f), PO (g), PO (Fock regions) (h), GO (i), UTD (j), FEM (k),

July 2011 FEKO User’s Manual

USING POSTFEKO 9-15

(a) (b) (c) (d) (e) (f)

(g)

(h) (i) (j) (k) (l) (m) (n)

Figure 9-26: Mesh highlighting options

windscreen active elements (l), planar green’s function aperture triangles (m) and numerical
Green’s function (n). The highlighting is activated by the associated button (Figure 9-24 e)
which will then change and show the icon corresponding to the selected highlighting option. To
deactivate, select the “none” (Figure 9-26 a) option.

The Result tab

The Result tab provides access to the visualisation setting of results. Determining which compo-
nents of a result to display is done on the result palette; the Result tab only provides control over
how that information is presented. To manage the display settings of a result, select it and note
which buttons become active.
Display settings for far fields, near fields, currents and rays are available. In the cases of currents
and near fields, arrows can be added to the view to help visualise the direction of the fields or
currents. Rays will only be available if it was requested that they be stored prior to running the
simulation.

(a) (b)
Figure 9-27: Result management group

The Manage group Figure 9-27 shows the Manage group buttons on the Result tab. With these
options, it is possible to make a duplicate of the selected result with all of its visualisation settings
(Figure 9-27 a). It is also possible to store a local copy of a result. This result will be stored in
the POSTFEKO session and becomes independent from any future changes made to the model.
Note that a copy of the stored result is displayed in the view as soon as the copy is made.

The Rendering group Figure 9-28 shows the Rendering group buttons on the Result tab. The
rendering group applies to far fields, near fields, currents and error estimates. In most cases, the
rendering options will be useful for near and far fields. The Grid button (a) enables visibility of
a mesh grid on top of the result, which helps give a sense of dimension to 3D results. When the
Surface (b) option is deselected, the coloured surface of a result is hidden. Discrete (c) colouring
removes the interpolated colouring of a surface and uses a predefined set of colours to represent

July 2011 FEKO User’s Manual

USING POSTFEKO 9-16

(a) (b) (c) (d)

(e) (f) (g) (h)

Figure 9-28: Result rendering group

Figure 9-29: Advanced rendering dialog for far field size

the surface. The Colour option (d) applies to isometric surfaces (for 3D near fields). Origin (e)
and Size (g) settings apply only to far fields. Note that the far field can be set manually in the
advanced dialog shown in Figure 9-29. Opacity (Figure 9-28 f) sets the amount of transparency:
For a setting of 0%, the element will be purely transparent and will not be visualised. At 100%,
no transparency will be applied. Extrusion (Figure 9-28 h) applies to Cartesian near field surfaces
that lie in a flat plane.

(a) (b) (c) (d) (e)

Figure 9-30: Result requests group

The Requests group Before a simulation is run, it is good practice to validate that data has
been requested in the correct locations. Figure 9-30 shows the Request points (a–c), Setting
(d) and Boundary (e) buttons which help ensure that near and far fields have been requested
correctly. Requests points will automatically be shown if no result data is present. Once data
becomes available, the result data will be shown and the request points will be hidden. This
behaviour corresponds to the Auto request points display (Figure 9-30 a). Request points can
also be forced on or off by choosing Display request points (Figure 9-30 b) or Don’t display
request points (Figure 9-30 c), respectively. Note that the Settings button (b) is now enabled.

Figure 9-31: Request points display settings

The visualisation of the points can be set using the display settings dialog shown in Figure 9-31.

July 2011 FEKO User’s Manual

USING POSTFEKO 9-17

Here, the display type, marker colour and marker size for the requested points can be set. For
near fields, the bounding box can also be shown by selecting the Boundary button (Figure 9-30
c).

(a) (b) (c)

Figure 9-32: Result contours group

Figure 9-33: Manually specify contour positions

The Contours group Contour lines are lines that indicate a specific value on a result display.
The Show contours button (Figure 9-32 a) toggles the visibility of the contour lines. The colour of
the contours (Figure 9-32 c) can be set to any value, or the colour can be linked to the magnitude
of the displayed value. Figure 9-33 shows the dialog used to specify the contour position accessed
trough the Contour positions button (Figure 9-32 b). An option is given to choose between the
Number of contours and to Specify the contour values. The latter option allows for any number of
contours that lie on user-specified positions. The positions can either be defined by its magnitude
value, or by specifying a percentage of the value range.

(a) (b) (c) (d)

Figure 9-34: Result arrows group

The Arrows group Arrows indicate the direction in which a current flows or the direction in
which a field is pointing. As these results change over time, arrows can only be plotted if the
magnitude of such a result is shown for a specific phase value. Under Quantity in the result
palette, Instantaneous should be selected and the phase should be specified. Once arrows are
displayed by using the Show arrows button (Figure 9-34 a), the colour (b) and size (c, d) can be
set with the remaining buttons. Note that the arrows are scaled by their magnitude by default.

July 2011 FEKO User’s Manual

USING POSTFEKO 9-18

(a) (b) (c) (d) (e) (f)

Figure 9-35: Result rays group

The Rays group Different aspects of rays may be displayed independently from one another.
Ray lines (Figure 9-35 a), Ray numbers (Figure 9-35 b), Group numbers (Figure 9-35 c) and
Intersections (Figure 9-35 d) may be displayed in any combination by selecting the appropriate
button. It is then also possible to colour the displayed ray lines by magnitude (Figure 9-35 f).
The visibility threshold (Figure 9-35 e) can be set to only allow the higher magnitude rays to be
displayed. This helps reduce clutter and only leave the most important rays in the view.

The Animate tab

(a) (b) (c) (d) (e)

Figure 9-36: Animation types

Animation can be performed by varying a property over time. The properties that can be an-
imated include: phase (Figure 9-36 a), frequency (Figure 9-36 b) and the camera angle (Fig-
ure 9-36 c–e). Phase and frequency animation requires a result to be present that varies over
these parameters. Camera angle animation only requires a geometry to be present. The camera
angle may be animated over only phi, only theta, or theta and phi simultaneously.
For frequency and phase animations, the results that will be affected must be added to the view
and be made visible. For camera angle animations, only the geometry is required. Ensure that
the correct animation type is chosen from the Type button on the Animate →Settings tab. By
pushing the Play button, the animation will commence. The speed with which the animation
variable changes can be set using the Faster and Slower buttons. Deselecting the Legend button
will remove the description at the bottom of the screen.
Animation settings: The advanced animation settings allow for more precise defi-
nition regarding the speed and resolution with which a variable animates. Each
property specifies how much the variable must change per second. For example,
a Phase (ωt/s)=30.00◦ means that the phase will increment by 30◦ every second,
which will produce a complete 360◦ loop in 12 seconds. The other properties work
in much the same way.
Note that for continuous frequency models, it is necessary to break the continuous run into
discrete steps. The combination of Frequency (points/s) and Continuous frequency (# of points)
will then determine the sampling resolution and animation speed.
Export animation: Exporting an animation is also possible. Movies can be saved in
*.avi, *.mov, or *.gif formats. Setting the quality affects the compression ratio
for the specified screen size. For very high-quality exports, it is good practice to
reduce the screen size to as small as is need and setting the Export quality to high.
Setting the frame rate will affect how “smooth” the animation appears. Export works
on the currently selected animation type and settings.

July 2011 FEKO User’s Manual

USING POSTFEKO 9-19

9.2.6 Using legends

A legend is a utility that explains what the data is that is being presented in a view. Legends
can be placed on both 2D and 3D views. However, they function somewhat differently from one
another, since the data which they represent are displayed differently.

Legend (2D): For 2D views, a legend is automatically added to the top of the graph
(beneath the graph title). To change the location of a legend, click on the Legend
button and select a better location from the list. It is also possible to remove the
legend or to manually specify the location from here. The entries will reflect all
displayed traces and their representations. Styling can be performed on the legend
on the Format tab. The font settings, background colour, border line styling and drop
shadow settings can be changed.

Trace text: Legend entries for each trace are automatically generated. This is gen-
erally sufficient for understanding what each trace represents. However, it is often
desirable to manually specify the description of a trace. To do this, click on the Trace
text button. The same option is available by right-clicking on a trace in the result
palette. Figure 9-37 shows the dialog that can be used to modify the legend entry
for a selected trace.

Figure 9-37: Chart text dialog

Manually setting the formatting and entries for the legend is only available for 2D legends. Spec-
ifying the legend location resizes the size of the graph to make room for the legend in that
location. For example, a graph will be wider if the legend is placed on top than if it is placed to
the right. When the overlay or manual positions are used, the graph is maximised and the legend
should be placed such that it does not obscure important data.

Legend (3D): For 3D views, legends can be placed in any of the four corners of a
view. Clicking on any of the four buttons provide a list of entities that the legend can
link to. The title of the legend will give an indication of what data is being displayed
and in what units. A colour key is also given to help interpret the different colours
that a legend value can possess.
Setting the individual range of a 3D value legend is possible by clicking on the Individual range
button. The value range is set to the minimum and maximum value of the displayed data by
default. Here the range can be set manually for linear and logarithmic data. There are several
settings that will affect the range limits. These can be accessed by clicking on the advanced
dialog launcher in the corner of the Legends group. Figure 9-38 shows the dialog, followed by
an explanation of the entries.

July 2011 FEKO User’s Manual

USING POSTFEKO 9-20

Figure 9-38: Advanced range limit settings

• Round off legend range and step size

Most data will have a large number of decimal values to accurately describe a given value.
This results in a legend range that is difficult to read and interpret. To mitigate this, the
values are rounded off and a range is selected that results in a legend that is more legible.
The actual data range will always fall within the rounded off range.

• Scale to peak instantaneous values

This is used when the magnitude of a result along with the instantaneous phase is shown.
The minimum and maximum value limits will remain constant for each phase step, meaning
that the definitions for the colours are always the same. This makes it simpler to see how
the magnitude changes over phase. When only the magnitudes at the given phase are of
interest, this box can be deselected so that the range limits synchronise with the currently
displayed data.

• Scale to vector magnitude

Often, especially with near field results, one wants to compare two components with each
other. In these cases it is easier to compare relative magnitudes of the components if they
are scaled to the same maximum. The total vector magnitude is used to scale all result
components, so that a relatively small component is easily discernible from a larger one.

• Scale to visible results of the same quantity

It is possible to have several near field or far field requests in a model. Each request will
have its own minimum and maximum value. However, POSTFEKO automatically scales
the displayed data to keep all results that have been added into a view into account. This
makes it clear what the values are relative to one another. Deselecting this option will
render each result according to its own minimum and maximum value.

• Scale only to selected frequency

This option is only enabled when discrete frequency data is available in a model. Here
the range limits are determined by the minimum and maximum values for all calculated
frequency points. For example, seeing the change in gain for a far field over frequency is
simpler when the field is scaled according to all calculated points.

• Scale to request slice dimensions

It is not always possible or desirable to plot all requested result data at the same time. If
it is required that the range values are determined only by the displayed slice, this option

July 2011 FEKO User’s Manual

USING POSTFEKO 9-21

should be deselected. This is useful if only the range of data at a specific slicing is of
interest.

All of the above-mentioned options are selected by default. Each setting is applied independently,
meaning that a wide variety of combinations is available to help display the data in the desired
manner.

9.2.7 Use the project browser

The project browser is an advanced tool that provides a project-wide, model-wide and detailed
view of a project and its contents. It is divided into three sections, namely the project browser,
the model browser and the details browser.
Project browser: The project browser can be accessed by enabling it on the View tab
by the Project button. It can be hidden again in the same way, or by resizing the
panel to zero width.
The Project Browser contains a tree of the models that have been loaded into the project. It also
contains a tree for any stored and imported data that has been pulled into the project session.
When selecting any of the contents of this window, the Model Browser becomes filled with data
relevant to that entry. This is the only place where models may be removed from a project.
Right-click on a model in the tree and select Remove model to remove all reference of it in the
project.
When the Model Browser is populated, more detailed information about a model is given. It
includes a summary of all solution configurations, as well as any data that could be combined
over the configurations. Expanding the tree for a configuration shows further details on the media
that was defined, the mesh, the solution entities that are present, result requests, infinite plane,
etc. By selecting any of these components, the Details Browser is filled with further information
that is relevant to that entry. If results have been calculated for the model, it is possible to add
the data to the current view by right-clicking on the component and sending it to the active view.
The Details Browser summarises the information relevant to a select model component. For
example, selecting a far field request in the Model Browser will give information about where the
requested points are and what the request specifications are. Each entry is handled differently
and shows a different set of information.

9.2.8 Manually specify axis properties, ranges and captions

On a 2D graph, it is often necessary to rename axes, titles or legend entries to better describe the
data that is being presented. The axis ranges are set to automatically determine the data ranges
and scale themselves accordingly. To manually specify these ranges is possible. Along with all
the other styling options, a graph can be fully customised to professionally present data.
Chart text: The graph title, footer and axis labels are automatically derived from the
contents of the graph. To change any of these entries, open the dialog by clicking
on the Chart text button and deselect the Auto box. The customised entry will then
be applied once OK or Apply is pressed. Figure 9-39 shows the dialog used to make
these changes.

July 2011 FEKO User’s Manual

USING POSTFEKO 9-22

Figure 9-39: Chart text dialog

Axis settings: For polar and Cartesian graphs, the axis settings can be modified to
have different ranges, grid spacing or numbering formats. To change any of these
properties, ensure that the correct axis is being edited, deselect the box indicating
that values will be determined automatically and enter the custom values. Note that
the zoom to extents feature will now be applied to these ranges. Figure 9-40 shows
the dialog for making these changes.

Figure 9-40: Axis settings dialog

For a 3D view, it is sometimes useful to clamp data between two values. This affects the colouring
of the result, where blue will correspond to the minimum value and red to the maximum. This
can help reveal changes in a result between two values that would otherwise be missed.

July 2011 FEKO User’s Manual

USING POSTFEKO 9-23

Individual range: Control is given over the result value ranges for both linear and
logarithmic scaling (i.e. dB). For linear scaling, either the minimum and maximum
values of the result can be used, or the range can be fixed to user-defined extents.
For dB, the range limits are automatically defined by the minimum and maximum
values of the data (Automatic range limits). The option is also available to use the
fixed range values from the linear scale settings (Use linear range limits (with auto
clamping), although a dB-specific fixed range can also be specified (Fixed range).
The range can also be specified by specifying the maximum dynamic range. Here
the maximum value of the result data will be used and the minimum value will be
the the maximum minus the value entered.

9.2.9 Using math traces

Any result plotted on a 2D graph provides the option of mathematically altering the trace by
using a mathematical equation. However, it is also possible to create a math trace. These traces
inherently contain no data and require other traces or mathematical equations to present infor-
mation. The only difference between data traces using equations and math traces, is that math
traces do not use the self keyword.

Figure 9-41: Expression editor for mathematical equations

Equations may be edited using the expression editor (Figure 9-41), which contains a list of all
defined functions, traces and constants that may be used. Complex mathematics may be used in
calculating results, but only scalar data can be displayed. If a trace cannot be displayed, then a
warning icon should appear next to the trace in the results palette. Hovering over the icon will
indicate the reason as to why the equation is not being shown.

9.2.10 Using math scripts

Scripting in POSTFEKO (see section 8.5.6) was introduced briefly in a different section. To sum-
marise, a script is a mechanism with which an internal POSTFEKO result structure (or dataset)
can be created or manipulated directly. In order to modify a script, the editor shown in Fig-
ure 9-42 is provided. Features such as syntax highlighting and code completion make editing
scripts simpler and more efficient. Scripts are written in the Lua language.

July 2011 FEKO User’s Manual

USING POSTFEKO 9-24

Figure 9-42: Editor for Lua scripts in POSTFEKO

Scripting allows for the manipulation and creation of data which can then be displayed in POST-
FEKO as an internal data entity. For example, an existing near field can be pulled into a script,
modified and returned. POSTFEKO can then use this data as though it was generated by the
kernel. An alternative example is that a theoretical or measured pattern can be generated or
imported into a script. This result can then be displayed and compared to simulated data.

(a) (b) (c) (d) (e) (f)

(g) (h) (i) (j) (k) (l)

Figure 9-43: Editor buttons

The script will be executed when the Run button (Figure 9-43 a) is pressed. Saving a script
(Figure 9-43 b) will keep the changes made to the script and also save the POSTFEKO session.
The editor can be cleared using the Clear all button (Figure 9-43 c) or printed using the Print
button (Figure 9-43 d). Undo and Redo is also available (Figure 9-43 e and f). The clipboard
can be utilised through use of the Copy, Cut and Paste buttons (Figure 9-43 g, h and i). Block
comments can also be managed using the Comment and Uncomment buttons (Figure 9-43 j and
k). Additional help can be accessed using the Help button (Figure 9-43 l).
By default, a script will be re-evaluated once any of the result files in the session are reloaded.
There are cases where this behaviour is not desired, such as when the script requires many
system resources or is not dependent on the data being reloaded. If the automatic re-evaluation
is disabled, the dialog must be opened and the script rerun manually for the results to update.

July 2011 FEKO User’s Manual

USING POSTFEKO 9-25

9.2.11 Generating reports

POSTFEKO is a useful tool to help analyse and present data in a useful format. Once this has been
achieved, it is often desired to use the processed results in a report or presentation. To help make
it easier to generate these reports, several tools are available in POSTFEKO. The data, images or
animations can be exported, a Quick report can be generated, or a pre-defined template report
can be populated.

Exporting and copying data, images and animations

(a) (b) (c) (d) (e) (f)

Figure 9-44: Exporting and copying

Figure 9-44 shows all of the exporting and copying operations on data (Figure 9-44 a–b), images
(Figure 9-44 c–e) and animations (Figure 9-44 f).

Quick reports

Figure 9-45: Generating quick reports

Quick report: Figure 9-45 shows the interface to the Quick report tool. A quick
report is a report that contains all of the images and headers from a POSTFEKO
session through a few quick clicks. Microsoft PowerPoint and Word documents may
be generated, as well as a PDF (Portable Document Format) report. For the MS Office
reports, MS Office 2003 or newer must be installed on the machine where the report
is being generated.
These reports are a good starting point for a document and can be styled and extended externally
as required. The PDF reports cannot be modified with most PDF viewers, but given the correct
software can also be extended.

July 2011 FEKO User’s Manual

USING POSTFEKO 9-26

Generating a report from a template

When a MS PowerPoint or Word template is provided, images can be replaced by using the
template reporting tool. This process involves three steps:

1. Create an MS PowerPoint or Word template document (i.e. a *.potx or *.dotx docu-

ment).

2. Set up a new report using the New from template option.

3. Generate the report using the Generate report option.

The first step involves using MS PowerPoint or Word to create the template. Create a document
in the appropriate application. Set up the document exactly as it should appear, with the correct
styling, headers, images etc. POSTFEKO can add images to the document in places where place
holders were inserted. To insert a place holder, insert a rectangle shape into the document. Add
text to the document and identify them using tags in the following format:

<feko:image:tag>,

where tag can be replaced with any text string to help identify the place holder.

New from template: These tags can be read by POSTFEKO. The recognised tags will
be listed and an opportunity will be presented to match each tag with one of the
views currently open in the session (Figure 9-46). To start configuring a template
report, click on the New from template option on the ribbon.

Figure 9-46: Configuring a template report.

Generate report: Once the report is configured, it will be listed in the Project browser
under Template reports. Note that no report has been generated yet. The report has
only been configured. To generate the report, click on Generate report and select the
desired option from the menu. A prompt for the file location and name will appear,
after which the report will be generated and saved to the chosen location.
For MS Office 2007 or newer, Word templates may also be constructed using “content controls”.
This gives more control over the placement and formatting of the image. This is a more advanced
feature and it will require activating the Developer tab in MS Word. Once the tab is active, add an
image content control and set the Tag field to any value and POSTFEKO will treat it accordingly.

July 2011 FEKO User’s Manual

USING POSTFEKO 9-27

9.3 Errors and warnings

Errors and warnings are presented in a variety of ways depending on the source and severity of
the notification. Notifications can be triggered by:

• changes made to model (*.fek) or result (*.bof) files from outside POSTFEKO;

• results and traces that are added to a view but do not have the necessary data for visuali-
sation;

• or a user action that either needs to be confirmed or is invalid.

In each case, there is a different mechanism for indicating the error/warning message. The
mechanisms include icons next to results that are in error, a dialog indicating that a workflow
error in POSTFEKO has occurred and then there is also a dialog that appears when changes are
made externally to the model or result files.

9.3.1 Warning icon

Figure 9-47: A warning icon on an erroneous result component.

The warning icon in Figure 9-47 shows that the far field cannot be displayed because of a missing
results file. The tooltip appears when the mouse hovers over the icon and gives an indication as to
why the result is missing. This type of notification occurs mainly when results cannot be plotted
due to missing or corrupt data, or when an invalid mathematical equation is used.

9.3.2 POSTFEKO workflow error and warning dialog

Figure 9-48: Error dialog for attempted invalid operations.

The error dialog in Figure 9-48 shows the second mechanism of informing the user of a problem.
This type of dialog is commonly found when invalid actions are attempted. In the example
shown, it was attempted to plot far field data on a Smith chart, which is not a valid operation.

July 2011 FEKO User’s Manual

USING POSTFEKO 9-28

Figure 9-49: Dialog for logging background errors.

9.3.3 Error and warning logging dialog

This dialog appears when files are modified externally such that notifications are required. For
example, if a simulation is run and warnings or errors have occurred, then the window in the
dialog in Figure 9-49 is appended with the appropriate message. Errors and warnings are sepa-
rated into two groups via a tab, so that they can be summarised separately. The same mechanism
will be used, for example, when the result (*.bof) files contain errors or warnings or are out of
synch with the model (*.fek) files.
When the OK button is clicked, the windows will be cleared. The next time an notification is
logged, the dialog will appear again in a fresh dialog. By checking the box next to Don’t show
errors and warnings during this session and clicking the OK button, the window will be cleared
and no more warnings or errors will be shown for the remainder of that project.

9.4 Getting Help

Additional information or explanations can be found in a variety of ways in the FEKO suite. A
help browser can be launched from within the POSTFEKO application, digital copies of the user
manual are included in the installation, examples are available on the FEKO website and a team
of technical support staff can be contacted for more advanced help with specific problems or
queries.
Figure 9-50 shows the home page of the Help browser. The browser contains information on how
the FEKO Suite works as well as links to the digital copies of the manuals (in a *.pdf format),
the FEKO website and contact details to get in touch with customer support.

The Help browser can be launched by clicking on the help button in the top
right-hand corner of the POSTFEKO screen...

...or by clicking on the help button in the corner of the start page, or by
pressing F1. Pressing F1 while a dialog is in focus will also jump to the
relevant section in the Help browser.

Digital copies of all help documentation can be accessed from the start page, as well as links to
the FEKO website (http://www.feko.info).

July 2011 FEKO User’s Manual

USING POSTFEKO 9-29

Figure 9-50: Home page of the Help browser.

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-1

10 Scripting guide for POSTFEKO

It is the intent of this chapter to provide a comprehensive guide to managing POSTFEKO results
within a Lua environment. For a complete guide on the Lua language, we refer you to the of-
ficial Lua Reference Manual (http://www.lua.org/manual/5.1/) and Programming introduction
(http://www.lua.org/pil/). Another useful resource is the community maintained Wiki that is
available at http://lua-users.org/wiki/.
The rest of the chapter deals with two principles:

• How to pull data into the scripting environment and how to return the results to a POST-
FEKO session. The Lua language is used as a base with which to generate data that can be
displayed in POSTFEKO, or to modify data that already exists in the session.

• How to use the language extensions that are made available within the Lua environment.
Once a handle on a result has been obtained, it must be possible to modify that data. The
tools that are available to analyse and modify results will be discussed.

10.1 The interface between POSTFEKO and the scripting environment

There are several keywords (other than the standard Lua keywords) that are recognised by the
editor. These are:

Table 10-1: Global keywords recognised in the Lua editor.

complex An object class that helps manage complex numbers.

dataset A namesepace that provides a set of tools for managing POSTFEKO data
structures in the Lua environment. A dataset contains all of the
information that is necessary for POSTFEKO to display and interpret
information correctly.
inspect A general function that displays the contents of any table that can be broken
down into basic components. Output can be seen in the output window of
the editor.
pf The main interface between the Lua scripting environment and POSTFEKO.
The pf namespace is used to pull results from a POSTFEKO session into the
Lua environment for further processing.
point An object class that helps manage points in 3D space.
printlist A general function that prints out the contents of key/value pairs. Output
can be seen in the output window of the editor.
p
i, j Predefined constants for −1

In addition, the print function is overloaded in the POSTFEKO editor environment. If a POST-
FEKO dataset is given to the print function, a short summary of the structure of the dataset
is displayed in the output window.

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-2

Figure 10-1: Auto-complete expansion of the pf namespace.

It may be useful to note that the auto-complete functionality of the editor will expand many of
the keywords to reveal more operations. Figure 10-1 shows the expansion for the pf names-
pace, revealing a list of sub-categories to explore. Some of these may also be expanded, as seen
in the second panel of the figure. The bottom level of the expansion is a function call (e.g.
pf.nearfield.getnames()).

10.2 Pulling data into the scripting environment

A common use for the scripting functionality is to modify existing POSTFEKO results. As such, it
is necessary to get a handle on a result (called a dataset) into the scripting environment. The
following code will print a list of all near field results in a session. The ouput is also indicated.
nf_names = pf.nearfield.getnames()
printlist(nf_names)
1: ’Horn.Configuration1.NearField1’
2: ’startup.Configuration1.NearFields’

To get a handle on either of the near fields in the session, the key string must be used in conjunc-
tion with the get function. The key string is in the format:
"[Model].[Configuration].[Request Name]"

The following line of code will make the near field from the Horn model available for processing
in the scripting environment under the variable name nf.
nf = pf.nearfield.get("Horn.Configuration1.NearField1")

This process can be applied to any result type. The result is a dataset, which will be explained
in the section pertaining to dataset structures (see section 10.3).

10.3 The structure of a dataset

In order to modify or generate a dataset, it is important to understand how they are con-
structed. The structure of a dataset contains three types of data:

• axes refer to the independent axes along which data can be plotted in a view.

• quantities refer to the types of data that can be plotted along a specified axis.

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-3

Figure 10-2: Visualisation of the dataset.

• metadata refers to additional data that pertains to a dataset that will affect how POST-
FEKO interprets the dataset.

The additional data field refers to the actual values of the results. The indexing of the data is
dependent on the structure of the axes and quantities that were defined. Since the data fields
should not be accessed directly, it is not discussed here. Rather, the discussion of the proper
access methods will follow (see section 10.4).

Table 10-2: Supported units.

Distance m
inch
feet
mile
mil
Frequency Hz
Time s
min
hr
Mass g
t
Temparature C
K
F
Current A
Charge C
Force N
Potential Difference V
Resistance Ohm
Conductivity S
Power W
Capacitance F
Inductance H

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-4

Table 10-2 shows all of the base units currently supported by POSTFEKO. For all axes and
quantities, the units need to be specified to help ensure that POSTFEKO knows how to man-
age the values. More complex units may be constructed out of the base unit from the table by
using the operators “/” and “ˆ”. For example, acceleration can be written as “m/sˆ2”.
Data can also consist of different types of data. Table 10-3 indicates the supported data types.

Table 10-3: Supported data types.

scalar Any real valued data.

complex Complex values; values containing real and imaginary components.
boolean Either a true or false value.

10.3.1 Axes

An axis is a dimension along which data can be calculated. They typically contain the set of
physical points where data is calculated, the frequency at which data is calculated, etc. A full list
is given in Table 10-4. Any of these axes may be added to a dataset.

Table 10-4: Axis types for the axes field.

Position axes x
y
z
theta
phi
rho
r
Frequency axis frequency
Other axes arbitrary
index
solution
undefined
time

As an example, we pull in a near field from a horn antenna and print out the different axes. Note
that the near field specifies three spatial dimensions and a frequency dimension.
nf = pf.nearfield.get("Horn.Configuration1.NearField1")
for index, axis in pairs(nf.axes) do
print("axis["..index.."]:",axis)
end
axis[1]: "frequency" axis with a single value at 1.645E+009Hz
axis[2]: "x" axis spanning from -0.26m to 0.26m with 21 values
axis[3]: "y" axis spanning from -0.2m to 0.2m with 21 values
axis[4]: "z" axis with a single value at 0.46m

Analysing the example, one can see that the near field was calculated at a single frequency and
at a single height. The other two spatial dimensions form a rectangle, making it clear that the

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-5

near field is a flat surface at a fixed height and frequency. Each axis also indicates the unit in
which the axis is measured. The spatial axes are measured in metre (m), while the frequency
axis is measured in Hertz (Hz). All of the supported units are given in Table 10-2. Note also that
the data type for all of the axes are scalar. In other words, all of the axis values in this instance
contain real values.

10.3.2 Quantities

A quantity is a value that is calculated at each point for each axis in a dataset. More than
one quantity can be calculated for any position on the axes; each one typically representing a
different type or component of a result. For example, the Ex , Ey and Ez components of an electric
near field would be stored as three separate complex quantities. All of these quantities are valid
at the same physical position and frequency. Looking again at the near field example of a horn
antenna, the structure of quantities are illustrated.
nf = pf.nearfield.get("Horn.Configuration1.NearField1")
inspect(nf.quantities)
[hfieldcomp1][+]
[unit] A/m
[dataType] complex
[hfieldcomp2][+]
[unit] A/m
[dataType] complex
[hfieldcomp3][+]
[unit] A/m
[dataType] complex
[efieldcomp1][+]
[unit] V/m
[dataType] complex
[efieldcomp3][+]
[unit] V/m
[dataType] complex
[efieldcomp2][+]
[unit] V/m
[dataType] complex

The quantities contained in this near field are the three components of both the electric and
magnetic field. These values are specified at every dimension for the specified axes values. In
order to specify a near field component, complex values must be used. The units are specified as
A
m
for the electric near field components and mV for the magnetic components. If we were to inspect
the entire dataset, a single data point would then contain six values.
[hfieldcomp1] -0.0016298448547029 + 0.0030584305657125i
[hfieldcomp2] 0.00019556510683489 + 5.7238782893743e-005i
[hfieldcomp3] 0.0023804214721832 - 0.00073565845419827i
[efieldcomp1] 0.33955688994892 - 0.15945928741613i
[efieldcomp2] -0.1306987855743 - 0.50674748783847i
[efieldcomp3] 0.10921303015844 - 0.13966547870835i

A dataset may contain any quantity. However, if the dataset is to be used as an internal type,
the minimum quantities required for POSTFEKO to interpret the data correctly must be present.
For instance, a near field must contain either a complete set of electric field components or a
complete set of magnetic field components in order to be valid. This restriction only applies to a
dataset that will be used as though it is a near field calculated by the FEKO kernel.

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-6

10.3.3 Metadata

Sometimes it is necessary to include additional data that is not relevant to a specific axis or
quantity. This information generally tells POSTFEKO to treat the values in a slightly different
way. For example, the x-, y- and z-axis is relative to a local coordinate system. The metadata
fields can be used to specify the local origin and axis rotations so that the axes values are
interpreted correctly.

10.4 Processing and modifying a dataset

A typical use for the scripting functionality is to modify existing results. The steps that will be
followed in this case would typically be:

1. Create and simulate a model so that results are available.

2. Create a math script:

(a) Pull a single result or multiple results into the script.

(b) Perform processing on the results in the scripting environment.
(c) Store the results of the script in a dataset that is accessible in the POSTFEKO session.

3. Display and process the results.

These steps assume that a valid dataset is pulled in and being manipulated. Therefore, it is not
necessary to create or modify any of the axes, quantities or metadata fields. The data stored
in the dataset must simply be accessed and modified. However, the data block in a dataset
cannot be accessed directly (see Figure 10-2). Instead, an indexing scheme is available with
which to reach an element. Since individually indexing each element can become cumbersome,
a library function is made available that iterates through all of the elements.

10.4.1 Indexing a single element in a dataset

This type of scheme is used when fine control is required when accessing a dataset. It is also the
most intuitive method of accessing the elements. Each axis in a dataset contains a set number
of values. By specifying the index of each value on an axis, the values can be accessed.
nearField[1][1][1][1].efieldcomp1 = nearField[1][1][1][1].efieldcomp1 * 2

The previous command will multiply the value of the Ex at the first frequency, at the first indexed
point in space. By iterating through all of the axes, it is possible to selectively modify specific
values based on the index of the axes. In the following example, all fields more than a 0.15m
from the z-axis is set to 0.

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-7

-- Create the near field dataset

nf = pf.nearfield.get("Horn.Configuration1.NearField1")
-- Iterate through all of the axes and modify the field values
axis1 = nf.axes[1]; axis2 = nf.axes[2]; axis3 = nf.axes[3]; axis4 = nf.axes[4]
for freq = 1, #axis1 do
for x = 1, #axis2 do
for y = 1, #axis3 do
for z = 1, #axis4 do
if ( math.sqrt(axis2[x]^2 + axis3[y]^2) >= 0.15 ) then
nf[freq][x][y][z].efieldcomp1 = 0 + 0*i
nf[freq][x][y][z].efieldcomp2 = 0 + 0*i
nf[freq][x][y][z].efieldcomp3 = 0 + 0*i
nf[freq][x][y][z].hfieldcomp1 = 0 + 0*i
nf[freq][x][y][z].hfieldcomp2 = 0 + 0*i
nf[freq][x][y][z].hfieldcomp3 = 0 + 0*i
end
end
end
end
end
return nf

10.4.2 Iterating through all elements in a dataset

For the individual element style of indexing, it is necessary to manually iterate over each element.
Expressions can become difficult to work with, so an alternative is provided. The library function
dataset.forallvalues provides a simpler method for iterating through a dataset.
nf1 = pf.nearfield.get("Horn.Configuration1.NearField1")
nf2 = pf.nearfield.get("modNF_indiv")
nfAverage = pf.nearfield.get("modNF_indiv")

function average(index, target, source1, source2)

target[index].efieldcomp1 = 0.5*(source1[index].efieldcomp1 + source2[index].efieldcomp1)
target[index].efieldcomp2 = 0.5*(source1[index].efieldcomp2 + source2[index].efieldcomp2)
target[index].efieldcomp3 = 0.5*(source1[index].efieldcomp3 + source2[index].efieldcomp3)
target[index].hfieldcomp1 = 0.5*(source1[index].hfieldcomp1 + source2[index].hfieldcomp1)
target[index].hfieldcomp2 = 0.5*(source1[index].hfieldcomp2 + source2[index].hfieldcomp2)
target[index].hfieldcomp3 = 0.5*(source1[index].hfieldcomp3 + source2[index].hfieldcomp3)
end
dataset.forallvalues(average, nfAverage, nf1, nf2)
return nfAverage

Note that in this example, it was never necessary to define a loop. Instead, a function is written
that explains what should happen to a given element. Any number of dataset sources can be
provided as source inputs. The result of the calculation is then stored in the target dataset.

10.5 Creating a custom dataset

Another use for scripting is for when custom data objects are required. In these cases, the user
would typically follow these steps:

1. Create a POSTFEKO session.

2. Create a math script. Note that the math script must not be of any specific type, other-
wise POSTFEKO will attempt to validate the axes to make sure that the expected dataset
structure is returned.

(a) Create a new dataset.

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-8

(b) Add all of the desired axes and quantities fields. In this case any quantity can
be defined, since POSTFEKO will not attempt to interpret them in any particular way.
axes are still confined to the pre-defined values, where the arbitrary axis can be
used for any unspecified axis type.
(c) Perform processing on the results in the scripting environment.
(d) Store the results of the script in a dataset that is accessible in the POSTFEKO session.

3. Display and process the results.

Consider the following example. Here the script looks at an existing near field. When the mag-
nitude of the field at any point exceeds 5 mV , the custom dataset stores that value. For all other
values it is zero.
nf1 = pf.nearfield.get("Horn.Configuration1.NearField1")
-- Create custom dataset
custom = dataset()
axisF = dataset.axis("frequency", "Hz", 1.645e9)
axisX = dataset.axis("x", "m", -0.26, 0.26, 21)
axisY = dataset.axis("y", "m", -0.2, 0.2, 21)
axisZ = dataset.axis("z", "m", 0.46)
dataset.addaxis(custom, axisF)
dataset.addaxis(custom, axisX)
dataset.addaxis(custom, axisY)
dataset.addaxis(custom, axisZ)
dataset.addquantity(custom, "above5V", "scalar", "V/m")

-- Populate the values

function process5Vthreshold(index, target, source1)
local magEx = source1[index].efieldcomp1:magnitude()
local magEy = source1[index].efieldcomp2:magnitude()
local magEz = source1[index].efieldcomp3:magnitude()
local totalE = math.sqrt(magEx^2 + magEy^2 + magEz^2)
if (totalE >= 5) then
target[index].above5V = totalE
else
target[index].above5V = 0
end
end
dataset.forallvalues(process5Vthreshold, custom, nf1)

return custom

Here the dataset called custom has the same spatial axes of the source near field. An arbitrarily
defined quantity was defined. For each position on every axis, a scalar value must be defined for
the dataset to be valid. The dataset was created entirely within the script. The axes and
quantities were added to the dataset and the data was populated. The result can also be
plotted on a 3D or 2D view, similar to any internal data type.

10.6 Structures for internal dataset types

When creating a script that returns a pre-defined result type, it is necessary to add certain min-
imum fields so that POSTFEKO can process it correctly. For each type, the minimum values will
be provided in the following sections.

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-9

10.6.1 Near field

Near field results must contain either a complete set of electric field components, a complete set
of magnetic field components, or a complete set of both in order to be valid. For each different
coordinate system, a different set of spatial axes are required. Table 10-5 will summarise the
required fields.

Table 10-5: Required properties for near fields.

Property name Property type Unit Description

frequency axes Hz Every near field requires a valid frequency
axis.
axis 1 axes Every near field requires three independent
axis 2 spatial axes. Depending on the coordinate
axis 3 system, these axes may vary. Table 10-6 will
give a breakdown of the required axes for
each system.
efieldcomp1 quantities V/m Required quantity set for electric fields. All
efieldcomp2 three complex values must be defined for a
efieldcomp3 complete electric field definition.
hfieldcomp1 quantities A/m Required quantity set for magnetic fields. All
hfieldcomp2 three complex values must be defined for a
hfieldcomp3 complete magnetic field definition.
origin metadata The local origin for the workplane around
which the far field is defined.
uvector metadata A point relative to the origin which indi-
cates in which direction the û-vector.
vvector metadata A point relative to the origin which indi-
cates in which direction the v̂-vector.
conical metadata A boolean flag indicating whether the near
field is defined in a conical coordinate system.
The value may be either true or false.
rhostepsize metadata A scalar value indicating the the increment
with which the cone’s rho axis increases. This
effectively controls the angle of the cone for
a conical system. Note that this field is only
required if Conical is true.

The conical coordinate system is a special case, since it is technically not a complete coordinate
system. As such, some additional information is required. For near fields defined in the conical
coordinate system, the following must also be provided under the metadata fields:

1. conical = true. This flag indicates to POSTFEKO that the dataset fields must be inter-
preted differently than other coordinate systems.

2. rhostepsize. The step size indicates the intervals with which the rho axis grows. Note
that the rho axis only contains a single value which corresponds to the starting point of the
cone.

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-10

Table 10-6: Coordinate system axes.

Coordinate system axis 1 axis 2 axis 3

Cartesian x y z
Spherical theta phi r
Cylindrical (x-axis) rho phi x
Cylindrical ( y-axis) rho phi y
Cylindrical rho phi z
Conical rho phi z

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-11

10.6.2 Far field

Table 10-7: Required properties for far fields.

Property name Property type Unit Description

frequency axes Hz Every far field requires a valid frequency axis.
theta axes deg Elevation angle component relative to the lo-
cal workplane. The vertical position sits at
θ = 0◦ .
phi axes deg Azimuthal angle component relative to the lo-
cal workplane.
incidenttheta axes deg The theta direction from where an incident
plane wave originates.
incidentphi axes deg The phi direction from where an incident
plane wave originates.
theta quantities V This is a complex value indicating the theta
component of the electric field in this theta
direction, or Eθ .
phi quantities V This is a complex value indicating the phi
component of the electric field in this phi di-
rection, or Eφ .
directivityfactor
quantities A scaling factor that will scale the magnitude
of a field value to the expected directivity in
a given direction. The formula that is used is
given below in Eq. 10-1.
gainfactor quantities A scaling factor that will scale the magnitude
of a field value to the expected gain in a given
direction. The formula that is used is given
below in Eq. 10-2.
realisedgainfactor
quantities A scaling factor that will scale the magnitude
of a field value to the expected realised gain
in a given direction. The formula that is used
is given below in Eq. 10-3.
rcs quantities A scaling factor that will scale the magnitude
of a field value to the expected radar cross
section for a given observation direction.
origin metadata The local origin for the workplane around
which the far field is defined.
uvector metadata A point relative to the origin which indi-
cates in which direction the û-vector.
vvector metadata A point relative to the origin which indi-
cates in which direction the v̂-vector.

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-12

Directivity

The formula used to calculate directivity is given. The directivity is a figure of merit indicating in
which direction the most energy will be radiated. Directivity is the power density radiated in any
direction versus an isotropic radiator which is radiating the same amount of energy.

D = directivityfactor × [Re(Eθ )2 + Im(Eθ )2 + Re(Eφ )2 + Im(Eφ )2 ] (10-1)

Gain

The formula used to calculate gain is given. Gain is calculated in the same way as directivity,
except that the input power is used rather than the radiated power. In other words, system losses
are taken into account.

G = gainfactor × [Re(Eθ )2 + Im(Eθ )2 + Re(Eφ )2 + Im(Eφ )2 ] (10-2)

Realised Gain

The formula used to calculate realised gain is given. Realised gain is calculated in the same way
as gain, except that the power which is reflected back to the input port is taken into account. In
other words, system losses and mismatch effects are included.

RG = realisedgainfactor × [Re(Eθ )2 + Im(Eθ )2 + Re(Eφ )2 + Im(Eφ )2 ] (10-3)

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-13

10.6.3 Source

Table 10-8: Required properties for sources.

Property name Property type Unit Description

frequency axes Hz Every source requires a valid frequency axis.
current quantities A Currents are complex values that are mea-
sured in Ampere.
admittance quantities S The reciprocal of impedance and is measured
in Siemens and is a complex value.
power quantities W The rate at which energy is expended. This is
equal to the current times the voltage and is
measured in Watts, which is a complex value.
type quantities A value must be given to help indicate what
type of source the dataset represents. A
value of 0 represents a voltage source, where
a value of 1 represents a current source.
impedance quantities Ohm The total opposition to current flow and the
ratio of the voltage to the current. It is mea-
sured using the complex value Ohm.
mismatchloss quantities The fraction of power that is reflected to or
transmitted to other ports; i.e. the amount of
power that does not enter the system. This is
a scalar value with no unit.
voltage quantities V The potential difference over the port which
is a complex value measured in Volts.

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-14

10.6.4 Load

Table 10-9: Required properties for loads.

Property name Property type Unit Description

frequency axes Hz Every loads datasetrequires a valid fre-
quency axis.
current quantities A Currents are complex values that are mea-
sured in Ampere.
impedance quantities Ohm The total opposition to current flow and the
ratio of the voltage to the current. It is mea-
sured using the complex value Ohm.
power quantities W The rate at which energy is expended. This is
equal to the current times the voltage and is
measured in Watts, which is a complex value.
voltage quantities V The potential difference over the port which
is a complex value measured in Volts.

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-15

10.6.5 Network

Table 10-10: Required properties for networks.

Property name Property type Unit Description

frequency axes Hz Every networks datasetrequires a valid fre-
quency axis.
arbitrary axes This is an axis that lists the port numbers for
the network.
current quantities A Currents are complex values that are mea-
sured in Ampere.
impedance quantities Ohm The total opposition to current flow and the
ratio of the voltage to the current. It is mea-
sured using the complex value Ohm.
power quantities W The rate at which energy is expended. This is
equal to the current times the voltage and is
measured in Watts, which is a complex value.
voltage quantities V The potential difference over the port which
is a complex value measured in Volts.

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-16

10.6.6 S-Parameter

Table 10-11: Required properties for scattering matrices (s-parameters).

Property name Property type Unit Description

frequency axes Hz Every loads datasetrequires a valid fre-
quency axis.
arbitrary axes The string values here represent the scatter-
ing matrix element definitions in the form
S x, y, which can be read as “the ratio be-
tween the voltage wave measured at port x
as a result of the voltage wave inserted at
port y.”
portflag quantities A value indicating whether the port is active
or passive (or both). If the value is 0, then the
port is treated as a passive port. If the value
is 1, then the port is considered to be adding
and absorbing energy from the model.
load quantities Ohm A complex component that is added to a
model that has the property of being able to
dissipate energy.
sparameter quantities Scattering parameters are complex values
that relate the voltage wave measured at a
port vs. the voltage wave that is inserted into
port.

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-17

10.6.7 Power

Table 10-12: Required properties for power.

Property name Property type Unit Description

frequency axes Hz Every power datasetrequires a valid fre-
quency axis.
activepower quantities W The amount of power that is being inserted
into the system. This power could be radiated
or absorbed by other ports.
efficiency quantities A percentage value indicating the relationship
between the active power and the power pro-
vided to the system.
powerloss quantities W The amount of power that is dissipated in the
system. This could refer to any power lost due
to lossy materials, etc.

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-18

10.6.8 Transmission and reflection coefficients

Table 10-13: Required properties for Transmission and Reflection coefficients.

Property name Property type Unit Description

frequency axes Hz Every transmission / reflection coefficient
datasetrequires a valid frequency axis.
CoPolarisedReflectionCoefficient
quantities A complex value giving the ratio between the
reflected wave vs. the incident wave. The co-
polarised component is given and is defined
in the direction of the polarisation angle of
the incident plane wave.
CoPolarisedTransmissionCoefficient
quantities A complex value giving the ratio between the
transmitted wave vs. the incident wave. The
co-polarised component is given and is de-
fined in the direction of the polarisation angle
of the incident plane wave.
CrossPolarisedReflectionCoefficient
quantities A complex value giving the ratio between the
reflected wave vs. the incident wave. The
cross-polarised component is given and is de-
fined in the direction of the polarisation angle
of the incident plane wave.
CrossPolarisedTransmissionCoefficient
quantities A complex value giving the ratio between the
transmitted wave vs. the incident wave. The
cross-polarised component is given and is de-
fined in the direction of the polarisation angle
of the incident plane wave.

July 2011 FEKO User’s Manual

SCRIPTING GUIDE FOR POSTFEKO 10-19

10.6.9 Specific Absorption Rates (SAR)

Table 10-14: Required properties for Specific Absorption Rates (SAR).

Property name Property type Unit Description

frequency axes Hz Every specific absorption rate dataset requires
a valid frequency axis.
hasAverageSAROverTotalVolume
quantities This is a boolean value (i.e. true
or false). If the flag is true, then
averageSAROverTotalVolume must
be specified.
averageSAROverTotalVolume
quantities W/kg Average SAR over the total volume of the
model. This quantity is a scalar value.
hasAverageSAROverRequestedVolume
quantities This is a boolean value (i.e. true
or false). If the flag is true, then
averageSAROverRequestedVolume must
be specified.
averageSAROverRequestedVolume
quantities W/kg Average SAR over a specifically requested
sub-volume of the model. This quantity is a
scalar value.
hasPeakSAR
quantities This is a boolean value (i.e. true or false).
If the flag is true, then massOfPeakSARCube,
airFractionOfPeakSARCube and
peakSARInCube must be specified.
massOfPeakSARCube
quantities kg The mass of the SAR cube in kilograms. The
value is a scalar and is typically 0.01 kg or
0.001 kg.
airFractionOfPeakSARCube
quantities Air fraction in percent for this specific cube.
The value is a scalar.
peakSARInCube
quantities W/kg Peak SAR in this sepcific cube. The value is a
scalar.

July 2011 FEKO User’s Manual

Part IV

Working with EDITFEKO

INTRODUCTION TO EDITFEKO 11-1

11 Introduction to EDITFEKO

The editor EDITFEKO was designed to simplify the process of generating and editing *.pre
(PREFEKO input) files. Though the *.pre file is a standard ASCII text file that may be created
using any available text editor, EDITFEKO is a basic text editor with customised functionality.
Note that EDITFEKO was also used as an editor for the optimisation definition in a text *.opt
file before the release of FEKO Suite 5.3. This function is no longer supported by EDITFEKO
and all optimisation-related options should be defined in the optimisation tools provided in the
CADFEKO interface (see section 7).
In this chapter, the general usage of the EDITFEKO tool, and the functions provided to aid in the
creation and editing of specific cards in a *.pre file are discussed. More information regarding
the usage of symbolic variables, advanced flow control using FOR loops and IF statements, adding
termination conditions and printing information to the screen is discussed in the chapter dealing
with PREFEKO (see section 17).
The model geometry and desired calculations are entered on separate lines of text in the *.pre
file, referred to as cards. (For models built in CADFEKO the complete geometry is imported using
one of these cards.) Each card may have a number of parameters which must appear in fixed
positions. The arrangement of the cards is also important and controls the order of the steps
that will be taken during simulation. (Some information regarding the card-order is found in
the chapters describing the geometry cards (see section 13) and control cards (see section 14).
More detailed information regarding the general structure of the *.pre file may be found in the
chapter on advanced modelling and solution control (see section 12.2)).
The EDITFEKO interface consists of an editor area (see Figure 11-1) where more than one file may
be opened — each in its own edit window. The toolbars at the top provide quick access to some
standard functions. The status bar on the bottom of the window contains licence information,
indicates when “Overwrite” is on, the line and column numbers at the current cursor position
and also gives an indication of the file modification status. In the standard PREFEKO mode the
button panel provides quick access to the card editor dialogs. When editing a TIMEFEKO *.tim
file, EDITFEKO currently provides only a plain text editor with no special tools or functionality.
The program may be started from a command prompt by entering
editfeko example.pre

where the optional parameter (here example.pre) specifies the file to be opened. If the param-
eter does not contain an extension, the extension pre is added. Also, if such a file does not exist,
it will be created. EDITFEKO can also be opened from inside CADFEKO and POSTFEKO.
If a file is opened in EDITFEKO a lock-file is created (with the name .filename.ext.lock) in
the same directory as the file. If you open the file in another instance of EDITFEKO at the same
time, it will give a warning (“A lock file already exists for this file. . . . Are you sure you want to
continue with this operation?”) with options to cancel or continue. Generally one should then
click Cancel and switch to the other instance of EDITFEKO. One should only continue here if
the lock file resulted after EDITFEKO was not closed properly (such as a power failure or system
crash).

July 2011 FEKO User’s Manual

INTRODUCTION TO EDITFEKO 11-2

Figure 11-1: The EDITFEKO interface

11.1 Toolbars

The toolbars in EDITFEKO are usually docked at the top of the EDITFEKO window, but they can
be dragged free or docked on any of the four sides. They can also be switched off by selecting
File → Preferences from the main menu and selecting the Appearance tab.
Note that the splitter between the editor and the dialog panel may be dragged to resize the panel.

11.1.1 File control toolbar

Used for file and print control. The functions of the buttons are:

1. Create a new *.pre file and switch to PREFEKO mode.

2. Create a new *.tim file.
3. Open a file (and switch to the appropriate mode).
4. Save the current file.
5. Print current file.

July 2011 FEKO User’s Manual

INTRODUCTION TO EDITFEKO 11-3

11.1.2 The edit and window toolbars

The edit toolbar provides standard edit commands while the window toolbar allows positioning
the edit windows in the editor area. The functions of the buttons are:

1. Undo. Multiple undo commands are supported.

2. Redo. When new text is added (by typing or adding a card) the redo chain is cleared and
the new text forms the last entry in the undo chain.
3. Cut.
4. Copy.
5. Paste.
6. Search for text.
7. Cascade all edit windows.
8. Tile windows vertically.
9. Tile windows horizontally.
10. Arrange windows as best as possible.

11.2 Menu items in EDITFEKO

11.2.1 File menu

The File menu allows opening and closing files and is similar to standard Windows applications.
Most of the options and the list of recently used files do not need any explanation.
The New sub menu item opens a sub menu where the user may choose *.pre for PREFEKO or
*.tim for TIMEFEKO files. EDITFEKO will enter the appropriate mode depending on the selected
option. The Save menu will ask for a file name if the file does not have one, otherwise it will save
(without confirmation) using the current file name. Various short-cut keys are presented next to
the respective menu items.
The file menu is also used to set user preferences (see section 11.3).

11.2.2 Edit menu

This menu provides general editing commands. Note that multiple level undo/redo is allowed.
Each change to the editor is recorded in the undo list. When the user clicks Undo the entry is
moved to the redo list and the change to the editor is undone. When new text is added (by typing
or adding a card) after a sequence of undo commands, the redo chain is cleared and the new text
is recorded at the end of the undo list at that time.

July 2011 FEKO User’s Manual

INTRODUCTION TO EDITFEKO 11-4

It is possible to search the editor for specific text, case sensitive or not. The search always starts
from the current cursor position. If it is desired to start the search from the top of the file,
the cursor must be moved there (using <Ctrl><Home>). Note that there is also a fast card
search. By right clicking on any button on the button panel, the next occurrence (starting from
the current cursor position) of the associated card is found. This search is always downwards,
and does not wrap back to the top of the file.
Note the Goto line option which is very useful to find lines reported by PREFEKO. (PREFEKO
always reports the line number where errors occur.)
The Comment and Uncomment item allow the user to comment out blocks of text or remove
the commenting indicators. This is often useful when debugging large files. The comments are
inserted as ** followed by a space and only removed if this sequence is present at the beginning
of the line.
For example, in the lines
** ** Select the label
** LA 1
*** Next comment

the comment characters will not be removed from the third line using this function. It was also
not generated by the Comment function. These commands are also available using the short-cut
keys <Alt><C> (for Comment) and <Alt><U> (for Uncomment) or by right clicking in the
editor and selecting from the resulting pop-up menu.

11.2.3 Accessing Geometry Control card definition panels

The Geometry cards and Control cards lists in the menu launch the card editor dialog panels of
the various cards when editing a *.pre file. The relevant panel may also be launched by clicking
on the appropriate buttons in the button panel. The cards in both the menu lists and on the
button panel are grouped by functionality.
The card representing the settings defined in the card definition panel will be added in the line
that the cursor was located when the panel was launched.
Pressing <F1> while the cursor is on a line containing a card definition will launch the relevant
card-definition panel for the card defined on that line.

11.2.4 Run menu

This menu is used to launch the various FEKO components. If another GUI component is started
from the Run menu, the relevant component of the currently active project is automatically
opened in the launched UI component. Note, however, that CADFEKO does not automatically
create a new model if a *.pre file with that name already exists. Hence if one runs CADFEKO
from EDITFEKO on a model for which only a *.pre file exists, CADFEKO will open with a new
nameless model. If this model is saved to the same name, it will overwrite the existing *.pre
file. Hence one should generally create models in CADFEKO and then run EDITFEKO. Note that
when running PREFEKO or FEKO, the current file is saved without confirmation.
Select Run → Component parameters to set the command line options for the various kernel
components. See section 6.16.1 for more information.

July 2011 FEKO User’s Manual

INTRODUCTION TO EDITFEKO 11-5

It is possible to launch the solution on a remote machine. If the Remote FEKO execution (see
section 18.2.3) item is checked, EDITFEKO launches the solution on the host specified in the
Remote execution group of the FEKO tab of the Component parameters dialog.
If Parallel FEKO execution is checked, the solution is launched on a parallel cluster as defined in
the Parallel execution group of the FEKO tab of the Component parameters dialog. Unless the
parallel job is also a remote one, the local machine must be the first host name in the configura-
tion. (See also section 18.2.2.)
If Farm out by OPTFEKO is checked, parallel optimisation are enabled. It is also possible to check
both the Parallel FEKO execution and the Farm out by OPTFEKO. The combination of the two
will result in farming during optimisations using parallel FEKO runs.

11.2.5 Window menu

This menu presents the standard Windows functionality to arrange or select the editor windows
on the main editor area.

11.2.6 Help menu

In addition to providing access to information regarding the version of EDITFEKO (About EDIT-
FEKO), this menu contains shortcuts to the EDITFEKO section of the FEKO help (Help → FEKO
help) and the FEKO User’s Manual and Getting Started Manual documents. These manuals are in
PDF format and a PDF viewer must be configured on the Preferences dialog (see section 11.3).
Contextual help can also be accessed by pressing <F1> while EDITFEKO is open.
The Check for updates item opens the Check for updates dialog. This polls the EMSS web-site
and informs the user if any updates have been added since the last check. (Since EDITFEKO does
not know if the user also has licences for other platforms, it gives an update message regardless
of which platform the update is intended for.) The update checker downloads one file containing
the list of updates. It does not automatically download anything else and does not send any
information to the web-site. If Activate automatic update checking is checked, EDITFEKO polls
the web-site each time it is started more than seven days after last checking the web-site. (Note
that this is off by default.) It is also possible to configure a proxy server if this is required to
obtain access to the internet. Click the Check button to poll the web-site. The Info tab applies to
the update checker, not the available update.

11.3 Preferences in EDITFEKO

Selecting File → Preferences opens the Configure EDITFEKO dialog. Various options are grouped
under separate tabs as follows.

11.3.1 Appearance tab

Here the user can

July 2011 FEKO User’s Manual

INTRODUCTION TO EDITFEKO 11-6

• Switch off the top toolbar.

• Set the printer font.

• Set the size of the screen fonts.

• Set the preferred card format.

In the column based card format, the fields are located at fixed positions and cannot be longer
than 5 (string and integer fields) or 10 (real fields) characters. In the colon separated format,
the different fields in each card are separated with :-characters. This format allows arbitrary
length entries. If EDITFEKO is using column based format, it will automatically switch to colon
separated format if any of the input fields are longer than the allowed length.

11.3.2 Options tab

Here one may

• Elect to show superuser options if this is allowed by the licence. The superuser mode in
FEKO is only used during program development and is not available to the general user.
If one tries to edit an existing card containing superuser parameters, EDITFEKO will also
prompt the user and switch to superuser mode if this is allowed. If the licence does not
allow superuser mode, superuser options will be removed.

• Turn on auto-save and/or backups and specify the auto-save interval.

• Specify the PDF reader required for the manuals.

11.4 Inserting and editing input cards and variables

11.4.1 Generating input cards

Cards can be entered and edited directly in the editor window. In most cases the user would
call one of the card editor dialogs by clicking on the appropriate button on the button panel or
selecting from the Geometry cards and Control cards menus.
The buttons are grouped according to functionality namely

• Definition of node points, labels, segmentation parameters and variables (# button).

• Cards to generate surfaces which will be meshed into triangles elements.

• Cards generating wire segments.

• Cards for generating dielectric / magnetic volumes to be subdivided into cuboids, or speci-
fication of dielectric regions.

• Cards used in connection with the PO formulation.

• Cards to define the MLFMM and the FEM methods and parameters.

July 2011 FEKO User’s Manual

INTRODUCTION TO EDITFEKO 11-7

• Cards used in connection with the UTD formulation.

• Modifications to the geometry, such as translation, scaling and mirroring as well as the
IN card which allows inclusion of an external file - typically a secondary *.pre file or
geometry from an external file.

• The EG card signifying the end of the geometry. Generally the buttons above it should be
used before the EG card and those below it should be after the EG card. Exceptions are the
DP and IN cards and the FOR–NEXT and IF-ELSE blocks.

• Cards dealing with finite medium properties and finite grounds.

• Cards dealing with loading of wire segments.

• General solution cards, including Non radiating networks, transmission lines, and solution
method parameters.

• Cards to specify the excitation type, frequency and power of the excitation.

• Cards used to control which data is written to the output files.

• Cards to specify where to calculate the fields, and other solution requests.

• The EN card which signifies the end of the input file.

A typical card editor dialog is shown below.

When a card editor dialog is selected from either the button panel or the menu, there is a button
to Add card. This adds the card without closing this dialog. This is useful when defining a number
of similar cards, for example, when specifying the first group of DP cards at the start of a model.
Often only one or two parameters differ between cards. Clicking OK adds the card and closes
the editor dialog. Pressing <Enter> has the same effect as clicking on the OK button. Similarly
<Esc> closes the card editor dialog without adding the card as if the Cancel button has been
clicked.
In the card editor dialogs, moving the mouse pointer over the labels may give more information
of the specific input field. If the user is uncertain about the meaning or units of a given field,
it is always advisable to move the mouse pointer to the label to determine if any additional
information is available.
All the input fields in FEKO have a fixed anticipated length. In the card editor dialogs it is
possible to enter longer strings than anticipated for the particular field. If longer strings or values
that the maximum anticipated number of characters are entered, EDITFEKO will revert to the
colon-delimited format (see section 11.3) in which there is no field-length limitation.

July 2011 FEKO User’s Manual

INTRODUCTION TO EDITFEKO 11-8

Editing an existing line

The editor dialogs can also be used to edit existing lines. The line at the current cursor position
may be edited either by pressing <F1> or by right clicking at any position in the editor and
selecting Edit line at cursor from the pop-up menu. Note that right clicking does not change the
cursor position. When the selection (or highlighted part) in the editor spans more than one line,
the line containing the cursor is edited. For multi-line cards the cursor can be anywhere in the
card.
EDITFEKO processes each card as if it does not contain errors. For some of the fields where the
user may select a number of options, the editor dialog’s display will default to a “common” value
if the input field is invalid. Thus it is advisable to close the card editor with OK even when it was
used just to check the validity of the line.
If the user types, for example, PS and then presses <F1>, EDITFEKO may treat it either as a
new card (with default options), or as an existing card with all options blank (zero). Setting all
options to zero does not always make sense and EDITFEKO thus treats this case as a new card
and uses the normal defaults. If the user wants to have a card with all options zero, he should
type the card name followed by a space before pressing <F1>.
When editing existing lines, the Add card button is not available as the card is edited and written
back to the editor.

Adding a comment label to an existing card

A comment or label field (see section 13.1) may be added to the end of the line representing a
card after a ‘**’ comment indicator. Label comments are automatically appended to all relevant
cards in *.pre files that are generated by CADFEKO according to the labels in the CADFEKO
model. The labels assigned to the EDITFEKO control cards are important in that they are used to
uniquely identify the excitations and solution requests referred to in an optimisation setup (see
section 7.1.4). In POSTFEKO the labels of solution request cards are also used in the legend text
when plotting results from that card.

11.4.2 Variable editor

The # button on the top right corner of the button panel launches the card editor to edit variables.
This is useful as it presents a list of functions and operations understood by FEKO and calculates
the value of the variable as it would be evaluated by PREFEKO at this point.
The functions, variables or operations may be selected from the three drop-down lists and a
special group for the “FILEREAD” function. The selected item is automatically placed at the
current cursor position. Note that variables and functions are highlighted after insertion, but
operations are not. If a function is selected while some part of the variable string is selected, the
selected text will be inserted as the argument of the new function. Variables, operators and the
“FILEREAD” function replace the currently selected text.

July 2011 FEKO User’s Manual

INTRODUCTION TO EDITFEKO 11-9

11.5 Important keystrokes

The arrow keys as well as <Pg Up> and <Pg Dn> behave in the normal fashion. The following
keys may be different to some other applications. A block may be selected using the mouse,

Table 11-1: EDITFEKO cursor movements

Key combination Description

<Ctrl><Left Arrow> Move a word left
<Ctrl><Right Arrow> Move a word right
<Ctrl><Page Up> Move to top of visible page
<Ctrl><Page Down> Move to bottom of visible page
<Home> Move to beginning of line
<End> Move to end of line
<Ctrl><Home> Move to beginning of file
<Ctrl><End> Move to end of file

or pressing <Shift> and using the normal movement keys. If a block is selected, it will be
overwritten when a key is pressed. The following list of short-cut keys are often used:

Table 11-2: EDITFEKO short-cut keys

Short-cut key Description

<Ctrl><C> or <Ctrl><Ins> Copy to clipboard
<Ctrl><X> or <Shift><Del> Cut (delete) to clipboard
<Ctrl><V> or <Shift><Ins> Paste (insert) from clipboard
<Ctrl><S> Save
<Ctrl><A> Save all files
<F1> Edit line
<Alt><C> Comment line(s)
<Alt><U> Uncomment line(s)
<Ctrl><F> Find
<F3> Find next
<Ctrl><R> Find and replace
<Alt><2> Run PREFEKO
<Alt><3> Run POSTFEKO
<Alt><4> Run FEKO
<Alt><6> Run OPTFEKO
<Alt><7> Run TIMEFEKO

July 2011 FEKO User’s Manual

PREFEKO LANGUAGE AND CONCEPTS 12-1

12 PREFEKO language and concepts

12.1 PREFEKO

PREFEKO is the FEKO Suite pre-processor. It is responsible for the creation of the *.fek input
file used by the solution kernel. The input file for PREFEKO is the (*.pre) file that is edited using
EDITFEKO (*.pre files can be edited in any text editor). For more information about launching
PREFEKO from a command line see chapter 17.
This chapter presents the language and concepts required to create *.pre input files. The dif-
ferent input cards (geometry and solution control) are discussed in the section dealing with
EDITFEKO geometry cards (see section 13) and section dealing with EDITFEKO control cards
(see section 14) respectively.

12.2 Structure of the *.pre input file

The *.pre file is a text file that contains a list of high level commands. These commands are
translated by the FEKO preprocessor (PREFEKO) to generate the input for the FEKO solver. The
*.pre file commands are referred to as cards, and each card performs a specific function.

12.2.1 Order of the cards in the *.pre file

The cards can be categorised into geometry cards, control cards (which define the field param-
eters to calculate) and execution flow cards. PREFEKO processes these cards to generate the
*.fek file.
When a model is saved in CADFEKO, CADFEKO writes a *.pre file that includes a card that
indicates that the meshed CADFEKO model (stored in a *.cfm file) should be imported. The
user may modify this card if required to change the import process. For example, the card may
be changed so that structures on a symmetry plane are not imported — the symmetry plane can
then be defined and an additional import card used to import the rest of the geometry. Applying
symmetry in this way may provide a more efficient simulation approach for certain advanced
cases.
Users are advised to use EDITFEKO to add and edit cards in the *.pre file. This will ensure
that each card is formatted correctly and that the required parameters are limited to the correct
values.

12.2.2 Format of cards in the *.pre file

The card format is described here mainly for users would like to generate FEKO input files auto-
matically. FEKO supports two different formats for cards - a colon-separated and a column-based
format. These formats may be mixed in a single input file. After the last column of a card (irre-
spective of the format) a comment may be added after the comment indicator (‘**’). For control
cards that define solution requests, this comment is used as a label for that card. The label of
a card is used by OPTFEKO to uniquely identify specific results, and is indicated in POSTFEKO

July 2011 FEKO User’s Manual

PREFEKO LANGUAGE AND CONCEPTS 12-2

to aid the identification of the solutions and blocks of interest when post-processing simulation
results.

Column based format The basic form of this card format is shown below. The upper numbers
indicate the columns. The name field (shown as ‘xx’) in columns 1 and 2 specifies the type
of the card. This is followed by five integer parameters I1 to I5 (for some cards, these input
fields may also contain strings such as node names) containing five digits each, and eight
real-value parameters R1 to R8 containing ten digits each.

1 6 10 15 20 25 30 40 50 60 70 80 90 100 110
xx I1 I2 I3 I4 I5 R1 R2 R3 R4 R5 R6 R7 R8

INT INT INT INT INT

STR STR STR STR STR REAL REAL REAL REAL REAL REAL REAL REAL

Colon separated format This is a less restrictive input format than the column-based format,
where the individual integer and real parameters are separated by a colon character. Unlike
the column-based format, when using this format, integer and real input fields are not
restricted to 5 or 10 characters respectively. In this card format, the card name is still
located in columns 1 and 2. The name is followed immediately in column 3 by a colon.
The rest of the card has no spacing limitations. The lines shown below demonstrate valid
syntax in this format.
DP: S1 : : : : : #x : #y : #z
BP: S2 : S2 : S3 : S4

All input and output parameters of FEKO are in SI units (e.g. lengths are in metres, potential in
volts, etc.) All angles are in degrees. FEKO includes various scaling options (see the SF, TG and
IN cards) so that dimensions may be entered in different units and scaled to metres.
The principal structure of the input file is shown below:

** Comments at the start of the input file

. . . Cards that define the geometry ** Comments
EG End of the geometry
... Control cards that define excitation, special solution op-
tions and indicate which quantities should be calculated **
Card labels / Comments
EN End of the input file
The description in the section dealing with EDITFEKO geometry cards (see section 13) gives
an overview of the cards related to the creation of geometry with detailed descriptions of the
individual cards (these cards include the IN card, which is used to control the import of external
meshes or CADFEKO models).
The description in the section dealing with EDITFEKO control cards (see section 14) gives an
overview and description of the control cards used to define materials, solution methods, excita-
tions and calculation requests for the model.

July 2011 FEKO User’s Manual

PREFEKO LANGUAGE AND CONCEPTS 12-3

12.3 Guidelines for creating geometry in EDITFEKO

12.3.1 Meshing guidelines regarding connectivity

As discussed in section 2.2.3, elements must be connected at edges or vertices to ensure electrical
connectivity. Most of these rules are automatically complied with when creating FEKO models in
CADFEKO. However, users must take care to conform to these rules when combining CADFEKO
models with EDITFEKO scripting (e.g. attaching an antenna modelled with geometry cards on an
aircraft meshed in CADFEKO), when creating the geometry solely in EDITFEKO, or when working
with imported meshes.
Note that cuboidal volume elements (used to model volume dielectrics — the DK card (see sec-
tion 13.10), the DZ card (see section 13.12) and the QU card (see section 13.40)) do not need
to be connected in this way.
When creating structures with scripting commands, wires are divided into segments that are
equal to or shorter than the specified segment length. For surfaces, the triangle edges along the
boundary of the surface are always equal to or shorter than the specified edge length. Thus,
meshing the same line with the same mesh size will always give the same number of exactly
equal divisions. The internal edges may, however, be longer than the specified edge length. This
is not necessarily the case with CADFEKO meshes where the specified mesh size is the average
size and the internal structure influences the placement of vertices along the surface boundaries.
When creating wire junctions as shown in Figure 12-1, it is important to ensure that the wire AB
have a vertex at point C. The best option is to construct this as two wires, one from A to C and
the other from C to B.
D
A B
C
Figure 12-1: Example of a wire structure

Similarly, where two surfaces touch the common edge must be part of both surfaces. For example,
the surface in Figure 12-2 should not be created as two rectangles ABFG and CDEF. If this is done,
it is highly unlikely that there will be ohmic connection along the line BF. There are a number of
ways to correctly create this structure. It can be created from the rectangles ABFG, CDHB and
BHEF or the quadrangles ABEG and BCDE. In both cases the contacting edges are common and
will be meshed correctly. The simplest way to mesh this structure is to create a single polygon
ABCD(H)E(F)G.

G F E

H
A B

C D
Figure 12-2: Example for a surface

A connection point between a segment and one or more triangles is only recognised when the
beginning or the end of the segment lies on the vertex or vertices of the triangles. In Figure 12-3

July 2011 FEKO User’s Manual

PREFEKO LANGUAGE AND CONCEPTS 12-4

Figure 12-3: Incorrect (left) and correct (right) connection between a segment and triangles

Figure 12-4: Incorrect (left) and correct (right) connection between curved edges

the left side is an incorrect and the right side a correct connection (here the segment is connected
to six triangles).
When curved structures (circles, cylinders, spheres) are modelled, a finer mesh may be used
along the curved edges to get a more accurate representation of the geometry. In this case the
same edge length should be used on both edges and the reference points should be identical. See
the example in Figure 12-4.

12.3.2 Reducing GUI mesh sizes

For very large models (or at very high frequencies) it is possible that the user interface com-
ponents cannot create and/or display the required mesh. To some extent, this problem can be
overcome by creating a mesh of larger elements in CADFEKO and using the RM card in EDIT-
FEKO to subdivide this further. The only restriction here is that the original mesh should use
much larger elements than the desired mesh. If this is not the case, the subdivision may result in
an unnecessary large number of elements.

12.4 Usage and concept of labels

In the scripting language, items are identified by their labels. These are then used to do opera-
tions or set electromagnetic properties on geometry. For instance in Figure 12-5 the label Feed
can be used to define the feed segment (voltage source), or the label MainRefl can be used to
instruct FEKO to use physical optics for the main reflector.

July 2011 FEKO User’s Manual

PREFEKO LANGUAGE AND CONCEPTS 12-5

Labels are set either directly in CADFEKO (see section 6.18.3), or by preceding geometry cards in
EDITFEKO with the LA card (see section 13.29). When importing certain mesh formats by means
of the IN card (see section 13.22) then labels can also be imported (for instance the NASTRAN
property gets converted into a FEKO label). In principle, FEKO labels can be:

• A positive integer number (including zero).

• Any valid expression, like 3*#i+2, can be used. These expressions are evaluated, and the
resultant numerical value is used as or in the label.

• A string of characters (valid are ‘a’-‘z’, ‘A’-‘Z’, ‘0’-‘9’ and the underscore ‘_’), optionally fol-
lowed by a variable (which starts with the ‘#’ sign (see section 12.5)). Such variables at
the end are evaluated and replaced by the corresponding numerical value (rounded to in-
teger). Note that string labels are case insensitive in FEKO, i.e. labels Roof and ROOF are
treated identically inside FEKO.

Figure 12-5: Example demonstrating the usage of labels (display of labels in colour with legend in POST-
FEKO)

So for instance these labels are valid

23
5*#k+#j/2
LeftWing
Front_Door
Part#i

while these labels are not valid

Left+Wing (invalid character ’+’)
-23 (negative integer)
Part_#i_#k (two variables)

With the CB card, FEKO allows the user to change labels (e.g. after having imported some geom-
etry). A powerful wild card globbing (expanding a non-specific label name containing a wildcard
character into a set of specific labels) is supported (see section 13.6). At certain FEKO cards the
user can specify label ranges and at other cards labels for created geometry can be derived from
other labels (eg. when using symmetry - SY card), and thus it is important for users to understand

July 2011 FEKO User’s Manual

PREFEKO LANGUAGE AND CONCEPTS 12-6

how the label algorithm works. The algorithm for this is that after having evaluated expressions
or having replaced variables, a label is split into the associated number and the remaining base
string. The associated number is split off from the back of the label, and if there are no digits,
this is set to zero. See Table 12-1 for examples of labels are split into the base string and the
associated number.

Table 12-1: Examples of splitting a label into its base string and associated number.

Label name Base string Associated number

1 1
Roof Roof 0
Part_17 Part_ 17

When incrementing labels, the base string is kept and the associated number is incremented.
There is just one exception — the label zero will always remain zero. See Table 12-2 for an
example when incrementing a label by two.

Table 12-2: Label incrementing example (increment by two).

Label name Incremented label

1 3
Roof Roof
Part_17 Part_19

When using ranges of labels, then for the range specification to be valid the base strings must
match, and the associated number must be in the correct range between the associated numbers
of start and end label. So for instance the label Part_12 is in the range of the labels Part_10
. . . Part_20, but the label Part_5 is outside this range. Using a range like Front . . . Back is
not valid (different base strings).

12.5 Symbolic variables

Instead of using numerical values in all the different cards, it is possible to use predefined vari-
ables. The name of a variable always consists of a #-sign followed by a string consisting of the
characters a-z, A-Z, 0-9 and the special character _. The following are valid variable names:
#height, #a, #STARTINGFREQUENCY, #a_1 or #P5_7f; while the following are not valid: #a?1
or #value2.1. There is no distinction between upper and lower case characters. For example,
#a and #A are interpreted as the same variable.

Variables

It is important to note that in CADFEKO variables are used without the #-sign whereas PREFEKO
requires the #-sign to distinguish between variables and functions (such as sin). When CADFEKO
writes the variables to the *.cfm file, it prepends the #-sign so that the variables can be used after
the IN card in the *.pre file. When using OPTFEKO in a model that has no CADFEKO geometry
defined, the *.cfm mesh import command must be removed, or variables that are defined in the

July 2011 FEKO User’s Manual

PREFEKO LANGUAGE AND CONCEPTS 12-7

*.cfm file must be excluded from the import, as these variable values will override the values
assigned by OPTFEKO if they are included.
Expressions and functions may be used when defining variables, so that direct calculations can be
carried out. The variables have to be defined before they can be used in the respective cards. It is
possible to use expressions like 2*#radius in the input fields subject to the maximum allowed
length (10 characters for real values, 5 characters for integer values). For larger expressions
additional variables have to be defined.
A variable is defined in the following way:
#2pi = 2*#pi
#vara = 1 + sqrt(2)
#varb = #vara * 2.3e-2 * (sin(#pi/6) + sin(rad(40)) + #vara^2)
#sum = #vara+#varb

Note that the # sign has to appear in the first column. Variables may also be defined from the
command line as described in the previous section.

Arrays

In addition to plain variables, arrays with integer indices like #a[5] or, more complex,
#am_0[3*#i+ceil(#r[2])]

are also supported. The expression between the square brackets must evaluate to an integer
number, which can also be negative. The implementation of using arrays is so that they do not
need to be allocated, however they need to be initialised. So for instance after having used the
instructions
!!for #i = 10 to 20
#array_a[#i] = 3*#i-10
#array_b[-#i] = 0
!!next

one will be able to use #array_a[10] or #array_a[17] or also #array_b[-12] in other

expressions. But trying to use for instance #array_a[5] or #array_b[0] will result in an error
message that an undefined variable is used.

12.6 Built-in functions

Any predefined expression or variable as discussed above, can be used in conjunction with any
of the following trigonometric-, Bessel-, miscellaneous-, coordinate- or data functions.
The following trigonometric functions as given in Table 12-3, are supported.

July 2011 FEKO User’s Manual

PREFEKO LANGUAGE AND CONCEPTS 12-8

Table 12-3: Trigonometric functions

SIN sine (argument in radians)

COS cosine (argument in radians)
TAN tangent (argument in radians)
COT cotangent (argument in radians)
ARCSIN arcsine (in radians)
ARCCOS arccosine (in radians)
ARCTAN arctangent (in radians)
ATAN2 this function has two arguments atan2(#y,#x)
it yields arctan(#y/#x) in the range −π. . .π
ARCCOT arccotangent
SINH hyperbolic sine
COSH hyperbolic cosine
TANH hyperbolic tangent

Table 12-4: Bessel functions

BESJ(n,x) Bessel function Jn (x) of integer order n ≥ 0 and real argument x

BESY(n,x) Neumann function Yn (x) of integer order n ≥ 0 and real argument x
BESI(n,x) Modified Bessel function of the first kind I n (x) of integer order n ≥ 0
and real argument x
BESK(n,x) Modified Bessel function of the second kind Kn (x) of integer order
n ≥ 0 and real argument x

July 2011 FEKO User’s Manual

PREFEKO LANGUAGE AND CONCEPTS 12-9

Table 12-5: Miscellaneous functions

SQRT square root

LOG logarithm to the base 10
LN natural logarithm
EXP exponential function
ABS absolute value
DEG convert radians into degrees
RAD convert degrees into radians
STEP step function, STEP(x) = 0 for x ≤ 0 and STEP(x) = 1 for x > 0
CEIL smallest integer value that is equal to or greater than the argument
FLOOR largest integer value that is equal to or smaller than the argument
MAX returns the largest of the two arguments — called as max(#a,#b)
MIN returns the smallest of the two arguments — called as min(#a,#b)
FMOD this function also has two arguments fmod(#a,#b) and returns the remainder
of the division #a/#b
RANDOM This function returns a random value in the range 0 . . . 1. If the argument X of
RANDOM() is -1, then a random number is returned. For any other argument
X in the range 0 . . . 1 this value is used to set the seed, and then a random
number is created using this seed. (Using the same seed allows one to create a
deterministic and reproducible random number series). If “RANDOM(-1)” is
called before any seed is set in the *.pre file, then the returned values are
random and not reproducible. (The internal seed is used based on the time
when PREFEKO is executed).

Table 12-6: Coordinate functions

X_COORD This function returns the x coordinate of a point previously defined by a DP

card, as discussed below.
Y_COORD Returns the y coordinate of a previously defined point similar to the function
X_COORD discussed below.
Z_COORD Returns the z coordinate of a previously defined point similar to the function
X_COORD discussed below.

The X_COORD, Y_COORD and Z_COORD, are used as followes: The name of the point, in quo-
tation marks, is passed as an argument to the function, for example
DP PNT01 1.234 0.4567 #z
#x = x_coord("PNT01")

will set the parameter #x equal to 1.234.

Data functions

The FILEREAD function reads a numerical value from an arbitrary ASCII file. The general syntax
is
fileread("Filename", Line, Column)

July 2011 FEKO User’s Manual

PREFEKO LANGUAGE AND CONCEPTS 12-10

and contains the file name, the line number to read from and the column to read. (The data
in the respective columns of any line are separated by one or more spaces or tab characters.)
For example, consider a data file containing a list of frequencies and a load impedance for each
frequency:
Frequency in MHz Re(load) in Ohm Im(load in Ohm)
100 22.54 -12.56
150 25.07 -6.54
200 27.42 0.23

The frequency and loading can be imported directly from this file
#numfreq = 3 ** Number of frequencies
!!for #i = 1 to #numfreq

** Define the frequency (conversion from MHz to Hz)

#freq = 1.0e6*fileread("datafile.dat", #i+1, 1)
FR 1 0 #freq

** Define the load

#Zr = fileread("datafile.dat", #i+1, 2)
#Zi = fileread("datafile.dat", #i+1, 3)
LZ 0 #Zr #Zi

** Computations ...

!!next ** End of frequency loop

Logical & mathematical operators

The following mathematical operators can be used in conjunction with any predefined expression
or variable:

() brackets
+ addition
- subtraction
* multiplication
/ division
ˆ powers, for example 2ˆ3=8

In addition to these functions, PREFEKO allows the use of logical operations. It supports the
function NOT() — which returns TRUE if the argument is FALSE and FALSE when the argument
is TRUE — and the delimiters >, <, >=, <=, =, <>, AND and OR. When boolean operations are
applied to variables, a value of 0 is taken as FALSE and everything else is interpreted as TRUE.
Similarly, in the result of a logical operation, FALSE is mapped to 0 and TRUE to 1.
PREFEKO also supports a logical function DEFINED(#variable) which returns TRUE if a the
variable #variable has been defined, and FALSE if not. This is useful in *.pre files used for
OPTFEKO, TIMEFEKO or ADAPTFEKO runs. These insert variables at the top of the file, but it
may be required to define the variable in the file for preview purposes. For example, if a *.pre
file will be optimised with respect to the variable #a, one may use
!!if (not(defined(#a))) then
#a = 200.0e-3
!!endif

The order of precedence is (evaluated first in the order listed) is:

July 2011 FEKO User’s Manual

PREFEKO LANGUAGE AND CONCEPTS 12-11

1. single number, expressions in brackets

2. function calls

3. + and - (when used as sign)

4. ˆ

5. * and /

6. + and -

7. >, <, >= and <=

8. = and <>

9. AND

10. OR

Some variables are predefined in PREFEKO, but may be overwritten by re-assignments. These
are shown in Table 12-7.

Table 12-7: Pre-defined variable list.

Name Value Description

#pi 3.14159265358979. . . The constant π
#eps0 8.85418781761 · 10−12 Dielectric constant "0 of free space
#mu0 4 π · 10−7 Permeability µ0 of free space
1
#c0 p
µ0 "0
The speed of light in free space
Ƶ
0
#zf0 "0
The intrinsic impedance of free space
#true 1 Used for logical true
#false 0 Used for logical false

There are three other special variables #!x, #!y and #!z which are very useful for the connection
of complex wire structures. The three variables specify the Cartesian coordinates of the end point
of the wire segment most recently defined. This enables the correct and easy connection of a
straight wire to a curved length of wire, as the next extract from an input file demonstrates:
CL .....
DP A #!x #!y #!z
#z = #!z + 0.5
DP B #!x #!y #z
BL A B

The following example demonstrates the use of variables.

July 2011 FEKO User’s Manual

PREFEKO LANGUAGE AND CONCEPTS 12-12

** A dielectric sphere in the field of an incident wave

** Define the variables

#r = 1 ** Radius of the sphere
#betrad = 1 ** Electrical size of the sphere
#epsr = 15 ** The relative dielectric constant
#maxlen = 0.7 ** The maximum edge length

** Define segmentation parameters

IP #maxlen

** The corner points

DP A 0 0 0
DP B 0 0 #r
DP C #r 0 0

** Select the medium

ME 1 0

** Generate an eighth of the sphere

KU A B C 0 0 90 90 #maxlen

** Use symmetry in all three coordinate planes

** yz-plane: ideal electrically conducting plane
** xz-plane: ideal magnetically conducting plane
** xy-plane: only geometrically symmetric
SY 1 2 3 1

** End of the geometry

EG 1 0 0 0 0

** Assigning the dielectric’s properties

DI #epsr 1.0

** Incident plane wave excitation

#freq = #betrad * #c0 / (2*#pi*#r)
FR 1 0 #freq
A0 0 1 1 1.0 0.0 -180.0

** Near fields along the z axis

FE 1 1 1 25 0 0.0 0.0 -1.98 0.0 0.0 0.04
FE 4 1 1 50 0 0.0 0.0 -0.98 0.0 0.0 0.04
FE 1 1 1 25 0 0.0 0.0 1.02 0.0 0.0 0.04

** End
EN

The use of variables makes the investigation of structures with varying geometry (e.g. variable
distance of the antenna in front of a reflector) an easy process, because only one variable needs
to be changed. It also allows FOR loops and IF conditions.

12.7 FOR/NEXT loops

Some cards in FEKO implicitly use loops (such as when an FR card with multiple frequencies is
used). This, however, does not always offer the flexibility which one may require, for example,
to change the material parameters inside the loop. Another example would be the use of a loop
to create a complex geometry.
For completely general loops, PREFEKO allows the construct
!!for #var = #start to #end step #delta

!!next

where a simple example would be

July 2011 FEKO User’s Manual

PREFEKO LANGUAGE AND CONCEPTS 12-13

** Loop for the relative permittivity

!!for #eps_r = 1 to 5 step 0.5

** Set material parameters

GF 0 #eps_r 1

** Compute fields etc.

FE

** End of loop
!!next

The syntax requirements of FOR/NEXT loops are:

• The !! characters must be located in the first two columns of the line. This is followed by
a number of optional spaces and the keyword for (it is not case sensitive, so also FOR or
For are accepted).
• The keyword for is followed by the name of the loop variable (starting with #).

• Next follows an expression for the initial value of the loop (a constant, variable or formula,
see the example below).

• This is followed by the keyword to and the terminating value of the loop variable (again a
constant, variable or formula).

• The default increment of the loop variable is 1, but it can be changed by using the keyword
step followed from an expression. Negative increments are allowed.
• The loop is terminated by a line of the form !!next (spaces are allowed between !! and
next but not before the !!). All instructions and input cards between !!for and !!next
are evaluated repeatedly inside the loop.

• Several loops can be nested as shown in the example below.

A more complicated example, illustrating some of the points above, is as follows

#end = 3+sin(4)
!!for #x1 = sqrt(5) + 2*3 to 2*#end step -#end/10
!! for #x2 = 1.23 to 2*#x1 ** this is the inner loop

#x3 = #x1 + #x2

DP ....
.... (more commands)

!! next
!!next

12.8 IF/ELSE/ENDIF constructs

This construct is used to allow different control cards under different conditions. The syntax
requirements of IF/ELSE/ENDIF constructs are:

• The !! characters must be located in the first two columns of the line. This is followed by
an arbitrary number of spaces, the expression to be evaluated and the keyword then (it is
not case sensitive, THEN or Then are also accepted).

July 2011 FEKO User’s Manual

PREFEKO LANGUAGE AND CONCEPTS 12-14

• The block is terminated by a line of the form !!endif (again spaces are allowed between
!! and endif but not before the !!).
• An optional line of the form !!else (again, the !! must be in the first two columns and
spaces are allowed before the keyword which is not case sensitive).

• All instructions and input cards between !!if and !!endif (or !!else if it is present) are
processed if the expression is TRUE. If it is present, all lines between !!else and !!endif
are processed if the expression is FALSE.

Examples, illustrating some of the points above, are as follows

!!if #a > 5 then

...

!!endif

or
#l = (#a+5 > 21) and (#a < 100)
!!if ( (3*#a+5 >= #x/2) and not(#l) ) then

...

!!else
!! if (sin(#x/10) > 0.5) then

...

!! else

...

!! endif
!!endif

12.9 EXIT command

PREFEKO also supports the command !!exit to stop execution. This can be useful for checks
like
!!if #a < 2*#b then
!! exit
!!endif

(see the various PRINT commands below to make this more informative).

12.10 PRINT commands

There are various commands to print strings (enclosed in double quotes) and floating point
numbers to the screen and *.out file respectively:

• !!print will print some text to the screen.

• !!print_warning will also print some text to the screen, but will indicate that the given
text is a warning, i.e. for console runs the string WARNING will be prepended, while for
GUI runs the colour of the output will change according to a warning.

July 2011 FEKO User’s Manual

PREFEKO LANGUAGE AND CONCEPTS 12-15

• !!print_error is similar to !!print_warning, but will handle an error message. Note

that processing continues so that multi-line error messages can also be printed. To stop
execution use the !!exit statement.

• !!print_to_out will write some text to the *.out file while the FEKO kernel is run.

These commands accept multiple arguments separated by commas. For example,

!!print "2*#b = ", 2*#b
!!if #a < 2*#b then
!! print_error "The value of #a is too small:", #a, " (exiting now)"
!! exit
!!endif

will print an error text and exit if the variable #a is smaller than two times variable #b. The line
!!print_to_out "This run was done with #b = ", #b

will print the value of #b to the *.out file at the position where it appears in the *.pre file.

12.11 Symbolic node names

Symbolic node names or named points can be constructed as either single points or as an array
of points. Arrays of points can then be referenced by only specifying the array name. This is
particularly handy when a large number of points are required.

Single node names

When defining or using node names, simple variable names of the form A#i can be used to define
the node. If a hash sign is found in a node point name, this hash sign and everything that follows
is interpreted as a variable string, evaluated and rounded to the nearest integer. Thus if we have
#k=15 and use or define a point P#k then this is equivalent to using P15 as point name. The
length of the node name string (before and after expansion) is still limited to 5 characters.
For instance, it would now be possible to define the points P1 to P20 inside a loop
!!for #k = 1 to 20
DP P#k
!!next

and use these either individually or inside another loop.

Node name arrays

Symbolic node name arrays can also be defined. These node arrays are particularly useful when
creating polygonal surfaces (PY card and PM card) with many points since only the node name
has to be specified instead of each individual point. Node name arrays allow expression of the
form A[2*#i+3] to be used to index the array.
A symbolic node name array can be defined in a loop
!!for #k = 1 to 20
DP P[2*#k+3]
!!next

and then later referenced only as P. Single node names can be referenced by indexing the array.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-1

13 Geometry cards

The following table lists all input cards that are used to create the geometry (i.e. the cards that
appear before the EG card in the *.pre file5 ). Most of these cards are processed by PREFEKO.
For example, PREFEKO processes the BP card and writes the triangle elements to the *.fek file
as input to FEKO.

Card Type of Excitation

BL creates a line
BP creates a parallelogram
BQ creates a quadrangle
BT creates a triangle
CB changes already assigned labels
CL creates a circular line using segments
CN changes the direction of the normal vector
DD
DK creates a dielectric or magnetic eighth of a sphere
DP defines a node point
DZ creates a cylindrical dielectric shell
EG defines the end of the geometry
EL creates a segment of an ellipsoid
FM set parameters related to the MLFMM
FO defines a Fock region
FP set parameters related to the FEM
HC creates a cylinder with a hyperbolic border
HE creates a coil from wire segments
HP creates a plate with a hyperbolic border
HY creates a hyperboloid section
IN reads an external include file containing mesh information
IP sets the parameter that defines the degree of meshing
KA defines the border of the PO area
KK creates a elliptical conical segment
KL sets the wedges in the PO area
KR creates a planar elliptical element
KU creates a spherical element
LA specifies the label for segments, triangles, polygonal plates, etc.
ME defines the medium
MB defines a modal port boundary condition
NU creates a NURBS surface from specified control points
PB creates a paraboloid
PE specifies the unit cell that will be used in periodic boundary condi-
tion calculations

5
In general all the geometry cards must appear before the EG card. Exceptions are the IN card (when including
*.pre files with control commands) and the DP and TP cards (when defining points for the AP card).

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-2

PH creates a flat plate with an elliptic hole

PM creates a polygonal shape that is meshed into triangles
PO applies the Physical Optics approximation
PY creates a polygonal plate for use with UTD
QT creates a dielectric or magnetic cuboid (meshed into tetrahedral el-
ements)
QU creates a dielectric or magnetic cuboid (meshed into cuboidal ele-
ments)
RM remeshing and adaptive mesh refinement
SF enters a scaling factor, with which all dimensions are multiplied
SY utilises symmetry in the construction of the geometry
TG transformation (i.e. translation and rotation) of the geometric struc-
tures
TO creates a toroid
TP transforms a point
UT parameters for the uniform theory of diffraction (UTD)
UZ creates a cylinder for use in the UTD region
VS specifies known visibility information (required when using physical
optics with multiple reflections)
WA define all active windscreen antenna elements
WG creates a parallelogram consisting of a wire grid
WR defines the dielectric windscreen reference plane
ZY creates a cylindrical element

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-3

13.1 ** card

The ** card is not a command, but defines a comment line. Everything that is found in this line
is ignored by PREFEKO.
It is possible to add a comment to the end of an existing line or card. For example,
** Definition of parameters
#lambda = 1.0 ** Wavelength
#radius = #lambda/2 ** Cylinder radius
#height = 2*#lambda ** Cylinder height

More information on the usage of comment lines and comments as card names may be found in
the comment card description under the control cards (see section 14.1).

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-4

13.2 BL card

This card is used to connect two points to form a line, which is then subdivided into wire seg-
ments.

Parameters:

Start point of wire: The start point of the wire, previously defined with the DP card.

End point of wire: The end point of the wire, previously defined with the DP card.

Set wire radius: If checked, the radius set at the previous IP card is overridden for the current
wire. This setting does not affect segments created after this card. Both radius
values are in m and are affected by the SF card scaling factor. If only the start
radius is specified, the wire will have a constant radius.

Radius at start . . .: The radius of the wire at the start point.

Radius at end . . .: The radius of the wire at the end point.

The points have to be defined by a DP card, prior to using this card. The wire radius is set by an
IP card preceding the BL card, but can be set locally. Note that by using different radius values
for the start and end points of the wire, a tapered wire can be created.

Figure 13-1: Sketch illustrating the use of the BL card

Examples of BL card usage:

The BL card can be used to create segmented wires as shown in Figures 13-2 and 13-3. In the
first example the radius is specified with an IP card and in the second an exaggerated taper is
specified at the BL card.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-5

Figure 13-2: Example of a BL card from demo_BL1.pre

Figure 13-3: Example of a BL card with a tapered radius from demo_BL2.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-6

13.3 BP card

A mesh of surface triangles in the shape of a flat parallelogram can be created with this card. In
general, this card is replaced by the PM card. This card should only be used when the user wants
to force the very regular meshing that this card produces.

Parameters:

S1, S2, S3, S4: The points S1 to S4 are the four corner points of the parallelogram. These
points should have been defined previously with DP cards.

Specify non-uniform . . .: Normally a parallelogram is segmented according to the edge length specified
with the IP card. In some cases, e.g. when creating small microstrip lines, it
may be desirable to use a finer segmentation in one direction. Check this item
if a finer segmentation is required in one direction. The mesh sizes are in m
and are scaled by the SF card.

Mesh size along sides a and c: Edges S1–S2 and S3–S4.

Mesh size along sides b and d: Edges S2–S3 and S4–S1.

The points are connected in the order that they appear in the BP card. Thus the user has to
ensure that the points describe a parallelogram. If this is not the case, then PREFEKO will abort
with the appropriate error message.
The direction of the normal vector (n̂) of the subdivided triangles is determined by the right hand
rule, through all the corners. This direction is only important when used with the Physical Optics
— PO card (see section 13.37) — or with dielectrics — ME card (see section 13.31) or with the
CFIE — CF card (see section 14.26).
Example of BP card usage:
Figures 13-4 and 13-5 show a plate with uniform meshing and a strip with non-uniform meshing
respectively, both created using the BP card.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-7

Figure 13-4: Example for the BP card from demo_BP1.pre

Figure 13-5: Example of a BP card with non-uniform meshing from demo_BP2.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-8

13.4 BQ card

A mesh of surface triangles in the shape of a flat quadrangle can be created with this card. Models
constructed using the BQ card can generally be simplified by using the PM card.

Parameters:

S1, S2, S3, S4: The points S1 to S4 are the four corner points of the quadrangle. These points
should have been defined previously with the DP card.

Specify non-uniform . . .: Normally a quadrangle is segmented according to the edge length specified
with the IP card. In some cases, e.g. when creating small microstrip lines, it
may be desirable to use a finer segmentation in one direction. Check this item
if finer segmentation is required along any edge. The mesh sizes are in m and
are scaled by the SF card.

Mesh size along side a: Edge S1–S2.

Mesh size along side b: Edge S2–S3.

Mesh size along side c: Edge S3–S4.

Mesh size along side d: Edge S4–S1.

The points have to be predefined using DP cards prior to the BQ card and are connected in the
order that they appear in the BQ card.
In principal the BQ card can create all types of quadrangles, including parallelograms. The
difference is that the BP card creates a regular subdivision.
The direction of the normal vector (n̂) of the subdivided triangles is determined by the right hand
rule through all the corners. This direction is only important when used with the Physical Optics
(PO card) or with dielectrics (ME card) or for the CFIE (CF card).
Examples of BQ card usage:
The BQ card can be used to create the mesh shown in Figure 13-6, or also for inhomogeneous
meshing as shown in Figure 13-7.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-9

Figure 13-6: Example of a BQ card from demo_BQ1.pre

Figure 13-7: Example of a BQ card with an inhomogeneous segmentation from demo_BQ3.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-10

13.5 BT card

A mesh of surface triangles in the shape of a flat triangle can be created with this card. In general,
this card is replaced by the PM card.

Parameters:

S1, S2, S3: The points S1 to S3 are the three corner points of the triangle. These points
should have been defined previously with the DP card.

Specify non-uniform . . .: Normally a triangle is segmented according to the edge length specified with
the IP card. In some cases it may be desirable to use a finer segmentation in
one direction. Check this item if finer segmentation is required along any edge.
The mesh sizes are in m and are scaled by the SF card.

Mesh size along side b: Edge S2–S3.

Mesh size along side c: Edge S3–S1.

Mesh size along side a: Edge S1–S2.

The direction of the normal vector (n̂) of the subdivided triangles is determined by the right hand
rule through all the corners. This direction is only important when used with the Physical Optics
(PO card) or with dielectrics (ME card) or for the CFIE (CF card).
Examples of BT card usage:
The meshes shown in Figures 13-8 and 13-9 were created with BT cards using uniform meshing
and non-uniform meshing respectively.

Figure 13-8: Example of a BT card from demo_BT1.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-11

Figure 13-9: Example of a BT card with non-uniform meshing from demo_BT2.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-12

13.6 CB card

This card is used to change or reassign the labels assigned to points, segments, triangles, cuboids,
polygonal plates, tetrahedral elements, etc.

Parameters:

Specify old/new label here: This selection allows to specify an old label and a new label in the corre-
sponding input fields.

Read list of old/new labels from file: This selection allows to read a list of old/new label pairs from an
external data file. See further details below.

Old label: All the structures with this label are relabelled.

New label: The new label for all the structures with the old label.

File name: The name of the file used when reading the label list from an external data file.

Renaming labels is especially useful when more labels are created by using symmetry (SY card)
or transformations (TG card) or an imported geometry from CADFEKO, and for example, edges
or wedges in the PO area are considered or any other properties shall be set by label (e.g. Skin
effect). Structures created after the CB card are not affected.
In order to make the renaming of a whole set of different labels simpler, the Old label field in the
CB card is also supporting wild cards ‘*’ (an arbitrary sequence of characters) and ‘?’ (a single
arbitrary character). So for instance to rename all these labels
Cube.Face1
Cube.Face2
Cube.Face3
Cube.Face4
Cube.Face5
Cube.Face6

to a new label CubeSurface one could use six CB cards, but with the wild cards this is much
simpler to use just one CB card and specify the old label as
Cube.Face?

or also as
Cube.*

(depending on what other labels are also in the model).

Note that such wild cards are only supported in the Old label field of the CB card. The New label
must be unique.
Another possibility to do a bulk renaming of labels is to read a label mapping table from an
external file, which follows the syntax of the ANSA package. In ANSA version 11, this file consists
of an arbitrary number of lines

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-13

old_label | new_label

(i.e. the old and new label entries separated by the | character). Alternatively, the ANSA version
12 format is also supported, where there is no | character to separate the old and new label,
but just white space (i.e. a space or tab character). Comments line are allowed in these files and
these are indicated using ‘**’ as the first characters of the line.
Some external meshing programs can for instance export a NASTRAN file along with such a
mapping table, and then by using the two commands
IN 3 3 "geometry.nas"
CB "geometry.txt"

one can get the model into FEKO with the proper names of the parts (i.e. the file geometry.txt
would then do a proper mapping of the NASTRAN property to the part name in the original CAD
program).

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-14

13.7 CL card

This card is used to create an arc consisting of wire segments.

Parameters:
S1: The centre point of the circle.

S2: A point perpendicular to the plane in which the circle lies and above its centre.

S3: The start point of the arc.

Subtended angle ϕ: The direction of the subtended angle, φ, is in the positive sense around the
axis S1–S3. A negative value will create the arc in the opposite direction.

Maximum length of . . .: The maximum length of the segments that make up the arc. If this field is left
empty, the value specified with the IP card is used. This length is in m and is
scaled by the SF card.

Set wire radius: If checked, the radius set at the previous IP card is overridden for the current
arc. This setting does not affect segments created after this card. Both radius
values are in m and are affected by the SF card scaling factor. If only the start
radius is specified, the arc will have a constant wire radius.

Radius at start . . .: The radius of the wire at the start point.

Radius at end . . .: The radius of the wire at the end point.

Scale second half axis: If this parameter is empty or is set to 1, a circular arc is created. If set to
b
a
, an elliptical arc is created. Here ab gives the ratio of the two half axes,
where a is the distance S1–S3. It is not recommended to generate elliptical
arcs with extremely small or extremely large axial ratios with a CAD system as
the distortion formulation used in PREFEKO may fail in these cases.

Quite often modelling the geometry of an arc requires shorter segments than those used for
straight wires. Thus the maximum segment length specified with the IP card can be overridden
along the arc by specifying a value in the field Maximum length of segments.
The radius of the arc is given by the distance between the points S1 and S3.
Examples of CL card usage:
The meshes shown in Figures 13-10 and 13-11 are created with the CL card. The first figure
shows the result of a wire arc with a uniform radius, and the second figure shows the result with
an exaggerated taper specified.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-15

Figure 13-10: Example of a CL card from demo_CL1.pre

Figure 13-11: Example of a CL card with tapered wire radius from demo_CL2.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-16

13.8 CN card

This card is used to reverse the normal direction of previously created triangles or polygonal
plates, for example after importing CAD data.

Parameters:

Reverse normal of . . .: In this group, the user selects to reverse the normals of either Triangles or
Polygonal plates.

Selection by: Here the user specifies whether the triangles / polygonal plates are identified
by their Label or absolute element Number.

Selection: The label / element number of the triangles / polygonal plates that must have
their normals reversed.

The normal direction is important in some cases, such as when defining dielectric surfaces. For
triangles, the normal vector is reversed by interchanging corners 1 and 3. For polygonal plates
the first point remains as is, but the corner points are listed in the opposite direction.
The CN card changes the normal of the affected triangles, but it does not change the settings
of the ME card (which medium is on which side of the triangle as determined by the normal
vector). For example, if the ME card is used to specify that the normal vectors of the triangles
point from medium 5 to medium 2 then the application of the CN card will effectively change
which medium lies on which physical side of the triangle.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-17

13.9 DD card

With this card a MoM model consisting of a dominant fixed static part and a smaller dynamic
part, may be solved more efficient. The factorisation of the static interaction matrix is written to
a *.ngf file and may then be reused to solve the MoM problem when a smaller dynamic part is
modified.

Parameters:

No data files (normal execution): When this option is selected, no data files are read or saved to a
*.ngf file.

Save *.ngf file: The results are saved to a *.ngf file.

Read the *.ngf file: The results are read from a *.ngf file.

Read *.ngf file if it exists, else create it: The *.ngf file is read if it exists, else a *.ngf file is created con-
taining the results.

Number of triangle labels: The number of element labels to be included in the fixed static part.

Labels of the elements: The labels of the elements which are to be defined as the fixed static part in
the model.

Note that only a single DD card is allowed in a model.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-18

13.10 DK card

This card is used to create an eighth of a sphere, meshed into cuboidal elements, for solutions
using the volume equivalence principle in the MoM. The meshing parameters as set at the IP card
are used, and the medium as set at the ME card is assigned to all created cuboidal elements.

Parameters:

S1: The centre of the sphere.

S2, S3, S4: Specify the three directions S1–S2, S1–S3 and S1–S4, that form the border of
the eighth of the sphere. They must be perpendicular to each other and all
three must have the same length (the sphere’s radius).

Maximum cuboid edge length: The maximum side length of cuboids along the curved edge (in m)
can be specified. This value is scaled by the SF card. If left empty, the value
specified with the IP card is used.

Choose the medium: Select here whether the sphere is dielectric or magnetic or both (this is always
with respect to the environment, e.g. if the relative permittivity " r of the cuboid
material differs from the environment, then this is a dielectric sphere).

Old format (with medium parameters): For a detailed description of this parameter please see the QU
card (see section 13.40).

Dielectric bodies treated with the volume equivalence principle (using cuboids) cannot be used
simultaneously with dielectric bodies treated with the surface equivalence principle or the FEM
or with special Green’s functions.
Example of DK card usage:
The DK card can be used to create a mesh of the eighth of a sphere as shown in Figure 13-12.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-19

Figure 13-12: Example for the DK card from demo_DK1.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-20

13.11 DP card

With this card points in space are defined. These points are used to define the extent and orien-
tation of other geometric entities and to locate excitations.

To avoid ambiguity each point is assigned a name (a 5 character string if column based format is
used). In the other cards (e.g. BL card) the points are referred to by their names.
Parameters:

Point name: Name of the point.

X, Y, Z coordinate: Cartesian coordinates of the point in m (is scaled by the SF card).

Nurb control point weight: The weight of the control point when this point is used with the NU card
(NURBS surfaces). If the field is empty, it defaults to 1.

In addition to its coordinates, each point is also assigned the current label (see LA card), so that
a group of points can be selected by label, for example when moving points with the TP card.
Point names may use the characters a-z, A-Z, 0-9 and the special character _ and no distinction
is made between upper and lower case characters. Thus P1a and p1A refers to the same point.
In addition, when defining or using node names, simple variable names of the form A#i are
allowed. The algorithm is that if a hash sign is found in a node point name, this hash sign and
everything that follows is interpreted as a variable string, evaluated and rounded to the nearest
integer. Thus if we have #k=15 and use or define a point P#k then this is equivalent to using P15
as point name.
For instance, it would now be possible to define the points P1 to P20 inside a loop.
!!for #k = 1 to 20
DP P#k .....
!!next

Unlike most other geometry cards, the DP card (as well as the TP card) may also be used in the
control section (after the EG card) of the *.pre file. This allows defining the points required by
the AP card in this part of the file.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-21

13.12 DZ card

This card is used to create a cylindrical shell, meshed into cuboidal elements, for solutions using
the volume equivalence principle in the MoM. The meshing parameters as set at the IP card are
used, and the medium as set at the ME card is assigned to all created cuboidal elements.

Parameters:

S1, S2: The start and end points, respectively, of the cylinder axis.

S3, S4: Points on the inside and outside, respectively, of the shell.

The angle ϕ: The angle of the cylindrical segment in degrees.

Maximum cuboid edge length on arc: Maximum edge length of the cuboids along the arc in m (is
scaled by the SF card). If this parameter is left empty, the value specified
with the IP card is used.

Choose the medium: Select here whether the cylindrical shell is dielectric or magnetic or both (this
is always with respect to the environment, e.g. if the relative permittivity " r
of the cuboid material differs from the environment, then this is a dielectric
cylinder).

Old format (with medium parameters): For a detailed description of this parameter please see the QU
card (see section 13.40).

Dielectric bodies treated with the volume equivalence principle (using cuboids) cannot be used
simultaneously with dielectric bodies treated with the surface equivalence principle or the FEM
or with special Green’s functions.
Example of DZ card usage:
Figure 13-13 is an example of a cylindrical shell created with the DZ card.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-22

C D

Figure 13-13: Example for the DZ card from demo_DZ1.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-23

13.13 EG card

This card indicates the end of the geometrical input. It is essential that this card is used.

Parameters:

Write geometrical output to: The geometry data of the segments and surface elements can be written
to the FEKO output file, a NASTRAN format file, an STL file, or any combination
thereof. The name of the NASTRAN or STL file will be the same as the FEKO
model, but with a *.nas or *.stl extension. Writing the geometry data to
the output file may lead to huge output files.

Send to standard output: If the field Nothing is selected, no messages are sent to the standard output
device (usually the screen). If the item Warnings, errors, progress messages is
selected, warnings, errors and messages that indicate the program’s progress
are sent to the standard output device.

Switch normal geometry checking off: If this item is checked, the verification of the geometry elements
will be switched off (see the discussion below this table).

Solution accuracy: This parameter can be set to force FEKO to use single precision for the storage
of some of the memory critical arrays. Single precision storage is the default
behaviour, and as compared to double precision the memory requirement is
then half. Using double precision is recommended when the FEKO kernel gives
a warning to switch to double precision (this might happen for instance at low
frequencies where an increased accuracy is required).

Activate low frequency stabilisation for MoM: If this item is checked, low frequency stabilisation for
MoM is activated.

The following should be noted regarding the export of the FEKO geometry to NASTRAN or STL:

• The STL export just dumps the data of all triangular patches of the FEKO model to an ASCII
formatted STL file. Any other geometry (wires, tetrahedra etc.) is not exported since the
STL format does not make provision for this. It should also be noted that the FEKO mesh
does not contain any information about the geometry that the triangle mesh elements were
created from. Thus there is no special grouping of elements based on regions or solid parts
in the STL file — this implies that the exported STL file does not represent a valid STL file
in the strict sense. However, the exported information is still useful in most cases.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-24

• For the NASTRAN export, the wide column format is used to ensure that all significant digits
are exported. Unlike the STL export, in NASTRAN all the various mesh elements used in
FEKO are present. However, some information is lost, for instance for wire elements the
thickness (wire radius) cannot be exported simply because the NASTRAN file format does
not make provision for this. Also the NASTRAN property is used to represent the FEKO
label. But since NASTRAN properties are just integer values and the FEKO label can be an
arbitrary string, a mapping is done so that from each FEKO label (see section 12.4) just the
associated number is used and exported as the NASTRAN property.

The Maximum identical distance is used to set the tolerance in the mesh. The mesh information
is created by the program PREFEKO, and stored in a *.fek file, in which all the triangles and
segments are described by their corner points. Due to rounding errors it is possible that, for
example, end points of connecting segments do not coincide. When searching for nodes, an
ohmic connection is made when the difference is smaller than the Maximum identical distance.
FEKO automatically checks for typical user errors that have been observed in the past. Examples
of errors are connecting a wire segment to the middle of another wire, where the connection
points do not coincide, or connecting surfaces that have different segmentation along the com-
mon edge (see section 2.2.3). Such errors are detected if the parameter Switch normal geometry
checking off is unchecked. The error detection routine should always be used. However, if the
same geometry is to be used a number of times, the error detection can be disabled by checking
this item.
If the surrounding medium is not vacuum, one can set the material parameters with the EG card
as shown above. Alternatively the parameters of the surrounding medium can be set with the
GF card which offers greater flexibility. For example, the GF card can be used to set the material
parameters (as an arbitrary function of frequency) inside a frequency loop which is not possible
with the EG card.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-25

13.14 EL card

A mesh of surface triangles in the shape of an ellipsoidal section can be created with this card.

Parameters:

S1: The centre point of the ellipsoid.

S2: A point, in the direction ϑ=0◦ in elliptical coordinates. The distance of the two
points S1 and S2 determines the half-axis of the ellipsoid in this direction.

S3: A point in the direction ϑ = 90◦ , ϕ = 0◦ in elliptical coordinates. The distance

of the two points S1 and S3 determines half of the axis of the ellipsoid in this
direction.

S4: A point in the direction of the third coordinate, i.e. the axes S4–S1, S3–S1
and S2–S1 must be perpendicular. The distance of the two points S1 and S4
determines half of the axis of the ellipsoid in this direction.

Begin angle ϑa : Start angle of the ellipsoid in degrees.

Begin angle ϕa : Start angle of the ellipsoid in degrees.

End angle ϑe : End angle of the ellipsoid in degrees.

End angle ϕe : End angle of the ellipsoid in degrees.

Maximum triangle edge length: Maximum length of the triangles along the curved edge in m (is scaled
by the SF card). If this parameter is left empty, the value specified with the IP
card is used.

Note that the angles ϑ and ϕ are defined in an elliptical, rather than in a spherical coordinate
system. For a Cartesian coordinate system with origin S1, x axis in direction of S3, y axis in the
direction of S4 and z axis in the direction of S2, a point on the surface of the ellipsoid is given as

a sin ϑ cos ϕ
   
x
~r =  y  =  b sin ϑ sin ϕ  (13-1)
   
z c cos ϑ

where the lengths a, b and c are the lengths of the ellipsoid’s three half-axes. (For example the
length a is the distance between the points S3 and S1).

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-26

The normal vector of the generated triangles always points outwards. The algorithm used for
the segmentation can fail if the ratio of the half-axis is too extreme, for example if the longest
half-axis is a factor 100 longer than the shortest. It is strongly advised to check the geometry
with POSTFEKO.
Example of EL card usage:
The mesh shown in Figure 13-14 is generated by using the EL card.

Figure 13-14: Example for the EL card from demo_EL.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-27

13.15 FM card

This card is used to instruct the FEKO solver to calculate the solution using the Multilevel Fast
Multipole Method (MLFMM) instead of the MoM on all structures in the simulation. A second
method, the Adaptive Cross Approximation method is also available and is applicable to low
frequency problems or when using a special Green’s function.

Parameters:

Use MLFMM: The Multilevel Fast Multipole Method method is used to calculate the solution.

Use adaptive cross-approximation (ACA): The ACA method is used to calculate the solution. This
method approximates the impedance matrix by constructing a sparse H-matrix
(only a few selected elements are computed).

Box size at fines level: The MLFMM is based on a hierarchical tree-based grouping algorithm, and
depending on the frequency and the model dimensions FEKO automatically
determines the number of levels in this tree and the size of the boxes at the
finest level. It is also recommended that this default box size of 0.23λ is kept.
When there is no convergence in the MLFMM, then advanced users might try
to slightly increase or decrease this box size by setting it manually (the input
is in terms of the wavelength).

Near field calculation method: The MLFMM method can use a fast near field calculation method (this
is the default), but under certain circumstances it is preferred to use the tradi-
tional integration method. The method to use can be selected here.

Far field calculation method: The MLFMM method can use a fast far field calculation method (this is
the default), but under certain circumstances it is preferred to use the tradi-
tional integration method. The method to use can be selected here.

Multilevel Fast Multipole Method (MLFMM)

Currently, the MLFMM may not be used in conjunction with the Multi-layer Greens function.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-28

Advantages of Adaptive Cross Approximation (ACA)

• ACA is done on the matrix level.

• No frequency breakdown like MLFMM.

• Not restricted to free space Green’s function like MLFMM.

• Iterative solution, each matrix-vector multiplication is very fast.

Note that the ACA can currently only be used sequentially.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-29

13.16 FO card

With this card an area is defined in which the surface current density is an approximation ac-
cording to the Fock theory.

Parameters:
Perfectly conducting cylinder: Select this option if the Fock area is/resembles a cylinder.

Perfectly conducting sphere: Select this option if the Fock area is/resembles a sphere.

Triangle labels: The label of the metallic triangles that form the surface of the Fock area (e.g.
the surface of the cylinder).

Axis start / Axis end point: These dialogs are only visible for Fock cylinders and should correspond to
the start and end points of the cylinder axis.

Sphere centre point: This dialog is only visible for Fock spheres and should correspond to the centre
of the sphere.

Formulation of the Fock currents: The type of process for the Fock currents, either the Method by
Daniel Bouche or the Method by Louis N. Medgyesi–Mitschang.

Decouple with moment method: Check this item to force FEKO to neglect the coupling between the
MoM and Fock regions, so that there is no feedback by which the Fock currents
may influence the current distribution in the MoM region. This option, which
is particularly applicable when the MoM and Fock regions are not in close
proximity, should result in a considerable reduction in computational effort
and storage space.

The radius of the cylinder or sphere does not have to be defined. It is determined by the distance
to the metallic triangles, with the label specified in Triangle labels.
The cylinder Fock currents can also be applied to cones (KK card, approximated by a staircase
construction of cylinders) and sections of a torus that resembles a cylinder (TO card). Although
the FO card is strictly only applicable to spherical and cylindrical surfaces, it is often a good
approximation on conical and toroidal surfaces.
It must be noted that the search for creeping rays on the Fock surface does not take into account
multiple Fock regions, i.e. one creeping ray can only exist in one Fock region. Therefore, when
for instance modelling a sphere and using symmetry, it is highly advisable to create part of the
sphere, then use the SY card to mirror it, and only then use one FO card which applies to the
whole sphere. When using SY or TG cards, they do operate on already existing Fock regions
defined above these cards and thus mirror or move them, but with the SY card the problem is
that after applying symmetry there exist multiple Fock regions. The creeping rays along geodesic
lines stop at the boundary of each Fock region.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-30

13.17 FP card

With this card options related to the FEM (Finite Element Method) can be set.

Parameters:

Decouple from MoM (use FEM absorbing boundary condition): Normally within the MoM / FEM hybrid
method these two techniques are fully coupled, and also the surface of the FEM
region is treated by the MoM. For certain applications when there is a larger
separation between the MoM and the FEM regions (e.g. human body with a
GSM base station antenna) this decoupling checkbox can be checked. Then
similar to switching off the coupling for the MoM/PO or MoM/UTD hybrid
methods, first the MoM region is solved for by neglecting the FEM domain,
and this MoM solution is then used as impressed excitation for the FEM. Also
the MoM is in this de-coupled case no longer used on the FEM surface, but
rather an absorbing boundary condition is applied there. The advantage of
this decoupling is a saving of memory and computation time.

Element order: When second order is selected, FEKO uses hierarchal tetrahedral elements with
LT/QN (linear tangential/quadratic normal) vector basis functions for the elec-
tric field inside the FEM region. On the boundary surface of the FEM region
CT/LN (constant tangential/linear normal) vector basis functions are used for
the equivalent electric and magnetic surface currents. When first order is se-
lected, then CT/LN basis functions are used everywhere, on the boundary and
inside the FEM region. When switching to first order, normally a finer mesh
is required to get the same solution accuracy compared to second order ba-
sis functions. Thus second order is the default. However, when having a fine
mesh (like modelling details of a biological structure), then one might consider
switching to first order only to reduce the number of unknowns.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-31

13.18 HC card

This card creates a cylinder with a hyperbolic border.

Parameters:

S1: The origin (where the asymptotes intersect) of the hyperbolic border.

S2: The pole of the hyperbolic border.

S3: The point where the hyperbolic arc and the straight edge intersect.

S4: The pole of the hyperbolic border at the opposite edge of the cylinder.

Normal vector directed: The normal vectors can be directed inward or outward.

Max. triangle edge length (at hyperbolic border): The maximum edge length on the hyperbolic border.

The hyperbolic border may require shorter mesh edges than those used for straight edges. Thus
the maximum segment length specified in the IP card may be overridden along the arc by setting
Max. triangle edge length (at hyperbolic border).
Example of HC card usage:
The cylinder with a hyperbolic border shown in Figure 13-15 is created using the HC card.

Figure 13-15: HC card examples — These examples are found in demo_HC1.pre.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-32

13.19 HE card

With this card a helical coil, consisting of wire segments, can be created.

Parameters:
S1: The start point of the coil’s axis.

S2: The end point of the coil’s axis.

S3: The start point of the windings.

Connect helix . . .: Create connections from the two ends of the coil to the axis (at points S1 and
S2). See also left side of Figure 13-16. If the connections are not generated,
point S3 is a connection point. See also the right side of Figure 13-16.

Coil orientation: Indicate whether a right- or left handed coil should be created.

Number of turns: In this field the number of turns for the helix is entered. It need not be an
integer number.

Maximum segment length: Maximum length of the segments, that are used for the windings in m (is
scaled by the SF card). If this parameter is left empty, the value specified with
the IP card is used.

Set (tapered) radius . . .: If this item is checked, a tapered wire radius can be set. Normally the wire
radius is set with the IP card. Checking this item overrides this radius for the
current helix without affecting the default for later segments (The radius is in
m and is affected by the SF card scaling factor.) The segments connecting to
the axis are not tapered and have radii corresponding to the start point and
end point respectively.

Radius at start . . .: The radius of the wire at the start of the coil.

Radius at end . . .: The radius of the wire at the end point of the coil.

Scale second half axis: If this parameter is empty or is set to 1, a helix with a circular cross section
is created. If set to ab , a helix with an elliptical cross section is created. Here
b
a
gives the ratio of the two half axes, where a is the distance S1–S3. It is
not recommended to generate elliptical helices with extremely small or ex-
tremely large axial ratios with a CAD system as the distortion formulation used
in PREFEKO may fail in these cases.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-33
.
Quite often modelling the geometry of the coil requires shorter segments than those used for
straight wires. Thus the maximum segment length specified by the IP card can be overridden
along the arc by setting Maximum segment length.
The windings are generated between the two points S1 and S2, that lie on the axis. The radius
of the coil is defined by the distance between the points S1 and S3. For elliptical cross sections
this is the length of one half axis and the other one is Scale second half axis times this length.
Example of HE card usage:
The two coils shown in Figure 13-16 are created using the HE card.

B1 B2

A1 A2
C1 C2

Figure 13-16: HE card examples — the coil on the right is coiled in the left handed direction. These
examples are found in demo_HE1.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-34

13.20 HP card

This card creates a plate with a hyperbolic border.

Parameters:

S1: The origin (point at which the asymptotes intersect) of the hyperbolic border.

S2: The pole of the hyperbolic border.

S3: The point where the hyperbolic arc and straight edge intersect.

Max. triangle edge length (at hyperbolic border): The maximum edge length on the hyperbolic border.

The hyperbolic border may require shorter mesh edges than those used for straight edges. The
maximum edge length specified in the IP card can be overridden on the hyperbolic arc by setting
Max. triangle edge length (at hyperbolic border).
Example of HP card usage:
The plate with a hyperbolic border shown in Figure 13-17 is created using the HP card.

Figure 13-17: HP card example. This example is found in demo_HP1.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-35

13.21 HY card

This card creates a hyperboloid section.

Parameters:

S1: The origin (point at which the asymptotes intersect) of the hyperboloid section.

S2: The pole of the hyperbolic border.

S3: A point that defines the outer border of the hyperboloid section.

Normal vector directed: The normal vectors can be directed inward or outward.

Subtended angle ϕ (in degrees): The subtended angle measured from S3.

Max. triangle edge length (at circular border): The maximum edge length on the circular border of the
hyperboloid section.

The hyperbolic border may require shorter mesh edges than those used for straight edges. Thus
the maximum segment length specified in the IP card may be overridden along the circular arc
by setting Max. triangle edge length (at circular border).
Example of HY card usage:
The hyperboloid section shown in Figure 13-18 is created using the HY card.

Figure 13-18: HY card example. This example is found in demo_HY1.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-36

13.22 IN card

This card is used to include external files. These files may be other *.pre files (which are
included as if they were part of the master file) or mesh data files containing wire segments,
triangles, quadrangles, tetrahedral volume elements and/or polygonal plates (in FEMAP neutral,
an ASCII format, NASTRAN, meshed AutoCAD *.dxf, NEC model, Concept model, STL, PATRAN
neutral, ANSYS *.cdb, ABAQUS *.inp or GiD *.msh mesh files).
Some fields are common to more than one option:

File name: The name of the file. This parameter is required for all import options. The file
name may contain directory names as well, for example,
../myfiles/include.inc
and will have different extensions for the various import options. Both \ and /
are allowed on Windows and UNIX systems.

Include segments: Check this item to include all wire segments that match the label selection.
(See below.)
Include triangles: Check this item to include surface triangles.
Include quadrangles: Check this item to include quadrangles. The quadrangles are subdivided into
triangles (along the shortest diagonal) during importation.
Include tetrahedral elements: Check this item to include tetrahedral elements (for FEM).

Include polygons: Check this item to include polygonal plates (for UTD).
Include node points: Check this item to include node points.
Include only node points for imported triangles and/or wires: If this item is checked, only the node points
which are used by the imported elements, are imported. This is useful if one
imports, for example, a few segments from a file containing a large number
of triangles. With this option one may then only import the points associated
with the segments — even if they have the same label as the ones associated
with triangles only.

Label selection: Most options allow label selective importing. (How the various layers / prop-
erties / names are converted to FEKO labels is discussed separately for each
import option.) One may Include all items, Include items with only a single
label or Include items with range of labels. If the first option is selected all ele-
ments are imported, irrespective of label. If the single label option is selected,
the Include structures with. . . field becomes active. Specify the label (as it will
be after conversion) in this field. If the range option is selected, the Up to. . .
field also becomes active. All elements with the label larger or equal to the
first and smaller or equal to the second field, are included. If the import option
does not support label selection, all elements are imported.

Scale factor: An optional constant scaling factor can be applied to the imported geometry.
This is necessary, for example, if separate CAD files with different units must be
imported, or if the *.pre is, for example, created using mm while the CAD file
is constructed using inches as unit. It should be noted that the scaling factor
specified here is applied in addition to any scaling factor that may be set with
the SF or TG cards.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-37

PREFEKO file

This option is used to include cards and commands (such as variable definitions etc.) from a
separate file. One may, for example, create a single file with an antenna which is then imported
into different environment models.

For this option, the file name is the only parameter. This file is included as if it was part of
the master *.pre file. These include files usually have the extension *.inc, but can have any
extension. The cards and instructions in the included files are processed as if they were part
of the main file. Therefore, points and labels defined in the included file remain valid in the
remainder of the main file. Note that it is also possible to use such an IN card in the control
section of the *.pre file (for instance to import some feed model).
When reading a PREFEKO file it is not possible to add a scaling factor to the IN card. In this case
the TG card must be used if the global (SF card) scaling option is not sufficient.
It is possible to use multiple nested levels of include files (i.e. one include file can include another
one and so on). It is also possible to specify together with the file name an absolute or relative
path like in
IN 0 "..\subdir\file.inc"

In such a case — if multiple levels of include files are used — it is first tried to find the include
file using the path relative to the location of the file where the IN card is used. If the include file
is not found there, then PREFEKO also tries to find the include file using the path relative to the
location of the main *.pre file which is processed.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-38

FEMAP neutral file

This option is used to import models generated by the commercial CAD meshing program FEMAP.
The models must be exported from FEMAP in the FEMAP neutral file format (*.neu).

This card supports all the parameters described in the general section of the IN card above.
The label selection uses the FEMAP layer numbers which are converted to FEKO labels. Wires
must be meshed into elements which are imported as segments, surfaces into triangles or quad-
rangles which are imported as FEKO triangles, and boundary surfaces are imported as polygonal
plates. The boundary surface must be bordered with line curves rather than edge curves.
The user can also elect to import points from the *.neu file. All points defined as such in
FEMAP are then available in PREFEKO as points (as if they were defined by DP cards) of the
form Pxxx where xxx is the point ID in FEMAP. This may be used, for example, when attaching
additional structures to a geometry partly created in FEMAP. In addition, the coordinate values of
the point are available as variables in PREFEKO. For example, the variables #p1234x, #p1234y
and #p1234z give the coordinates of the FEMAP point with ID 1234. Note that points are not
included by default.
It should be remembered that it is not possible to specify a wire radius in FEMAP. Thus the
wire radius must be specified by an IP card preceding the IN card. Similarly, when specifying
the surface of a dielectric, the IN card must be preceded with the correct ME card (completely
analogous to the case without FEMAP).
POSTFEKO should be used to verify the included geometry.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-39

Import special ASCII data file

This option is used to import meshes stored in the geometry data file as specified below.

The ASCII format supports segments, triangles, tetrahedra and polygonal plates, but all other
(non-selection) parameters discussed in the general section of the IN card above apply. In this
case the label is specified directly in the file and no conversion is required.
Dielectric triangles or metallic triangles which form the surface of a dielectric, are created by
preceding the IN card with the appropriate ME card. (In exactly the same way as is the case
without the IN card.)
The data of the segments, triangles and polygonal plates are given in an ASCII file, formatted as
shown below. There is no need to adhere to specific columns, the data fields merely have to be
separated by one or more spaces.
nk nd ns np nt
x(1) y(1) z(1) (String_name)
x(2) y(2) z(2) (String_name)
...
x(nk) y(nk) z(nk) (String_name)

d1(1) d2(1) d3(1) 0 (Label)

d1(2) d2(2) d3(2) 0 (Label)
...
d1(nd) d2(nd) d3(nd) 0 (Label)

s1(1) s2(1) 0 0 (Label)

s1(2) s2(2) 0 0 (Label)
...
s1(ns) s2(ns) 0 0 (Label)

nnp(1) p1(1) p2(1) p3(1) ... (Label)

nnp(2) p1(2) p2(2) p3(2) ... (Label)
...
nnp(np) p1(np) p2(np) p3(np) ... (Label)

t1(1) t2(1) t3(1) t4(1) (Label)

t1(2) t2(2) t3(2) t4(2) (Label)
...
t1(nt) t2(nt) t3(nt) t4(nt) (Label)

The meaning of the above is:

nk Number of nodes

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-40

nd Number of triangles

ns Number of segments

np Number of polygonal plates

nt Number of tetrahedral volume elements (defaults to 0 if not specified)

x(i) x coordinates of node i in m (is scaled by the SF card)

y(i) y coordinates of node i in m (is scaled by the SF card)

z(i) z coordinates of node i in m (is scaled by the SF card)

d1(j) Number (index) of the first vertex of triangle j

d2(j) Number (index) of the second vertex of triangle j

d3(j) Number (index) of the third vertex of triangle j

s1(k) Number (index) of the starting point of segment k

s2(k) Number (index) of the end point of segment k

nnp(m) Number of corner points in polygon m

p1(m) Number (index) of the first corner of polygon m

p2(m) Number (index) of the second corner of polygon m

p3(m) Number (index) of the third corner of polygon m

t1(m) Number (index) of the first corner of tetrahedron m

t2(m) Number (index) of the second corner of tetrahedron m

t3(m) Number (index) of the third corner of tetrahedron m

t4(m) Number (index) of the third corner of tetrahedron m

... Number (index) of the additional corners of polygon m

String_name Optional string name of the point. It must be a string of up to five characters, similar to
the point name of the DP card. If a point is named, it can be used in any card following
the IN card.

Label Specifying the label as the last parameter of any structure is optional. If no label is
specified, the value defined at the last LA card will be used. Note that if a label or range
of labels is specified (with parameters after the file name), this LA card label will be
used to determine if a structure is included or not.

The radius of segments must be specified by an IP card before the IN card. It is recommended to
check the geometry with POSTFEKO.
Example:
The structure in Figure 13-19, consisting of 5 node points and 3 triangles with label 7 (no seg-
ments or polygonal plates), may be imported from the following data file

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-41

5 3 0 0
3.0 0.0 1.0
4.0 2.0 1.0
2.5 3.0 2.5
0.0 3.0 4.0
1.0 0.0 3.0
1 2 3 0 7
1 3 5 0 7
3 4 5 0 7

P1 = (3, 0, 1) P4
P2 = (4, 2, 1) y
P3 = (2.5, 3, 2.5)
P4 = (0, 3, 4)
P5 = (1, 0, 3)
P5 P3

P2
P1

Figure 13-19: Example for IN card

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-42

Import NASTRAN file

With this option, PREFEKO can import a model from a NASTRAN file. It supports both 8-character
and 16-character wide column based files as well as comma separated files. NASTRAN files in
cylindrical, spherical coordinate systems, as well as files in Cartesian coordinate systems are
supported. Only the keywords GRID, CTRIA3, CQUAD4, CTETRA, CBAR and CROD for nodes,
triangles, quadrangles (divided into two triangles along the shortest diagonal), tetrahedral ele-
ments, and segments are processed.

NASTRAN does not support polygonal plates, but all other parameters in the general section of
the IN card above apply. The label selection uses the NASTRAN properties which are converted
to FEKO labels.
As when importing FEMAP neutral files, the wire radius must be set with the IP card preceding
the IN card, and an ME card must be used when specifying dielectric surfaces in the same way as
when the IN card is not present.
The user can also import points from the NASTRAN file. The points defined in the NASTRAN
file will then be available in PREFEKO as points (as if they were defined by DP cards) of the
form Nxxx where xxx is the index of the grid point. This may be used, for example, to attach
additional structures to the geometry. In addition, the coordinate values of the point are available
as variables in PREFEKO. For example, the variables #n1234x, #n1234y and #n1234z give the
coordinates of the NASTRAN grid point with index 1234. Note that points are not included by
default.
Since grid points do not have an associated property, points are imported irrespective of their
label, but they may be limited to those used for the imported geometry.
Each line in the 8-character column based format consists of one keyword such as “GRID” starting
in column 1. From column 9 onwards follow 9 input fields with widths of 8 characters each. Thus
input field 1 uses columns 9 to 16, input field 2 uses columns 17 to 24 etc. The ninth (and last)
input field ends at column 80. Below is a very simple NASTRAN example file consisting of a
plate (property 1; subdivided into eight triangles) and a rod (property 2; subdivided into two
segments).

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-43

|1 |9 |17 |25 |33 |41 |49

ID XXXXXXXX,YYYYYYYY
CEND
BEGIN BULK
GRID 1 0.0 0.0 0.0
GRID 2 0.50000 0.0 0.0
GRID 3 1.00000 0.0 0.0
GRID 4 0.0 0.50000 0.0
GRID 5 0.50000 0.50000 0.0
GRID 6 1.00000 0.50000 0.0
GRID 7 0.0 1.00000 0.0
GRID 8 0.50000 1.00000 0.0
GRID 9 1.00000 1.00000 0.0
GRID 10 0.50000 0.50000 2.00000
GRID 11 0.50000 0.50000 1.00000
CROD 9 2 5 11
CROD 10 2 11 10
CTRIA3 1 1 4 5 8
CTRIA3 2 1 4 8 7
CTRIA3 3 1 5 6 9
CTRIA3 4 1 5 9 8
CTRIA3 5 1 1 2 5
CTRIA3 6 1 1 5 4
CTRIA3 7 1 2 3 6
CTRIA3 8 1 2 6 5
ENDDATA

For the node points FEKO also supports 16 character wide input fields. The keyword GRID in
columns 1 to 4 is followed by a star and three spaces. The node ID is then in columns 9 to 24,
the x coordinate in columns 43 to 56, y in columns 57 to 72 and z in columns 9 to 24 of the next
line.
For example
|1 |9 |25... |43 |57 |73 |81

GRID* 1 50.000000000 -18.480176926 1

* 1 23.222875595
GRID* 2 50.000000000 -18.480176926 2
* 2 -13.410394669

For the comma separated format, the individual entries are separated by commas
GRID,1,0,-238.533,186.7983,0.000000,0
GRID,2,0,-244.777,214.3057,172.9991,0
GRID,3,0,288.0060,115.1831,339.8281,0
GRID,4,0,356.2201,50.15516,0.000000,0
CTRIA3,1,1,1,2,3,,0.0,,
CTRIA3,2,1,1,2,4,,0.0,,

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-44

Import AutoCAD DXF file

This card allows importing *.dxf models. The *.dxf file must comply with the release 12 DXF
format specifications. It should contain meshed- or closed- polyline surfaces (see the discussion
below) and lines (that will be segmented by PREFEKO as discussed below).

In addition to the file name, label selection and scale factor discussed in the general section of
the IN card above, the DXF import option supports the following element selection:

Include segments: Same as above.

Include meshed polylines (triangles and quadrangles): Check this item to include meshed polylines (tri-
angles and quadrangles) from the DXF file into the model.

Include node points: Check this item to include node points from the DXF file into the model.

Include closed polylines (meshed into triangles): Check this item to include closed polylines from the
DXF file into the model. These structures will be meshed by PREFEKO.

Layers named n or LAYER_n (where n is an integer number) in the *.dxf model are converted to
label n in FEKO. For all structures for which no label is defined in this format, the label specified
with the last LA card preceding the IN card is used. (If no such LA card is in effect, the default is
label 0.) This label is used in the label selection.
As for the other meshed CAD formats, dielectric triangles or metallic triangles which form the
surface of a dielectric, are created by preceding the IN card with the appropriate ME card.
PREFEKO only processes the geometry information in the section of the file between the keywords
ENTITIES and ENDSEC.
The user can import points (i.e. vertices of polylines and start/end points of lines) from the DXF
file. The points defined in the DXF file will then be available in PREFEKO as points of the form
Qxxx where xxx is a consecutive point index. In addition, the coordinate values of the point are
available as variables in PREFEKO. For example, the variables #q1234x, #q1234y and #q1234z
give the coordinates of point 1234. Note that points are not included by default. The imported
node points will have the label (i.e. converted layer) of the corresponding LINE or POLYLINE
structure (the layer of the VERTEX block for polylines is not used). A label range selection at the
IN card may be applied so that only the points with a correct layer will be imported.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-45

Segments are imported from blocks defined by the keyword LINE.

0
LINE
8
LAYER_01
.
.
10
-0.0538
20
0.0
30
8.134
11
5.110
21
2.857
31
0.0
0
... (next keyword)

The group code 8 at some point below LINE indicates that the next line contains the layer name.
In this case, the layer will be converted to label 1. The line will be imported and segmented if this
label lies in the required range. (If not, PREFEKO will search for the next occurrence of LINE.)
Next the x, y and z components of the start point follow the group codes 10, 20 and 30; and
those of the end point follow the codes 11, 21 and 31.
Here the start and end points are (x, y, z) = (-0.0538, 0.0, 8.134) and (5.110, 2.857, 0.0)
respectively. If any of the coordinate group codes are absent (such as in a 2D model), the related
coordinate is set to zero. The block is terminated by the group code 0. The wire is segmented
according to the maximum segment length specified by the IP card, and the segments also have
the radius specified by this card.
Meshed surfaces are imported from blocks denoted with the keyword POLYLINE. This block
contains the layer name (following the group code 8 as before; if there is no group code 8 before
the first VERTEX, the label specified with the last LA card will be used) and a number of VERTEX
structures. There can be an arbitrary number of VERTEX structures, but there should be at least
four.
The POLYLINE structure is terminated by the keyword SEQEND.
0
POLYLINE
8
LAYER_02
.
.
VERTEX
.
.
VERTEX
.
.
VERTEX
.
.
0
SEQEND

There are two types of vertices. The first type defines points in space

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-46

0
VERTEX
8
LAYER_02
.
.
10
7.919192
20
3.393939
30
0.0
.
.
0
... (next keyword)

where the x, y and z components of the point follow the group codes 10, 20 and 30. The layer
information is ignored.
The second type of vertex is a “linker”.
0
VERTEX
8
LAYER_02
.
.
70
128
71
4
72
2
73
1
74
3
.
.
0
... (next keyword)

which defines a triangle or quadrangle by specifying the indices (starting from 1 in the order the
non-linker vertices are specified) of the vertices which form its corner points. Vertices are defined
as linkers by setting a value of 128 in the group code 70 field. For linker vertices the coordinates
are ignored. Note that some old *.dxf versions do not contain linker vertices — they cannot be
imported. (Usually they do not contain mesh information.)
The four integer numbers after the group codes 71, 72, 73 and 74 give the indices of corners of
the triangle or quadrangle. (In the case of a triangle one of these is absent.) PREFEKO divides
each quadrangle into two triangles along the shortest diagonal.
In addition to being able to import meshed polylines, closed polylines can also be imported. These
will be meshed into triangular patches during the import according to the meshing parameters
set at the IP card.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-47

Import NEC model file

PREFEKO also supports importing wire geometry from NEC6 models. Note that NEC models
usually consist of wire grid surfaces and it would be more efficient to convert the models to
FEKO surfaces, but this cannot be done automatically.

For this option only the file name, label selection and Include segments are supported.
The label selection uses the NEC tags which are converted to FEKO labels. This applies to the
tag when the element is defined. If the tag is modified after the inclusion (for example with the
GM card) the elements with the modified tag are also included. The type selection parameter x
is also supported, but it may only have the value 1 for wire segments.
The NEC import filter considers only the geometry cards CM, CE, GA, GW, GM, GR, GS, GX and
GE. A warning is given if other cards are encountered. If the model contains multiple geometries
only the first one is read.

6
G.J. Burke and A.J. Poggio, “Numerical Electromagnetics Code (NEC) — Method of Moments,”
Lawrence Livermore Laboratory, January 1981.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-48

Import Concept geometry file

With this option one may import CONCEPT7 geometry files.

Since CONCEPT uses two different files for wires and surface elements, the type of element
selection is obligatory and determines the type of geometry file to be read. The only other
parameters supported here are the file name and scale factor. (The CONCEPT file does not
contain labels.) If the concept model contains quadrangles, they are divided into triangles.
Since wires don’t have a radius in the model files, the radius is specified with a preceding IP card.
Likewise, the elements don’t have labels, and the label as specified at the last LA card before the
IN card is used. If there is no LA card, the label defaults to zero.
As for the CAD models, dielectric triangles or metallic triangles which form the surface of a
dielectric, are created by preceding the IN card with the appropriate ME card.
The CONCEPT files for wires are as follows
number_of_wires
x_start y_start z_start x_end y_end z_end [number_of_wires times]

where the first integer specifies the number of wires followed by the coordinates of the start and
end point of each wire. The file is completely free format — the values are just separated by
white space. The surface file is
number_of_nodes number_of_patches
x y z [number_of_nodes times]
p1 p2 p3 p4 [number_of_patches times]

again using free format. The values x, y and z specify the node coordinates and p1, p2, p3 and
p4 specify the corner nodes of the triangles (in this case p4 is 0) and quadrangles.

7
A MoM code developed at the University of Hamburg-Harburg, Germany.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-49

Import STL file

PREFEKO can also import STL — both ASCII and binary — files. STL files supports only triangular
patches and these are all imported. Also, since the STL file makes no provision for any labels,
label selection is not supported. The scale factor is supported.

An example of an ASCII STL file is

SOLID CATIA STL PRODUCT
FACET NORMAL -4.602166E-01 -1.858978E-01 -8.681260E-01
OUTER LOOP
VERTEX 4.789964E-01 -8.440244E-01 2.878882E-01
VERTEX 4.764872E-01 -8.439470E-01 2.892018E-01
VERTEX 4.783065E-01 -8.414296E-01 2.876983E-01
ENDLOOP
ENDFACET
FACET NORMAL -4.601843E-01 -1.859276E-01 -8.681367E-01
OUTER LOOP
VERTEX 4.764872E-01 -8.439470E-01 2.892018E-01
VERTEX 4.761175E-01 -8.425569E-01 2.891001E-01
VERTEX 4.783065E-01 -8.414296E-01 2.876983E-01
ENDLOOP
ENDFACET
ENDSOLID

For the description of binary STL files, please see:

http://www.ennex.com/~fabbers/StL.asp

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-50

Import CADFEKO mesh file

CADFEKO exports the mesh to a *.cfm file which is imported by an IN card in the default *.pre
file created by CADFEKO. The options are similar to those of the other formats that PREFEKO
can import. For the import of *.cfm-files (in addition to the mesh element inclusion/exclusion
options provided) provision is made for the inclusion/exclusion of the variables that are defined
in the *.cfm-file during the import process. Imported variables can then be refered to in other
PREFEKO cards.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-51

Import PATRAN neutral file

PREFEKO also supports importing PATRAN files.

PATRAN does not support polygonal plates, but all other parameters in the general section of the
IN card above apply. The label selection uses the PATRAN material ID’s which are converted to
FEKO labels. Only the following PATRAN neutral packet types are imported:

01: Node data (all coordinates are interpreted in the global rectangular frame, local coordinate
frames are not supported).

02: Element data. The shapes 2(bar), 3(tri), 4(quad) and 5(tet) are allowed. Quadrangles are
automatically subdivided into triangles along the shortest diagonal.

99: End of file flag.

Other packet types are ignored.

As when importing *.neu files, the wire radius must be set with the IP card preceding the IN
card, and an ME card must be used when specifying dielectric surfaces in the same way as when
the IN card is not present.
The user can also import points from the PATRAN file similar to importing points from FEMAP
or NASTRAN files. The points defined in the PATRAN file will then be available in PREFEKO as
points (as if they were defined by DP cards) of the form Txxx where xxx is the index of the
grid point. This may be used, for example, to attach additional structures to the geometry. In
addition, the coordinate values of the point are available as variables in PREFEKO. For example,
the variables #t1234x, #t1234y and #t1234z are set to the coordinates of the point with index
1234. Note that points are not included by default. Since points do not have an associated
property ID, points are imported irrespective of their label.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-52

Import ANSYS CDB file

PREFEKO also supports importing geometry from ANSYS *.cdb files. By default, when exporting
such files from ANSYS, the BLOCKED option is used. PREFEKO only understands this BLOCKED
syntax, the UNBLOCKED version is not supported. Also regarding the element type, only the
ANSYS element types 200 (filaments, triangles, tetrahedral elements, and brick elements) as
well as element type 16 (pipe16, wire with a finite radius) are supported.

The selection of polygonal plates and quadrangles are not supported, but all other (non-selection)
options discussed in the general section of the IN card is supported. It also supports an additional
selection option:
Include cuboidal volume elements: Check this item to include cuboidal elements to be used with the
volume equivalence principle in FEKO.

The component name from the CMBLOCK is converted to the FEKO label. Since the FEKO label
must be an integer value, only component names which are integer strings (for example, 15) or
end with an underscore followed by an integer string (for example, FEED_7) will be converted to
FEKO labels (15 and 7 in the examples above). In all other cases (for example, for a component
name PATCH) the FEKO label will be set to zero.
Note unlike most of the other CAD import formats supported by the IN card, the ANSYS CDB
file makes provision for a wire radius of the segments of type pipe16 (real constant from the
associated RLBLOCK). This is then used during the import and any setting at the IP card is
ignored (the IP card radius is still used for filaments of element type 200). For dielectric bodies,
one must use an ME card to specify the element type and medium indices. The ANSYS field for
the material number cannot be used, since for triangles FEKO requires two such material indices
(medium on each side).
The user can also import points from the ANSYS CDB file similar to importing points from FEMAP
or NASTRAN files. The points defined in the ANSYS CDB file will then be available in PREFEKO
as points (as if they were defined by DP cards) of the form Cxxx where xxx is the index of the
grid point. This may be used, for example, to attach additional structures to the geometry. In
addition, the coordinate values of the points are available as variables in PREFEKO. For example,
the variables #c1234x, #c1234y and #c1234z are set to the coordinates of the point with index
1234. Note that points are not included by default.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-53

Import ABAQUS mesh file

PREFEKO also supports importing mesh files from ABAQUS *.inp files. The various options for
the card panel are as described globally or at the other import options:

Here for ABAQUS mesh files, the ABAQUS element set (ELSET command) is converted to the
FEKO label.
Points are imported similar to the case for NASTRAN files, i.e. they are available in PREFEKO
as points (as if defined by DP cards) of the form Nxxx where xxx is the index of the grid point.
The coordinate values of the point are available as variables of the form #n1234x, #n1234y and
#n1234z. Note that points are not included by default.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-54

Import GiD mesh

PREFEKO also supports importing mesh files from GiD *.msh files.

For GiD mesh files, the material type or layer property is converted to the FEKO label during the
import.
The import options are similar to the options for NASTRAN and ABAQUS mesh imports. The
following should be noted regarding the support of special elements in the GiD mesh.

• Hexahedral elements are not supported in the FEKO import of GiD meshes.

• Nodes that only contain 2 coordinates are interpreted as x and y coordinates on the z=0
plane.

• Quadrilateral elements are divided into triangle elements along the shortest diagonal dur-
ing import.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-55

13.23 IP card

With this card a number of parameters that define the meshing parameters or also the wire radius
are set.

Parameters:

Radius of wire segment: Segment radius in m (it is scaled by the SF card if used).

Maximum triangle edge length: Maximum edge length of triangular elements in m (it is scaled by the
SF card).

Maximum wire segment length: Maximum segment length for wire segments in m (it is scaled by the
SF card).

Maximum cuboid edge length: Maximum edge length of cuboidal volume elements for dielectrics (vol-
ume equivalence principle of the MoM) in m (it is scaled by the SF card).

Maximum tetrahedral edge length: Maximum edge length of tetrahedral volume elements (FEM) in m
(it is scaled by the SF card).

The IP card only affects the commands following it, i.e. it has to be declared prior to the cards
that define segments, triangles or cuboids.
It is possible to use more than one IP card in a file. This is necessary when a finer mesh is
required in certain parts or when different radii occur in the geometry. For any command (e.g.
the BL card) the previous IP card is applicable.
Regarding the meshing, certain rules apply relating the element size (see section 2.2.2) to the
wavelength etc.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-56

13.24 KA card

With this card two points are joined, which form the border of the PO area. On this edge, fringe
wave currents are taken into account.

Parameters:

Start point of edge: Start point of the edge.

End point of edge: End point. The direction of an edge is arbitrary, i.e. it does not matter which
edge point is chosen as the end or start point of the edge.

Label of triangles . . .: Label of the PO triangles adjacent to the PO border, i.e. the edge correction
current from this edge is applied to all triangles with this label.

Note that the surface must be flat, i.e. all triangles with the label specified here must lie in the
same plane.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-57

13.25 KK card

A mesh of surface triangles in the shape of a conical section can be created with this card. The
cone can also be distorted to have an elliptical cross section. Cones or conical sections with
included angles that vary from the top to the bottom, or that do not start in a specified plane,
can also be created.

Parameters:

S1: The start point of the axis of the cone (the centre of the base).

S2: The end point of the axis (the tip of the cone, or the centre point of the circle
when creating a conical section.

S3: A point on the radius of the base.

S4: If this parameter is defined, the cone is cut off at the top; if not the cone has a
sharp tip. This point must be in the plane given by S1, S2 and S3.

Start angle at the bottom: The angle ϕ from the plane S2–S1–S3 at which the bottom of the cone
should start.

End angle at the bottom: The angle ϕ from the plane S2–S1–S3 at which the bottom of the cone
should end.

Start angle at the top: The angle ϕ from the plane S2–S1–S3 at which the top of the cone should
start.

End angle at the top: The angle ϕ from the plane S2–S1–S3 at which the top of the cone should end.

Maximum edge length (bottom): The maximum edge length of the triangles along the base arc — in
the plane containing S1 — of the cone. (This value is in m and is scaled by the
SF card). If this parameter is left empty, the value specified with the IP card is
used.

Maximum edge length (top): This value only applies if S4 is specified and gives the maximum edge
length of the triangles along the top arc — in the plane containing S2 — of the
cone. (This value is in m and is scaled by the SF card). If this parameter is left
empty, the value specified with the IP card is used.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-58

Normal vector directed: The triangles can be created so that the normal vector points Outward (away
from the axis) or Inward (to the inside of the cone).

Scale second half axis: If this parameter is empty or is set to 1, a cone with a circular cross section is
created. If set to ab , a cone with an elliptical cross section is created. Here ab
gives the ratio of the two half axes, where a is the distance S1–S3. It is recom-
mended to generate elliptical cones with extremely small or extremely large
axial ratios with a CAD system as the distortion formulation used in PREFEKO
may fail in these cases.

The fineness of the mesh on the shell’s surface is determined by the maximum edge length spec-
ified by the last IP card prior to the KK card. Along the arcs, accurate modelling of the geometry
may require finer segmentation and the values Maximum edge length (bottom) and Maximum
edge length (top) specify the maximum edge length along the corresponding arcs. Maximum
edge length (top) is only used when a truncated cone is created. If either of these values is not
specified the length specified with the IP card will be used on the corresponding arc.
Examples of KK card usage:
All of the following meshes are created using the KK card. These examples show a sharp cone, an
oblique elliptical cone, a conical section with different angles at the top and bottom and a conical
section where the start angle is not in the plane S2–S1–S3.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-59

Figure 13-20: Example for the KK card from demo_KK1.pre

Figure 13-21: Example of a cone with an elliptical cross section from demo_KK2.pre

Figure 13-22: Example of a cone with different subtended angles at the top and at the bottom from
demo_KK3.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-60

Figure 13-23: Example of a conical section where the start angle is not in the plane defined by S1, S1 and
S3 (demo_KK4.pre)

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-61

13.26 KL card

This card defines a wedge for which correction terms are added to the PO currents on the two
surfaces connected to it.

Parameters:

K1 : The start point of the edge of the wedge.

K2 : The end point of the edge of the wedge.

P0 : A point on the O side of the wedge.

PN : A point on the N side of the wedge.

Label of the O side triangles: The label of the PO triangles that are adjacent to the wedge on the O side.
This means that the corresponding correction term for the O side is assigned
to the PO triangles that have this label.

Label of the N side triangles: The label of the PO triangles that are adjacent to the wedge on the N side.
This means that the corresponding correction term for the N side is assigned
to the PO triangles that have this label.

Note that the wedge must be between flat surfaces, and that all triangles with the label specified
here must lie in the same plane.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-62

13.27 KR card

This card creates a mesh of surface triangles in the shape of circular region with or without a
hole. It is also possible to distort it to an elliptical region.

Parameters:

S1: The centre point of the circle.

S2: A point that is situated at any distance perpendicular to and above the plane
of the circle.

S3: A point on the outer arc.

S4: If there is a value present for this parameter, then a circular ring is created. S4
must lie between S1 and S3.

The angle ϕ: The angle subtended by the arc in degrees.

Maximum edge length (outer): The maximum edge length of the triangles along the outer edge of the
arc in m (is scaled by the SF card). If this parameter is left empty, the value
specified with the IP card is used.

Maximum edge length (inner): When a disk with a hole is created, the maximum edge length for the
triangles along the inner edge of the arc in m (is scaled by the SF card). If this
parameter is left empty, the value specified with the IP card is used.

Scale second half axis: If this parameter is empty or is set to 1, a circular disk is created. If set to
b
a
, an elliptical disk is created. Here ab gives the ratio of the two half axes,
where a is the distance S1–S3. It is recommended to generate elliptical disks
with extremely small or extremely large axial ratios with a CAD system as the
distortion formulation used in PREFEKO may fail in these cases.

The circle’s plane is perpendicular to the line S1–S2. This length is arbitrary. The radius of the
disc is given by the length between the points S3 and S1. The area that is to be subdivided (the
shaded region in the figure) is generated by sweeping the edge S3–S1 around the axis S1–S2
through ϕ degrees in the mathematically positive sense. For ϕ = 360◦ a circle is obtained.
The fineness of the mesh is determined by the maximum edge length specified by the last IP
card prior to the KR card. Along the arcs, accurate modelling of the geometry may require finer
segmentation and the maximum edge length values specify the maximum edge length along the
outer and inner (if applicable) arcs respectively. If any of these values are not specified the length
specified with the IP card will be used on the corresponding arc.
The normal vectors of the triangles on the disk all point in the direction from S1 to S2.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-63

Examples of KR card usage:

The following example meshes are all created using the KR card. Shown is a circular plate, a flat
circular ring and a flat elliptical ring.

Figure 13-24: First example for the KR card from demo_KR1.pre

Figure 13-25: Second example for the KR card from demo_KR2.pre

Figure 13-26: Example of an elliptical disk with a hole from demo_KR3.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-64

13.28 KU card

This card creates a mesh of surface triangles in the shape of a spherical section.

Parameters:

S1: The centre of the sphere.

S2: A point that indicates the ϑ = 0◦ direction in a spherical coordinate system.

The distance between S1 and S2 is the radius of the sphere.

S3: A point that indicates the ϑ = 90◦ , ϕ = 0◦ direction in a spherical coordinate

system. The distance S1–S3 must be equal to the distance S1–S2.

Normal direction: The triangles can be created so that the normal vectors point Outward (away
from the centre of the sphere) or Inward (towards it).

Begin angle ϑa : The start angle ϑa in degrees of the spherical segment.

Begin angle ϕa : The start angle ϕa in degrees of the spherical segment.

Begin angle ϑe : The end angle ϑe in degrees of the spherical segment.

Begin angle ϕe : The end angle ϕe in degrees of the spherical segment.

Maximum triangle edge length: The maximum length of the triangles along the curved edges in m (is
scaled by the SF card). If this parameter is left empty, the value specified with
the IP card is used.

A complete sphere may be created with ϑa = ϕa = 0, ϑe = 180◦ and ϕe = 360◦ .

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-65

An example of KU card usage:

The spherical segment shown in Figure 13-27 is generated using the KU card.

Figure 13-27: Example for the KU card from demo_KU1.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-66

13.29 LA card

With this card, labels are assigned to segments, triangles, polygonal plates, cuboids, UTD cylin-
ders and points.

Parameters:

Label for following . . .: The label assigned to all segments, triangles, etc. defined in cards following
this one.

In order to select the position of the feed (Ax cards)8 , the location of impedance loading (LD, LS,
LP and LZ cards) or structures on which to apply the skin effect (SK cards), each segment, trian-
gle, etc. is assigned a label. Other applications include the selective transformation or copying of
parts of geometry (TG card), and outputting of currents on selected elements (OS card). Labels
are also used to define triangles on which to apply physical optics (PO card (see section 13.37)).
All elements, etc. that are created after the LA card (e.g. by a BP card), are assigned the value
specified in the dialog as label. A different label is only assigned by a new LA card. All structures
created before the first LA card (or if no LA card is present), is assigned the default label which
is 0 (number zero).
Label names (see section 12.4) can be an arbitrary string using the characters A-Z, the digits 0-9
or also the underscore _. Labels may be manipulated using label increments (see section 12.4)
and referenced using label ranges (see section 12.4).

8
The definition Ax stands for any of the control cards A0, A1, A2, . . ..

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-67

13.30 MB card

With this card, a modal port boundary condition may be applied on the boundary of a finite
element method (FEM) region. A modal port essentially represents an infinitely long guided
wave structure (transmission line) connected to a dielectric volume modelled with FEM.

Parameters:

Name of the modal port: The label of the modal port.

S1: Point S1 situated on the FEM modal boundary.

S2: Point S2 situated on the FEM modal boundary.

S3: Point S3 situated on the FEM modal boundary.

Note that the modal port is only available in conjunction with FEM applied to dielectrics. A FEM
modal port can only be applied on the boundary of a FEM region, situated on a planar surface.
The technology behind a modal port is a two dimensional FEM eigensolver which computes the
eigenvalues (modal propagation constants) and eigenvectors (modal electric field distribution)
for the associated infinitely long guided wave structure.
The memory requirement can, for a modal port, be estimated from the number of tetrahedral
faces on the modal port and the order of the solution. The eigensolver by default uses second
order basis functions. Changing the solver settings to use first order basis functions for the FEM
(FP card) will also apply to the modal port analysis. The number of unknowns for a first order
solution is roughly double the number of modal port triangles, and for a second order solution,
7 times. The memory requirement scales with N 2 , where N is the number of unknowns.
The user should take note that memory and time scaling may become an issue with finely meshed
modal port geometries. Note that when meshing modal ports, the default is to use second order
basis functions on modal ports. Hence, a coarser mesh can be used than on the FEM/MoM
boundary (where first order basis functions are always used).

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-68

13.31 ME card

When solving the fields in dielectric objects by means of the MoM/MLFMM surface, volume cur-
rent methods or by means of the FEM or VEP, this card must be used to distinguish the different
media or to create segments and metallic triangles within the dielectric. Furthermore, this card
is used to switch between the generation of metallic triangles and triangles that represent the
surface of the dielectric. Another special case is when metallic triangles represent the surface of
a dielectric object (e.g. a dielectric that has been coated with metal).

Parameters:

Metallic triangles in a homogeneous medium: If this option is selected, then all the surface structures
between this card and the next ME card are assumed to be fully contained
inside the medium specified in the Medium for triangles/segments dialog.

Triangles representing the surface of a dielectric body: If this option is selected, then all the surface struc-
tures created between this card and the next ME card are assumed to define
the boundary between two media. Note that the user needs to provide the
names of the medium on both sides of the triangles. The normal vector points
from medium A to medium B. Thus if we have a dielectric body of medium
DIELECTRIC, constructed so that all the triangle normals point outward, we
set medium A to for instance DIELECTRIC and B to 0 (the number zero always
represents the outer free space region).

Metallic triangles representing the surface of a dielectric body: If this option is selected, then all the sur-
face structures created between this card and the next ME card are assumed
to define a metallic boundary between two media. The selection of the sides is
the same as for the non-metallic case discussed above.

Planar Green’s function aperture triangles in a homogeneous medium: If this option is selected, then all
the surface structures between this card and the next ME card are assumed to
be fully contained inside the medium specified in the Medium for triangles/seg-
ments dialog.

Finite element method (FEM): If this option is selected, the tetrahedral mesh elements will be solved
with the FEM.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-69

Volume equivalence principle - dielectric: If this option is selected, dielectric tetrahedral mesh elements
will be solved with the VEP.

Volume equivalence principle - magnetic: If this option is selected, magnetic tetrahedral mesh elements
will be solved with the VEP.

Volume equivalence principle - dielectric and magnetic: If this option is selected, dielectric and mag-
netic mesh elements will be solved with the VEP.

All the wire segments that follow this card are assigned the properties of the medium in which
they are found. Triangles are treated differently — it depends upon whether they are metallic
triangles or triangles on the boundary of a dielectric object. Here the properties of the media are
assigned to the respective sides.
All triangles and segments before a ME card represent metallic structures in free space. This is
also the case when an input file does not have a ME card.
When using the FEM and meshing structures into tetrahedral elements or when using the volume
equivalence principle (VEP) in connection with the MoM/MLFMM and meshing into cuboidal
volume elements, then the selection for type of triangle is not relevant. The specified medium
will be used (medium A if there are multiple media input fields).
The medium name can be an arbitrary string using the characters A-Z, the digits 0-9 or also the
underscore ‘_’. See FEKO User Manual (section 12.4) for details, also with respect to ranges etc.
Note that the outer medium must always be medium 0 (the number zero).
The use of the ME card to create a simple dielectric sphere is shown in example_04 (see the
Script Examples, note that the normal vectors of the sphere point outwards from medium 1 —
the dielectric — to medium 0 — free space). In addition example_23 shows the more com-
plex example of a dielectric cone (medium 1) mounted on top of a metallic cylinder shown in
Figure 13-28. There are three types of triangles involved:

• Metallic triangles in free space (Medium 0) on the bottom and side of the cylinder

• Metallic triangles also forming the border surface of the dielectric body on the lid of the
cylinder (which is also the basis of the dielectric cone)

• Dielectric triangles forming the surface of the dielectric body (the boundary between medium
1 — the inner dielectric region — and medium 0 — the free space outer region) on the top
surface of the cone

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-70

Figure 13-28: Example of a dielectric cone on top of a metallic cylinder, to demonstrate the use of the ME
card.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-71

13.32 NU card

Surface triangles representing a NURBS surface are created using this card.

Parameters:

Degree (p) of Bezier curve in u-direction: The degree of the Bézier curve in the û-direction.

Degree (q) of Bezier curve in v-direction: The degree of the Bézier curve in the v̂-direction.

Both p and q must be in the range from 1 to 4 where 1 is linear, 2 quadratic, and so on. The
control points are entered in the table, more or less representing their physical relation. There
are p+1 rows and q+1 columns.
It is possible to create a triangular Nurbs surface. In this case all control points on one side must
be identical (use the same point). The weights of the control points are specified at the DP card.
Note that for higher order Bézier curves, the surface does not pass through the control points
except those on the corners.
Examples of NU card usage:
The “saddle point” shape in Figure 13-29 and the linear-quadratic shape in Figure 13-30 are
generated with NU cards using 4 and 6 control points respectively.
NURBS may also be used to generate flat surfaces with curved edges. The section of a circular
plate with a square hole in Figure 13-31 is generated using a NU cards.
Z

AB
BA
AA

BB

Figure 13-29: “Saddle point” example using the NU card from demo_NU1.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-72

AB

AA

BA
AC

BC

BB

Figure 13-30: Linear-quadratic example using the NU card from demo_NU2.pre

Figure 13-31: A flat surface with curved edges created with an NU card in demo_NU3.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-73

13.33 PB card

This card can be used to generate a section of a parabolic reflector as shown in the figure on the
card.

Parameters:

S1: The centre of the paraboloid.

S2: A point perpendicular to the base plane and at any distance above the centre
point.

S3: A point on the outer edge of the paraboloid, but on the base plane.

S4: A point in the plane S2–S1–S3, directly above S3 on the edge of the paraboloid.

Subtended angle ϕ: The angle subtended by the arc of the parabolic reflector in degrees.

Maximum triangle edge length: Maximal edge length of the triangles along the outer edge of the arc in
m (is scaled by the SF card). If this parameter is left empty, the value specified
with the IP card is used.

The radius R of the paraboloid is derived from the distance between the points S1 and S3, as can
be seen in the figure in the card. The height is determined by the distance between points S3 and
S4. The focal point f is determined by:
R2
f = . (13-2)
4h

Example of PB card usage:

The parabolic reflector as shown in Figure 13-32 can be generated by using the PB card.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-74

Figure 13-32: Example for the PB card from demo_PB1.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-75

13.34 PE card

This card defines the unit cell for a periodic boundary condition (PBC) calculation. The phase
change between the cells are specified using the PP card.

Parameters:

One dimension: One dimensional periodicity is used when the unit cell is repeated along a line
(one dimension). See the image for a definition of the unit cell.

Two dimensions: Two dimensional periodicity is used when the unit cell is repeated to form a
surface. See the image for a definition of the unit cell.

S1 : Name of the point S1.

S2 : Name of the point S2.

S3 : Name of the point S3. (Only required for two dimensional periodicity.)

The lattice vectors (û1 and û2 ) do not have to be orthogonal. Geometry may also cross the
periodic boundary as long as it lines up with the geometry edge on the opposite side of the unit
cell. These features allow a large number of problems to be solved.
There are some restrictions on the current implementation of periodic boundary conditions.
These are listed below:

• PBC is not available in conjunction with volume equivalence dielectrics (VEP).

• Triangles and wires are not allowed to cross, but are allowed to touch the the cell boundary.

• PBC can only be used with the free space Green’s function.

• PBC can only be used with MoM (both sequential and parallel), but not with the MLFMM,
PO, UTD or FEM techniques.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-76

13.35 PH card

This card creates a triangular or quadrangular plate with a circular or elliptical hole as shown in
the card.

The hole can be used, for example, to attach a cylinder (ZY card) to the plate and it can be filled
with the KR card.
Parameters:

S1: The corner where the hole is located (also the hole’s centre).

S2: The second corner of the plate.

S3: The third corner of the plate. If this field is left empty, a triangular plate is
created.

S4: The fourth corner of the plate.

S5: A point on the line S1–S2 that forms the starting point of the circle or ellipse
bordering the hole. The special case where S5 is identical to S2 is supported,
but due to the resulting geometry yields triangles with very small angles

Max. edge length on curve: The maximum edge length of the triangles along the edge of the hole in m
(is scaled by the SF card). If this parameter is left empty, the value specified
with the IP card is used.

Scale second half axis: If this parameter is empty or is set to 1, a circular hole is created. If set to ab , an
elliptical hole is created. Here ab gives the ratio of the two half axes, where a
is the distance S1–S3. It is not recommended to generate the plate with a CAD
system if the elliptical hole has an extremely small or extremely large axial
ratio as the distortion formulation used in PREFEKO may fail for such cases.

Examples of PH card usage:

The PH card can be used to create the rectangular plate shown in Figure 13-33. Note the ex-
tremely narrow triangles at the corners as mentioned above. Figures 13-34 and 13-35 show,
respectively a quadrangular and a rectangular plate with the same elliptical hole. The triangular
plate is obtained by leaving the field S3 empty.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-77

Figure 13-33: Example using the PH card from demo_PH1.pre

A E B

Figure 13-34: Example of a quadrangular plate with an elliptical hole (demo_PH2.pre)

A E B

Figure 13-35: Example of a triangular plate with an elliptical hole (demo_PH3.pre)

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-78

13.36 PM card

A surface mesh of triangles in the shape of a polygon is created by using this card. This card also
allows the specification of interior mesh points. This card should generally be used in favour of
other cards that create flat surface meshes with straight edges.

Parameters:

Specify points table/s: This option allows data entry in a table with a maximum of 13 points.

Use point/variable arrays: This option allows data entry using arrays without a limit on the number
of corner points.

Specify non-uniform . . .: Check this item to enable the table in which edge lengths for each edge can
be entered.

Specify internal points: Check this item to enable the table in which internal points can be specified.

Corner points: Enter the points, previously defined with the DP card. The points can be en-
tered in the rows or as point (node name) arrays.

Internal points: Any points internal to the structure where mesh points are required can be
entered here. This is particularly useful for the connection points between
surfaces and wires. The points can be entered in the rows or as point (node
name) arrays.

Edge length: The mesh length on each edge can be set separately in this table. The edge
lengths can be entered in the rows or as variable arrays. When using the table
entry, any blank entries in this table will be meshed with the parameters set
in the IP card. There may not be more entries in this table than in the first

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-79

one. When using the array entry method, the variable array has to be the same
length as the corner point array.

The polygon is defined by entering the points (previously defined with the DP card) on the corners
of the polygon. A maximum of 13 corner points are allowed using the table entry method, but
there is no restriction on the number of points when using the array entry method. The points
are connected in the order that they are entered in the PM card. The array entry method allows
specifying the number of elements, say n, in the array that should be used. Only the first n
elements of the array is then used. Concave corners are allowed. The user can also specify a
smaller or larger mesh size along certain edges.
Examples of PM card usage:
Shown in Figure 13-36 is a plate with a concave corner created with the PM card — note the
finer mesh along the edges from B to C and C to D.
A plate with three internal points, generated using a PM card is shown in Figure 13-37. Note that
there are node points at Q1, Q2 and Q3.

A F

B
C

D E

Figure 13-36: Example for the PM card from demo_PM1.pre

P3

P4
Q3

P5 Q1 Q2

P1
P2

Figure 13-37: Example for the PM card with internal mesh points from demo_PM2.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-80

13.37 PO card

With this card the application of the physical optics approximation is possible.

Parameters:

Use PO on all surfaces with label: together with (optionally) up to label are used to specify the label
or range of labels of all metallic / dielectric triangles that are treated with the
physical optics approximation. If the second field is left blank, only the label
specified in the first field is considered. See also LA and CB cards and also
general discussion of label ranges (see section 12.4).

Do full ray tracing: Normal, complete ray tracing is carried out.

Assume all surfaces to be illuminated: The ray tracing is switched off to save computational time. The
assumption is made that all triangles on which the PO approximation is made
are illuminated by the source and the moment method area. The side in rela-
tion to the normal vector is automatically determined.

Full ray tracing, illumination only from outside: Full ray tracing is done, but metallic triangles can only
be lit from the side to which the normal vector is pointing. See note below.

Use symmetry in ray tracing: If full ray tracing is done, then symmetry can be used to reduce the com-
putational time required to determine the shading. For electric and magnetic
symmetry, this speedup is always used. If geometrical symmetry is used, then
this item should be checked to utilise symmetry. It is possible to e.g. define half
a plate and create the other half through geometric symmetry. An asymmetric
object may then be placed in front of the plate. In this case symmetry should
not be used in the ray tracing.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-81

Decouple with moment method: When this item is checked, the coupling between the MoM region
and the PO region is neglected. The implication is that the currents in the PO
region has no effect on the current distribution in the MoM region. This option,
which should lead to some saving in computational time and storage space, is
especially useful when the PO region and the MoM is not directly adjacent.

Use large element PO formulation: When this item is checked, electrically large triangular patches for
PO are allowed.

Use multi-level boxing to speed up ray tracing: The ray tracing required for the physical optics is accel-
erated by recursively subdividing the problem domain, a so called “multilevel
tree”. It must balance memory requirement against speedup — both increase
as the number of levels increases. The number of levels is determined by spec-
ifying the number of triangles at the lowest level. The user can specify not to
use this algorithm, for the program to determine maximum triangles/box or to
specify this number manually. When a number is specified manually, it should
be greater than 2 and at least a factor 10 less than the number of triangles in
the problem. In general these options should only be set by advanced users.

Save/read PO shadowing information: For the PO formulation the information which triangles are il-
luminated and which are shadowed from the sources is required. Normally
FEKO computes this each time, but there is an option to store this to a file and
load again to save time when the geometry stays constant (say for multiple
runs using different frequencies). The options are:

• No *.sha files (normal execution): Shadowing information is not stored

or read — the default behaviour.
• Save shadowing to a *.sha file: The PO shadowing information is written
to a *.sha file for later reuse.
• Read shadowing from a *.sha file: The PO shadowing information is read
from the *.sha file, i.e. the ray-tracing part is skipped. For large models
this can result in considerable time saving.
• Read *.sha file if it exists, else create it: If a *.sha file exists, the PO
shadowing information is read from this file. Otherwise the information
is calculated and saved in a *.sha file for later use.

Use multiple reflections: When this item is checked, multiple reflections are considered for the ray
tracing. The number of reflections that must be considered is set in the Num-
ber of reflections dialog. This parameter determines the number of reflections
to be taken into account for triangles with labels in the specified range. (For
example, the Number of reflections must be at least 2 to calculate the scatter-
ing from a dihedral and at least 3 for a trihedral.) Increasing the number of
reflections that must be considered significantly increases computation time,
and this should only be done based on physical considerations.

Visibility information: The visibility information related to multiple reflections can be saved to re-
duce the computation time for future runs. There are four options that can be
selected with respect to saving the multiple reflection visibility information:

• No *.vis files (normal execution): Visibility information is not used or

stored — the default behaviour.
• Save visibility to a *.vis file: The PO visibility information is stored in a
*.vis file for later reuse.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-82

• Read visibility from a *.vis file: The PO visibility information is read from
the *.vis file, i.e. the calculation of the visibility information is skipped.
For large models this can result in considerable time saving.
• Read *.vis file if it exists, else create it: If a *.vis file exists, the PO
visibility information is read from this file. Otherwise the information is
calculated and saved in a *.vis file for later use.

The physical optics (PO) approximation can only be used for certain structures. Structures where
the antenna is situated in front of a reflector are well suited. Then PO can be used for the triangles
that form the reflector. This results in a large reduction in computational time and memory for
electrically large objects.
Note that the ray tracing options and the number of reflections can be specified on a per label
basis, by using multiple PO cards. All other parameters can only be specified once. For the global
parameters, the values of the last PO card will be used.
Using Full ray tracing, illumination only from outside has two main applications:

• Acceleration of the PO ray tracing with closed bodies (the normal vector must then point
outward), since the dot product of the normal and propagation vectors can be used to
quickly determine if a triangle is to be used in the ray tracing. In this case the closed model
must be constructed with the normals pointing outward.

• In, for example the MoM/PO hybrid method on a closed body, the MoM region (such as an
antenna) can be prevented from illuminating the PO region from inside.

A basis function that has been assigned to an edge between two triangles will only be solved with
the PO, if the PO approximation has been declared for the labels of both triangles.
The metallic PO region must be perfectly conducting, i.e. no losses are allowed. Dielectric coat-
ings (see CO card) and thin dielectric sheets (see SK card) can, however, be treated with the PO
approximation.

Large element PO

The following needs to be considered when regarding mesh size for large element PO:

• The triangular patch must be a large smooth area with no discontinuities of incident field
(e.g. close to a point source). The surface containing the triangular patch should also not
contain any sharp corners.

• The shadowing is done on the triangular patch level, i.e. smaller mesh elements are re-
quired close to the shadow boundaries.

• For near field calculations (electric or magnetic) or potentials, the maximum triangle mesh
size is limited to 2λ.

• If only far field calculations are requested, the size of the mesh elements is not limited.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-83

• If near fields calculations are requested, a warning will be given if the mesh size is set equal
to or greater than 2λ2 . An error will be given if the mesh size is set equal or greater than
6λ2 .

When using large element PO, the following limitations apply:

• As in the case of normal PO, large element PO may not be used in conjunction with UTD or
MLFMM.

• The PO solution only or the MoM/PO hybrid solution is supported, but it is then required
that the MoM and PO regions are decoupled.

• In order to get one wave propagation factor kn in the PO region, sources (e.g. MoM basis
functions) must be “close” together.

• Large element PO triangles must be PEC. Note that large element PO may be used in con-
junction with normal PO. Connections between normal PO and large element PO are al-
lowed.

• No edge/wedge correction terms are allowed for labels where large element PO is used.

• No extraction of singular terms for near field computations, i.e. near fields are inaccurate
closer than λ/20 . . .λ/10 from surfaces.

• The export of surface currents and visualisation in POSTFEKO are not supported for large
element PO regions.

• Multiple reflections are not allowed in the model if a large PO region is present. When
multiple reflections are present, ray-launching GO should be the preferred method.

• Periodic boundary conditions are not allowed to be used in conjunction with large element
PO.

The following are supported for large element PO:

• A BO ground for example a PEC plate or a real ground.

• Symmetry may be used (also for ray-tracing).

• Shadowing effects but not when located inside of individual large triangles.

• Parallel processing (i.e. parallel ray tracing, parallel field computations etc.)

• Near field (E, H and potentials) computations as well as far field computations including
RCS are allowed.

• All impressed sources my be used in conjuction with the large element PO.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-84

13.38 PY card

This card defines (by specifying the corner points) a polygonal plate surface to which the UTD
formulation is applied.

Parameters:

Specify points table: This option allows data entry in a table with a maximum of 26 points.

A: The first corner point of the polygon.

B: The second corner point of the polygon, etc.

Use points array: This option allows data entry using arrays without any limit on the number of
corner points.

Array name: The name of the point (node name) array.

Number of points in the array: The number of elements, say n, of the point
(node name) array to use. Only the first n elements of the array will be
used.

A maximum of 26 corner points are allowed using the table entry method, but there is no re-
striction on the number of points using the array entry method. The points are connected in the
order that they are entered in the PY card. The corner points have to be defined prior to the PY
card by a DP card.
Example of PY card usage:
This card can be used to generate the polygon (in this case a triangle) shown in Figure 13-38.
Note that this triangle is not meshed, as the result would be if the BT or PM cards had been used.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-85

Figure 13-38: Example for the PY card from demo_PY1.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-86

13.39 QT card

This card is used to create a dielectric or magnetic cuboid, meshed into smaller tetrahedral
volume elements (for solutions using the FEM or VEP). The meshing parameters as set at the
IP card are used, and the medium as set at the ME card is assigned to all created tetrahedral
elements.

Parameters:

S1: First corner of the cuboid.

S2: Opposite corner of the cuboid if aligned with the principle planes, otherwise
one of the corners adjacent to the first corner.

S3: Optional third corner of the cuboid, adjacent to the first.

S4: Optional fourth corner of the cuboid, adjacent to the first.

Note that when using the FEM, in the same model metallic structures are allowed (metallic
surfaces also inside the FEM region or on the FEM boundary, wires only outside). But using
dielectric bodies inside the MoM region at the same time is not supported.
Example of QT card usage:
The dielectric cuboid shown in Figure 13-39 is generated using a QT card.

Figure 13-39: Example for the QT card from demo_QT1.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-87

13.40 QU card

This card is used to create a dielectric or magnetic cuboid, meshed into smaller cuboidal vol-
ume elements, for solutions using the volume equivalence principle in the MoM. The meshing
parameters as set at the IP card are used, and the medium as set at the ME card is assigned to all
created cuboidal elements.

Parameters:

S1: First corner of the cuboid.

S2: Opposite corner of the cuboid if aligned with the principle planes, otherwise
one of the corners adjacent to the first corner.

S3: Optional third corner of the cuboid, adjacent to the first.

S4: Optional fourth corner of the cuboid, adjacent to the first.

Choose the medium: Select here whether the cuboid is dielectric or magnetic or both (this is always
with respect to the environment, e.g. if the relative permittivity " r of the cuboid
material differs from the environment, then this is a dielectric cuboid).

Old format (with medium parameters): Up to and including FEKO Suite 4.3 for cuboids the material
parameters were specified directly at the QU card. This is the old card format.
From FEKO Suite 5.0 onwards the concept of ME/DI cards is used to define
the material by name and to set the material parameters. When checking this
option, the panel layout will change to the old format so that the material pa-
rameters can be entered (depending then on the selection whether dielectric
or magnetic cuboid). FEKO then uses a compatibility mode and creates artifi-
cial media with names QU_MED_xx (xx is an index). It is not recommended
to use the old card format for new models. When working on old models and
pressing F1 in EDITFEKO on an existing QU card, the old format panel will be
opened automatically.

Dielectric bodies treated with the volume equivalence principle (using cuboids) cannot be used
simultaneously with dielectric bodies treated with the surface equivalence principle or the FEM
or with special Green’s functions.
Example of QU card usage:
The dielectric cuboid, shown in Figure 13-40 is generated using a QU card.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-88

Figure 13-40: Example for the QU card from demo_QU1.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-89

13.41 RM card

The RM card provides a sophisticated remeshing and adaptive mesh refinement facility. Most
types of meshes (surface mesh with triangular patches, wire segment mesh, cuboidal volume
elements) created by any option supported in FEKO (e.g. direct creation in PREFEKO with cards,
but also import from NASTRAN, FEMAP, PATRAN etc.) can be used as a basis, and one can then
apply either a local or a global mesh refinement. Unfortunately in FEKO Suite 5.4 there is still a
restriction that tetrahedral volume elements as used for the FEM cannot be refined with the RM
card.
A local mesh refinement refers to a point or a line as reference, or also to a complex cable harness
geometry, which is defined directly by importing the corresponding *.rsd file from CableMod or
CRIPTE.
Note that similar to other FEKO cards, the RM card applies only to what follows in the *.pre file
after the line where the RM card has been read. So for instance if one wants to import a mesh
from a NASTRAN file via the IN card and do a mesh refinement during the import, then one first
has to use the RM card, then followed by the IN card.
Multiple RM cards can be used, for instance if there are multiple areas in a model where the
mesh shall be refined locally. Or also if we use a mesh refinement with respect to one point, the
mesh size increases linearly with distance, and by adding another RM card with a global mesh
refinement option, a threshold can be set.

Parameters:

Remove all existing . . .: Clear all previously defined remeshing rules (i.e. the behaviour is as if no
RM card was read). This option is useful if after having imported a structure
using mesh refinement, one wants to import another structure or create objects
directly in PREFEKO, and for these new structures no mesh refinement shall be
used. If this option is checked, all the other parameters are ignored.

Start a new remeshing rule: Set a new remeshing option (previously read RM cards will be discarded).

Add a remeshing rule: Add a remeshing rule to the already defined ones (i.e. existing RM card rules
will be kept, the new rule will be added to these).

Global mesh refinement: Global mesh refinement using the specified finer mesh size.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-90

Local mesh refinement for a point: Here an adaptive mesh refinement is performed to obtain a finer
mesh close to a point. The point must have been defined before with a DP
card, and its name is passed in the input field for the reference point. Note
that this point can be located arbitrarily in space, there is no need for this to
be on the meshed structure.

Local mesh refinement for a line: Here an adaptive mesh refinement is performed to obtain a finer
mesh close to a line. The line is defined by its start and end point. These two
points must have been created before with DP cards. Multiple simultaneously
active RM cards can be used to perform a mesh refinement with respect to an
arbitrary polygonal shaped line, composed by multiple straight line segments.

Local mesh refinement for a cable harness: With this option one can perform a local mesh refinement
close to a cable harness. The cable harness geometry is specified by a Ca-
bleMod/CRIPTE *.rsd file. The file name of this must be entered into the
respective input field (visible only when this option has been selected).

Mesh polygonal plates: As a special feature, the RM card also allows to mesh unmeshed polygonal
plates (which are used in FEKO for the UTD) during the import. This can be
very useful if e.g. a UTD model is imported from FEMAP using then boundary
surfaces, and instead of the UTD a MoM or MLFMM or PO solution shall be
conducted (where a mesh is required).

Reference point: When using local mesh refinement with respect to a point, then here the name
of this point is entered (the point must have been specified before at a DP
card).

Start point of line: When using local mesh refinement with respect to a line, then here the name of
the start point of the line is entered (the point must have been specified before
at a DP card).

End point of line: When using local mesh refinement with respect to a line, then here the name of
the end point of the line is entered (the point must have been specified before
at a DP card).

CableMod/CRIPTE *.rsd file: When using local mesh refinement with respect to a cable harness, then
here the file name of the CableMod/CRIPTE *.rsd file is specified.

Global finer mesh size: When a global mesh refinement is used, then this is the new mesh size which
shall be applied. Mesh coarsening is not supported, only mesh refinement. So
when the new mesh size is larger than the existing mesh size, simply no mesh
refinement will be done.

Distance D1: Reference distance d1 for the mesh refinement, discussed below.

Mesh size at D1: Mesh size s1 at the reference distance d1 , discussed below.

Distance D2: Reference distance d2 for the mesh refinement, discussed below.

Mesh size at D2: Mesh size s2 at the reference distance d2 , discussed below.

The mesh sizes specified for the global or local mesh refinement apply to all types of geometry
(i.e. triangles, wires, cuboidal volume elements etc.) in the same manner. This is not a principal
restriction. If different refinement options are desired say for wires and triangles, one can use

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-91

one RM card, create or import say just triangles, and then use another RM card and after this
create or import just wires etc.
When the actual remeshing is done, then for each created or imported mesh element (e.g. a tri-
angular surface patch element) PREFEKO internally loops over all the RM cards which are active
at this time, and determines the local mesh length for each RM card, and the smallest of these is
then used for the actual mesh refinement of the specific mesh element under consideration.
If one RM card specifies a global mesh refinement, then the local mesh size is readily given by the
global finer mesh size. If one does local mesh refinement with respect to a point, then first the
distance of the mesh element to this point is determined. Similarly for a line or a cable harness,
the shortest distance from the mesh element to this line or cable is determined. If we assume
that this distance is d, then the local mesh size s is given by the equation
s2 − s1
s = s1 + (d − d1 ) . (13-3)
d2 − d1

This means that for a distance d=d1 we get the mesh size s=s1 , and for the distance d=d2 the
mesh size is s=s2 . For any other distances (smaller than d1 , in between d1 and d2 , or also larger
than d2 ) a linear interpolation is used by means of the formula above. Thus the linear increase of
the mesh size also continues for larger distances, but one should keep in mind that the RM card
can only do a mesh refinement and no mesh coarsening, i.e. as soon as for larger distances the
remeshing option exceeds the already used mesh size of the original model, simply nothing will
happen. Although not required, it is often useful to set the mesh size s2 identical to the global
already existing mesh size, then the parameter d2 readily controls the region where a local mesh
refinement is desired (i.e. for distances d larger than d2 the original mesh will be kept).
It shall also be mentioned here that if a CableMod or CRIPTE *.rsd file is imported, in order to
evaluate distance d between each mesh element and the cable harness in the right base unit (if
an SF card scaling factor is set this can be for instance mm), the cable harness coordinates have
to be scaled accordingly. Thus the SF scaling factor must be known before the RM card can be
used. PREFEKO will give an error if an SF card is read and a RM card was processed before. The
user must then just move the SF card in front of the RM card in the *.pre file.
Examples of RM card usage:
A first example is shown in Figure 13-41 with the original mesh on the left hand side and on
the right hand side the result of a local mesh refinement with respect to a point is given. For
the example in Figure 13-42 a local mesh refinement with respect to two lines is used (i.e. two
simultaneously active RM cards).

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-92

Figure 13-41: Example of local mesh refinement with respect to a point from demo_RM3.pre

Figure 13-42: Example of local mesh refinement with respect to lines demo_RM5.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-93

13.42 SF card

With this card the scaling of the geometric data is possible. This is useful for specifying models
in a convenient unit (e.g. cm) and specifying a scaling factor once, since internally FEKO uses all
dimension related values in metres.

Parameters:

Modify all dimension related values: If this item is checked all geometrical dimensions are scaled. If
unchecked some coordinate values are not scaled (for example the positions of
near field calculations, see the list below). This should only be unchecked for
backwards compatibility with old input files.

Multiply dimensions with: The scale factor. For example, if this is set to 0.001, all dimensions are
entered in mm.

Only one SF card is allowed in the input file. This is global and can be positioned anywhere.
(However, since it is a geometry card it must be before the EG card.)
If Modify all dimension related values is unchecked, the following is scaled:

• Coordinates of the corner points of the triangular surface elements

• Coordinates of the corner points of the segments

• Radii of the segments

• Coordinates of the corner points of the cuboids

• Radii of the all the layers when the Green’s function for a homogeneous or layered dielectric
sphere is used (see section 14.39)

• Thickness of the layers when the Green’s function for a planar, multilayered substrate is
used (see section 14.39)

• Coordinates of the corner points of the polygonal plates

• Coordinates, radii and dimensions of UTD cylinders

• Coordinates of the corner points of tetrahedral volume elements

• Thickness of dielectric surface elements (see section 14.59)

• Radius and thickness of a wire coating (see section 14.29)

• Coordinates of wedges and edges in the PO region

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-94

• Coordinates of the Fock region

• Transmission line length and end point coordinates (see section 14.61)

• The dimensions of the aperture used in the AP card and the amplitudes of the A5 and A6
dipoles which depend on the incremental areas.

• The variable Maximum identical distance, specified with the EG card (see section 13.13),
which controls whether two points are considered to lie at the same point in space

If it is checked all geometrical dimensions and coordinates are scaled. This includes, in addition
to the parameters listed above, the following:

• Named points defined by the the DP card.

• The coordinates of the source point specified in the excitation cards A1, A2, A3, A4, A5,
A6, A7 (if the selection is not made by label).

• Coordinates of the origin of the radiation pattern specified with the AR card.

• Coordinates of the start and end points of the impressed currents for the AI and AV source
cards, as well as the wire radius specified with these cards.

• Radii of the coaxial feed in the A3 card.

• Radius of the approximated connecting segment in the A4 and L4 cards.

• Positions where the near field is calculated with the FE card.

• Offset in the near field calculation (see section 14.51).

Note that if, for example, all data is specified in mm with the scaling set to 0.001, all input
values are interpreted as mm. This also applies to the segmentation parameters (IP card) (see
section 13.23) and possible translations (TG card) (see section 13.44).

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-95

13.43 SY card

Here symmetry can be used to generate the geometry and to reduce computation time.

Parameters:

Select symmetry for the plane x = 0: The type of symmetry, if any, in the yz-plane.

Select symmetry for the plane y = 0: The type of symmetry, if any, in the xz-plane.

Select symmetry for the plane z = 0: The type of symmetry, if any, in the x y-plane.

Label increment for the new structures: After they are mirrored, the labels of the new elements are
incremented with the value specified in this field. Label 0 is, however, not
incremented. The corresponding new elements will also have label 0. If this
field is empty or set to 0, the labels are not incremented, i.e. the new elements
will have the same label as the one they were created from.

All the conducting and/or dielectric triangles, segments, cuboids, tetrahedral volume elements,
wedges, edges, Fock regions and polygonal plates that have been declared before the SY card,
are mirrored. Furthermore, the second and third corners of the triangles are swapped, so that
the direction of the normal vector is retained. Likewise the corners of image polygonal plates are
rearranged to retain the normal direction. (The first corner point of the original polygon becomes
the last corner of the mirror image.)
Sources are not mirrored. If, for example, a Hertzian dipole is placed on one side of the symmetry
plane, the user must also place the correct image on the opposite side of the symmetry plane.
Multiple SY cards can be used and it is possible to mirror around more than one plane at once.
A detailed description of the different types of symmetry (geometric, electric and magnetic sym-
metry) is given in FEKO User Manual (section 24.1).

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-96

13.44 TG card

With this command the already entered geometric elements (triangles, segments, etc.) can be
translated, rotated, mirrored and/or scaled. It is also possible to duplicate structures.

Parameters:

Number of copies: The number of copies to make, for example if set to 3 the selected elements
will be rotated, translated, mirrored and scaled 3 times such that there will be
a total of 4 structures. If set to 0, the existing elements, are rotated, translated,
mirrored, scaled and the number of elements remains the same.

Use label selection: If this option is not checked, then the TG card applies to all the previously
defined geometry. If this option is checked, then a label selective processing is
possible.

Copy structures starting from: together with ending at label can be used to apply the TG card only to
a selected part of the structure. The TG card is applied only to those elements
whose label lies within the range set here (see also LA and CB cards and also
the general discussion of label ranges (see section 12.4)). If the second field
is left empty, only structures with the label set in the first field are considered.
Note that certain element types on the specified label(s) can be excluded from
the selection lower in the card.

Label increment for the new structures: Each newly generated structure will be assigned a label that is
incremented by this value from that of the original structure. An exception is
the label 0 which is retained.

Include: This group can be used to specify which element types (provided they satisfy
the label criterion) are rotated / translated.

Rotation around the x-axis: Angle of rotation α x around the x axis in degrees.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-97

Rotation around the y-axis: Angle of rotation α y around the y axis in degrees.

Rotation around the z-axis: Angle of rotation αz around the z axis in degrees.

Translation along the x-axis: Translation ∆ x in the x direction in m (scaled by SF card).

Translation along the y-axis: Translation ∆ y in the y direction in m (scaled by SF card).

Translation along the z-axis: Translation ∆z in the z direction in m (scaled by SF card).

Mirror about plane at x equal to: The geometry is mirrored around a plane at x equal to a constant
specified. If no value is specified, no mirroring around the plane is performed.

Mirror about plane at y equal to: The geometry is mirrored around a plane at y equal to a constant
specified. If no value is specified, no mirroring around the plane is performed

Mirror about plane at z equal to: The geometry is mirrored around a plane at z equal to a constant
specified. If no value is specified, no mirroring around the plane is performed.

Scale factor: The scaling factor γ, with which the structures must be scaled. (If left empty,
it defaults to 1.)

• For wire segments the wire radius is scaled as well as the coordinates of
the start and end points.
• The scaling factor γ is applied after the translations/rotations have been
conducted, i.e. the new coordinates after the translation/rotation will be
scaled. This means that the effective translation is the value specified at
the TG card multiplied by the scaling factor. (If this is not desired, then
two different TG cards may be used - the first applying only a scaling and
the second performing the translation only).

When an SY card (symmetry) is used before the TG card, the TG card resets the symmetry if
the new structures invalidates the symmetry. Cases where the symmetry is not reset is when, for
example, the plane z = 0 is a symmetry plane and the TG card specifies rotation about the z axis
for a symmetrical selection of elements. In this case the symmetry is retained.
Translation, rotation, mirroring and scaling are performed as a single transformation. The order
is rotate, translate, scale and then mirror.
If more than one copy is made, successive points are generated from the previous point using the
same relation.
With a TG card the simultaneous rotation around multiple axes as well as translation in multiple
directions is possible. A point (x, y, z), for example the corner point of a triangle, is transformed
to a new point
∆x
     
xT x
 yT  = γ M ·  y  + γ  ∆ y  (13-4)
     
zT z ∆z
with the rotation matrix
 
 cos α y cos αz − cos α y sin αz sin α y 
M =  cos α x sin αz + sin α x sin α y cos αz
 cos α x cos αz − sin α x sin α y sin αz − sin α x cos α y 
 (13-5)
sin α x sin αz − cos α x sin α y cos αz sin α x cos αz + cos α x sin α y sin αz cos α x cos α y

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-98

Multiplication by the rotation matrix M effectively rotates a point first by an angle αz around
the z axis, then by an angle α y around the y axis and finally by an angle α x around x axis. It is
important to note that the second rotation around the y axis represents the global y axis. This
is also equivalent to rotating α x around the x axis, then rotating α y around the new y 0 axis and
finally rotating αz around the new z 00 axis.
The transformation angles as used by FEKO in this order are generally referred to as Kardan
angles as opposed to the also commonly used Euler angles. If the rotation shall be performed in
the other order (i.e. first around the x axis, then around the y axis and finally around the z axis),
then one can simply use multiple consecutive TG cards. But since the same rotation algorithm
is also used at other FEKO cards (for instance AC or AR) where one cannot use multiple cards,
a short PREFEKO code segment shall be given here which illustrates how the angles can be
converted:
** Desired rotation angles so that we rotate first around x, then y, and
** then around z
#a1 = 30 ** Angle in deg. around x axis
#b1 = 60 ** Angle in deg. around y axis
#c1 = 90 ** Angle in deg. around z axis

** Precompute some sin() and cos() terms

#ca1 = cos(rad(#a1))
#cb1 = cos(rad(#b1))
#cc1 = cos(rad(#c1))
#sa1 = sin(rad(#a1))
#sb1 = sin(rad(#b1))
#sc1 = sin(rad(#c1))

** Some auxiliary terms resulting from equating the transformation matrices

#cc2 = #cb1*#cc1/(sqrt((#cb1*#cc1)^2+(#ca1*#sc1-#sa1*#sb1*#cc1)^2))
#cb2 = #cb1*#cc1/#cc2
#ca2 = #ca1*#cc2/#cc1
#sa2 = #cc2*(#sa1*#cc1-#ca1*#sb1*#sc1)/(#cb1*#cc1)
#sb2 = #sa1*#sc1+#ca1*#sb1*#cc1
#sc2 = #cc2*(#ca1*#sc1-#sa1*#sb1*#cc1)/(#cb1*#cc1)

** Finally compute the angles which must be used in FEKO in the TG card
** for the rotation order first around z, then around y, and then around x
#a2 = deg(atan2(#sa2,#ca2))
#b2 = deg(atan2(#sb2,#cb2))
#c2 = deg(atan2(#sc2,#cc2))

The file card based version of example_18.pre (see the Script Examples) gives an example of
an application of the TG card.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-99

13.45 TO card

Using this card a surface mesh in the form of a toroidal segment can be generated.

Parameters:

S1: The centre of the toroid.

S2: A point that is perpendicular and is situated an arbitrary distance above the
plane of the toroid.

S3: The start point of the axis of the toroid.

S4: A point on the surface of the toroidal segment. It must be in the plane S2–S1–
S3.

The angle ϕ: The angle of rotation around the axis S1–S2.

The angle α: The angle of rotation around the axis of the toroid, see the figure displayed in
the card.

Max edge length, ϕ direction: The maximum edge length along the curved edge in the ϕ direction in
m (is scaled by the SF card). If this parameter is left empty, the value specified
with the IP card is used.

Max edge length, α direction: The maximum edge length along the curved edge in the α direction in
m (is scaled by the SF card). If this parameter is left empty, the value specified
with the IP card is used.

Normal vector directed: The triangles can be created so that the normal vectors point Outward (out-
ward, away the ring axis of the toroid) or Inward.

Scale second half axis: If this parameter is empty or is set to 1, a toroid with a circular cross section
is created. If set to ab , an elliptical toroid is created by distorting the entire
geometry along the second half axis (orthogonal to the axis S1–S3) with the
factor ab where a is the distance S1–S3. It is not recommended to generate
toroids where the elliptical cross section has extremely small or extremely large
axial ratios with a CAD system (such as FEMAP) as the distortion formulation
used in PREFEKO may fail in these cases.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-100

A complete toroid is obtained by using the parameters ϕ = 360◦ and α = 360◦ .

Examples of TO card usage
The toroidal segment, which is shown in Figure 13-43, is generated using a TO card. This card
can also be used to generate the toroidal segment with an elliptical cross section as shown in
Figure 13-44.

Figure 13-43: Example for the TO card from demo_TO1.pre

Y D
C

Figure 13-44: Example for the TO card with an elliptical cross section from demo_TO2.pre. Note that it
is stretched in the direction of the y axis, i.e. it is elliptical in the ϕ-plane. It is also elliptical
in the α-plane when ϕ=90◦ , but it is circular in the α-plane when ϕ=0◦ .

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-101

13.46 TP card

With this card points (previously defined with the DP card) can be translated, rotated and/or
scaled (relative to the origin).

Parameters:

Use label selection: If this option is not checked, then the TP card applies to all the previously
defined points. If this option is checked, then a label selective processing is
possible.

Move all points starting from label: together with ending at label specify the label range of points that
must be translated, rotated or scaled. See the general discussion of label ranges
(see section 12.4). If the second field is left empty, only structures with the
label set in the first field are considered.

Label increment for moved points: Each transformed point will be assigned a label that is the label of
the original point incremented by this value. The exception are points with
label 0 — their label is not incremented, it remains 0.

Rotation around the x-axis: Angle of rotation α x around the x axis in degrees.

Rotation around the y-axis: Angle of rotation α y around the y axis in degrees.

Rotation around the z-axis: Angle of rotation αz around the z axis in degrees.

Translation along the x-axis: Translation ∆ x in the x direction. (All three translation distances are
affected by the scaling factor set with the SF card.)

Translation along the y-axis: Translation ∆ y in the y direction.

Translation along the z-axis: Translation ∆z in the z direction.

Scaling in: In this group the user may choose whether scaling is in the x-direction, y-
direction or z-direction or a combination of these.

Scale factor: The scaling factor γ, with which the point is scaled after rotation and transla-
tion (if the parameter R7 is not specified, it defaults to γ=1).

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-102

If a point is rotated around more than one axis with a single card, it is rotated first by an angle
αz around z axis, then by α y around the y axis and finally by α x around the x axis. A more
detailed description of the transformation can be found in the description of the TG card (see
section 13.44).
In an exception to the rule that all geometry cards must appear before the EG card, this card (as
well as the DP card) can be used after the EG to define points for use in the AP card.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-103

13.47 UT card

With this command the parameters for the Uniform geometric theory of diffraction (UTD) for
polygonal plates and Geometrical optics (ray launching) (GO) are defined.

Uniform geometric theory of diffraction (UTD)

Parameters:

Type: The method (UTD/GO) that is to be applied can be chosen.

Max no. of ray interactions: This parameter gives the maximal number of ray-interactions (i.e. reflec-
tion and diffraction combined). If, for example, the parameter is set to 3, a ray
can have 3 reflections, or 2 reflections and a diffraction. If set to 0, only direct
rays are taken into account.

Write debug information to *.dbg: If this item is checked a debug file (extension *.dbg) is generated.
This file contains large amounts of information and should only be used when
debugging.

Export ray data for later viewing: When this item is checked the ray information is exported to the
*.bof and to a special *.ray file, so that the ray paths can be displayed in
POSTFEKO. The ray information can become very large, and thus it should
only be exported if specific ray paths are to be examined.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-104

The following abbreviations are used in the *.ray file and POSTFEKO:

B: Diffraction at an edge
D: Diffraction at a corner (of a wedge)
E: Diffraction at a corner (of an edge)
K: Diffraction at a wedge
Q: Source point
R: Reflection
S: Observation point
C: Creeping wave
T: Tip diffraction

Select ray contributions: Determines which ray contributions to take into account:

• GO (direct and reflected rays, shadowing): Geometrical optics (ray launch-

ing) (GO), i.e. direct and reflected rays and shadow regions are taken into
account.
• Edge and wedge diffracted rays: Diffraction on edges and wedges are
taken into account. The ray may include an arbitrary number of reflec-
tions, but only one diffraction. (The total number of interactions — the
number of reflections plus one for the diffraction — must not be larger
than specified in the Max no. of ray interactions field.)
• Corner diffraction terms: Corner diffraction.
• Double diffractions and diffraction/reflection: Double diffraction on edges
and wedges and combinations of reflections are taken into account. Sin-
gle diffraction rays are not included in this item.
• Creeping waves: Creeping waves on curved surfaces.
• Cone tip diffraction: Tip diffraction at the tip of a cone.

Uncoupled with moment method: This item specifies whether the coupling from the UTD region to the
MoM region should be considered. This option should only be used when the
UTD and MoM regions are not close together.

Increasing the type and number of ray interactions increases accuracy but computation time as
well. The user should therefore make a compromise between the number of ray interactions and
the ray contributions. Choices made in this card should be made on physical considerations to
get optimal use from the UTD formulation.
The following restrictions apply for the hybrid MoM/UTD:

• No dielectric bodies or dielectric ground.

• Only perfectly conducting flat polygonal plates or a single cylinder allowed in the UTD
region.

• No UTD and PO at the same time.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-105

Geometrical optics (ray launching)(GO)

Parameters:

Type: The method (UTD/GO) that is to be applied can be chosen

Use geometrical optics (ray launching) on all surfaces with label: The label to which the geometrical op-
tics (ray launching) should be applied

(optionally) up to label: The label up to which the GO should be applied. GO will not be applied to
labels outside of the range between this field and the label in the previous field.

Max no. of ray interactions: This parameter gives the maximum number of ray-interactions (i.e. re-
flection and diffraction combined). If, for example, the parameter is set to 3,
a ray can have 3 reflections, or 2 reflections and a diffraction. If set to 0, only
direct rays are taken into account.

Write debug information to *.dbg: If this item is checked a debug file (extension *.dbg) is generated.
This file contains large amounts of information and should only be used when
debugging.

Export ray data for later viewing: When this item is checked the ray information is exported to the
*.bof and to a special *.ray file, so that the ray paths can be displayed in
POSTFEKO. The ray information can become very large, and thus it should
only be exported if specific ray paths are to be examined.

Select ray contributions: Determines which ray contributions to take into account:

• GO (direct and refl. rays, shadowing, diel. trans): direct, reflected and
transmitted rays are taken into account.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-106

Ray launching settings: The specific settings for the ray launching used in the geometrical optics (ray
launching) can be set here.

• Automatic determination: The ray launching parameters will automati-

cally be determined during the solution process
• Manually set angular / spatial increments The angular / spatial incre-
ment angles for the ray launching are manually specified by the user.

Manually set angular / spatial increments

Angular increment (spherical ray launching): Theta-direction (degrees): Specific Theta increment
angles can be specified for the ray launching proce-
dure in the geometrical optics (ray launching) so-
lution.
Phi-direction (degrees): Specific Phi increment angles can be specified fro
the ray launching procedure in the geometrical op-
tics (ray launching) solution.
Spatial increment (parallel ray front): u-direction (metres): This parameter sets the distance be-
tween the individual rays in the parallel ray front
in the u-direction.
v-direction (metres): This parameter sets the distance between the indi-
vidual rays in the parallel ray front in the v-direction.

Uncoupled with moment method: This item specifies whether the coupling from the UTD region to the
MoM region should be considered. This option should only be used when the
UTD and MoM regions are not close together.

The main advantages of ray-launching GO:

• Suitable for large arbitrarily shaped objects (also curved).

• Dielectrics may be modelled with the ray-launching GO method.

• Support of a plane wave as a excitation (required for the calculation of far field RCS).

• Multiple reflections more efficient than UTD or PO.

When no UT card is used, the following default values apply:

• Max no. of ray interactions: 3.

• Write debug information to *.dbg: unchecked.

• Export UTD ray data for later viewing: unchecked.

• Select ray contributions includes.

– GO (direct and reflected rays, shadowing).

– Edge and wedge diffracted rays.
– Corner diffraction terms.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-107

13.48 UZ card

With this card a cylinder is created for the UTD region.

Parameters:

S1: The start point of the cylinder axis.

S2: The end point of the cylinder axis.

S3: A point on the radius of the cylinder, the angle S2–S1–S3 must be 90◦ .

The angle ϕ: The angle of the cylinder segment. Currently only an angle of 360◦ is sup-
ported.

S1 side end-cap: Select a flat end cap or a semi-infinite end on the side of S1.

S2 side end-cap: Select a flat end cap or a semi-infinite end on the side of S2.

The cylinder in Figure 13-45 was created using a UZ card. Note the absence of discretisation.

Figure 13-45: Example for UZ card from demo_UZ1.pre

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-108

13.49 VS card

This card specifies known visibility information (required when using physical optics with multi-
ple reflections) to reduce the time required to calculate it.

To accurately compute multiple reflections the code needs to determine which basis functions are
visible to each other. Since this applies to all the PO triangles it may be very time consuming for
large problems. The time required to determine the visibility may be greatly reduced if the user
can inform the code that certain triangles are hidden from each other and others are visible to
each other.
Parameters:

Triangles labelled: The label of the source triangles.

are visible from: All triangles with the label specified in the field Triangles labelled are visible
from all triangles with label(s) indicated in the fields below.

are not visible from: All triangles with the label specified in the field Triangles labelled are not visible
from all triangles with label(s) indicated in the fields below.

(specify range of labels): If this item is unchecked, only a single label is specified (in the field triangles
with label). If checked, the card applies to all triangles with labels in the
range from the value specified in the triangles with label field to that in the to
triangles with label field.

Note that visibility is reciprocal, i.e. if all triangles with label n are visible from all triangles with
label m, all triangles with label m are visible from all triangles with label n as well.
Basis functions cannot illuminate each other if all the triangles they are attached to lie in the
same plane.
The VS card should only be used if the user can specify the visibility beyond any doubt and if it
applies to all triangles of that label. If no information is specified for a specific combination of
labels/triangles, full ray tracing will be executed.
Consider the structure shown in Figure 13-46 consisting of four flat plates and a cylindrical
section. The two plates lying at 45 degrees to the coordinate system (labelled 1 and 3), are half
as wide as the plates with labels 0 and 2. Thus some triangles with label 2 are visible to some
triangles with label 0, but not all.
We have to specify which triangles are visible / hidden from all triangles with label 0 first, then
those visible from label 1 and so on. The VS cards for this example would be as follows:

• Triangles labelled 0 are not visible from triangles with label 0.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-109

Z
0
1
2
3
4

0 1
X

Figure 13-46: Structure used to demonstrate the use of VS cards from demo_VS1.pre

• Triangles labelled 0 are visible from triangles with label 1.

• Triangles labelled 0 are not visible from triangles with labels 3 to 4.

• Triangles labelled 1 are not visible from triangles with label 1.

• Triangles labelled 1 are visible from triangles with label 2.

• Triangles labelled 1 are not visible from triangles with labels 3 to 4.

• Triangles labelled 2 are not visible from triangles with label 2.

• Triangles labelled 2 are visible from triangles with label 3.

• Triangles labelled 3 are not visible from triangles with label 3.

• Triangles labelled 3 are visible from triangles with label 4.

Since all the triangles with label 0 lie in the same plane, they cannot illuminate each other. Thus
the first card states that label 0 is hidden from label 0.
All triangles with label 1 are visible from all triangles with label 0. This is specified by the second
VS card. Since some triangles with label 2 are visible from some triangles with label 0 while
others are hidden, we cannot specify any information for this combination of layers. However,
the plate with label 2 shadows all triangles with labels 3 and 4 and we may specify that these
are hidden. This is done with the third VS card. Note that this card specifies a range of hidden
labels.
Next we must specify which triangles are visible (or hidden) from all triangles with label 1. As
for label 0, triangles with label 1 are not visible to each other — specified by the fourth VS card.
All triangles with labels 0 and 2 are visible from all triangles with label 1. Since we have already
specified the visibility between labels 0 and 1, we do not specify it again. The fifth VS card then
specifies that label 2 is completely visible from label 1. As for label 0, both labels 3 and 4 are
hidden completely which completes the first six VS cards,

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-110

Next we look at label 2. As before we need not consider labels lower than 2. Also the label is
hidden from itself as indicated by VS card number seven. Next we state that label 3 is visible, but
we cannot specify anything about label 4 as only some of these triangles will be visible.
Similarly VS cards 9 and 10 states that label 3 is not visible to itself and fully visible to label 4.
Finally we must consider the case for triangles with label 4. All visibility with layers 0 to 3 has
been specified and may not be specified again. Unlike the previous flat plates, layer 4 is curved
and some triangles may indeed illuminate other triangles with the same layer. However, not all
other triangles will be illuminated (this is only possible for a doubly concave surface), so that we
cannot specify any information for label 4.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-111

13.50 WA card

Use this card to define all active windscreen antenna elements. This would include all elements
in close proximity to the finite glass structure and can consist of either segments or triangles (all
defined by labels).

Note that the the three cards WR (see section 13.52), WA (see section 13.50) and WD (see
section 14.63) should be used together to create windscreen antenna models. These three cards
respectively define the windscreen reference surface, windscreen active elements (antenna) and
the windscreen layered media definition.
Parameters:

Windscreen name: Name of the windscreen.

Use windscreen modelling for label: Start of the label range.

(optionally) up to label: End of the label range (optionally).

Offset from reference (Offset A): Offset of the specified label geometry w.r.t. the reference windscreen
triangles.

The antenna elements will be limited to laying tangentially w.r.t. the windscreen surface. Using
a defined offset from the reference plane (in the direction of the reference plane normal), these
elements can then be positioned at the exact required location. This offset is specifically needed
because of the limitations of a finite mesh (compared to a smooth surface) in combination with
curvature in the model.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-112

13.51 WG card

With this card a wire grid in the shape of a parallelogram can be generated.

Parameters:

S1, S2, S3, S4: The four corners of the parallelogram in consecutive order.

Generate wires on the outside edges: If the No item is selected, only the wires inside the parallelogram
are generated whereas if the Yes item is selected all the wires are generated.
This option is important when two adjacent parallelograms are generated, as
the segments along the sides must not be generated twice.

Length of the grid gaps: The maximum segment length is given by the IP card. This parameter is an
integer number and specifies the density of the grid. If, for example, this is set
to 2 the wires only cross at every second segment.

Example of WG card usage:

The wire grid seen in Figure 13-47 is generated using the WG card.

Figure 13-47: Examples of the WG card from demo_WG1.pre — the one on the right has a value of 2 in
Length of the grid gaps.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-113

13.52 WR card

Use this card to define the dielectric windscreen reference plane. Geometrically this surface is
not part of the EM model and is used simply to determine the curvature factor between the two
elements on the windscreen.

Note that the the three cards WR (see section 13.52), WA (see section 13.50) and WD (see
section 14.63) should be used together to create windscreen antenna models. These three cards
respectively define the windscreen reference surface, windscreen active elements (antenna) and
the windscreen layered media definition.
Parameters:

Windscreen name: The name of the windscreen.

Use as reference all surfaces with label: The start of the label range.

(optionally) up to label: The end of the label range (optionally).

The windscreen reference triangles are defined by label and since this plane also forms the zero
reference w.r.t. the defined windscreen antenna elements and glass layers, it should be noted
that they should all have their normals pointing in the same direction.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-114

13.53 ZY card

With this card a surface mesh in the form of a cylindrical segment can be generated.

Parameters:

S1: The start point of the axis.

S2: The end point of the axis.

S3: A point on the corner of the cylindrical segment.

Normal vector directed: The triangles can be created so that the normal vector is points Outward or
Inward.

The angle ϕ: The angle in degrees, which is subtended by the cylindrical arc.

Maximum edge length on arc: Maximum edge length of the triangles along the curved side in m (is
scaled by the SF card). If this parameter is left empty, the value specified with
the IP card is used.

Scale second half axis: If this parameter is empty or is set to 1, a circular cylinder is created. If set to
b
a
, an elliptical cylinder is created. Here ab gives the ratio of the two half axes,
where a is the distance S1–S3. It is not recommended to generate elliptical
cylinder with extremely small or extremely large axial ratios with a CAD system
as the distortion formulation used in PREFEKO may fail in these cases.

For an orthogonal cylinder (i.e. the lines S1–S2 and S1–S3 are perpendicular), the segmented
area (shaded in the figure of the cylinder) is obtained by rotating the point S3 around the axis
S1–S2 through ϕ. For ϕ=360◦ a full cylinder is obtained.
An oblique cylinder (i.e. the circular or elliptical rim is not perpendicular to the axis) can also
be created. Then S1–S2 still represents the axis, but the top and bottom planes of the rims are
defined by planes perpendicular to the plane defined by the three points S1, S2, S3, and parallel
to the line S1–S3.
Examples of ZY card usage:
The cylindrical section (Figure 13-48), elliptical cylinder (Figure 13-49) and non-orthogonal
cylinder (Figure 13-50) below were all created with ZY cards.

July 2011 FEKO User’s Manual

GEOMETRY CARDS 13-115

Figure 13-48: Example for the ZY card from demo_ZY1.pre

Figure 13-49: Example of an elliptical cylinder from demo_ZY2.pre

Figure 13-50: Example of a non-orthogonal elliptical cylinder from demo_ZY3.pre

July 2011 FEKO User’s Manual

CONTROL CARDS 14-1

14 Control cards

The following table summarises all the PREFEKO and FEKO control cards. These cards should
not be used in the geometry section of the *.pre input file, i.e. they should only be used below
the EG card. The control cards are used to specify, for example, the frequency and the type of
excitation. They also determine the required calculations (such as the locations for near field
calculations and the directions for far field calculations, etc.).

Card Description
** characters used to indicate a comment line
Ax type of excitation (e.g. an incident plane wave or a voltage source)
AB creates a modal port excitation
BO through the use of the reflection factor coefficient a ground plane
can be inserted
CA defines a cable path section for the cable irradiation computation.
CD defines a specific cable cross-section (single, coaxial, ribbon and
bundle/multi-cable)
CF set the type of integral equation for perfectly conducting metallic
surfaces
CG the algorithm used to solve the matrix equation is selected
CM field calculation for CableMod and CRIPTE (coupling into transmis-
sion lines) or PCBMod (coupling into a PCB)
CO inserts a dielectric and/or magnetic surface on the elements
CS defines a cable path section and the centre/reference location to
which a cable cross-section is applied
DA creates additional files for the results
DI enters the properties of the dielectric medium
DL defines a layered medium
EE calculates the error estimates
EN indicates the end of the input file
FE calculates the near fields
FF calculates the far fields
FR sets the frequencies at which the calculations are to be carried out
GF sets the Green’s functions
L2 defines a complex load on a vertex
L4 adds a load between a metallic triangle and the ground plane for the
planar multilayer Green’s function
LC defines a cable load
LD defines a distributed load, consisting of resistance, inductance and
capacitance
LE defines a load on the edge between surface triangles
LF impress a complex impedance between two points inside a FEM
mesh
LN defines a complex load to any non-radiating network port that is not
connected to geometry

July 2011 FEKO User’s Manual

CONTROL CARDS 14-2

LP defines a parallel circuit (resistance, inductance and capacitance)

load
LS defines a series circuit (resistance, inductance and capacitance) load
LZ defines a complex load on a segment
NW define a linear non-radiating network
OF offset i.e. displacement of the origin when calculating the near fields
OS saves the surface currents in a file
PP defines the phase for periodic boundary condition calculation
PS sets general control parameters
PW defines the radiating power of a transmitting antenna
RA defines an ideal receiving antenna
SA used to calculate SAR in dielectric media
SH used to define solid and braided cable shields
SK takes a finite conductivity into account through the skin effect of
ohmic losses; also for thin dielectric layers
SP calculates the S-parameters for the active sources
TL specifies a non-radiating transmission line
TR calculate reflection and transmission coefficients for an incident
plane wave on a planar structure
WD defines the dielectric properties of the windscreen glass layers

As mentioned above the control cards form the second part of the input file (see section 12.2).
Control cards are processed line by line and only affect other cards and calculations specified
below them in the input file. (Information specified in a control card is not available before that
line is processed.) Any number of control cards can be used, but they should adhere to a basic
sequence. Thus, for example, the frequency (FR card) and the type of excitation (Ax card) must
be defined before the near fields can be calculated with an FE card.
In addition, a sensible order for the control cards can result in a considerable reduction in the
computation time. For example, for an SK card the whole matrix has to be recalculated, while an
Ax card only redefines the right hand side of the matrix. A summary of the dependencies is given
in the table below: There are also other dependencies. If the matrix elements are recomputed

Action For the cards

recalculates matrix elements BO, CF, CO, DI, FR, GF, L4, LD, LE, LP, LS,
LZ, SK, SP, TL
recalculates the right hand side Ax, BO, DI, FR, GF
resolves the matrix equation CG

then the matrix equation has to be solved again. The actual calculation is started by the CM, FE,
FF, OS, RA, SA and SP cards. All other cards are read and the data stored.
When solving for a number of frequencies all the control cards following the FR card (up to, but
excluding, the next FR card or EN card) are read into a buffer. For each frequency all these cards
are read and processed. The computation time can be reduced significantly by using the cards
in the correct order. If, for example, a structure needs to be investigated at two frequencies and

July 2011 FEKO User’s Manual

CONTROL CARDS 14-3

with two different excitations then the control cards can be organised in either of the following
ways:

...

FR FR card for the two frequencies

Ax first excitation
...

Ax second excitation
...

EN end of the input file

or

...

FR FR card for the two frequencies

Ax first excitation
...

FR FR card for the two frequencies

Ax second excitation
...

EN end of the input file

In the first format, the two excitations are processed one after the other for each frequency.
(The cards are executed in the order FR Ax . . . Ax . . . FR Ax . . . Ax . . . EN, where the second
FR indicates execution of the frequency loop for the second specified frequency.) In the second
format the first excitation is treated for both frequencies before treating the second excitation for
both frequencies. (Here the cards are executed as FR Ax . . . FR Ax . . . FR Ax . . . FR Ax . . . EN.)
For the computational requirements, one finds that in the first case the matrix elements have to
be calculated twice (it has to be completely recalculated each time the frequency is changed) and
in the second case four times.
The computation time is not only influenced by the control cards, but also by what has to be
solved for. In the following example, the structure has to be solved at a number of frequencies
and for the ideal (conducting) and non-ideal (conduction with losses) cases:

...

FR FR card for multiple frequencies

FE calculation of the near fields

July 2011 FEKO User’s Manual

CONTROL CARDS 14-4

SK include skin effect

FE calculation of the near fields
EN end of the input file

The three control cards FE, SK and FE are written to the buffer and are worked through in the
loop for the different frequencies. At the first frequency the FE card initiates the field calculation.
Because a SK card has not yet been read, ideal conductivity is assumed. Then the SK card is read
and losses are taken into account when the second FE card is run. Thus the first frequency pass is
finished. At the next frequency pass the cards FE, SK, and FE are read again, but the losses from
the SK card are still active from the first pass. The SK card is thus useless and the two FE cards
calculate the same things twice.
The correct input order is:

...

FR FR card for multiple frequencies

SK skin effect switched off
FE calculation of the near fields
SK skin effect switched on
FE calculation of the near fields
EN end of the input file

Then the four cards SK, FE, SK and FE are calculated for all the frequencies.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-5

14.1 ** card

The comment lines (see section 13.1) can be used to add comment lines before or after the EG
card.
The ‘**’ indicator may also be used to name the sources, loads, cables, transmission lines and a
number of solution requests (e.g. near fields, far fields, currents, SAR, receiving antennas) after
the EG card.
These names are then used by the FEKO kernel during the solution (e.g. for the screen process
output or in the *.out file), but are also used for post-processing purposes.
Such names are also critical for any solution entities that are to be used or referred to in the
optimisation setup (see section 7), as these labels are used to uniquely identify specific data
in the output of the FEKO solution for consideration during the optimisation goal evaluation
process.
The names of specific cards are set by adding a ** comment at the end of the line in which the
card is defined (for multi-line cards, this comment should be added at the end of the first line of
the card).
For example, with the card sequence
...
A1 0 Feed 1 0 ** Top right array element

FE .... ** Cut plane 1 (through left mast)

FE .... ** Cut plane 2 (vertical through ground)

A voltage source is defined with the name Top right array element and two near field
computation requests with the names Cut plane 1 (through left mast) and Cut plane
2 (vertical through ground). These names can then be referred to in the optimisation
goal definitions (see section 7.1.4) and are used in POSTFEKO and in the FEKO *.out file.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-6

14.2 Ax Cards

This card defines the type of excitation as well as other parameters regarding the excitation. The
following possibilities are available:

Card Type of Excitation

A0 A linear polarised plane wave incident on the structure.
A1 Excitation by means of a voltage gap on a segment (i.e. impressed
electric field strength along a segment).
A2 Excitation by means of a voltage gap at a node i.e. between two
segments.
A3 Excitation by means of a magnetic ring current (TEM-frill) on a seg-
ment. Thus a coaxial feed can be modelled.
A4 Special vertical pin excitation, e.g. for a patch antenna on a planar
substrate with a ground plane (coaxial probe excitation mode).
A5 An electric Hertzian dipole is used as excitation. The position and
orientation in the space are arbitrary.
A6 A magnetic Hertzian dipole is used as excitation. The position and
orientation in the space are arbitrary.
A7 Excitation by means of a voltage gap on an edge between two trian-
gles. This card has been generally replaced by the AE card.
AC This card reads the geometry and current distribution (possibly for
more than one frequency) from an *.rsd file created by a trans-
mission line simulation program (e.g. CRIPTE or CableMod) or by a
PCB simulation tool (e.g. PCBMod) or by export with the OS card in
FEKO. The excitation is due to the electromagnetic fields radiated by
these line currents.
AE The AE card is an excitation between triangle edges similar to the
A7 card, however the AE card permits the simultaneous excitation
of several edges.
AF Excitation by an impressed line current in the FEM region.
AI Excitation by an impressed line current.
AK Excitation by means of a voltage source connected to a radiating
cable.
AN Excitation by means of a voltage source connected to a non-radiating
network port.
AP Excitation by means of equivalent sources in an aperture (array of
electrical and magnetic Hertzian dipoles).
AR Excitation by an antenna with a given radiation pattern.
AS Excitation by means of impressed radiating spherical modes.
AV Excitation by an impressed line current similar to the AI card, but
the endpoint of the current is electrically connected to a conducting
surface.
AW Excitation by an impressed mode on a waveguide port.

The different ways to realise a voltage source are summarised in Figures 14-1 and 14-2. The
~i .
impressed electric field strength is indicated by E

July 2011 FEKO User’s Manual

CONTROL CARDS 14-7

A1 card: Voltage source on a segment

Ei

A2 card: Voltage source on a node between segments

Ei

A3 card: TEM-frill on a segment

magnetic current loop

Figure 14-1: Possible ways to realise a voltage source on a wire segment

A4 card: Vertical pin approximation (dielectric substrate)

Ei

A7 card: Voltage gap on an edge

Ei
AE card: Voltage gap along edges

Ei

Figure 14-2: Possible ways to realise a voltage source in connection with triangles

July 2011 FEKO User’s Manual

CONTROL CARDS 14-8

More than one excitation is also allowed. Thus one may, for example, generate an elliptically
polarised plane wave by super-imposing two out of phase linearly polarised plane waves with
different amplitudes. It is also possible to feed an antenna with two different voltage sources.
For this purpose the parameters New source and Add to sources are available in each Ax card.
This parameter indicates whether the current excitation is additional (Add to sources) or not
(New source). When New source is selected, only the current excitation will be used and the
excitations prior to the current one will be erased.
For the excitations A1, A2, A3, A4 and A7 it is possible to select the feed element through the
label. Alternatively the position of the feed is specified in Cartesian coordinates. FEKO then
searches for a segment or an element at this position. This comparison of the position uses the
tolerance parameter Maximum identical distance (EG card (see section 13.13)).
The excitations are described in detail in the following sections.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-9

14.3 A0 card

This card realises excitation by a linearly polarised incident plane wave.

Parameters:

New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

Polarisation: This group sets the behaviour of the polarisation vector. If either of the rotating
options is selected, the Ellipticity must be specified below.

Rotation about axis: The rotation of the plane wave about the x, y and z axis.

Number of ϑ angles: If more than one direction of incidence is to be examined, then this parameter
indicates the number of incident angles in the ϑ direction. If this field is left
empty or set to 0, a value of 1 will be used.

Number of ϕ angles: If more than one direction of incidence is to be examined, then this parameter
indicates the number of incident angles in the ϕ direction. If this field is left
empty or set to 0, a value of 1 will be used.

Magnitude: ~ 0 of the incident field in

Magnitude of the field strength E V
.
m

Phase: ~ 0 of the incident field in degrees.

Phase of the field strength E

Initial ϑ value: Angle of incidence ϑ of the plane wave in degrees, see the figure in the card
above.

Initial ϕ value: Angle of incidence ϕ of the plane wave in degrees, see the figure in the card
above.

Polarisation angle η: See the figure in the card above. It is the angle — in a right-handed sense
~0 .
when viewing in the incident direction — from −ϑ̂ to E

July 2011 FEKO User’s Manual

CONTROL CARDS 14-10

Increment in ϑ: If more than one direction of incidence is to be examined, ϑ is incremented by

this value for each new angle of incidence.

Increment in ϕ: If more than one direction of incidence is to be examined, ϕ is incremented by

this value for each new angle of incidence.

Ellipticity: This field is only applicable when elliptical rotation is selected under Polarisa-
tion above and gives the ellipticity of the rotating polarisation. It must be larger
than 0 (linear polarisation) and smaller or equal to 1 (circular polarisation).

The direction of incidence β̂0 is specified by the incidence angles ϑ and ϕ. The polarisation angle
η (measured from the negative of the spherical coordinate system unit vector ϑ̂) and the field
~ 0 are defined as indicated in the figure in the card above.
strength vector E
The electric field strength of the incident field is then given by

~ i (~r ) = E~ 0 + jv( E~0 × β̂0 ) · e− j β~0~r

” —
E (14-1)

where v is the ellipticity. For linear polarisation, ellipticity is 0; for right hand rotating, ellipticity
is equal to the value in the Ellipticity field; for left hand rotating, ellipticity is the negative of the
value in the Ellipticity field. The incident magnetic field is given by
1
~ i (~r ) =
H ~i
β̂0 × E (14-2)
ZF
with Z F the wave impedance in the surrounding free space medium.
It should be noted that the incident power density (which is required, for example, for RCS
computations) is given by
~ 0 |2
1 |E
~i =
S (1 + v 2 ) . (14-3)
2 ZF

It is possible to vary the direction of incidence. This is particularly useful when e.g. determining
the monostatic radar cross section. The two parameters Initial ϑ value and Initial ϕ value indicate
the direction of the first wave. The direction of incidence is varied in the ϑ direction by the
increment Increment in ϑ and in the ϕ direction by Increment in ϕ. In each direction these two
angles are examined and a total number of incident waves equal to the product of these two
angles are examined.
If an A0 card with either Number of ϑ angles or Number of ϕ angles larger than 1 is read, all the
following control cards up to, but excluding, the next Ax, FR or EN card will be read into a buffer.
All these cards are then processed in a loop, over all the different angles of incidence. If e.g. the
monostatic radar cross section is to be calculated for ϑ=90◦ and 0◦ ≤ ϕ ≤ 180◦ , the following
command is used:
...
A0 0 1 181 1.0 0.0 90.0 0.0 0.0 0.0 0.1
FF 2
EN

In this demonstration file, the FF card is read into the buffer and processed 181 times. Through
the use of the parameter Fields calculated only in incident direction in the FF card the far field is
calculated in the direction of the incident wave.
If more than one direction of incidence is to be examined, the right hand side of the linear
equation system is changed, but the matrix remains unchanged. Thus it makes sense, by using

July 2011 FEKO User’s Manual

CONTROL CARDS 14-11

the CG card, to use Gauss elimination (default if a CG card is not used) which performs a LU-
decomposition of the matrix. When the direction of incidence is varied, then only the relatively
fast backward substitution has to be performed.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-12

14.4 A1 card

With this card a voltage source is placed on a segment.

Parameters:

New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

Select segment: When this item is selected, the Apply source to last segment with label field be-
comes active. Here one specifies the label of the segment to which the source
must be applied. If more than one segment has this label, the source is applied
to the last segment with this label. Alternatively, one may select the item Set
source position, then the feed segment is determined by specifying the Carte-
sian coordinates in the Segment centre group. These values are in m and are
scaled by the SF card if Modify all dimension related values is checked.

Reverse feed orientation: By default, the vector of the voltage lies in the direction from the start of the
segment to its end (the direction in which it was created). When this option
is checked, the vector of the voltage lies in the opposite direction. (This is
the direction of the current flow through the segment. The internal EMF —
electromagnetic force — of the impressed voltage source is in the opposite
direction.)

Magnitude of source: Magnitude of the voltage source in V.

Phase of source: Phase of the voltage source in degrees.

S-parameter impedance: The port impedance of this excitation is used in connection with S-parameter
and realised gain calculations . If this field is empty or 0, the value specified at
the SP card is used. (This value is only used if the S-parameters are requested
with an SP card.)

July 2011 FEKO User’s Manual

CONTROL CARDS 14-13

14.5 A2 card

With this card a voltage source is placed at a node between two segments or between a segment
and a triangle, ground plane or polygonal plate. It is mostly used to feed wires attached to plates,
etc.

Parameters:

New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

Select segment: When this item is selected, then the Source label field becomes active. Here
one specifies the label of the segment which shall be fed (either start or end
point as determined by the corresponding check box). The excitation has to
be located at a node, either between two segments, or between a segment and
a triangle, ground plane or polygonal plate. Only one segment with this label
should be declared. If there is more than one segment with this label then only
one node will be fed.

Set source position: When this check box is activated, then the feed node is determined by specify-
ing its Cartesian coordinates in the Coordinates of node group. These values
are in m and may be scaled by the SF card. The source orientation is always
like the basis function defined over this node (see the discussion below under
Use default feed direction.

Source at start of segment: This option is only available when selecting the feed segment by label. If
set, it indicates that the feed location is at the start of the wire segment with a
matching label.

Source at end of segment: This option is only available when selecting the feed segment by label. If
set, it indicates that the feed location is at the end of the wire segment with a
matching label.

Use default feed direction: This option is only available when selecting the feed segment by label. If
set, it indicates that the positive feed direction is like the basis function setup
in FEKO. For wire / surface junctions (UTD plates, BO ground, or meshed
triangle surfaces), this direction is away from the wire onto the surface. For

July 2011 FEKO User’s Manual

CONTROL CARDS 14-14

wire connections between two segments, this direction is from the segment
with the lower index to the segment with the higher index.

Positive feed direction like wire segment orientation: This option is only available when selecting the
feed segment by label. If set, then the positive feed direction is like the orien-
tation of the wire segment with the specified label.

Negative feed direction like wire segment orientation: This option is only available when selecting the
feed segment by label. If set, then the positive feed direction is opposite to the
orientation of the wire segment with the specified label.

Magnitude of source: Magnitude of the voltage source in V.

Phase of source: Phase of the voltage source in degrees.

S-parameter impedance: The port impedance of this excitation is used in connection with S-parameter
and realised gain calculations . If this field is empty or 0, the value specified at
the SP card is used. (This value is only used if the S-parameters are requested
with an SP card.)

There may not be more than two segments connected to the node for nodes between segments.
For nodes between a segment and a triangle, UTD plate or an infinite plane only one segment is
allowed to connect at the node. Note that the vector direction of the feed is the direction of the
current flow through the node. The internal EMF — electromagnetic force — of the impressed
voltage is in the opposite direction.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-15

14.6 A3 card

This card realises excitation by a magnetic ring current (TEM-frill) on a segment. This will give
an accurate model of a coaxial feed, but requires both the inner and outer radii.

Parameters:

New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

Select segment: When this item is selected, the Apply source to last segment with label field
becomes active. Here one specifies the label of the segment on which the
TEM-frill is placed. If more than one segment has this label, the source is
applied to the last segment with this label. Alternatively, one may select the
item Set source position, then the feed segment is determined by specifying
the Cartesian coordinates in the Segment centre group. The position values
are in m and are scaled by the SF card if Modify all dimension related values is
checked.

Reverse feed orientation: By default, the vector of the excitation points in the direction of the segment
- from the start point to the end point as the segment was created. When this
option is checked, the orientation of the excitation is reversed.

Magnitude of source: Magnitude of the voltage source in V.

Phase of source: Phase of the voltage source in degrees.

Inner conductor radius: Radius of the inner conductor of the coaxial feed.

Outer conductor radius: Radius of the outer conductor of the coaxial feed.

S-parameter impedance: The port impedance if this excitation is used in an S-parameter calculation.
If this field is empty or 0, the value specified at the SP card is used. (This value
is ignored if no SP card is used.)

The excitation is not, as in the A2 card, an impressed electric field strength, but is a magnetic
ring current.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-16

As a rule of thumb, the radius of the inner conductor must be the same as the radius of the
segment and that the outer radius should be 2 to 3 times the size of the inner. If an impedance Z
is desired, then the following relation can be used

Outer conductor radius

 
Z ≈ 60 Ω · ln (14-4)
Inner conductor radius

to determine the outer radius. For Z = 50 Ω the outer conductor radius should be equal to 2.3
times the inner conductor radius.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-17

14.7 A4 card

This card creates a coaxial attachment feed approximation for use in connection with the Green’s
function for planar substrates with a metallic ground plane (GF card (see section 14.39)).

Parameters:
New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

Select triangle: When this item is selected, the Source label field becomes active. Here one
specifies the label of the triangle to excite. The feed point is at the centroid
of the triangle (see Figure 14-3). If there are more than one triangle with this
label, the excitation is placed on the one with the highest element number.
Alternatively, the user may select the item Set source position, and specify
the Cartesian coordinates of the feed point in the Coordinates of node, group.
(These values are in m and optionally scaled by the SF card.) FEKO will excite
the triangle whose centroid is closest to the specified coordinate.

Transform impedance to ground plane: If unchecked the impedance is computed directly at the excita-
tion point. If checked, an inductive approximation of the feed pin is used and
the impedance is transformed to the ground plane.

Magnitude of source: Absolute value of the excitation current |~I0 | in A. The positive current direction
is the positive z direction.

Phase of source: Phase of the current ~I0 in degrees.

Radius of the connection pin: The radius of the coaxial probe feed pin in m. This value is optionally
scaled by the SF card.

S-parameter impedance: The port impedance of this excitation is used in connection with S-parameter
and realised gain calculations . If this field is empty or 0, the value specified at
the SP card is used. (This value is only used if the S-parameters are requested
with an SP card.)

An example of the excitation is shown in Figure 14-3. A typical application of the A4 card is given
in example_30b.pre (Scripting examples). It is of course possible to discretise the vertical pin
(into segments) and feed one of the segments with a voltage source (A1 card). The advantage of
the A4 card is that there are no vertical currents, which results in substantially simpler Green’s
functions and a significant reduction in computing time.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-18

Figure 14-3: Excitation of a patch antenna with a vertical pin

July 2011 FEKO User’s Manual

CONTROL CARDS 14-19

14.8 A5 card

This card specifies excitation by an electric Hertzian dipole.

Parameters:

New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

Magnitude of I dl: Absolute value of the complex amplitude I · l in Am.

Phase of I dl: Phase of the complex amplitude I · l in degrees.

x, y, z position: Coordinates of the position of the dipole in m. These values are optionally
scaled by the SF card.

ϑ angle: Orientation of the dipole in space — ϑ (in degrees) is the angle between the
dipole and the z axis.

ϕ angle: Orientation of the dipole in space — ϕ (in degrees) is the angle between the
projection of the dipole onto the plane z = 0 and the x axis.

The dipole moment of the electric dipole is given by

Il
p= (14-5)
jω

The power radiated by the dipole in a free space environment is given by

β02 · Z F 0 · |I l|2 ω2 µ20 |I l|2

P= = (14-6)
12π 12πZ F 0

FEKO, however, considers the properties of the medium in which the dipole is located, as well
as the coupling of the dipole with surrounding structures or other sources (for example other
Hertzian dipoles in an array antenna), when calculating the power radiated by the Hertzian
dipole.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-20

14.9 A6 card

This card specifies excitation by an elementary magnetic Hertzian dipole.

Parameters:
New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

Electric ring current: Use the model of an electric ring current for the magnetic dipole (the moment
m = µIA where I is the loop current and A the enclosed surface).
Im l
Magnetic line current: Use the model of a magnetic line current (the moment m = jω
where I m is
magnetic line current and l the length of the dipole).

Magnitude of . . .: The magnitude of the dipole. For the electric ring current it is I · A in Am2 . For
the magnetic line current it is I m · l in Vm.

Phase: Phase of the complex amplitude in degrees.

x, y, z position: Coordinates of the position of the dipole in m. These values are optionally
scaled by the SF card.

ϑ angle: The angle between the dipole and the z axis in degrees.

ϕ angle: The angle between the projection of the dipole onto the plane z = 0 and the x
axis in degrees.

The power radiated by the dipole in a free space environment is given by

β04 · Z F 0 · |I A|2
P= (14-7)
12π
FEKO, however, considers the properties of the medium in which the dipole is located as well
as the coupling of the dipole with surrounding structures or other sources (for example other
magnetic dipoles in an aperture approximation — see the AP card) when calculating the power
radiated by the Hertzian dipole.
Even though the two formulations, electric ring current and magnetic dipole, result in the same
near and far fields, if the dipole moment m is the same, the radiated potentials are different. The
electric ring current model gives rise to a magnetic vector potential A~, while the magnetic dipole
model results in an electric vector potential F~ as well as a magnetic scalar potential Ψ.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-21

14.10 A7 card

This card is used to specify a voltage source on an edge between two triangles or at a connection
edge between a single triangle and a PEC ground plane or UTD plate. The AE card is substantially
simpler to use and should be used for all new models. (The A7 card is supported only for
compatibility with FEKO input files that were created before the AE card became available.)

Parameters:

New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

Select segment: If this item is selected, a triangle with label specified in the Element label field
is searched for. The excitation is placed on the edge that lies opposite to the
first corner of the triangle. Once again the label must be unambiguous, i.e.
if possible only one triangle must have this label. If there is more than one
triangle with this label then only one will be fed. Alternatively, when selecting
the item, Set source position, the feed edge is determined by specifying its
Cartesian coordinates in the Coordinates of edge centre. These coordinates are
in m and optionally scaled by the SF card. The edge must lie between two
triangles or between a triangle and a ground plane or UTD plate.

Magnitude of source: Absolute value of the voltage source in V.

Phase of source: Phase of the voltage source in degrees.

If two triangles are connected to the edge, the basis function between these triangles is excited.
The vector direction of the voltage source lies in the same direction as the basis function associ-
ated with this edge. (This is the direction of the current flow through the edge. The internal EMF
(electromagnetic force) of the impressed voltage source is in the opposite direction.)
In certain special cases there may be only one triangle connected to the edge. If the edge lies
in the plane of a polygonal plate or a PEC ground plane (specified with a GF or BO card), the
excitation is placed on the appropriate basis function connecting the triangle to the plate/plane.
The positive feed direction is then towards the edge.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-22

14.11 AB card

This card is used to create a FEM modal excitation. The modal port will be excited with the
fundamental mode of the associated, infinitely long guided wave structure of the modal port. If
no excitation is defined, then the modal port will act as a passive port (sink) for fields incident
on the port.

Parameters:

New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

Name of the modal port: Label of the modal port.

Magnitude of source: Magnitude of the impressed node in V/m.

Phase (degrees): Phase of the impressed node in degrees.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-23

14.12 AC card

This card inputs data from a *.rsd file containing the geometry of a transmission line or PCB
structure and the current distribution along this line or on the PCB for one or more frequencies.
Such a *.rsd file is created, for example, by the transmission line simulation programs CableMod
or CRIPTE or by the PCB code PCBMod or by a current export with the OS card in FEKO. The
excitation is due to the electromagnetic fields radiated by these line currents (the CM card allows
the treatment of electromagnetic fields coupling into lines).

Parameters:

New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

No action: No execution, do not read the *.rsd file — this option is used to specify the
end of the frequency loop, see below.

Model transmission line with Hertzian dipoles: The line geometry, frequency and currents are read from
the *.rsd file, and the line is modelled with an array of Hertzian dipoles (see
the A5 card). The number of dipoles per line segment is specified with the pa-
rameter Number of dipoles per transmission line. Note that this model is only
valid if the line segments do not make electrical contact with any conducting
surface. (All the segments in the *.rsd file must be of the type intern and not
loaded.)

Model line with impressed line currents: The line geometry, frequency and currents are read from the
*.rsd file, and the line is modelled with a continuous current distribution us-
ing one AI card per line segment. (The AI cards are created automatically by
PREFEKO when importing the *.rsd file.) If a line segment has a loaded end-
point it is automatically modelled by an AV card to allow the electrical contact.
The radius of the impressed current element can be set in the parameter Ra-
dius of impressed current. This parameter is optional and is passed on to the
AI and AV cards. (See the description in these cards.) If the parameter is zero
or empty a current filament approximation is used.

Source translation (directions): When importing transmission line currents from CableMod or CRIPTE
or PCB currents from PCBMod, then the coordinate system of these programmes
as represented in the *.rsd file is used and the source is imported at this po-
sition in FEKO. Here an offset can be specified which translates the source in
x, y, or z direction. Standard FEKO units are used for these offsets (i.e. me-
tres, but scaled accordingly if a factor is set at the SF card). Note that the
units as specified in the *.rsd file are not applicable here for the translation
parameters (only to the import of the data).

July 2011 FEKO User’s Manual

CONTROL CARDS 14-24

Rotation about axes: Like the translation described above, an imported source can here be rotated
and thus positioned arbitrarily. The rotation angles are in degree, and the same
definition as also used at the AR (see section 14.19) or TG (see section 13.44)
cards applies.

File name: The name of the *.rsd file.

Use adaptive frequency sampling: Only read the minimum and maximum frequency from the *.rsd
file and obtain a continuous solution in this frequency band using adaptive
frequency sampling. If this option is used, only one AC card is permitted in the
*.pre file (and no FR cards).
Maximum number of discrete frequency points: When using adaptive frequency sampling, the maxi-
mum number of sample points can be specified here. See also the discussion
on adaptive frequency sampling at the FR card.

Minimum frequency stepping: When using adaptive frequency sampling it could be necessary to spec-
ify the minimum allowable frequency between samples. See also the discussion
on adaptive frequency sampling at the FR card.

If the imported *.rsd file contains currents for several frequencies, the option New source must
be chosen as the AC card then results in a frequency loop and currents with different frequencies
cannot be superimposed. (If it is not chosen, PREFEKO will give an error). The frequency is
defined in the *.rsd file, thus the preceding FR cards are ignored when processing an AC card.
All commands following the AC card in the FEKO input file (for example FF, FE, OS, GF, BO, . . .)
are processed within a frequency loop through all the frequencies in the *.rsd file. The loop
is terminated by any of the following three cards (these cards are not included in the loop they
terminate).

• AC (importing a new *.rsd file, or using the No action option)

• FR (manually setting a new frequency)

• EN (end of the FEKO input file)

For example, if a *.rsd file must be read and the near field calculated for each frequency, the
input file may look as follows
AC ... ** Read the *.rsd file
FE ... ** Calculate the near field
EN ** End

However, if one wants to analyse, for example, a metal plate, which is excited first by an im-
pressed line current and then also by a plane wave (in each case the near fields and the currents
on the plate must be written to the output file), the input file would be
** Excitation by a line current
AC ... ** Read the *.rsd file
FE ... ** Calculate the near field
OS ... ** Output the currents
** Excitation by a plane wave
FR ... ** Set the frequency and terminate
AC loop
A0 ... ** Specify plane wave excitation
FE ... ** Calculate the near field
OS ... ** Output the currents
** End
EN

July 2011 FEKO User’s Manual

CONTROL CARDS 14-25

14.13 AE card

This card specifies an excitation at an edge between triangular surface elements similar to the A7
card, but the AE card has the advantage that the location of the feed point and the positive feed
direction are substantially easier to specify. In addition it is possible to specify a feed edge which
contains a number of triangle edges as shown in Figure 14-4.

Negative side Positive side

Ei
Figure 14-4: Example of the use of the AE card.

Parameters:

New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

Excite edge: This specifies how the edge is determined:

• between regions with multiple labels: The excitation is placed on the edge
between the regions with labels specified in Negative side and Positive
side. Use the field Maximum number of labels on a side to increase the
number of rows available in the table. The positive source direction is
from the Negative side towards the Positive side side (see Figure 14-4).
• connected to ground/UTD: Excite the edges of metallic triangles with the
labels specified in the Negative side or Positive side. The edges of these
triangles are connected to UTD surfaces or to a PEC ground plane (as
specified with a BO or GF card).
• of microstrip between two points: This is a special microstrip line feed.
The excitation is placed on all edges on a line between points (previously
specified with DP cards). The points are specified in the Start point of
edge and End point of edge dialogs. A GF card with a conducting ground
must be present.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-26

Meshed surface represents positive feed side: By default, the feed direction is such that the meshed sur-
face represents the negative feed side. The vector direction of the current is
then towards the UTD or ground. When this option is checked, the feed orien-
tation is reversed.

Magnitude of source: Magnitude value of the voltage source in V.

Phase of source: Phase of the voltage source in degrees.

S-parameter impedance: The reference impedance to use for this excitation it is included in an S-
parameter calculation requested by the SP card. If this field is empty or zero,
the impedance specified at the SP card is used. (This value is only used if an
SP card is applied to this source.)

The positive source direction as used above is in the direction of the current flow through the
edge. The internal EMF (electromagnetic force) of the impressed voltage source is in the opposite
direction.
It should be noted that the edge between the surfaces with labels Negative side and Positive side
does not have to be straight. One may, for example, excite two half cylinders against each other.
If an impedance must also be applied to the edge, the AE card can be combined with the LE card.
For the case where the item, between regions with multiple labels, is selected, more than two
surfaces may be connected to a feed edge. Examples for this are given in Figure 14-5 (one
surface fed against three others), Figure 14-6 (two surfaces fed against two other surfaces), and
in Figure 14-7 (permitted feed scenarios as indicated in the caption of this figure).

Figure 14-5: Example of a feed edge where more than one surface is connected on one side of the feed

Figure 14-6: Example of a feed edge where more than one surface is connected on both sides (negative
and positive) of the feed

July 2011 FEKO User’s Manual

CONTROL CARDS 14-27

Figure 14-7: Example of a feed edge of three surfaces with different labels, where for instance label 2
could be fed against labels 1 and 3, but also label 1 could be fed against 2 and 3, or 3 could
be fed against 1 and 2.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-28

14.14 AF card

With this card an uniform electric current filament is impressed between two arbitrary points
inside of the FEM region (i.e. it does not have to coincide with tetrahedral edges). This can be
used to excite for instance a patch antenna. See also the example DRA_finite in the Examples
Guide.

Parameters:

New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

Amplitude of source: Amplitude in A of the impressed line current.

Phase of source: Phase of the current in degrees.

x, y, z coordinate: Coordinates of the start and end points in m. (Note that all the coordinate
values are optionally scaled by the SF card.)

S-parameter impedance (Ohm): The port impedance of this excitation is used in connection with S-
parameter and realised gain calculations . If this field is empty or 0, the value
specified at the SP card is used. (This value is only used if the S-parameters
are requested with an SP card.)

The following restrictions apply when using the impressed current elements of AF type inside a
FEM region:

• The electric current source would usually be connected to metallic surfaces at its terminals,
but this is neither enforced nor checked in FEKO. If a physical connection to a metallic
structure is required, then the user should ensure that the feed terminals are also attached
in the discretised model.

• Input impedance is computed for this source from the line integral over the electric field
solution between the terminals of the source. The length of the impressed current should
therefore be small compared to the shortest wavelength in the band of interest to render a
reasonable accuracy.

• An intrinsic limitation of this model is that no radius is taken into account, therefore the
field is singular in the vicinity of the filament, and this affects the accuracy of the computed
input impedance of the source.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-29

14.15 AI card

With this card an impressed current source is specified. The current varies linearly between the
values at the start and end points, see Figure 14-8.

I2

I1 r2

r1

Figure 14-8: Impressed line current with a linear current distribution

Parameters:

New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

Amplitude: Amplitude |I1 | in A of the current at the Start point, ~r1 , and End point, ~r2 .

Phase: Phase of the current at the start and end points in degrees.

x, y, z coordinate: Coordinates of the start and end points in m. (Note that all the coordinate
values are optionally scaled by the SF card.)

Radius of impressed current: This parameter is optional. If specified, and different from zero, this
value gives a finite wire radius for the impressed current element. FEKO then
assumes that the current is uniformly distributed on the wire surface and uses
the exact wire integral. If this parameter is not specified, the current filament
approximation is used. (This value is optionally scaled by the SF card.)

July 2011 FEKO User’s Manual

CONTROL CARDS 14-30

The following restrictions apply when using the impressed current elements:

• It is not possible to attach the impressed current to a wire segment in the FEKO model. (If
the impressed current is making electrical contact with a triangular surface current element,
the AV card should be used.)

• When modelling dielectric bodies with the surface equivalence method, the current ele-
ment must be in the free space medium, i.e. outside the dielectric bodies. (The material
parameters of this medium can, however, be set with the EG and/or GF cards).

• When used with the spherical Green’s function, the current element must be outside the
dielectric spheres.

• The current segments may be joined with each other and with the AV card to form long
paths and/or closed loops. The point charges which arise when the current does not go to
zero at an end point or when there is a current discontinuity at a connection point, are not
taken into consideration. This is required to model, for example, the case where radiating
lines are terminated in a non-radiating structure. If these charges must be considered
explicitly, the line current should be modelled by a row of Hertzian dipoles (see the A5
card). Note, however, that the constant line charge along the current segment is correctly
taken into account.

• If several of these current elements are used, the total radiated power (required to cal-
culate, for example, the far field gain/directivity) can only be calculated accurately if the
mutual coupling between segments is taken into account. Due to neglecting the point
charges at the ends of the segments, the coupling cannot be determined accurately. If exact
values of the radiated power are required, it should be determined by integrating the far
field (see the FF card). It should be noted that, for example, the computed near and far
fields (the actual field strength values), the induced currents, coupling factors and losses
are computed correctly.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-31

14.16 AK card

With this card a voltage source is applied to a radiating cable

Parameters:

New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

Ground: When this option is selected the cable connector pin is connected to ground
(reference pin), denoted by Pin 0.

Connector label: The connector label uniquely identifies the connection of a cable path section.

Pin number: The pin number identifies the conductor associated with the cable path section
as defined by the Connector label in the CS card (see section 14.30).

Magnitude of source: Magnitude of the voltage source in V.

Phase of source: Phase of the voltage source in degrees.

S-parameter impedance: The port impedance of this excitation is used in connection with S-parameter
and realised gain calculations . If this field is empty or 0, the value specified at
the SP card is used. (This value is only used if the S-parameters are requested
with an SP card.)

Adding a voltage source to a cable requires knowledge of the conductor to cable connector pin
relation. The following rules apply:

• Pin 0 always corresponds to the geometry or meshed model

• Use increasing pin numbers starting at the innermost conductor of a simple cable type and
counting towards the outside

• Use increasing pin numbers in the order of cross-section definition inside of a bundle

Note that loading (excitation) of a cable path is allowed only between connector pins within the
same sub-circuit and its reference pin (i.e. loading between elements are not allowed over a
shield of a cable). Pin numbering used for cable modelling can best be described by examples.

Example 1: An insulated single conductor above ground consisting of the following:

July 2011 FEKO User’s Manual

 14
15
CONTROL CARDS 14-32
16 17 Core: Pin 1 Core: Pin 1
Bundle2

11
10
Ground: Pin 0 Ground: Pin 0

12 Core: Pin 1
Cores: Sub-circuit 1: Single core (Pin 1) with geometry or meshed model as reference (Pin 0).

1 2 n Core: Pin 1
Example 2: Coaxial cable above ground consisting of the following: Ground: Pin 0

Ground: Pin 01:

Sub-circuit Outside of shield (Pin 2) with geometry or meshed model as reference (Pin 0).
Ground: Pin 0 Core: Pin 1
Sub-circuit 2: Coaxial core (Pin 1) with the inside of shield (Pin 2) as reference.

3
Core: Pin 1
Ground: Pin 0
14 Shield: Pin 2
15
Ground: Pin 0

6 17
Example 3: Ribbon cable with n cores consisting of the following:
Bundle2

11 Sub-circuit 1: Cores 1..n (excluding Pin i) with Pin i as reference.

0

12 Cores:

1 2 n

Ground: Pin 0
Example 4: Complex bundle (mixed) cable consisting of the following:

Bundle1 (shielded): Single1, Single2, Coax1, Ribbon1, Coax2, Bundle2 and Bundle3

Ribbon1: 3 cores

Bundle2 (not shielded): Single3, Single4 and Single5

Bundle3 (shielded): Single6 and Coax3

For example, loading (excitation) for sub-ciruit 2 is only allowed between 1, 2, the outside of the
shield of coaxial cable 4, 5, 6, 7, the outside of the shield of coaxial cable 9, 10, 11, 12 and the
outside of the shield of the coaxial cable 16 and the inside of the shield of the bundle 17.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-33
Cores:

1 2 n

Mixed cable: Bundle1

Single1

Coax1 1 Bundle3

3 4
14
Single2 13 15

2 16 17
7 Bundle2
Ribbon1
6 11
5 Coax2
10

12
8 9

Ground: Pin 0

Table 14-2: Allowed loading between connector pins within the same sub-circuit and its reference pin

Sub-circuit Pin numbers Reference pin

number
1 17outsid e 0 (geometry/meshed
model)
2 1, 2, 4outsid e , 5, 6, 7, 9outside , 10, 11, 12 , 17inside
16outsid e
3 3 4inside
4 8 9inside
5 13, 15outsid e 16inside
6 14 15inside

July 2011 FEKO User’s Manual

CONTROL CARDS 14-34

14.17 AN card

With this card a voltage source can be added to any port of a general non-radiating network that
does not have a connection to geometry.

Parameters:

New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

Network name: The network or transmission line name, with the network port number, uniquely
identifies the connection terminals.

Network port number: The network port number, with the network/transmission line name, uniquely
identifies the connection terminals.

Magnitude of source: Magnitude of the voltage source in V.

Phase of source: Phase of the voltage source in degrees.

S-parameter impedance: The port impedance of this excitation is used in connection with S-parameter
and realised gain calculations . If this field is empty or 0, the value specified at
the SP card is used. (This value is only used if the S-parameters are requested
with an SP card.)

Adding of a voltage source to an internal network port (i.e. not connected to geometry) is
completely defined using an AN-card. The connection is specified by the combination of the
network name and port number that will be excited.
Note that for excitation of network ports connected to geometry, the A1 (voltage source on a
segment), A2 (voltage source on a node), A3 (magnetic frill excitation) and AE (edge excitation)
sources are supported.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-35

14.18 AP card

With this card a planar, cylindrical or spherical aperture of measured or calculated field values
is converted into an equivalent array of electric and magnetic dipoles. The card is processed by
PREFEKO and replaced by A5 and A6 cards in the *.fek file.

Parameters:
New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

. . . field data . . .: In this group one selects between Load field data from *.efe/*.hfe file (i.e. from
*.efe or *.hfe files previously calculated with FEKO), Load field data from
ASCII text file (the format of these files are documented below) or The field
data follows in the *.pre input file (described below).

. . . aperture: Here one may select a planar, cylindrical or spherical aperture. For a planar
aperture one may elect to use only electric or magnetic fields (this radiates in
both directions, see comment below).

Also sample along edges: This item is used to determine if dipoles should lie on the edges of the aper-
ture or not (see Figures 14-9 to 14-12). When checked the outer dipoles lie
on the edges, when unchecked the dipoles lie half an increment away from the
edges. Dipoles should not lie on the edges of two apertures that have a com-
mon side, otherwise two dipoles may have the same location and polarisation.
If this is the case the power calculation in FEKO will fail.

S1, S2 and S3: These define the orientation of the aperture and should be clear from the figure
on the dialog. For a planar aperture it defines the position of the origin and the
direction of the û2 and û3 directions respectively. (The field data is assumed to
vary first along the û2 direction.) For cylindrical and spherical apertures they
define the origin and the direction of the local z and x axes respectively. All
angles are relative to these axes (see Figures 14-11 and 14-12).

July 2011 FEKO User’s Manual

CONTROL CARDS 14-36

E-field file name: The input file name from which the electric field data must be read. These may
either be a *.efe or a text file (with units V/m).

H-field file name: The input file name from which the magnetic field data must be read. These
may either be a *.efe or a text file (with units A/m).

Start from point number: The number of the first field point to be used for the aperture. If set to 1,
field values are read from the start of the file, for larger values the first point
number-1 values (*.efe and *.hfe files) or lines (text files) are ignored. This
may be used, for example, if the data file contains the field values for more
than one aperture. This corresponds to the line number if all non-data lines
are stripped from the file.
The Start from point number field is not used if the field data is obtained from
the *.pre input file.

Number of points along . . .: These two fields specify the number of field points along each of the two
axes of the aperture.

Amplitude scale factor: A constant by which the amplitudes of all the dipoles in the aperture are scaled.

Phase of aperture: A constant phase added to all dipoles in the aperture.

The aperture is based on the equivalence principle. This states that the sources and scatterers
inside a given volume can be removed, and modelled by placing the equivalent currents J~s = n̂× H ~
and M ~ s = −n̂ × E
~ on the enclosing surface. The vector n̂ is a unit vector, normal to the surface,
and points towards the exterior region. The fields in this region are the same as the original
fields, while those in interior region are zero.
Field values are read from the data files (with a possible offset specified with Start from point
number) or the *.pre input file and converted to equivalent electric (magnetic fields) and mag-
netic (electric field) dipoles at these points. Note that all angles are read from the data, but no
distance values. Thus for planar apertures the positions are calculated entirely from the specified
points (S1, S2 and S3). For cylindrical apertures S1 and S2 specify the extents of the aperture
along the local ẑ direction and S1–S3 specifies the direction of the x axis as well as the radius
of the cylinder. The points are placed at the ϕ-values listed together with the field data. For
spherical apertures, S1–S2 specifies the direction of the z axis and S1–S3 the x axis. S2 and S3
must lie on the same radius which is also the radius of the field points. In this case both ϑ and ϕ
are read with the data.
Figures 14-9 and 14-10 show the application of the equivalence principle to a planar aperture.
In both figures there are the same number of field points along the two orthogonal directions.
When the Also sample along edges item is checked, the first point lies at S1 with the following
points in the direction of S2 as shown by the indices in Figure 14-9. When it is unselected, the
pattern is as shown in Figure 14-10. The normal vector is calculated from n̂ = û2 × û3 with û2
and û3 as defined in the figure.
Figure 14-11 shows the dipole locations for a cylindrical aperture created from a data file con-
taining field values for ϕ from 20◦ to 80◦ in 10◦ increments and 5 values in the z direction. When
the Also sample along edges item is checked, the samples extend up to the edges of the aperture,
the points and the effective aperture are shown in figure (a). When it is unchecked, samples do
not lie on the edges as shown in figure (b). Note that, when using identical input data as for the

July 2011 FEKO User’s Manual

CONTROL CARDS 14-37

case when the item is checked, the z positions of the samples changed, while in the ϕ direction
the size of the effective aperture is increased by 5◦ on both sides.
Figure 14-12 shows the dipole locations for a spherical aperture created from field values for ϑ
from 40◦ to 80◦ with 10◦ increments and ϕ from 20◦ to 80◦ also with 10◦ increments. In this
case the aperture increases in size in both directions when the Also sample along edges item is
checked.
For planar apertures, the data must vary first along the û2 direction. For cylindrical and spher-
ical apertures PREFEKO will determine which coordinate is incremented first and write out the

(N3-1)*N2+1 N3*N2
S3

3*N2+1

û3 2*N2+1 2*N2+1 3*N2

N2+1 N2+2 2*N2

1 2 3 4 5 6 N2-1 N2
S1 û2 S2

Figure 14-9: Location of the equivalent dipoles on a planar aperture where the Also sample along edges
item is checked.

S3 (N3-1)*N2+1 N3*N2

3*N2+1

û3 2*N2+1 2*N2+1 3*N2

N2+1 N2+2 2*N2

1 2 3 4 5 6 N2-1 N2

S1 û2 S2

Figure 14-10: Location of the equivalent dipoles on a planar aperture where the Also sample along edges
item is unchecked.

Z Z
S2 S2

S1 S1

S3 S3
j j
X X
(a) (b)

Figure 14-11: Location of the equivalent dipoles on a cylindrical aperture.

(a) Also sample along edges checked (b) Also sample along edges unchecked.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-38

Z Z
S2 S2
J J

S1 S1

S3 S3
j j
X X
(a) (b)

Figure 14-12: Location of the equivalent dipoles on a spherical aperture.

(a) Also sample along edges checked (b) Also sample along edges unchecked.

dipoles accordingly.
The dipole amplitude is the product of the surface current and the incremental area between
samples. In addition, the amplitude of the dipoles on the sides (when the Also sample along
edges item is checked) are reduced by a factor of 2 and those on the corners by a factor of 4 so
that the effective aperture of integration has the same size as the specified aperture.
A fully closed surface can be created by specifying 6 planar apertures or a spherical one. The
surface equivalence principle can be applied to this surface by reading both electric and magnetic
fields for each plane. (For planar apertures the user should specify 6 AP cards, each using both
electric and magnetic fields. If separate cards are used for the electric and magnetic fields the
radiated power is not calculated correctly.) The normal vector must point to the exterior region,
normally this is outward. (For planar apertures created form *.efe and *.hfe files, the sample
order determines the directions of û2 and û3 which, in turn, determines the normal vector n̂ =
û2 × û3 . If this is pointing into the cube, an additional 180◦ phase shift is obtained by setting
Phase of aperture (degrees) to 180. This changes the sign of the field radiated by the aperture
which, when interacting with the remaining sources, will result in the correct total fields in the
desired region.) All surfaces and scatterers inside the body must be removed and those outside
retained.
For planar apertures (for example, the opening of a horn antenna), one may use the mirror
principle if the field at the edges can be neglected. This results in a duplication of the magnetic
current and cancellation of the electric current. Thus it is sufficient to read only the electric
fields and scale by the factor Amplitude scale factor=2. In this case any sources or structures
in the region towards which the normal is pointing, should also be subjected to the mirroring
(i.e. the structures should be electrically mirrored by using the SY card). Further, it should be
remembered that the fields will only be correct in the direction that the normal vector points
to. The symmetric fields in the other half-space will not be equal to the fields of the original
problem. Note that FEKO takes this into account and divides the total radiated power by two
when calculating the power radiated by a planar aperture containing only electric or magnetic
fields.
When the data is read from an ASCII format text file, each line in the file represents one point and
the values are space delimited. For planar apertures, each line (point definition) must contain

July 2011 FEKO User’s Manual

CONTROL CARDS 14-39

the following four parameters: the absolute value and phase (in degrees) of the field component
in the û2 direction followed by the absolute value and phase in the û3 direction (see the example
below). It is required that the data is in a format in which the position first increments along the
û2 direction. For cylindrical apertures each line (point definition) must contain the following five
parameters: the angle ϕ (in degrees), then the absolute value and phase of the ϕ̂ component,
and then the absolute value and phase of the ẑ component. For spherical apertures it must have
six parameters: the angles ϑ and ϕ followed by the absolute value and phase of the ϑ̂ and ϕ̂
components.
When the data is read from the *.pre input file, the data must be in the normal column based
input format with the result that FOR loops etc. may be used. The four field components are the
same as for the text data, and must be entered in columns 10 characters wide from columns 51 to
90. The angle ϕ must be in columns 40 to 49 and ϑ in columns 30 to 39 when they are required.
If both electric and magnetic fields are required, all the electric fields are given first, followed by
the same number of magnetic fields.
Example of AP card usage:
As an example, consider an open ended X-band waveguide radiating through a hole in a large
ground plane, as shown in Figure 14-13. Away from the aperture the plane z = 0 is perfectly
conducting, i.e. the tangential electric field is zero, while the magnetic field is not — thus we will
use electric symmetry.

z
y
P3
û3
P1
û2
P2 x
Infini
te gr
ound
plane
on th
e pla
ne z=
0

Figure 14-13: Example of an open waveguide as an implementation of the AP card

For this example the field is considered to be purely y directed (i.e. it has only a ŷ, or û3 , com-
ponent). The field is assumed to be constant in the y direction and to have a cosine distribution
in the x direction (i.e. the û2 axis).
With Number of points along û2 =5 and Number of points along û3 =3 — in practice more points
may be required — the data file will be as follows:

July 2011 FEKO User’s Manual

CONTROL CARDS 14-40

0.0 0.0 0.0 0.0

0.0 0.0 0.707 0.0
0.0 0.0 1.0 0.0
0.0 0.0 0.707 0.0
0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0
0.0 0.0 0.707 0.0
0.0 0.0 1.0 0.0
0.0 0.0 0.707 0.0
0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0
0.0 0.0 0.707 0.0
0.0 0.0 1.0 0.0
0.0 0.0 0.707 0.0
0.0 0.0 0.0 0.0

The zero values will not result in any dipoles, but they must be in the data file to allow correct
indexing. The *.pre file will contain the following section:
...

** Only electric fields --- use electric symmetry in the z=0 plane
SY 1 0 0 2

** Define the corner points of the aperture

#wx = 0.02286
#wy = 0.01016
DP P1 -#wx/2 -#wy/2 0
DP P2 #wx/2 -#wy/2 0
DP P3 -#wx/2 #wy/2 0

** The geometry ends after the corner nodes have been defined
EG 1 0 0 0 0

** Specify the frequency

FR 1 0 9.375e9

** Specify the AP card as a new source

** The amplitude factor of 2.0 is due to the use of the equivalence principle
AP 0 3 P1 P2 P3 1 5 3 2.0 0.0 "Guide.dat"

...

This will generate nine x directed magnetic dipoles of the correct magnitude in the *.fek file.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-41

14.19 AR card

With this card the radiation pattern of an antenna is used as an impressed source. The pattern is
read from a data file or defined in the *.pre file below the AR card.
This card has a variety of uses, for example, importing measured radiation patterns, synthesising
arrays from the individual patterns of the elements, realising radiation only within certain sectors,
etc. In the MoM/UTD hybrid it is possible to simulate, for example, the antenna on its own and
to save the far field in a *.ffe file. This field is then imported and used as source in the UTD
part which may greatly speed up the ray tracing computation as there is now only one source
point.

Parameters:

New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

Read pattern data from: In this group the user must select one of four options:

• a *.ffe file: Read the radiation pattern from an *.ffe format file (which
may be created with the DA and FF cards).
• an external data file: Read the radiation pattern from an ASCII file (the
format of this file is described below).
• after this line in the *.pre file: The radiation pattern is specified in the
*.pre file following the AR card (the format is described below). With
this option one can make use of the FOR loops to generate patterns from
known functions.
• use last pattern defined at previous AR card: When using multiple AR
cards (i.e. different radiating antennas in the same model) then it is quite
common that the shape of the pattern is identical. If this is the case, then
one can load the pattern just once and at subsequent AR cards use this

July 2011 FEKO User’s Manual

CONTROL CARDS 14-42

option. Then the last defined pattern will be used and memory can be
saved (no need to store it again). Note that one is still able to set the
antenna position, orientation, and amplitude and phase individually.

Source amplitude scale factor: The field strength values in each direction is determined from the data.
This parameter is used to scale the amplitude of the entire pattern by a constant
value.

Phase of source: This parameter specifies a constant additional phase for the entire pattern.

Source position: In this group the x, y and z coordinates of the source point (i.e. the position
where the antenna is placed) are entered in m. This value is affected by the
scale factor of the SF card if used.

Rotation about the axes: In this group the angles with which the imported pattern is rotated around
the x, y and z axes are entered in degrees. We will refer to these fields as α x ,
α y and αz in the rest of this discussion.

File name: The name of the *.ffe or ASCII input file.

Start from point number: This parameter is only relevant when the data is read from an external file,
and gives the line number of the first line to read from the input file. If the data
must be read from the beginning of the file, the value in this field should be set
equal to 1. This parameter is used when the *.ffe file contains more than one
pattern. For example, if the file contains the pattern at various frequencies, the
correct pattern can be selected by setting this field to the appropriate value for
each frequency.
If the *.ffe file is of a newer format and contains header data in addition to
the data blocks, the point number refers to the actual point number. This is the
same as the line number if all blank lines, comment lines and header lines are
stripped from the file.

Number of ϑ points: The number of ϑ angles in the pattern.

Number of ϕ points: The number of ϕ angles in the pattern.

The radiation pattern of the antenna must be specified in spherical coordinates (ϑ, ϕ) with the
phase centre located at the origin of the local coordinates (as used in the pattern data). If this
is not the case, the phase of the far fields will not be correct. (For example, if a *.ffe file is
exported with FEKO to be used with the AR card, the origin should be shifted with the OF card to
the phase centre of the antenna.) The vertical and horizontal components of the complex electric
field EϑF F and/or EϕF F must be specified at discrete angles (ϑ i , ϕ j ) with i and j larger or equal to 1
and smaller or equal to the number of points specified for the respective angles. The field values
are entered in Volts and the actual far fields are calculated from

e− jβR
Eϑ = EϑF F · (14-8)
R
or
e− jβR
Eϕ = EϕF F
, · (14-9)
R
with R the distance to the field point and β the complex propagation constant in the free space
medium (see the EG and GF cards). These formulas are used for all distances R, i.e. also in

July 2011 FEKO User’s Manual

CONTROL CARDS 14-43

the near field. However, FEKO tests whether the far field conditions are met (by calculating the
directivity and equivalent aperture) and gives an appropriate warning if this is not the case.
The permissible range of the angles ϑ i is 0◦ . . .180◦ and they must be in ascending order, i.e.
ϑ i+1 > ϑ i . However, the angles do not have to be equidistantly spaced. (Thus, for example, for a
highly directive antenna, a denser grid can be used close to the main beam direction.) The same
applies to the angles ϕ j where the permissible range is 0◦ . . .360◦ . For field angles outside the
start and end values defined in the data (i.e. for ϑ < ϑ1 , ϑ > ϑmax , ϕ < ϕ 1 or ϕ > ϕ max ), the field
strengths EϑF F and EϕF F are set to zero, so that a sector radiator can be realised. The values at field
angles within the defined range are determined by bilinear interpolation. To realise a complete
radiation pattern, rather than a sector radiator, the angles should be defined so that ϑ1 = 0◦ , ϑmax
= 180◦ , ϕ 1 = 0◦ and ϕ max = 360◦ .
The radiation pattern, specified in the local spherical coordinate system (ϑ, ϕ) of the antenna, is
read and initially placed at the origin of the global coordinate system in which the *.pre file is
constructed. The pattern is now rotated by an angle αz around the z axis, by α y around the y
axis and by α x around the x axis. (The rotation is identical to the rotation executed by the TG
card (see section 13.44) and the rotation matrix M is applicable to both the TG and AR cards.)
Finally the pattern is shifted to the specified location.
If the AR card is used simultaneously with a ground plane (BO card), FEKO includes the influence
of the ground plane on the radiation pattern. The imported pattern must therefore be the free
space radiation pattern of the antenna (in the absence of the ground plane). If this is not the
case the influence of the ground plane is considered twice.
The use of the PW card to specify the radiated power is allowed. The field amplitudes |EϑF F | and
|EϕF F | will be scaled accordingly. Multiple radiation patterns can be used simultaneously, and also
with other sources such as an incident plane wave. In such a case, the coupling is not considered
when the radiated power is determined.
The AR card cannot be used with special Green’s functions for a layered sphere or for a layered
substrate.
The format of the data depends on the value of the parameter Read pattern data from:

• an *.ffe file (*.ffe file)

With this option, the radiation pattern is read from a *.ffe file created with FEKO (using
the DA and FF cards). All the data of the radiation pattern (angles and field values) are
determined from the file. The user should ensure that, for example, the frequency is correct.
If an antenna is analysed with FEKO, the far field can be exported to the *.ffe file using
the commands (for 5◦ angle increments)
DA 0 0 1 0 0
FF 1 37 73 0 0 0 5 5

Note 37 points are used for ϑ and 73 for ϕ to ensure that the radiation pattern is closed
(see also the comment above).
This can then be imported as a source into another model with the command
AR 0 1 1 37 73 1.0 0.0 0.0 0.0 0.0 ...
... 0.0 0.0 0.0 "file.ffe"

• an external ASCII file

With this option, the data is read from the specified external data file. Each line contains 6
space delimited data fields in the following order:

July 2011 FEKO User’s Manual

CONTROL CARDS 14-44

– The angle ϑ in degrees

– The angle ϕ in degrees
– Amplitude of the field strength EϑF F in V
– Phase of the field EϑF F in degrees
– Amplitude of the field strength EϕF F in V
– Phase of the field EϕF F in degrees

The inner loop should be with respect to the angle ϑ so that the order of the lines is as
follows
ϑ1 ϕ1 ···
ϑ2 ϕ1 ···
ϑ3 ϕ1 ···
.. ..
. . ···
ϑ I4 ϕ1 ···
ϑ1 ϕ2 ···
ϑ2 ϕ2 ···
.. ..
. . ···
ϑ I4 ϕ I5 ···

• after this line in the *.pre file

This case is similar to reading an external ASCII file, except that the data is read directly
from the *.pre input file. The six data fields mentioned for the case of an ASCII file must
appear in the 6 columns of 10 characters, starting at character 31 and ending at character
90 in the lines following the AR card. When this option is selected, the card dialog shows
additional input fields where the user can specify these values for each point. The data
lines may be separated by comment lines (EDITFEKO, however, does not support this) and
FOR–NEXT loops may be used. Even when using FOR loops one may use the card dialog
in EDITFEKO to generate a typical line.

We conclude this description with an example of a sector radiator. We want to realise an ideal
sector radiator, which radiates 10 Watt horizontal polarisation in the angular region defined by
-70◦ ≤ ϕ ≤ 70◦ and 75◦ ≤ ϑ ≤ 105◦ . Since the angle range of the imported pattern must be
positive, one may define separate sources for the regions 0◦ ≤ ϕ ≤ 70◦ and 290◦ ≤ ϕ ≤ 360◦ . A
more elegant solution is to define a single pattern in the range 0◦ ≤ ϕ ≤ 140◦ and rotate it by
-70◦ around the z axis. The complete radiation pattern is defined in the following input file (note
that only horizontal polarisation, i.e. EϕF F , is required):

July 2011 FEKO User’s Manual

CONTROL CARDS 14-45

** Application example for the AR card: Sector radiator

** No other structures considered

EG 1 0 0 0 0

** Set the frequency

FR 1 0 100.0e6

** Specified radiated power

PW 1 10.0

** Define the sector radiator

AR 0 3 2 2 1.0 0.0 0.0 0.0 0.0 0.0 0.0 -70
** Theta Phi E_theta E_Phi
75 0 0 0 1 0
105 0 0 0 1 0
75 140 0 0 1 0
105 140 0 0 1 0
** Check: Compute the full 3D radiation pattern with 5 deg stepping
FF 1 37 73 0 0.0 0.0 5.0 5.0

** End
EN

FEKO determines a directivity of 10.1 dBi. The radiation pattern is easily validated by calculating
the far field as shown with the FF card in the last step. The result is shown in Figure 14-14.

Figure 14-14: 3D radiation pattern of the sector radiator.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-46

14.20 AS card

This card specifies an excitation by means of impressed spherical modes, which are either radi-
ating (i.e. propagating in positive r direction to infinity, with r being the radius in a spherical
coordinate system) or incident onto a structure (i.e. then propagating towards the origin r = 0).

This excitation option can thus be used for both the synthesis of an arbitrary electromagnetic field
(sum of the modes weighted with complex mode coefficients), and also for the determination of
the response (i.e. induced voltage or power at a load) of a receiving antenna due to the incident
modes (leading to the so-called generalised scattering matrix).
Parameters:

New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

Propagation direction: This determines if the spherical waves propagate Inward (the model is illumi-
nated with modes propagating towards r = 0, i.e. spherical Hankel function
of the first kind zn(3)= h(1)
n ) or Outward (the modes radiate towards r = ∞, i.e.
spherical Hankel function of the second kind zn(4) = h(2)
n ). This option is only
available for when the modes are entered in the *.pre file and not when the
modes are imported from a TICRA file (*.sph) in which case the outward
propagating direction is used.

Mode data source: The spherical modes can be entered directly in the *.pre file or it can be
imported from a TICRA file (*.sph). Note that multiple spherical modes can
be created as a single source. The imported spherical modes may be scaled
and the phase given an offset by entering values for the Scale magnitude by
and Offset phase (deg) fields.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-47

x, y, z position: The coordinates of the origin r = 0 of the mode in m. (These values are
optionally scaled by the SF card.)

ϑ angle: The angle between the dipole and the z axis in degrees.

ϕ angle: The angle between the projection of the dipole onto the plane z = 0 and the x
axis in degrees.

Rotation about axes: The rotation of the spherical mode source about the x, y and z axes.

Number of modes: The number of modes that are entered in the *.pre file must be specified.

Traditional index smn: If this option is checked, the user can specify TE-mode (s = 1) or TM-mode
(s = 2) and the indices m and n in the group below. Here n is the mode index
in radial direction and must be in the range 1, 2, . . . ∞ and m is the mode index
in the azimuth direction ϕ. We do not distinguish between even and odd modes
(with cos(mϕ) and sin(mϕ) angular dependencies), but rather use the angular
dependency e jmϕ . Thus the index m can also be negative, but it must be in the
range −n . . . n.

Compressed index j: With this option, a compressed one-dimensional mode numbering scheme is
used. The Mode index j is then specified in the field below. Here

j = 2 [n (n + 1) + m − 1] + s (14-10)

where s = 1 for TE-modes and s = 2 for TM-modes. This unified mode num-
bering scheme allows the computation of an extended scattering matrix (with
network and radiation ports). This index j then represents a unique port num-
ber in the scattering matrix.
p
Magnitude of the mode in W : Absolute value of the complex amplitude of this specific spherical
mode (due to thepapplied p normalisation of the spherical modes, the unit of
this amplitude is W = VA).

Phase of the mode: The phase of the complex amplitude of this spherical mode in degrees.

The implementation of the spherical modes at the AS card follows closely the references: J. E.
Hansen, Spherical Near-field Antenna Measurements, Peter Peregrinus Ltd., London, 1988 and B.
C. Brock, Using Vector Spherical Harmonics to Compute Antenna Mutual Impedance from Measured
or Computed Fields, Sandia National Laboratories, Report SAND2000-2217-Revised, April 2001.
One must realise that Hansen assumes a complex time dependence of e−iωt , while FEKO always
uses the positive sign e jω t .
In FEKO, using the modal coefficients Q csmn the electric and magnetic field strength is represented
in a spherical coordinate system by
4 X
X ∞ X
2 X n
~ (r, ϕ, ϑ) =
E Q(c) F~ (c) (r, ϕ, ϑ) (14-11)
smn smn
c=3 s=1 n=1 m=−n

4 X ∞ X
2 X n
j X (c)
~ ϕ, ϑ) =
H(r, Q(c) F~ (r, ϕ, ϑ) . (14-12)
smn 3−s,m,n
ZF c=3 s=1 n=1 m=−n

Here s, m and n are the mode indices with s = 1 indicating the TE-mode and s = 2 the TM-mode
(see also the discussion above), and c represents the propagation direction: c = 3 is inward and

July 2011 FEKO User’s Manual

CONTROL CARDS 14-48

c = 4 is outward. The term Z F denotes the wave impedance of the medium under consideration,
β below is the corresponding wavenumber.
(c)
The spherical wave functions Fsmn are given by
(c)
F~1mn (r, ϑ, ϕ) = M~ (c)
mn
Æ  m
Z
= β 2πF p 1 m
− |m|
 n(n+1)

· 0 · e~r (14-13)
+zn(c) (β r) sin ϑ Pd
jm |m|
(cos ϑ) e jmϕ e~ϑ
n n o 
(c) ∂
−zn (β r) ∂ ϑ Pn (cos ϑ) e
d|m| jmϕ
e~ϕ

and
(c)
F~2mn (r, ϑ, ϕ) = N
~ (c)
mn
Æ  m
Z
= β 2πF p 1 m
− |m|
 n(n+1)
n(n+1) (c)
· βr
zn (β r) Pd n
|m|
(cos ϑ) e jmϕ e~r (14-14)
© n o
+ β1r ∂ (β∂ r) β r zn(c) (β r) ∂∂ϑ Pd
¦
|m|
n
(cos ϑ) e jmϕ e~ϑ

+1 ∂ β r z (c) (β r)
¦ © jm
β r ∂ (β r) n
Pd|m|
sin ϑ
(cos ϑ) e jmϕ e~
n ϕ

with the associated Legendre function

r
2n + 1 (n − |m|)!
|m|
Pd (cos ϑ) = Pn|m| (cos ϑ) (14-15)
n
2 (n + |m|)!
and the spherical Bessel functions

zn(3) (β r) = h(1)
n
(β r) = jn (β r) + j yn (β r)
(4) (2) (14-16)
zn (β r) = hn (β r) = jn (β r) − j yn (β r) .

It should be noted that the Legendre polynomial Pn|m| (cos ϑ) as used in FEKO follows the defini-
tions of Abramowitz/Stegun (also used like this in Numerical Recipes) or also Harrington. The
formulas used in other references (e.g. Stratton or Hansen) have an extra factor (−1)m included.
This is not considered in FEKO, and thus the mode coefficients Q csmn might differ from those
computed according to Hansen (there is also of course the other time dependency, see earlier
discussion).
Theoretically the index n runs in the range 1, 2, . . . , ∞. For any practical application, one will
have to consider a finite number of modes only, i.e. limit the range n = 1 . . . N . Some rules of
thumb exist for the selection of N . For instance when representing the pattern of an antenna by
spherical modes one can use the upper limit
r0
N ≈ β r0 = 2π , (14-17)
λ
where β is the wavenumber, λ the wavelength, and r0 denotes the radius of the smallest sphere
enclosing the antenna. In critical cases, one might also rather use

N ≈ β r0 + 10 (14-18)

or p
N ≈ β r0 + 3 3
β r0 . (14-19)

July 2011 FEKO User’s Manual

CONTROL CARDS 14-49

When using the compressed numbering scheme with one index j, any upper limit N for n will
with the largest values m = N and s = 2 translate into an upper limit

J = 2 [N (N + 1) + N − 1] + 2 = 2 N (N + 2) (14-20)

for j (i.e. j = 1 . . . J then). So for instance for an antenna with enclosing radius r0 = λ4 (then
β r0 = 1.57) when using the last of the three rules of thumb above, one would need roughly
N ≈ 5 or J ≈ 70 modes, respectively. For r0 = λ these limits become already N ≈ 12 and
J ≈ 336, and for r0 = 5λ one has to use N ≈ 41 and J ≈ 3526 modes.
The origin r = 0 of the modes can be set arbitrarily in space by means of the Source position
field. However, no provision has been made to allow a rotation, i.e. in spherical coordinates the
distance r of some point is measured with respect to the mode origin, but the angles ϑ and ϕ
are always defined with respect to the global coordinate system. The reason why a rotation is
not supported here for the AS card (as opposed to the A5 or A6 elementary dipoles) is that the
modes inherently contain this radiation in an arbitrary direction. For instance, the TM-mode
n = 1, m = 0 represents a z oriented electric dipole (i.e. vertical), while the TM-modes n = 1,
m = ±1 represent horizontally aligned dipoles. An arbitrary pattern orientation can be obtained
by a suitable superposition of different modes.
The modes have been normalised such that each mode has a constant power flow through any
spherical surface (either inwards or outwards). In principle one could use the PW card for this,
but then power normalisation works only if there is not more than one mode active at the same
time (when using the PW card, just the total radiated power of all the modes is determined,
and then each mode is scaled with the same factor, so that the total radiated power is correct,
but here we enforce a specified power for each individual mode). The power for each mode is
independent of the mode indices

P = 0.5 |Magnitude of the mode|2 (14-21)

p
(unit is correctly Watt since the amplitude has a unit W ). Since the modes are orthogonal, if
multiple AS cards are active at the same time, the powers of the individual modes can just be
added. Any other power corrections (such as due to metallic elements being in the vicinity) are
not taken into account in FEKO.
If an AS excitation is used in connection with multiple different media, it should be noted that
we assume outward propagating modes (when Outward is selected under Propagation direction)
to originate from the Source position, i.e. the source is located in the medium where Source
position is located, and its contribution will be zero in all other media. For inward propagating
modes (when Inward is selected under Propagation direction) we assume them to originate at
infinity, in the free space medium 0, and such modes only contribute to this medium with index
0. In connection with the UTD, only outward propagating modes are allowed (they have a well
defined source point), while inward propagating modes are not supported (neither a source point
nor an incidence direction can be assigned to such modes).
When computing the far field with the FF card, then outward propagating modes are included
normally, which is important when synthesising antenna patterns by means of spherical modes.
However, for inward propagating modes the far field limit for R → ∞ of the field strength with
the factor
e− jβ R
(14-22)
R

July 2011 FEKO User’s Manual

CONTROL CARDS 14-50

split off does not exist (similar to the non-existent far field for an incident plane wave). Thus such
inward propagating modes are excluded from any far field computation. This is not a problem,
since normally inward propagating modes are applied when computing the generalised antenna
scattering matrix, i.e. the response of a receiving antenna to some incident mode. Then one looks
at these quantities:

• Induced voltage or power at the antenna terminals for the network ports (i.e. no far field
computation).

• Field scattered back, and decomposition of this field into spherical modes (far field ports).
Here one needs the far field computation, but similar to an RCS computation with an
incident plane wave, only the scattered far field is of interest, which can be obtained from
the FF card without problems.

Note that all the modes (inwards and outwards propagating) are correctly included when doing
a near field computation with the FE card (also for very large distances).
As an application example, we consider the TE-mode n = 5 and m = 0 and compute the far field
pattern:
** Create the far field radiation pattern of a spherical mode

** No geometry
EG 1 0 0 0 0

** Set the frequency

FR 1 100e6

** Spherical mode indices

#s = 1 ** 1 = TE-mode, 2 = TM-mode
#n = 5 ** mode order n (with n=1, 2, 3, ...)
#m = 0 ** mode order m (with m=-n ... n)

** Excitation
AS 0 4 #s #m #n 1 0

** Compute the full far field pattern

FF 1 91 37 0 0 0 2 10

** End
EN

The resulting pattern is shown in Figure 14-15. From the FEKO output file one can see the correct
radiated power of 0.5 Watt as obtained from the far field integration:
Integration of the normal component of the Poynting vector in the angular
grid DTHETA = 2.00 deg. and DPHI = 10.00 deg. ( 3367 sample points)
angular range THETA angular range PHI radiated power
-1.00 .. 181.00 deg. -5.00 .. 365.00 deg. 5.13889E-01 Watt
0.00 .. 180.00 deg. 0.00 .. 360.00 deg. 5.00001E-01 Watt

July 2011 FEKO User’s Manual

CONTROL CARDS 14-51

Figure 14-15: 3D radiation pattern of a spherical TE-mode with n = 5 and m = 0.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-52

14.21 AV card

With this card an impressed current source is specified similar to that of the AI card, but with
the AV card the end point makes electrical contact with a conducting surface as shown in Fig-
ure 14-16. The current varies linearly between the value at the start point and that at the end
point. At the connection point special singular functions are used for the surface current density
on the triangles to allow continuous current flow.

Parameters:

New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

Specify end point coordinates below: The coordinates of the end point ~r2 are known and specified with
the x, y, z coordinate fields. This point must coincide with a corner point of
one or more triangles.

Connect to closest triangle(s) vertex: The coordinates of the end point ~r2 are not known. In this case
the x, y, z coordinate fields of the End point are not used. FEKO searches
through all the metallic triangles for the corner point that is closest to the start
point ~r1 of the current element. This is then the end point ~r2 . For both cases
FEKO automatically searches for all the triangles making electrical contact with
the end point.

Amplitude: Current amplitude (in A) at the Start, ~r1 , and End, ~r2 , points.

Phase: Phase of the current at the start point in degrees.

x, y, z coordinate: Coordinates of the start and end points in m. (Note that all the coordinate
values are optionally scaled by the SF card.)

Radius of impressed current: This parameter is optional. If specified, and different from zero, this
value gives a finite wire radius for the impressed current element. FEKO then
assumes that the current is uniformly distributed on the wire surface and uses
the exact wire integral. If the parameter is not specified, the current filament
approximation is used. (This value is optionally scaled by the SF card.)

The following restrictions apply when using the impressed current elements making electrical
contact with conducting surfaces:

• All the restrictions given in the discussion of the AI card also apply in this case.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-53

• The start point of the impressed current segment may be connected with AI cards or further
AV cards. If there is a current discontinuity at this point, the resulting point charge is not
considered (see the discussion given with the AI card). Line charges along the current path
and surface charges on the triangles are correctly taken into account. At the connection
point ~r2 a continuous current model is used so that a point charge is not possible here.

I1

r1
im
pre
sse
dc
urr I2
ent
metallic
r2 triangles

Figure 14-16: Impressed line current with a linear current distribution and electrical contact to conducting
triangles

July 2011 FEKO User’s Manual

CONTROL CARDS 14-54

14.22 AW card

With this card a waveguide port excitation by an impressed mode on a rectangular, circular, or
coaxial waveguide can be modelled or the impressed travelling modes in all waveguides of a
multi-port network can be imported from a *.fim file.

Specify the source in the *.pre file

With this option a waveguide port excitation by means of an impressed mode on a rectangular,
circular, or coaxial waveguide, can be modelled.
Parameters:

New source: A new excitation is defined which replaces all previously defined excitations.

Add to sources: A new excitation is defined which is added to the previously defined excita-
tions.

Label of the port triangles: Label of the triangular mesh elements in the mesh which represent the
waveguide port. (If multiple solutions are defined in the same *.pre file, then
the usage of the waveguide ports with respect to the label/s to which it/they
are applied must be consistent for all solutions.)

Medium inside the triangles: The label of the medium inside the modelled waveguide.

Rectangular: A rectangular waveguide cross section is used, which is defined by three points
S1, S2, and S3 as follows: S1 is an arbitrary corner point, and S2 and S3 are
two corner points which define the waveguide sides û1 (from S1 to S2) and
û2 (from S1 to S3). The direction in which the mode is launched is given by
û3 = û1 × û2 .

July 2011 FEKO User’s Manual

CONTROL CARDS 14-55

Circular: A circular waveguide cross section is used. The point S1 denotes the centre
of the circular port, and the point S2 specifies the radius and start point for
the angular dependency. A further point S3 must be perpendicular above the
centre of the circular plate, so that the direction from S1 to S3 indicates the
direction in which the waveguide modes are launched.

Coaxial: Here a feed of a coaxial waveguide with circular cross sections of both the inner
and outer conductor can be specified. The point definitions are the same as for
the circular waveguide, except that an additional point S4 must be defined
between S1 and S2 which specifies the radius of the inner conductor.

Excite fundamental mode: Select this option to automatically excite the fundamental mode of the
waveguide. When this option is selected, the mode type and its indices (m
and n) can not be specified since they are determined automatically.

TE-mode: If this option is checked, a T Em,n mode (also referred to as H m,n ) is used as
excitation. This option is only available when “Excite fundamental mode” has
not been selected.

TM-mode: If this option is checked, a T Mm,n mode (also referred to as Em,n ) is used as
excitation. This option is only available when “Excite fundamental mode” has
not been selected.

TEM-mode: If this option is checked (only available for the coaxial waveguide since TEM
modes don’t exist in rectangular/circular waveguides), a TEM mode is used as
excitation. This option is only available when “Excite fundamental mode” has
not been selected.

Mode index m: The index m of the T Em,n or T Mm,n mode which is impressed at the port. Note
that for a rectangular waveguide the index m is related to the û1 direction (i.e.
from point S1 to S2). For a circular/coaxial waveguide, m denotes the angular
dependency. This option is only available when “Excite fundamental mode”
has not been selected.

Mode index n: The index n of the T Em,n or T Mm,n , mode which is impressed at the port. Note
that for a rectangular waveguide the index n is related to the û2 direction (i.e.
from point S1 to S3). For a circular/coaxial waveguide, n denotes the radial
dependency. This option is only available when “Excite fundamental mode”
has not been selected.

Magnitude. . .: Absolute value of the complex amplitude of the impressed mode. For a TE-
mode the unit is A/m, for a TM-mode the unit is V/m; for a TEM-mode the
unit is V. Note that an amplitude of zero can also be specified. In this case a
waveguide port is acting purely as a passive port (e.g. as waveguide termina-
tion), and no wave is launched.

Phase: The phase of the impressed mode in degrees.

Rotation angle ϕ 0 : This option is available for circular and coaxial modes only and indicates the
rotation angle in degrees by which a mode is rotated anti-clockwise with re-
spect to the reference direction (point S2).

Modal expansion: At a waveguide port a specific mode is used as impressed excitation. How-
ever, due to discontinuities in the model, also higher order modes can result
and will be propagating backwards through the port (applies to both active

July 2011 FEKO User’s Manual

CONTROL CARDS 14-56

and passive ports). The maximum modal expansion indices taken into account
during the calculation can be determined automatically by the kernel or speci-
fied manually. The included modes must be sufficient to capture the resulting
field distribution of the problem. Note that the mesh across the waveguide port
must be fine enough to represent the potential rapid field variation of included
higher order modes. Also note that an increased number of higher order modes
included in the model may have a significant impact on the run-time. If speci-
fied manually then the input values denote the maximum mode indices m and
n which will be used to expand the backwards travelling waves. If determined
automatically, then all propagating modes will be included, as well as evanes-
cent modes that decay faster than 1/e at a tenth of a wavelength away from
the waveguide port.

Passive port only: The waveguide port can be marked as passive only so that it will not be consid-
ered during S-parameter calculations. In this case the port is acting purely as
a passive waveguide termination, and the coupling to and from this port will
not be calculated.

Import modes from FEST3D file (*.fim)

With this option, impressed forward travelling modes in all waveguides of a multi-port network
can be imported from a *.fim file.

Parameters:

Position: A named point indicating the translation of the imported model in the FEKO
coordinate system. Note that this point must have been previously defined with
a DP card.

Rotation around the x-axis: This specifies the rotation of the imported model around the x-axis in
degrees.

Rotation around the y-axis: This specifies the rotation of the imported model around the y-axis in
degrees.

Rotation around the z-axis: This specifies the rotation of the imported model around the z-axis in
degrees.

File name: The name of the *.fim file.

In order to model a waveguide port excitation by an impressed mode, the cross section of the
waveguide at the port location must be meshed into metallic triangles with a unique label. The

July 2011 FEKO User’s Manual

CONTROL CARDS 14-57

propagation direction is given by the unit vector û3 , see the small graphics in the AW card panel
above.
In general, specific meshing rules exist in FEKO relating the triangular patch size to the wave-
length (see section 2.2.2). When meshing the cross section of a waveguide to define a waveguide
port, the mesh size must be small enough to capture the field distribution of the highest mode
(m, n) which is included in the expansion. FEKO checks this automatically and gives a warning
for coarse meshes or an error if the mesh size is too large. One must then either refine the mesh
just at the port or reduce the maximum mode indices used in the expansion.
The following restrictions apply when using a waveguide port excitation:

• Waveguide ports are available for models containing metallic objects (wires and surfaces
and wire / surface junctions, including PO) and dielectrics (solved using the SEP, FEM) or
dielectric coatings and thin dielectric sheets.

• Special Green’s functions may not be used in conjunction with waveguide port feeds.

• When using waveguide ports, then UTD is not allowed in the same model. Note, however,
that in FEKO it is possible (using the AR or AP cards) to decompose a model (say a horn
antenna in front of a reflector) into different sub-problems. If this approach is followed,
then waveguide feeds may be used. See Example_35 in the Scripting Examples guide for
an illustration of this decomposition technique.

The reflection coefficient at each waveguide port (S11 ) is always calculated and available for
display in POSTFEKO on an S-parameter graph, even when no S-parameter calculation has been
requested. Requesting S-parameters with the SP card is supported for waveguide ports. Multiple
ports (active and/or passive) can be present in the model. S-parameters are directly based on
the waveguide impedance of the specific mode under consideration. The reference impedance as
specified at the SP card is not used for waveguide ports.
Examples for the application of waveguide feeds can be found in the Examples Guide, there
waveguide_step (S-parameter computation for a waveguide discontinuity) and in the Script
Examples example_08 (horn antenna), and example_34 (coaxial cable).
In order to rule out any possible doubts and ambiguities regarding the waveguide mode defini-
tions, we give here the explicit expressions of the modes as used in FEKO. This implementation
follows closely the conventions in S. Ramo, J. R. Whinnery, and T. van Duzer, Fields and Waves in
Communication Electronics, John Wiley & Sons, Inc., 3rd ed., 1994.
Rectangular waveguide expressions
Local Cartesian coordinates (u1 , u2 , u3 ) are assumed. The factor e jωt± jβu3 is omitted for brevity,
where β is the complex modal propagation coefficient. In the expressions below, let a be the
dimension of the waveguide port in uˆ1 (distance between S1 and S2), and b the dimension of
the waveguide port in uˆ2 (distance between S1 and S3). The modal cutoff coefficient is the same
for TM- and TE-modes, and is given by,

mπ ‹2  nπ ‹2 1/2
βc = + . (14-23)
a b

July 2011 FEKO User’s Manual

CONTROL CARDS 14-58

For transverse magnetic (TM) modes the axial magnetic field component vanishes, and the axial
electric field component for the T Mm,n mode is expressed as,
mπu1 nπu2
EuTM (u1 , u2 ) = Âsin sin (14-24)
3 a b
for m = 1, 2, . . . , and n = 1, 2, . . . Â is the complex amplitude of the impressed mode in V/m. The
remaining field components are,
j⠁ mπ ‹ mπu1 nπu2
EuTM (u 1 , u 2 ) = ZTM H TM
(u1 , u 2 ) = − Âcos sin (14-25)
1 u2 βc2 a a b

j⠁ nπ ‹ mπu1 nπu2
EuTM (u1 , u2 ) = −ZTM HuTM (u1 , u2 ) =− Âsin cos (14-26)
2 1 βc2 b a b
β
ZTM = (14-27)
ω"

For transverse electric (TE) modes the axial electric field component vanishes, and the axial
magnetic field component for the T Em,n mode is expressed as,
mπu1 nπu2
HuTE3 (u1 , u2 ) = Âcos cos (14-28)
a b
for m = 0, 1, 2, . . . , and n = 0, 1, 2, . . . (but not both m and n zero). Â is the complex amplitude
of the impressed mode in A/m. The remaining field components are,
j⠁ nπ ‹ mπu1 nπu2
Eu1 (u1 , u2 ) = ZTE Hu2 (u1 , u2 ) = 2
TE TE
ZTE Âcos sin (14-29)
βc b a b

j⠁ mπ ‹ mπu1 nπu2
EuTE2 (u1 , u2 ) = −ZTE HuTE1 (u1 , u2 ) = − ZTE Âsin cos (14-30)
βc2 a a b
ωµ
ZTE = (14-31)
β

Circular waveguide expressions

Local cylindrical coordinates (r, ϕ, z) are assumed, with the z-axis on the waveguide axis (S1–S3).
The factor e jωt± jβz is omitted for brevity, where β is the complex modal propagation coefficient.
The expressions below are valid for the fields inside a circular waveguide, i.e. for r ≤ a, where a
is the radius of the waveguide port. Jm (x) is the m th order Bessel function of the first kind, and
Jm0 (x) denotes the derivative with respect to the argument.
For transverse magnetic (TM) modes the axial magnetic field component vanishes, and the axial
electric field component for the T Mm,n mode is expressed as,

EzTM (r, ϕ) = ÂJm (βc r) cos m(ϕ − ϕ0 )

 
(14-32)

for m = 0, 1, 2, . . . , and n = 1, 2, . . . . Â is the complex amplitude of the impressed mode in V/m,

and ϕ 0 is the rotation angle. The modal cutoff coefficient β c is the n th zero of Jm (β c a) = 0. The
remaining field components are,
jβ
E rTM (r, ϕ) = ZTM HϕTM (r, ϕ) = − ÂJm0 (βc r) cos m(ϕ − ϕ0 )
 
(14-33)
βc

July 2011 FEKO User’s Manual

CONTROL CARDS 14-59

jβ m
EϕTM (r, ϕ) = −ZTM H rTM (r, ϕ) = ÂJm (βc r) sin m(ϕ − ϕ0 )
 
(14-34)
βc2 r
β
ZTM = (14-35)
ω"

For transverse electric (TE) modes the axial electric field component vanishes, and the axial
magnetic field component for the T Em,n mode is expressed as,

HzTE (r, ϕ) = ÂJm (βc r) cos m(ϕ − ϕ0 )

 
(14-36)

for m = 0, 1, 2, . . . , and n = 1, 2, . . . . Â is the complex amplitude of the impressed mode in A/m,

and ϕ 0 is the rotation angle. The modal cutoff coefficient β c is the n th zero of Jm0 (β c a) = 0. The
remaining field components are,
jβ m
E rTE (r, ϕ) = ZTE HϕTE (r, ϕ) = ZTE ÂJm (βc r) sin m(ϕ − ϕ0 )
 
(14-37)
βc2 r

jβ
EϕTE (r, ϕ) = −ZTE H rTE (r, ϕ) = ZTE ÂJm0 (βc r) cos m(ϕ − ϕ0 )
 
(14-38)
βc
ωµ
ZTE = (14-39)
β

Coaxial waveguide expressions

Local cylindrical coordinates (r, ϕ, z) are assumed, with the z-axis on the waveguide axis (S1—
S3). The factor e jωt± jβz is omitted for brevity, where β is the complex modal propagation co-
efficient. The expressions below are valid for the fields inside a coaxial waveguide, i.e. for
ri ≤ r ≤ ro , where ro is the radius of the outer conductor and ri is the radius of the inner
conductor of the coaxial waveguide port. Jm (x) and Nm (x) are m th order Bessel functions of the
first and second kind, respectively and Jm0 (x) and Nm0 (x) denote the derivatives with respect to
the argument.
p
The fundamental mode in a coaxial waveguide is a TEM wave and propagates with β = ω µ".
The axial electric and magnetic field components are zero for a TEM-mode, and the transverse
field components have a static field distribution,

Â
E rTEM (r, ϕ) = ZTEM HϕTEM (r, ϕ) = − (14-40)
r
µ
Ç
ZTEM = (14-41)
"

 is the complex amplitude of the impressed mode in V.

For transverse magnetic (TM) modes the axial magnetic field component vanishes, and the axial
electric field component for the T Mm,n mode is expressed as,

Jm (βc r) Nm (βc r)
 
Ez (r, ϕ) = Â −
TM
+ cos m(ϕ − ϕ0 )
 
(14-42)
Jm (βc ro ) Nm (βc ro )

for m = 0, 1, 2, . . . , and n = 1, 2, . . . . Â is the complex amplitude of the impressed mode in V/m,

and ϕ0 is the rotation angle. The modal cutoff coefficient βc is the n th root of the transcendental

July 2011 FEKO User’s Manual

CONTROL CARDS 14-60

characteristic equation, Nm (βc ri )Jm (βc ro ) = Nm (βc ro )Jm (βc ri ), enforcing the boundary condition
that Ez must be zero at ri and ro . The remaining field components are,

E rTM (r, ϕ) = ZTM HϕTM (r, ϕ)

jβ
 0
J (β r) Nm0 (βc r)
 (14-43)
= − β Â − J m(β cr ) + cos m(ϕ − ϕ0 )
 
c m c o Nm (βc ro )

EϕTM (r, ϕ) = −ZTM H rTM (r, ϕ)

jβ m

J (β r) Nm (βc r)
 (14-44)
= β 2 r  − J m(β cr ) + sin m(ϕ − ϕ0 )
 
c m c o Nm (βc ro )

β
ZTM = (14-45)
ω"

For transverse electric (TE) modes the axial electric field component vanishes, and the axial
magnetic field component for the T Em,n mode is expressed as,

Jm (βc r) Nm (βc r)
‚ Œ
HzTE (r, ϕ) = Â − + cos m(ϕ − ϕ0 )
 
(14-46)
Jm0 (βc ro ) Nm0 (βc ro )

for m = 0, 1, 2, . . . , and n = 1, 2, . . . . Â is the complex amplitude of the impressed mode in A/m,

and ϕ0 is the rotation angle. The modal cutoff coefficient βc is the n th root of the transcendental
characteristic equation, Nm0 (βc ri )Jm0 (βc ro ) = Nm0 (βc ro )Jm0 (βc ri ), enforcing the boundary condition
that the derivative of Hz normal to the two conductors must be zero at the inner and outer radii.
The remaining field components are,

E rTE (r, ϕ) = ZTE HϕTE (r, ϕ)

(14-47)
 ‹
jβ m J (β r) Nm (βc r)
= β 2 r ZTE Â − J 0m(β cr ) + sin m(ϕ − ϕ0 )
 
c m c o Nm0 (βc ro )

EϕTE (r, ϕ) = −ZTE H rTE (r, ϕ)

(14-48)
 0 ‹
jβ J (β r) Nm0 (βc r)
= β ZTE Â − J 0m(β cr ) + cos m(ϕ − ϕ0 )
 
c m c o Nm0 (βc ro )

ωµ
ZTE = (14-49)
β

July 2011 FEKO User’s Manual

CONTROL CARDS 14-61

14.23 BO card

With this card a ground plane (at z = 0) can be specified for all computations following the BO
card. The reflection coefficient method is used.

Parameters:

No reflection coefficient ground: No ground plane (or use one of the other ground plane options —
such as the GF card for an exact model of real ground, or the SY card for a
perfect ground plane). This option is used to switch off the reflection ground
if the effect of different grounds are considered in a single input file.

Reflection coefficient ground plane: Use the reflection coefficient ground plane approximation with the
material parameters specified in the remaining input fields.

Perfectly electric conducting ground: Use an ideal electric ground in the plane z = 0. In this case the
remaining parameters are ignored.

Perfectly magnetic conducting ground: Use an ideal magnetic ground in the plane z = 0. The remaining
parameters are also ignored in this case.

Material label: The label of the medium to be used, as defined in the DI card.

It should be noted that it is not possible to calculate the fields below the ground plane, i.e. it
is not possible to calculate the fields in the region z < 0. In addition all structures must be in
the region z > 0. If calculations inside the ground are required, for example when there are
structures below ground, the exact Sommerfeld integrals (GF card) must be used.
When using a perfect electric or magnetic reflection coefficient ground plane, structures can be
arbitrarily close to the ground (while remaining above it). Segment end points and triangle edges
lying in the plane of the ground plane will make electrical contact with a perfect electric ground
plane. For a perfect magnetic ground the boundary condition forces the current to zero at this
point.
If real ground parameters are used, the reflection coefficient approximation is more accurate for
λ
structures further from the ground plane. Typically structures should not be closer than about 10
(FEKO will give a warning if this is the case).
A dielectric ground (real earth) can only be used with bodies treated with MoM, MLFMM, PO,
FEM, or the hybrid MoM/PO, i.e. the hybrid MoM/UTD method cannot be used in the presence
of a real ground.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-62

14.24 CA card

This card is used to define a section of a shielded cable which is used for irradiation (i.e. comput-
ing induced currents and voltages at the cable terminals) due to external sources. Transmission
line theory is applied, i.e. no need to discretise the cable as with the MoM. A section is defined
as a straight part of a cable (one cable can consist of multiple sections). Note that when new
models are created, it is recommended to use the SH card (shield definition), CD card (cable
cross-section definition), CS card (cable path section definition) and LC cards (cable loads).

Parameters:

Remove all existing cable paths: If checked, all previously defined cable paths are removed. All the
other input parameters are ignored.

New cable path: Defines a new cable path, all previously defined cable paths are replaced.

Add to existing cable paths: An additional cable path is defined (i.e. the previously added ones will be
kept).

Location: The location of the cable path section can be specified as two points or im-
ported from a NASTRAN file.

Define points: The Start point and the End point of the cable path section are
defined by point names. These points must have been defined previously
with a DP card (or by an external import etc.).

July 2011 FEKO User’s Manual

CONTROL CARDS 14-63

Import from NASTRAN file: The name of the NASTRAN file and the property
ID of the segments that have to be imported are required to import the
cable path section.

Cable type: This specifies the type of cable. There are two possibilities: User defined cable
or a pre-defined cable from the internal FEKO cable database:
If User defined cable is selected in the drop-down list, then the user has to
enter all the cable properties in the Cable properties section which will then be
enabled. The units of the individual parameters are included in the description.
Note that the Outer radius of cable shield will be scaled by any active SF card.
Also keep in mind that most of these parameters may depend on the frequency
and thus one might use variables or expressions.
If a pre-defined cable type is selected from the drop-down list, then the section
containing the cable properties will be disabled, and all required parameters
will be retrieved automatically from an internal FEKO cable database. There
are several commonly found shielded cable types (up to now all coaxial cables)
included.

Port termination: This section is used to define the ports (i.e. the two ends of the cable path
section defined by this CA card). For each port the user can decide if it is
terminated by an (internal and external) impedance or if this end of the cable
path section shall be connected to another cable path section.
The termination impedance value fields are only enabled if the corresponding
port is terminated. The internal impedance terminates the inner conductor
against the cable shield and the outer impedance terminates the cable shield
against the ground plane.
Note that a complete cable path has to be terminated on both ends. (Detailed
information is given below.)

Sampling point density: The cable path section will be subdivided into small segments for the com-
putation of the induced currents and voltages. The electric and magnetic field
strengths will be evaluated at each segment’s centroid, so this setting influ-
ences the accuracy of the computed result, but also the computation time. The
setting Automatic determination will choose the segment length automatically
(which should be adequate for most cases). If the maximum separation dis-
tance is specified, then this value will override the automatic mechanism. Note
that this manual value will be scaled by any active SF card.

Figure 14-17: Complete scenario of a cable path

July 2011 FEKO User’s Manual

CONTROL CARDS 14-64

Transmission line theory (TLT) is used in conjunction with the field calculation using the method
of moments (MoM) to compute the voltage coupled-in at the termination impedances of a cable
close to a conducting (metallic) ground. The cable itself is not taken into account when comput-
ing the external field distribution and it does not affect the field distribution at all, i.e. from a
“field-viewpoint” of the scenario, the cable is not present at all. This is also the reason for the re-
duced number of unknowns in comparison to a full MoM solution: the cable itself is not modelled
in the geometry and therefore not meshed into (wire) segments and it is therefore not necessary
to introduce a very fine mesh on the ground plane underneath the cable. A further advantage of
the transmission line approach is that multiple cable scenarios can be investigated in the same
model (e.g. different cable positions, etc.) without repetition of a time-consuming solution of the
whole model (e.g. MoM or MLFMM). Analysing another cable is similar to computing the near
field at another point.
The term cable path refers to the complete cable from its start point to its end point. Thus the
cable path can consist of a single or multiple cable path sections. Each cable path section is then
again subdivided into the segments which are used for the computation. Each complete cable
path has to be terminated on both ends.
An arbitrary number of cable path sections can be defined. The complete scenario is shown in
Figure 14-17. It consists of the cable path connecting two (virtual) enclosures with the termi-
nation impedances. The cable is illuminated by an external electromagnetic field (as caused by
sources and other radiating structures in the model) which couples into the cable and causes
the currents and voltages in the internal termination impedances. For the calculation to work
properly the segment direction vector ~s and the ground vector ~g must be (almost) perpendicular.
This figure also shows the cable path, the cable path sections (1 . . . N) and the segments.

Figure 14-18: Possible ways to define cable paths

Figure 14-18 shows the setup of some cable paths. There are three cable paths in total (distin-
guished by the line style):

Table 14-3: Cable paths in figure 14-18.

Path Description
1 Cable path from point 1 to 4 consisting of the cable path sections A, B, C
2 Cable path from point 5 to 4 consisting of the cable path sections D, E, F, G
3 Cable path from point 8 to 9 consisting of the single cable path section H

Note that even if there are crossings (section B and F) or sections using the same points (e.g.
sections B, C, D and E regarding point 5) there will be no conducting connection between sec-
tions belonging to different cable paths. As the cable paths will be assembled automatically by

July 2011 FEKO User’s Manual

CONTROL CARDS 14-65

searching and matching the points’ coordinates, the order of the CA cards determines how the
cable path gets built. (The search is always started at the first CA card defining a termination
impedance and then the cards will be processed in the order they appear in the input file.) For
example, in order to get the situation as shown in Figure 14-18 the CA cards have to be in the
following order:

CA Defining section A (start cable path 1)

CA Defining section B (continue cable path 1)
CA Defining section C (end cable path 1)
CA Defining section D (start cable path 2)
CA Defining section E (continue cable path2)
CA Defining section F (continue cable path2)
CA Defining section G (end cable path2)
CA Defining section H (single segment cable path 3)

or (note the changed relative cable path number):

CA Defining section A (start cable path 1)

CA Defining section C (continue cable path 1)
CA Defining section B (end cable path 1)
CA Defining section H (single segment cable path 2)
CA Defining section E (start cable path 3)
CA Defining section D (continue cable path 3)
CA Defining section F (continue cable path 3)
CA Defining section G (end cable path 3)

but not in the following order:

CA Defining section A (start cable path 1)

CA Defining section B (continue cable path 1)
CA Defining section D (end cable path 1)
CA Defining section H (single segment cable path 2)
CA Defining section E (start cable path 3)
CA Defining section C (continue cable path 3)

July 2011 FEKO User’s Manual

CONTROL CARDS 14-66

CA Defining section F (continue cable path 3)

CA Defining section G (end cable path 3)

In the last case the cable path section D will be connected to the cable path section B which is
not intended! (Cable path 1 would then start at point 1 and end at point 5 consisting of sections
A, B, D and cable path 3 will then start at point 4 and end at point 4 consisting of the sections C,
E, F, G.) Additionally in this case the cable path 3 will form a closed loop, but one has to keep in
mind, that even in this case the two ends of the cable path are not connected at point 4!
Current limitations of the cable irradiation analysis using the CA card:

• For the built-in cable types, the frequency range is limited as this data is based on measure-
ments only available for a certain frequency range. Currently the frequency range from
10 kHz up to 500 MHz is supported for all those cable types. (An error is given by FEKO
if one tries to set a frequency which is not in that range.) This restriction is not applied
for user defined cables, since it is assumed that the user has supplied the required cable
parameters for the frequency under consideration.

• The cable must be homogeneous, i.e. the cable parameters may not vary for the sections
belonging to the same cable path. (This is enforced for all cables.)

• Currently only single conductor coaxial cables are supported. A multi-conductor cable
cannot be modelled.

• Cables cannot be used in connection with UTD, but any other method is possible to repre-
sent the external configuration (e.g. a car body modelled with MoM or MLFMM).

• A reference plane acting as ground is required for the cable coupling algorithm. This is
currently implemented in such a way that only metallic triangles and perfectly conducting
ground planes (PEC-ground defined by a BO-card) are considered. FEKO will give an error
if the special Green’s functions are used in connection with a cable analysis.

• Connections of cables with wires or crossings with wires are not allowed. Even if a cable
path starts or ends at a point where also a wire segment starts or ends or if a cable path
crosses a wire, there will be no electrical connection established between the cable and the
wire at such a point.

Note also that this card is only intended to compute the coupling into the cable from an external
field strength. Thus no additional voltage or current sources are supported at the ports.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-67

14.25 CD card

With this card a specific cable cross-section can be defined.

Single conductor

A single cable is defined by specifying the cable cross-section.

Parameters:

PEC: Select this option to set the core of the cable to PEC.

Core material label: The label of the metallic (as defined in the DI card) to be used for the core.

Core radius: The radius of the core is specified.

Insulation material label: The label of the material (as defined in the DI card) to be used as the insu-
lation material.

Thickness: The thickness of the material to be used as insulator.

Coaxial cable

Three types of coaxial cables are supported: coaxial cables defined by characteristics, dimensions
and an internally defined database of cables.

• Coaxial cables defined by cable properties.

Shield label: The label of the metallic (as defined in the DI card) to be used for the
shield.
Outer radius: The outer radius of the shield.
Magnitude of the characteristic impedance (Ohm): The magnitude of the characteristic impedance
of the cable.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-68

Phase of the characteristic impedance (degrees): The phase of the characteristic impedance of
the cable.
Real part of the propagation constant (m−1 ): Specify the real part of the propagation constant.
Imaginary part of the propagation constant (m−1 ): Specify the imaginary part of the propaga-
tion constant.

• Coaxial cables defined by the geometry dimensions

Shield label: The label of the metallic (as defined in the DI card) to be used for the
shield.
Shield outer radius: The outer radius of the shield.
PEC: Select this option to set the core of the cable to PEC.
Material label: The label of the metallic (as defined in the DI card) to be used for the
core.
Core radius: The radius of the core.
Number of insulation layers: The number of the insulation layers in the cable.
Material label: The label of the material (as defined in the DI card) to be used for the
insulation layers.
Thickness: The thickness of the insulation layers.

• Internally defined database of cables. A pre-defined coaxial cable type can be selected from
the list. Several commonly found shielded cable types are included in this list.

Ribbon cable

With this option one can define ribbon cables. Note that currently only round cores are supported.
Parameters:

July 2011 FEKO User’s Manual

CONTROL CARDS 14-69

PEC: Select this option to set the core of the cable to PEC.

Material label: The label of the metallic (as defined in the DI card) to be used for the core.

Core radius: The radius of the core.

Number of cores: The number of cables which constitute the ribbon cable.

Core spacing: The spacing between the adjacent cores.

Return path number: The number of the cable specified as the the return path.

Insulation material label: The label of the material (as defined in the DI card) to be used for the
insulation layer.

Thickness: The thickness of the insulation layer.

Bundle (mixed) cable

Bundles allow for the construction of complex multi-core cables based on existing cables. These
cables are built from all types of cables already defined - single conductor, coaxial cable, rib-
bon cables and multi-cables. Note that currently only single conductor cables may be bundled
together.
Parameters:

Number of cables: The number of cables which constitute the bundle cable.

Cable: The reference number of the cable.

Offset X: The x offset for the respective cable from the cable path.

Offset Y: The y offset for the respective cable from the cable path.

Shield label: The label of the metallic (as defined in the DI card) to be used as the shield.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-70

Insulation material label: The label of the material (as defined in the DI card) to be used as the insu-
lation.

Outer radius: The outer radius of the cable.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-71

14.26 CF card

Set the type of integral equation for perfectly conducting metallic surfaces.

Parameters:

Type of integral equation for metallic surfaces: Here one can chose the EFIE (electric field integral equa-
tion) and also the CFIE (combined field integral equation). See the comment
below for more details.

Apply to all labels: The selected type of integral equation is applied globally to all metallic sur-
faces, irrespective of their label.

Apply to single label only: Here the selection of the type of integral equation applies to a single label
only, which is entered into the field From label.

Apply to label range: Here the selection of the type of integral equation applies to a range of labels,
which is entered into the fields From label and to label.

Factor of CFIE: In the CFIE formulation electric and magnetic terms are combined with a factor.
Leaving this input field empty will select the default value of 0.2, which should
normally be used.

The EFIE is the default in FEKO if no CF card is used or for labels where no setting is made at the
CF card. It is the most general formulation and can be applied to both open and closed bodies.
The CFIE can only be used in connection with closed objects, and the advantage is that the
conditioning of the system of linear equations is better. In particular in connection with MLFMM
the convergence can be improved if the CFIE is used for closed parts of an object. Note that the
CFIE can be used in conjunction with EFIE (on the same object) as long as all triangle normals
point away from the zero field region.
For the CFIE in addition to the fundamental restriction that the surface must be closed, these
further conditions apply:

• The normal vector must point outwards (i.e. from the closed field free region into the
domain of interest where there are sources and fields shall be computed).

• Using symmetry in order to reduce the memory or run-time is not supported in FEKO (it
will be switched off automatically).

• The CFIE formulation for the MoM cannot be used together with the MoM/PO, MoM/UTD,
or MoM/FEM hybrid methods.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-72

• The CFIE formulation can be used only with the free space Green’s function (i.e. using the
spherical or planar multilayer Green’s functions is not supported).

• When using the CFIE, dielectric bodies (solved using SEP, FEM or MoM/MLFMM) may be
present in the model, though all of the CFIE surfaces must be perfectly conducting (no
coating or Skin effect etc.).

Note that multiple CF cards can be used in order to specify for instance that the CFIE shall be
used at multiple distinct labels which do not form a range. The setting for Factor for CFIE is
global (not per label), the value read from the last CF card will be used.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-73

14.27 CG card

Here the method used to solve the matrix equation may be chosen. Normally the CG card should
not be used. FEKO has been written to choose the optimal solution technique and also the
corresponding pre-conditioners and options etc. for different types of problems automatically.
These algorithms should work in all cases, but they might not be optimal for specific MLFMM or
FEM configurations (which rely on iterative solvers). Thus for these solutions, advanced users
might after consultation with FEKO technical support, apply the CG card in such special cases.
But one should realise that convergence of the iterative techniques cannot be assured, and also
when using an inappropriate preconditioner the memory requirement might be much higher
than required. Also users should take care to reconsider any CG card settings in models that are
derived from models containing the CG card.

Parameters:

In the first drop-down list, the matrix solution method is chosen:

• Default solver selection (recommended) When this option is selected,

then FEKO will automatically select a suitable solver along with all its
required parameters. The choice depends on whether FEKO is executed
sequentially or in parallel, but also which solution method is employed
(e.g. direct LU decomposition solver for the MoM while an iterative solver
is used for MLFMM or FEM). This option has, regarding the solver type,
the same effect as not using a CG card, but still allows the user to change
the default termination criteria for the iterative solver types, or to make
other CG card selections with respect to the pre-conditioning.
• Gauss Elimination (LINPACK routines) Use Gauss elimination from the
LINPACK routines.
• Conjugate Gradient Method (CGM)

July 2011 FEKO User’s Manual

CONTROL CARDS 14-74

• Biconjugate gradient method (BCG)

• Iterative solution with band matrix decomposition
• Gauss elimination (LAPACK routines) Use Gauss elimination from the LA-
PACK routines.
• Block Gauss algorithm (matrix saved to disk) The block Gauss algorithm
is used (in case the matrix has to be saved on the hard disk, i.e. a sequen-
tial out-of-core solution is performed).
• CGM (Parallel Iterative Method)
• BCG (Parallel Iterative Method)
• CGS (Parallel Iterative Method)
• Bi-CGSTAB (Parallel Iterative Method)
• RBi-CGSTAB (Parallel Iterative Method)
• RGMRES (Parallel Iterative Method)
• RGMRESEV (Parallel Iterative Method)
• RCGR (Parallel Iterative Method)
• CGNR (Parallel Iterative Method)
• CGNE (Parallel Iterative Method)
• QMR (Parallel Iterative Method)
• TFQMR (Parallel Iterative Method)
• Parallel LU-decomposition (with ScaLAPACK routines) The parallel LU-
decomposition with ScaLAPACK (solution in main memory) or with out-
of-core ScaLAPACK (solution with the matrix stored on hard disk). This
is the default option for parallel solutions and normally the user need not
change it.
• QMR (QMRPACK routines)
• Direct sparse solver Direct solution method for the FEM (no precondi-
tioning).

Maximum number of iterations: The number of iterations limit for the iterative techniques.

Stopping criterium for residuum: Termination criterion for the normalised residue when using iterative
methods. (Terminate with convergence when the normalised residue is smaller
than this value.)

Stop at maximum residuum: For the parallel iterative methods, the solution is terminated when the
residuum becomes larger than this value. (Terminate with divergence when
the normalised residue is larger than this value.)

Preconditioners: In this list the preconditioner is selected from:

• Default preconditioner selection (recommended) When this option is se-

lected, then FEKO will automatically select a suitable preconditioner along
with all its required parameters. The choice depends on whether FEKO
is executed sequentially or in parallel, but also which solution method is
employed (e.g. MLFMM or FEM). This option has regarding the precon-
ditioner the same effect as not using a CG card, but still allows then to
make other selections with respect to the solver (e.g. residuum).

July 2011 FEKO User’s Manual

CONTROL CARDS 14-75

• No preconditioning: No preconditioning is used.

• Scaling the matrix A: Scaling the matrix [A], so that the elements on the
main diagonal are all normalised to one.
• Scaling the matrix [A]H [A]: Scaling the matrix [A]H [A], so that the ele-
ments on the main diagonal are all normalised to one.
• Block-Jacobi preconditioning using inverses: The inverses of the precon-
ditioner are calculated and applied during every iteration step. For per-
formance reasons Block-Jacobi preconditioning using LU-decomposition
is recommended.
• Neumann polynomial preconditioning: Self explanatory.
• Block-Jacobi preconditioning using LU-decomposition: Block-Jacobi pre-
conditioning where for each block a LU-decomposition is computed in
advance, and during the iterations a fast backward substitution is ap-
plied.
• Incomplete LU-decomposition: Use an incomplete LU-decomposition of
the matrix as a preconditioner.
• Block-Jacobi preconditioning of MLFMM one-level-up: Special precondi-
tioner for the MLFMM, where additional information is included into the
preconditioner.
• LU decomposition of FEM matrix: An LU decomposition of the FEM ma-
trix is used as preconditioner.
• ILUT decomposition of FEM matrix: An incomplete LU decomposition
with thresholding of the FEM matrix is used as preconditioner.
• Multilevel ILUT/Diagonal decomposition of the FEM matrix: Self ex-
planatory.
• Multilevel ILUT/ILUT decomposition of the FEM matrix: Self explanatory.
• Multilevel LU/Diagonal decomposition of the FEM matrix: Self explana-
tory.
• Sparse Approximate Inverse (SPAI) preconditioner: Preconditioner which
can be used in connection with the MLFMM.
• Complete LU-decomposition: Use a complete LU decomposition of the
matrix (MLFMM/ACA) as a preconditioner.

Options for the BCG: This applies to the Biconjugate Gradient Method. Options are:

• Fletcher’s method
• Jacob’s method
• Fletcher’s method, pre-iteration using Fletcher’s method
• Fletcher’s method, pre-iteration using Jacob’s method
• Jabob’s method, pre-iteration using Fletcher’s method

Block size: The block size to be used for LU-decomposition with LAPACK as well as for the
Block-Jacobi preconditioning. When nothing is specified (i.e. the input field
is left empty), then appropriate standard values are used for LAPACK and the
block preconditioners.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-76

Threshold value for ILUT: This is the thresholding value used for the FEM in connection with the ILUT
preconditioners.

Level-of-fill: This is used by the MLFMM during the iterative matrix solution in connection
with the incomplete LU preconditioner. The range of this parameter is between
0 and 12. FEKO will choose the value for the best preconditioning, but if the
size of the incomplete LU preconditioner is too large to fit into memory, it can
be reduced by reducing the level-of-fill. It should be noted that a lower level-
of-fill might result in a slower convergence or even divergence in the iterative
solution.

Fill-in level per row: This is used by the FEM during the iterative matrix solution in connection
with the incomplete LU preconditioners with thresholding. It sets a limit on
the number of entries per row that will be included in the incomplete LU-
decomposition of the preconditioner matrix.

Stabilisation factor (FEM): This applies only to the incomplete LU preconditioners of the FEM and can
be used to get better convergence for the FEM in critical cases (the value range
is between 0 and 1).

Save/read preconditioner: For the incomplete LU preconditioners used with the FEM one has here
the option to save run-time by computing the preconditioner only once and
write to a *.pcr file and then for a subsequent run just read again from this
file. Since the FEM pre-conditioners depend only on the FEM matrix even for
combined MoM/FEM hybrid methods, this method is useful for instance when
only the MoM part of a problem setup has changed.

Iterative solutions are used for instance for the MLFMM or the FEM. There might then be sit-
uations where during the iterations the residuum is decreasing but not reaching the specified
stopping criterion. Instead of waiting very long until the maximum number of iterations has been
reached, the user can press Ctrl-C or Ctrl-Break (under Windows) or send the SIGINT/SIGTERM
signals (under UNIX) so that FEKO will stop with the iterations and resume with the further
processing (e.g. far field computations) using the solution associated with the best residuum ob-
tained so far. To really interrupt a FEKO job Ctrl-C or Ctrl-Break must be pressed a second time
(or the corresponding signal must be sent once more).

July 2011 FEKO User’s Manual

CONTROL CARDS 14-77

14.28 CM card

This card is used to couple FEKO with the transmission line simulation programs CableMod or
CRIPTE or the PCB tool PCBMod to calculate the coupling of electromagnetic fields into trans-
mission lines. (The AC card is used for the case of radiation by these lines.)

The only parameter of this card is the file name of a *.rsd file created by CableMod, CRIPTE or
PCBMod (enclosed in double quotation marks and starting at or after column 91). The *.rsd file
contains geometry of the line. With the CM card FEKO calculates the electric and magnetic near
field at points along the line and write these to a *.isd file for further processing by CableMod,
CRIPTE or PCBMod. (The *.isd file also contains additional data required by CableMod, CRIPTE
or PCBMod for example the frequencies that were used during the solution.)
The complete geometry (without the transmission line) as well as the frequency and excitation
(Ax cards) must be defined in FEKO.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-78

14.29 CO card

This card specifies a dielectric or magnetic coating of wire segments or triangular surface ele-
ments. The coating applies to all calculations following the CO card.
Note that before a coating can be applied, the material properties for the coating need to be
defined in the DI card. A coating with the specific material properties are then applied by refer-
encing the label of the material in the CO card. A layered medium are defined by referencing the
labels of the materials to be used for the individual layers in the DL card.

Parameters:

Label of elements to coat: All segments or triangles with this label are coated.

No coating: No coating present (as if the relevant label has no CO card). This is used to
remove wire coatings from earlier solutions.

Wire coating (Popovic formulation): In this case, the radius of the metallic core is changed internally
to model the change in the capacitive loading of the wire and a corresponding
inductive loading is added. The only restriction of this method is that the loss
tangents of the wire coating and of the surrounding medium must be identical
(for instance both media could be lossless).

Wire coating (volume equivalence principle): Here the radius of the metallic wire is retained. The effect
of the dielectric layer is accounted for by a volume polarisation current. The
only restriction of this method is that the layer may not be magnetic in nature
(i.e. the relative permeability µ r as well as the magnetic loss tangent tan δµ of
the coating must be the same as those of the surrounding medium).

Electrically thin surface coating: This option adds multilayer dielectric / magnetic coatings to the sur-
face triangles with the specified label. The layers may have different permit-
tivity and permeability, but the total coating must be electrically (i.e. relative
to the wavelength in the coating) as well as geometrically thin (see the next
item).

Dielectric / magnetic surface coating: This option adds electrically thick multilayer dielectric / mag-
netic coatings to the surface triangles with the specified label. Here it is only
required that the total coating must be geometrically thin, i.e. it must be thin
relative to the triangle size (and thus also to the free space wavelength) as well
as the radius of curvature of the surface.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-79

Material label: Label of the material (as specified in the DI card) which will be used as a
coating.

Layered dielectric label: Label of the layered dielectric medium (as specified in the DL card) which will
be used as a coating..

Thickness of . . .: For surface coatings, the thickness hn of each respective layer; for wire coatings
this is the radius of the coating less the radius % of the wire-core. This value is
in m and is scaled by the SF card.

Wire radius: Wire radius % of the metallic wire, without layers, in m (it is scaled by the SF
card). This overrides the values specified with the IP card. This field is only
applicable to wire coatings.

When using the Popovic formulation for wire coatings, the following restrictions apply:

• The loss tangent tan δ of the layer (which is calculated from the conductivity σ and the
relation tan δ = ωεσε ) has to be identical to the loss tangent of the surrounding medium
r 0
(specified with the EG card, usually free space)

• Due to the change in the radius of the metallic core, no SK card should be active for
the same label, otherwise the skin effect and/or the ohmic losses refers to the wire with
changed radius.

• For pure dielectric layers (i.e. the relative permeability µ r as well as the magnetic loss
tangent tan δµ of the layer equal those of the surrounding medium) the option Wire coating
(Volume equivalence principle) is recommended.

Note that for wire coatings, no surface triangles with the same label are allowed. Likewise for
surface coatings, no segments with same label are allowed.
If Dielectric / magnetic surface coating (the electrically thick coating) is used, it must remain
consistent for the whole FEKO run. (It cannot be on for one solution and off for the next.) It is,
however, allowed to change the thickness and the medium parameters of the coating.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-80

14.30 CS card

This card is used to define a path along which the cable runs thus also providing the centre/ref-
erence location to which a cable cross-section definition is applied.
A path my be specified using data points in the *.pre file or loading a cable path from a NAS-
TRAN file.

Parameters:

Remove all existing cable paths: If checked, all previously defined cable paths are removed. All the
other input parameters are ignored.

New cable path: Defines a new cable path, all previously defined cable paths are replaced.

Add to existing cable paths: An additional cable path is defined (i.e. the previously added ones will be
kept).

Path section label: The label of the path section.

Cross-section definition label: The label of the defined cross-section to be applied to the path section.

Cable type: Irradiating When this option is selected, only the effect of external fields cou-
pling into a cable will be considered.
Radiating When this option is selected, the effect of the currents radiating in
the cables will also be considered.

Solution method: Multiconductor transmission line (MTL) When this option is selected, the
model will be solved with the multiconductor transmission line theory
which is also hybridised with MoM or the MLFMM. The cable path should
be within λ5 of the conducting surface.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-81

Hybrid MoM/MTL When this option is selected, the model will be solved by
means of the hybrid MoM/MTL solver. Any arbitrary cable path may be
defined.

Path: Specify points in *.pre file The points defining the cable path are specified
in the *.pre file. These points must have been defined previously with
a DP card. The number of points defining the cable path section in the
*.pre file is set by the Number of points option. The Path points (as
defined with a DP card) are specified to define the cable path section.
Import from NASTRAN file The NASTRAN File name is required to import
the cable path section. The NASTRAN segment property ID of the seg-
ments are also required to be able to import the cable path section.

Connector labels: The start and end points of a cable path section are uniquely identified using
the Connector at start and Connector at end labels. Currently cable path labels
can be joined if they share a node, use the same cable connector and the same
cable cross-section definition. Thus, after combining/reducing the number of
cable path sections, each cable path consists of a section of cable with a unique
cross-section. Note that there is no support for splitting of cable sections or
combining of cable sections using loads.

Sampling point density: The cable path section will be subdivided into small segments for the com-
putation of the induced currents and voltages. The electric and magnetic field
strengths will be evaluated at each segment’s centroid, so this setting influ-
ences the accuracy of the computed result, but also the computation time. The
setting Automatic determination will choose the segment length automatically
(which should be adequate for most cases). If the maximum separation dis-
tance is specified, then this value will override the automatic mechanism. Note
that this manual value will be scaled by any active SF card.

Export cable parameters to *.out file: When this item is checked the cable parameters such as induc-
tance/capacitance matrices and transfer impedance/admittances are exported
to the *.out file.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-82

14.31 DA card

With this card some data like near fields or S-parameters can be exported to additional ASCII
files. The card allows to switch this export on and off, and affects only cards for the computation
of, for example, near fields or S-parameters that follow the DA card. By default no such export
files are created.
Furthermore, normally FEKO writes all the results to the ASCII *.out file. For large datasets
(e.g. full far field patterns at many frequency points) this *.out file can become very large. In
order to display results in POSTFEKO only the binary output file (*.bof file) is required, and
therefore to reduce the size of the *.out file it is possible to switch off certain results in that file
(a header is still written so that one knows what type of computations were requested).

Parameters:
Write electric fields to: The electric field strength can be exported into a *.efe file. The output of this
result type to the *.out file can also be deactivated here.

Write magnetic fields to: The magnetic field strength can be exported into a *.hfe file. The output of
this result type to the *.out file can also be deactivated here.

Write far fields to: The far field can be exported into a *.ffe file. The output of this result type
to the *.out file can also be deactivated here.

Write currents/charges to: The currents can be exported into a *.os/*.ol file. The output of this
result type to the *.out file can also be deactivated here.

Write residue . . .: The residue from the iterative algorithm used to solve the matrix equation is
stored in a *.cgm file.

Write S-parameters . . .: The S-parameters (see the SP card) are written to a file in Touchstone *.snp
format (v1.0). The n here gives the number of ports.

Write spherical wave expansion . . .: A spherical wave expansion of the far field as computed by FEKO is
exported to an SWE file (extension *.sph) which can be imported into GRASP
from TICRA (code for reflector antenna modelling). Note that the far field
computation with spherical wave expansion must be requested subsequently
(FF card (see section 14.37)).

July 2011 FEKO User’s Manual

CONTROL CARDS 14-83

Write near fields to SEMCAD *.dat file: The near fields can be exported into a SEMCAD *.dat file. The
output of this result type to the SEMCAD *.dat file can also be deactivated here.

Write error estimates to the *.out file: The error estimates can be exported to the *.out file.

Write generalised S-parameter matrix to FEST3D *.chr file: The generalised S-parameter matrix (GSM)
for waveguide ports are exported to a FEST3D *.chr file.

More than one DA card is allowed in one input file. Thus, using the following sequence of control
cards, with the appropriate options only certain blocks will be saved to the data files:
DA ... ** Write near fields on
FE ...
DA ... ** Write near fields off
FE ...

With this sequence, the electric fields calculated with the first FE card can be written to the *.efe
file, but not those of the second FE card.

14.31.1 General File Format

The structure of the following file formats share a common structure and will be explained each
in turn: *.efe, *.hfe, *.ffe, *.ol and *.os. The general structure can be described as follows:

Header Block The header block contains general information on the file. It includes information
such as the file type, format, how and when the file was written, etc. Only one header block
may be created and must be located at the top of the file. Header lines are denoted by two
hash symbols (“##”), followed by the key / value pairs allowed by each file type for the
Header Block sections. The format is then “##Key: Value” for each header block line.
Comments Comments may be placed anywhere in the file. They are denoted by two asterisks
(“**”), indicating that the rest of that line must be ignored.
Solution Block Any number of solution blocks (typically one per request per frequency) can now
follow. A solution block can further be broken down into:
Solution Block Header This section contains information that describes the data block
and includes information such as the frequency, coordinate system, the request name,
column headers, etc. Solution block headers are denoted by a single hash symbol
(“#”), followed by the key / value pairs allowed by each file type for the Solution
Block Header sections. The format is then “#Key: Value” for each solution header
block line.
Data Block The data block contains space delimited values. Values are given in scientific
notation (e.g. 1.23E-001).

The column headers are a part of the solution block header and must be presented in the follow-
ing format:
#no of header lines: M
#"Column 1: Line 1" "Column 2: Line 1" ... "Column N: Line 1"
#"Column 1: Line 2" "Column 2: Line 2" ... "Column N: Line 2"
...
#"Column 1: Line M" "Column 2: Line M" ... "Column N: Line M"

Note that they differ from other solution block header lines in that they do not have a key / value
pair format. Column header lines still start with a single hash (“#”), but is then followed by the
column titles surrounded in quotation marks.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-84

14.31.2 *.efe / *.hfe file: Electric / Magnetic near fields

The following fields are available in the Header Block:

Key Required Description

File Type Yes Describes the type of the file. For electric near fields, the
value must be: Electric near field and for mag-
netic near fields: Magnetic near field
File Format No Denotes the file syntax version (e.g. “1.0”). If not present
it defaults to version 1.0 (files pre-dating Suite 6.1). The
initial new file format version will then be “2.0”.
Source No Denotes the base filename of the source where this data
comes from.
Date No Date of data export in format "YYYY-MM-DD hh:mm:ss"
(i.e. 24-hour format)

The following fields are available in the Solution Block Header:

Key Required Description

Request Name No The explicit name given to that solution request (as de-
noted in the *.pre file). If none is specified, then during
importing such files in POSTFEKO use a default name of
"request_N" (where "_N" is replaced with a number for
each unnamed request).
Frequency Yes Frequency in Hz for which the following data was mea-
sured / computed.
Coordinate No Coordinate system in which the axes are defined:
System
• Cartesian [default]
• Cylindrical
• Cylindrical (x axis)
• Cylindrical (y axis)
• Conical

Origin No Origin of the data coordinate system in form “(x, y, z)”

(always in Cartesian coordinates; based on global ori-
gin). If no origin is provided, assume (x, y, z) = (0, 0, 0).
u-Vector No Indicates a point on the u-axis relative to the Origin. If
none is specified, it is assumed that the u-axis coincides
with the x-axis.
v-Vector No Indicates a point on the v-axis relative to the Origin. If
none is specified, it is assumed that the v-axis coincides
with the y-axis.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-85

No. of [$$$] Yes The number of samples in each axis direction. The [$$$]
Samples term is replaced by:

• X/U
• Y/V
• Z/N
• Phi
• Theta
• Rho
• Radius
• Cuboid

Result Type No For electric near fields (*.efe), this specifies whether the
output is:

• Electric Field Values [default]

• Magnetic Vector Potential
• Gradient of Scalar Electric Potential

For magnetic near fields (*.hfe), this specifies whether

the output is:

• Magnetic Field Values [default]

• Electric Vector Potential
• Gradient of Scalar Magnetic Potential

No. of No Number of header lines to read. The column header lines

Header Lines must follow this line. If this value is not specified, assume
a value of “1”
Yes For the header lines (i.e. the column titles), the format in
the column header format section (see section 14.31.1)
should be used. The specific column headers for the dif-
ferent near field coordinate systems are also listed.

The default column headers will all have the same structure. The column headers will depend
on the coordinate system that was defined, which can be found in the Coordinate System
value. For each column, the text $$$ will be replaced with the appropriate value depending on
the Result Type. The column headers for each coordinate system is now given:
Cartesian:
"X" "Y" "Z" "Re($$$x)" "Im($$$x)" "Re($$$y)" "Im($$$y)" "Re($$$z)" "Im($$$z)"

Cylindrical (X axis):
"Rho" "Phi" "X" "Re($$$rho)" "Im($$$rho)" "Re($$$phi)" "Im($$$phi)" "Re($$$x)" "Im($$$x)"

Cylindrical (Y axis):
"Rho" "Phi" "Y" "Re($$$rho)" "Im($$$rho)" "Re($$$phi)" "Im($$$phi)" "Re($$$y)" "Im($$$y)"

July 2011 FEKO User’s Manual

CONTROL CARDS 14-86

Cylindrical:
"Rho" "Phi" "Z" "Re($$$rho)" "Im($$$rho)" "Re($$$phi)" "Im($$$phi)" "Re($$$z)" "Im($$$z)"

Spherical:
"Radius" "Theta" "Phi" "Re($$$r)" "Im($$$r)" "Re($$$theta)" "Im($$$theta)" "Re($$$phi)" "Im($$$phi)"

Conical:
"Rho" "Phi" "Z" "Re($$$rho)" "Im($$$rho)" "Re($$$phi)" "Im($$$phi)" "Re($$$z)" "Im($$$z)"

All of the coordinate system descriptions given contain a section of text containing $$$. This
text is not meant to be included in the file, but rather to serve as a place holder. The contents of
the $$$ place holder will depend on the type of data being presented and the type of file being
written. For any of the above systems, replace $$$ with:

(*.efe) Electric near field files

E for electric field values
A for magnetic vector potential values
grad(PHI) for gradient of the scalar electric potential values
(*.hfe) Magnetic near field files
H for magnetic field values
F for electric vector potential values
grad(PSI) for gradient of the scalar magnetic potential values

July 2011 FEKO User’s Manual

CONTROL CARDS 14-87

14.31.3 *.ffe file: Far fields

The following fields are available in the Header Block:

Key Required Description

File Type Yes Describes the type of the file. For far fields, the value
must be: Far field.
File Format No Denotes the file syntax version (e.g. “1.0”). If not present
it defaults to version 1.0 (files pre-dating Suite 6.1). The
initial new file format version will then be “2.0”.
Source No Denotes the base filename of the source where this data
comes from.
Date No Date of data export in format "YYYY-MM-DD hh:mm:ss"
(i.e. 24-hour format)

The following fields are available in the Solution Block Header:

Key Required Description

Request Name No The explicit name given to that solution request (as de-
noted in the *.pre file). If none is specified, then during
importing such files in POSTFEKO use a default name of
"request_N" (where "_N" is replaced with a number for
each unnamed request).
Frequency Yes Frequency in Hz for which the following data was mea-
sured / computed.
Origin No Origin of the data coordinate system in form “(x, y, z)”
(always in Cartesian coordinates; based on global ori-
gin). If no origin is provided, assume (x, y, z) = (0, 0, 0).
u-Vector No Indicates a point on the u-axis relative to the Origin. If
none is specified, it is assumed that the u-axis coincides
with the x-axis.
v-Vector No Indicates a point on the v-axis relative to the Origin. If
none is specified, it is assumed that the v-axis coincides
with the y-axis.
No. of [$$$] Yes The number of samples in each axis direction. The [$$$]
Samples term is replaced by:

• Phi
• Theta

July 2011 FEKO User’s Manual

CONTROL CARDS 14-88

Result Type No For far fields (*.ffe), this specifies whether the output is:

• Gain [default]
• Directivity
• RCS
• Unknown

Incident Yes This is a (Theta,Phi) pair indicating the angle where an

Wave infinite plane source is originating from. Note that this
Direction field is only required if the Result Type is RCS.
No. of No Number of header lines to read. The column header lines
Header Lines must follow this line. If this value is not specified, assume
a value of “1”
Yes For the header lines (i.e. the column titles), the format in
the column header format section (see section 14.31.1)
should be used. The specific column headers for far fields
are also listed following this table.

The default column headers will all have the same structure. The column headers will depend on
the far field type that was defined, which can be found in the Result Type value. The following
column headers are defined for each result type:
Gain:
"Theta" "Phi" "Re(Etheta)" "Im(Etheta)" "Re(Ephi)" "Im(Ephi)" "Gain(Theta)" "Gain(Phi)" "Gain(Total)"

Directivity:
"Theta" "Phi" "Re(Etheta)" "Im(Etheta)" "Re(Ephi)" "Im(Ephi)" "Directivity(Theta)" ...
... "Directivity(Phi)" "Directivity(Total)"

RCS (i.e. either a bistatic or monostatic radar cross-section problem):

"Theta" "Phi" "Re(Etheta)" "Im(Etheta)" "Re(Ephi)" "Im(Ephi)" "RCS(Theta)" "RCS(Phi)" "RCS(Total)"

Unknown (i.e. the results do not pertain to an antenna problem, nor to an RCS problem):
"Theta" "Phi" "Re(Etheta)" "Im(Etheta)" "Re(Ephi)" "Im(Ephi)"

July 2011 FEKO User’s Manual

CONTROL CARDS 14-89

14.31.4 *.ol file: Surface charge density

The following fields are available in the Header Block:

Key Required Description

File Type Yes Describes the type of the file. For surface charge density,
the value must be: Charges.
File Format No Denotes the file syntax version (e.g. “1.0”). If not present
it defaults to version 1.0 (files pre-dating Suite 6.1). The
initial new file format version will then be “2.0”.
Source No Denotes the base filename of the source where this data
comes from.
Date No Date of data export in format "YYYY-MM-DD hh:mm:ss"
(i.e. 24-hour format)

The following fields are available in the Solution Block Header:

Key Required Description

Request Name No The explicit name given to that solution request (as de-
noted in the *.pre file). If none is specified, then during
importing such files in POSTFEKO use a default name of
"request_N" (where "_N" is replaced with a number for
each unnamed request).
Frequency Yes Frequency in Hz for which the following data was mea-
sured / computed.
No. of [$$$] Yes The number of mesh elements that follow. The [$$$]
Samples term is replaced by:

• Electric Charge Triangle

• Segment Charge

No. of No Number of header lines to read. The column header lines

Header Lines must follow this line. If this value is not specified, assume
a value of “1”
Yes For the header lines (i.e. the column titles), the format in
the column header format section (see section 14.31.1)
should be used. The specific column headers for far fields
are also listed following this table.

For the column headers of charges, the only difference is an optional additional field that stores
the total surface area of a triangle vs. the total length of a segment.
Triangles:
"Num" "X" "Y" "Z" "Re(Q)" "Im(Q)" [optional] "Surface Area"

Segments:
"Num" "X" "Y" "Z" "Re(Q)" "Im(Q)" [optional] "Length"

July 2011 FEKO User’s Manual

CONTROL CARDS 14-90

14.31.5 *.os file: Surface current density

The following fields are available in the Header Block:

Key Required Description

File Type Yes Describes the type of the file. For surface charge density,
the value must be: Currents.
File Format No Denotes the file syntax version (e.g. “1.0”). If not present
it defaults to version 1.0 (files pre-dating Suite 6.1). The
initial new file format version will then be “2.0”.
Source No Denotes the base filename of the source where this data
comes from.
Date No Date of data export in format "YYYY-MM-DD hh:mm:ss"
(i.e. 24-hour format)

The following fields are available in the Solution Block Header:

Key Required Description

Request Name No The explicit name given to that solution request (as de-
noted in the *.pre file). If none is specified, then during
importing such files in POSTFEKO use a default name of
"request_N" (where "_N" is replaced with a number for
each unnamed request).
Frequency Yes Frequency in Hz for which the following data was mea-
sured / computed.
No. of [$$$] Yes The number of mesh elements that follow. The [$$$]
Samples term is replaced by:

• Electric Current Triangle

• Magnetic Current Triangle
• Segment Current

No. of No Number of header lines to read. The column header lines

Header Lines must follow this line. If this value is not specified, assume
a value of “1”
Yes For the header lines (i.e. the column titles), the format in
the column header format section (see section 14.31.1)
should be used. The specific column headers for far fields
are also listed following this table.

The column headers for currents are dependent on the element type.
Triangles (Electric currents):
"Num" "X" "Y" "Z" "Re(Jx)" "Im(Jx)" "Re(Jy)" "Im(Jy)" "Re(Jz)" "Im(Jz)" ...
... "Abs(Jcorn1)" "Abs(Jcorn2)" "Abs(Jcorn3)" ...
... "Re(Jx_c1)" "Im(Jx_c1)" "Re(Jy_c1)" "Im(Jy_c1)" "Re(Jz_c1)" "Im(Jz_c1)" ...
... "Re(Jx_c2)" "Im(Jx_c2)" "Re(Jy_c2)" "Im(Jy_c2)" "Re(Jz_c2)" "Im(Jz_c2)" ...
... "Re(Jx_c3)" "Im(Jx_c3)" "Re(Jy_c3)" "Im(Jy_c3)" "Re(Jz_c3)" "Im(Jz_c3)"

July 2011 FEKO User’s Manual

CONTROL CARDS 14-91

Triangles (Magnetic currents):

"Num" "X" "Y" "Z" "Re(Mx)" "Im(Mx)" "Re(My)" "Im(My)" "Re(Mz)" "Im(Mz)" ...
... "Abs(Mcorn1)" "Abs(Mcorn2)" "Abs(Mcorn3)" ...
... "Re(Mx_c1)" "Im(Mx_c1)" "Re(My_c1)" "Im(My_c1)" "Re(Mz_c1)" "Im(Mz_c1)" ...
... "Re(Mx_c2)" "Im(Mx_c2)" "Re(My_c2)" "Im(My_c2)" "Re(Mz_c2)" "Im(Mz_c2)" ...
... "Re(Mx_c3)" "Im(Mx_c3)" "Re(My_c3)" "Im(My_c3)" "Re(Mz_c3)" "Im(Mz_c3)"

Segments:
"Num" "X" "Y" "Z" "Re(Ix)" "Im(Ix)" "Re(Iy)" "Im(Iy)" "Re(Iz)" "Im(Iz)"

July 2011 FEKO User’s Manual

CONTROL CARDS 14-92

14.31.6 Other supported file formats

The structure of the remainder of the data files is described below:

*.cgm file In this file the number of iterations is given and the resulting residue from the iterative
solving process of the matrix equation.

*.snp file The Touchstone S-parameter file name contains the number of ports in the model. The
extension is *.s1p for a 1-port, *.s2p for a 2-port and so on. The file contains a header
(following the # character) which specifies the frequency unit, the parameter type, the data
format and the normalising impedance for all the ports. This is followed by the data lines
(which may be repeated for multiple frequencies):
1-port: f, |S11 |, ∠S11
2-port: f, |S11 |, ∠S11 , |S21 |, ∠S21 , |S12 |, ∠S12 , |S22 |, ∠S22
3-port: f, |S11 |, ∠S11 , |S12 |, ∠S12 , |S13 |, ∠S13
|S21 |, ∠S21 , |S22 |, ∠S22 , |S23 |, ∠S23
|S31 |, ∠S31 |S32 |, ∠S32 , |S33 |, ∠S33
4-port: f, |S11 |, ∠S11 , |S12 |, ∠S12 , |S13 |, ∠S13 , |S14 |, ∠S14
|S21 |, ∠S21 , |S22 |, ∠S22 , |S23 |, ∠S23 , |S24 |, ∠S24
|S31 |, ∠S31 , |S32 |, ∠S32 , |S33 |, ∠S33 , |S34 |, ∠S34
|S41 |, ∠S41 , |S42 |, ∠S42 , |S43 |, ∠S43 , |S44 |, ∠S44
where |S11 | is the absolute value and ∠S11 the phase (in degrees) of the given parameter.
Note that the 2-port file is formatted on a single line and in a different order than for cases
with more ports. This is consistent with the Touchstone file format specification version
1.0.

*.sph file This file is using the native SWE file format of TICRA as used in their code GRASP.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-93

14.32 DI card

This card can be used to define the frequency dependent or independent material characteristics
of a dielectric medium, metallic medium or an impedance sheet. The DI card is used for the
MoM/MLFMM when using the surface current or volume current methods, or also for the FEM.
Parameters:

Material label: Label of the material to be defined.

Define dielectric medium: A dielectric medium can be defined by specifying frequency independent
parameters, a frequency dependent dielectric model or loading points from a
file.

Define metallic medium: A metallic medium can be defined by specifying frequency independent pa-
rameters or loading points from a file.

Define impedance sheet: An impedance sheet can be defined by specifying the frequency independent
real and imaginary part or loading points from a file.

Source: The medium is defined by means of a wideband model or loading points from
a file (see section 6.1.1).

Define dielectric medium

With this option a user specify the dielectric model used to define a dielectric medium.

Dielectric modelling: Constant (Frequency independent)

Relative permittivity Relative permittivity ε r of the medium.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-94

Dielectric loss tangent Dielectric loss tangent tanδ of the medium (this is al-
ternative way to specify the conductivity σ — the two loss terms are
related by tanδ= ωεσ ε and have different frequency behaviour).
r 0
1
Conductivity Conductivity σ in Ωm
of the medium.

Dielectric modelling: Debye relaxation

Relative static permittivity Relative static permittivity εs of the medium

High frequency dielectric constant High frequency dielectric constant ε∞ of
the medium.
Relaxation frequency The relaxation frequency f r of the medium.

Dielectric modelling: Cole-Cole

Relative static permittivity Relative static permittivity εs of the medium

High frequency dielectric constant High frequency dielectric constant ε∞ of
the medium.
Relaxation frequency The relaxation frequency f r of the medium.
Attenuation factor Attenuation factor α of the medium.

Dielectric modelling: Havrillak-Negami

Relative static permittivity Relative static permittivity εs of the medium

High frequency dielectric constant High frequency dielectric constant ε∞ of
the medium.
Relaxation frequency The relaxation frequency f r of the medium.
Attenuation factor Attenuation factor α of the medium.
Phase factor Phase factor β of the medium.

Dielectric modelling: Djordevic-Sarkar

Variation of relative permittivity Variation of the relative permittivity ∆ε of

the medium.
Relative high frequency permittivity Relative high frequency permittivity ε∞
of the medium.
1
Conductivity Conductivity σ in Ωm
of the medium.
Lower limit of angular frequency The lower limit of the angular frequency
for the medium, ω1 .
Upper limit of angular frequency The upper limit of the angular frequency
for the medium, ω2 .

Dielectric modelling: Specify points in the *.pre file.

Specify points in the *.pre file. The dielectric properties of the medium are
specified in the *.pre file.

Dielectric modelling: Load points from file (linear interpolation)

A dielectric medium may also be defined by importing data points from file
(see section 6.1.1).

July 2011 FEKO User’s Manual

CONTROL CARDS 14-95

Define magnetic medium

With this option a user specify the magnetic model used to define a dielectric or metallic medium.

Magnetic modelling: Constant (Frequency independent)

Relative permeability Relative permeability µ r of the medium.

Magnetic loss tangent Magnetic loss tangent tan δµ of the medium specified
in Set properties for medium number (the complex permeability is then
given by µ = µ0 µ r (1 − j tan δµ )).

Magnetic modelling: Constant (Frequency independent)

Non magnetic The relative permeability µ r of the medium is set equal to 1

and the magnetic loss tangent tanδµ is set equal to 0.

Magnetic modelling: Specify points in the *.pre file

Specify points in the *.pre file. The magnetic properties of the medium are
specified in the *.pre file.

Magnetic modelling: Load points from file (linear interpolation)

A magnetic medium may also be defined by importing data points from file
(see section 6.1.1)

July 2011 FEKO User’s Manual

CONTROL CARDS 14-96

Define impedance sheet

With this option an arbitrary user defined complex surface impedance Zƒ can be used in FEKO. It
must be noted that the impedance boundary condition for the MoM (also then for MLFMM etc.)
has certain limitations regarding the range of validity. FEKO uses whatever the user has specified
as surface impedance, and it is the responsibility of the user to ensure that the application of an
IBC is still warranted for the specific configuration (dependent on the impedance value as such,
but also frequency, radius of curvature of the structure etc.).
Ω
For surfaces the unit of Zƒ is Ω. For wire structures, the value used by FEKO is in units of m and
results from the surface impedance expression by dividing it by π% where % represents the wire
radius.

Ω
Real part: The real part of the surface impedance Zƒ in Ω (for triangles) or in m
for wires.
Ω
Imaginary part: The imaginary part of the surface impedance Zƒ in Ω (for triangles) or in m
for wires.

Load points from file (linear interpolation): An impedance sheet may also be defined by importing
from file (see section 6.1.1).

July 2011 FEKO User’s Manual

CONTROL CARDS 14-97

14.33 DL card

This card is used to define an isotropic or anisotropic layered medium by specifying the label of
the material to be used for each layer.
Parameters:

Layered medium type: In this group the user must select one of two options:

• Isotropic layered medium

• Anisotropic layered medium

Isotropic layered medium

Parameters:

Number of layers: The number of isotropic layers defined for the layered medium.

Thickness of this layer: The thickness of the current layer in m (if an SF card is present, this is always
scaled).

Material label (dielectric): The label of the material to be used for the current layer (as defined in the
DI card).

Anisotropic layered medium

Parameters:

Number of layers: The number of anisotropic layers defined for the layered medium.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-98

Thickness of this layer: The thickness of the current layer in m (if an SF card is present, this is always
scaled).

Angle of principal direction: The angle (in degrees) from which the principal direction is obtained.

Material in principal direction: The material label of the material to be used in the principal direction.

Material in orthogonal direction: The material label of the material to be used in the orthogonal di-
rection.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-99

14.34 EE card

This card requests an a-posteriori error indicator whereby FEKO can test the solution against an
unconstrained physical test. The result is to give an indication of the region where local mesh
refinement should be considered.

Parameters:

No error estimation: No error estimation output.

Error estimation on all mesh elements: Output error estimation on all mesh elements.

Error estimation on mesh elements with specified label(s): Output error estimation on mesh elements
with labels specified by user.

Error estimation on mesh elements with label range: Output error estimation in the label range starting
on Start label and ending on End label.

Only error estimation on triangles: Output only error estimation on triangles.

Only error estimation on segments: Output only error estimation on segments.

Only error estimation on cuboids: Output only error estimation on cuboids.

Only error estimation on tetrehedra: Output only error estimation on tetrahedra.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-100

14.35 EN card

This card indicates the end of the input file. It is essential and has no parameters.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-101

14.36 FE card

This card controls the calculation of the near fields.

Parameters:

First drop-down list: Select what to calculate:

• No field calculation: Field is not calculated.

~.
• Electric field values: Calculate the electric field, E
~
• Magnetic field values: Calculate the magnetic field, H.
• Both electric field and magnetic field values: Calculate both electric and
magnetic fields.
• Electric field and SAR values in cuboids: The electric field and SAR values
in the dielectric volume elements. For this option, no other parameters
are required.
~.
• Magnetic vector potential: Calculate the magnetic vector potential, A
• Gradient of the scalar electric potential: Calculate the gradient of the
scalar electric potential, ∇ ϕ.
• Electric vector potential: Calculate the electric vector potential, F~ .
• Gradient of the scalar magnetic potential: Calculate the gradient of the
scalar magnetic potential, ∇ ψ.

Calculate only the scattered part of the field: When this item is checked, only the scattered part of the
field / potential is written to the output file. Otherwise the total field / po-
tential, the sum of the scattered and source contributions, are written to the
output file. Note that depending on the used formulation in FEKO the region
where the incident field as computed from the impressed sources is present
might be different (for instance when using the surface equivalence formula-
tion in the MoM to model dielectric bodies, then each source acts as incident
field only in the medium where this source is located).

Coordinate system: In this dialog, the coordinate system for the calculation of the requested fields
is specified.
If Specified points is selected here, additional dialogs for the Number of field

July 2011 FEKO User’s Manual

CONTROL CARDS 14-102

points and the Coordinates are shown. One must then enter the coordinates of
each point.
For all other cases, additional dialogs for Starting values, Increment and No. of
points are shown — each with three values for the three components of the
respective coordinate system.

Use old output format: If this item is checked, the old format of the near field is used in the *.out
output file. This should only be used for compatibility with third party post
processors. (POSTFEKO cannot extract SAR values from near fields in this
format).

Note that all coordinates are in metres and all angles in degrees. Scaling with the SF card is only
applicable when the option Modify all dimension related values is selected (default behaviour
and highly recommended) in the SF card — in this case coordinates must be in metre after
scaling.
Potentials cannot be computed with the FE card if UTD or PO is used. Also only the free space
Green’s function is supported - not the Green’s functions for layered spheres or multilayered
planar media.
If the total potentials are requested, the potentials for the sources are added. These are not
available for a plane wave (A0 card) or an impressed radiation pattern (AR card) and FEKO will
~ and the
give an error. For a magnetic dipole (A6 card) the electric ring current model yields A
magnetic current yields F~ and ∇ ψ; all the other potentials are zero.
~ and ∇ ϕ are written to the
If one requests *.efe and/or *.hfe files with the DA card, then A
*.efe file, while F~ and ∇ ψ are written to the *.hfe file.
If a ground plane is used, a calculation of the near fields in the ground plane is not possible. The
observation points in the area z < 0 are not taken into account.
It should be noted that the coordinates may have an offset (OF card). Thus the near field on the
surface of a sphere can be calculated, with the centre of the sphere not being located at the origin
of the coordinate system.
The different Coordinate systems are described on the following pages.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-103

• Cartesian coordinates x, y, z

Figure 14-19: Field calculation in the Cartesian coordinate system

Observation Point:  
x
~r =  y  (14-50)
 
z
Unit vectors of the coordinate system:
     
1 0 0
x̂ =  0  ŷ =  1  ẑ =  0  (14-51)
     
0 0 1

• Cylindrical coordinates around z axis ρ, ϕ, z

Figure 14-20: Field calculation in the Cylindrical coordinate system

Observation Point:
ρ cos ϕ
 

~r =  ρ sin ϕ  (14-52)
 
z
Unit vectors of the coordinate system:

cos ϕ − sin ϕ
     
0
ρ̂ =  sin ϕ  ϕ̂ =  cos ϕ  ẑ =  0  (14-53)
     
0 0 1

July 2011 FEKO User’s Manual

CONTROL CARDS 14-104

• Spherical coordinates r, ϑ, ϕ

Figure 14-21: Field calculation in the Spherical coordinate system

Observation Point:
r sin ϑ cos ϕ
 

~r =  r sin ϑ sin ϕ  (14-54)

 
r cos ϑ
Unit vectors of the coordinate system:

sin ϑ cos ϕ cos ϑ cos ϕ − sin ϕ

     

r̂ =  sin ϑ sin ϕ  ϑ̂ =  cos ϑ sin ϕ  ϕ̂ =  cos ϕ  (14-55)

     
cos ϑ − sin ϑ 0

• Cylindrical coordinates around the x axis r, ϕ, x

Figure 14-22: Field calculation in the Cylindrical coordinate system around the x axis

Observation Point:  
x
~r =  ρ cos ϕ  (14-56)
 
ρ sin ϕ
Unit vectors of the coordinate system:
     
0 0 1
ρ̂ =  cos ϕ  ϕ̂ =  − sin ϕ  x̂ =  0  (14-57)
     
sin ϕ cos ϕ 0

July 2011 FEKO User’s Manual

CONTROL CARDS 14-105

• Cylindrical coordinates around the y axis r, ϕ, y

Figure 14-23: Field calculation in the Cylindrical coordinate system around the y axis

Observation Point:
ρ cos ϕ
 

~r =  y (14-58)
 

−ρ sin ϕ
Unit vectors of the coordinate system:

cos ϕ − sin ϕ
     
0
ρ̂ =  0 ϕ̂ =  0 ŷ =  1  (14-59)
     
 
− sin ϕ − cos ϕ 0

• Conical coordinates around the z axis ϕ, z

Figure 14-24: Field calculation in the Conical coordinate system

This option is similar to the field calculation in cylindrical coordinates around the z axis,
where the radius r changes with the height z
∆r
r(z) = r0 + · (z − z0 ), (14-60)
∆z
and z lies within the range z0 . . . z0 + nz · ∆z.
Observation Point:  € 
∆r
Š
r 0 + ∆z
· (z − z 0 ) · cos ϕ
∆r
 € Š 
~r =  r0 + ∆z · (z − z0 ) · sin ϕ
 
 (14-61)
z
Unit vectors of the coordinate system:

cos ϕ − sin ϕ
     
0
r̂ =  sin ϕ  ϕ̂ =  cos ϕ  ẑ =  0  (14-62)
     
0 0 1

July 2011 FEKO User’s Manual

CONTROL CARDS 14-106

14.37 FF card

This card controls the calculation of the far fields in spherical coordinates.

Parameters:

No calculation: If this item is checked, no calculation is done.

Fields calculated as specified below: The far field is calculated with the specified settings.

Fields calculated only in the incident direction: The far field is calculated only in the incident direction
(used, for example, to calculate monostatic RCS).

Only integrate field over area given below: The far field is calculated but it is not written to the output
file in order to limit its size. This option is meaningful when the individual
values of the field strength (such as directivity and gain) are not required, but
the total radiated power has to be calculated from the integral of the Poynting
vector (see the discussion below), or if one is just interested in the modal
coefficients. If a *.ffe file has been requested with the DA card, the field
values used in this integration will be written to the file.

Calculate only the scattered part of the field: When this item is checked, the field radiated by the im-
pressed sources (such as Hertzian dipoles) is not included. This option is only
meaningful if only the scattered field is required. Normally one would not
check this item so that the total field is calculated. This includes all source
contributions except plane wave excitations.

Directivity / Gain: Select which quantity is required.

Number of ϑ points: The number of observation points in the ϑ direction. An empty field will be set
to 1.

Number of ϕ points: The number of observation points in the ϕ direction. An empty field will be set
to 1.

Initial ϑ: The angle ϑ0 in degrees of the first observation point.

Initial ϕ: The angle ϕ0 in degrees of the first observation point.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-107

ϑ increment: Increment ∆ϑ in degrees of the angle ϑ.

ϕ increment: Increment ∆ϕ in degrees of the angle ϕ.

Compute spherical mode coefficients: Check this item if the spherical mode coefficients of the far field
should be calculated with the FF card. The Maximum mode index N should be
specified if this option is chosen. More information on this option is supplied
below.

Calculate far field for array (for PBC calculations): Check this item if the far field is to be calculated
for an array of elements. The Number of elements in û1 direction (PE card)
and the Number of elements in û2 direction (PE) card should be specified if
this option is chosen.

Information regarding the output format and derived quantities available from fields calculated
at an FF card can be found in FEKO User Manual (section 19.6).
When calculating the monostatic radar cross section for a number of directions of incidence, the
parameter Fields calculated only in the incident direction is necessary, otherwise Fields calculated
as specified below can be used.
When using the FF card with Number of ϑ points, Nϑ , and Number of ϕ points, Nϕ , both larger
than 1, the Poynting vector is integrated over the two spherical segments:

• ϑ0 − 21 · ∆ϑ ≤ ϑ ≤ ϑ0 + (Nϑ − 12 ) · ∆ϑ and ϕ0 − 21 · ∆ϕ ≤ ϕ ≤ ϕ0 + (Nϕ − 21 ) · ∆ϑ

• ϑ0 ≤ ϑ ≤ ϑ0 + (Nϑ − 1) · ∆ϑ and ϕ0 ≤ ϕ ≤ ϕ0 + (Nϕ − 1) · ∆ϕ

In the case of an antenna the power provided by the voltage sources must be equal to the radiated
power over the whole sphere. The total radiated power can be calculated using for instance the
following commands:
** Far field integration in angular increments of #delta (in degrees)
#delta = 5
#nt = 180 / #delta + 1
#np = 360 / #delta + 1
FF 3 #nt #np 0 0 0 #delta #delta

The output in the *.out file then reads for example

Integration of the normal component of the Poynting vector in the
angular grid DTHETA = 5.00 deg. and DPHI = 5.00 deg. ( 2701
sample points)
angular range THETA angular range PHI radiated power
-2.50 .. 182.50 deg. -2.50 .. 362.50 deg. 5.60499E-03 Watt
0.00 .. 180.00 deg. 0.00 .. 360.00 deg. 5.52821E-03 Watt

If the problem is symmetrical, it is not necessary to carry out the integration over the complete
sphere. If there are three planes of symmetry (as for a simple dipole in free space) the integration
only needs to be done over an eighth of the sphere. The power then has to be multiplied by 8.
If a ground plane has been specified, the calculation of the far fields below the ground plane is
not possible. Observation points with z < 0 will thus be ignored.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-108

Spherical mode coefficients

can be computed while doing a far field computation. The calculation of these coefficients is
based on the far field values and they are written to the FEKO output file.
Doing the integration requires a two-dimensional far field computation (i.e. both Number of ϑ
points and Number of ϕ points larger than 1) over the whole sphere (i.e. 0 ≤ ϑ ≤ 180◦ and
0 ≤ ϕ ≤ 360◦ ) and the angular increments should be chosen according to the desired accuracy
and number of modes (i.e. a finer stepping is required for higher modes). It is suggested to do
some initial convergence study (e.g. with increments 5◦ and then 1◦ ) to see the sensitivity of the
results.
Spherical modes have three indices s, m, and n with s = 1 for TE-modes, s = 2 for TM-modes, m =
−N . . . N , and n = 1 . . . N (see also more detailed description at the AS card (see section 14.20),
including a one-dimensional compressed indexing scheme j = 1 . . . J and the normalisation of
the modes etc.). The input parameter Maximum mode index N determines the maximum mode
index N , i.e. a total number of J = 2N (N + 2) modes will be computed.
The modes origin is the same as for the far field computation in general (i.e. this is the global
origin unless an OF card has been used to specify an offset). It should also just be mentioned that
due to the nature of the far field (propagating towards r = ∞) all computed mode coefficients
refer to spherical cylinder functions zn(c) with type c = 4 (see more details at the AS card).
It shall be mentioned that when spherical modes are computed with the FF card, the DA card
can also be used (must be in front of the FF card) in order to request that this spherical mode
expansion is exported to an SWE file (extension *.sph) which can be imported into the computer
code GRASP from TICRA. GRASP is a reflector antenna modelling code, and by means of this SWE
file export one can for instance model a horn antenna as feed in FEKO and then export this feed
structure and use in GRASP. Note that the gain in GRASP will only be correct if the radiated power
in FEKO has been set to 4π Watts. The spherical mode expansion coefficients Q smn as used by
FEKO are described at the AS card (see section 14.20). In GRASP a slightly different convention
is used:

1
• An additional factor p
8π
is used

• The coefficients are conjugate complex (i.e. GRASP assumes an e−iωt time dependency)

• The index m is exchanged with −m (since in FEKO the ϕ dependency is defined differently
than in GRASP, e jmϕ versus e− jmϕ )

These conversions are done automatically by FEKO when exporting the SWE file, so that this can
readily be imported into GRASP.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-109

14.38 FR card

This card sets the frequency/frequencies (in Hz), at which the solution will be obtained.

The solution can be done for a single frequency, a loop of discrete frequencies (linear or multi-
plicative stepping), or a continuous solution in a given frequency band with adaptive frequency
interpolation with the option to set the convergence accuracy. For a continuous solution, only
one FR card is allowed. Specific quantities of interest to the user may also be selected to be
included for adaptive sampling. Unselected quantities will be calculated at the discrete solution
frequency points.
Parameters:
Single frequency: Only a single frequency will be analysed.

Discrete frequency points: If this item is selected the following parameters are applicable:

• Number of frequency points: For a discrete frequency sweep, the number

of frequency samples must be larger than 1.
• Frequency scale: In this group either Linear or Multiplicative scaling is
selected. If Linear is selected, then consecutive frequencies differ with
a fixed value, i.e. the new frequency is the previous value plus the fre-
quency increment. If Multiplicative is selected, then consecutive frequen-
cies differ by a constant factor, i.e. the new frequency is the previous
value multiplied by the frequency factor.
• Specify by: If Frequency increment / factor is selected, the user speci-
fies the increment or factor mentioned in the previous item. This, the
number of frequencies and the start frequency then determine the end-
ing frequency. If Ending frequency is selected, the user specifies this and
the increment / factor is calculated.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-110

Continuous data: Select this item to use an adaptive frequency interpolation technique to ob-
tain a continuous representation of the results in the given frequency band
(see FEKO User Manual (section 21)). When using this feature, the remaining
parameters have the following meaning:

• Adaptive frequency sampling convergence accuracy: This allows the user

to manually set the adaptive frequency sampling convergence accuracy.
Note that the default setting should be used for most cases, but in special
cases where a structure with many resonances is being modelled a higher
convergence accuracy should be used.
• Max number of sample points: Maximum number of discrete frequency
points in this frequency band at which FEKO may be executed (limitation
to avoid convergence problems). If left empty, the default value of 1000
will be used.
• Select quantities to include for adaptive frequency sampling: The quan-
tities selected will be be included in the adaptive frequency sampling.
• For CableMod/CRIPTE/PCBMod/Touchstone . . ., Number of discrete fre-
quencies . . .: This field is only relevant when the CM or SP card is used
to create an *.isd or *.snp file respectively. The results are written to
the *.isd or *.snp file for the number of discrete frequencies specified
in this field.
• Starting / ending frequency: Defines the frequency range.
• Min. frequency stepping: Minimum increment between adaptive sam-
ples, see the note below.

In order to obtain a continuous frequency response, the adaptive frequency interpolation tech-
nique obtains the solution at a set of discrete frequency points. They are automatically placed,
for example using large frequency increments in regions with a smooth behaviour of the results,
and much finer frequency increments close to resonances. Sometimes, for example when using a
frequency dependent mesh, the FEKO results versus frequency may contain small discontinuities.
In these cases the adaptive algorithm cannot converge. (It will continue to refine the frequency
increment as it tries to fit a smooth curve through the discontinuity and will only stop when
the Max. number of sample points is reached.) One may avoid this by setting the Frequency in-
crement to the minimum allowable separation distance between neighbouring frequency sample
points. The value of the Frequency increment must be smaller than the resolution required to
solve, for example, sharp resonances. If left empty, the default is:

1
(Ending frequency − Starting frequency) (14-63)
10000

If a discrete loop with more that one frequency is required then either the frequency increment
or the ending frequency must be specified, but not both. If the end frequency is specified, the
frequency increment is calculated from:

• for a linear frequency scale (additive increments):

f2 − f1
∆f = (14-64)
Nf − 1

July 2011 FEKO User’s Manual

CONTROL CARDS 14-111

• for a multiplicative frequency scale (multiplicative increments):

1
f2
 
N f −1
f ac = (14-65)
f1

where ∆ f is the frequency increment (linear stepping), f ac the increment factor (multiplicative
stepping), f1 the start frequency, f2 the ending frequency and N f the number of frequencies.
When writing results at discrete frequencies to a *.isd file, the frequency increment when a
linear frequency scale is used is calculated similar to the case for ∆ f as shown above.
If more than one frequency is to be examined, then all the control cards up to the next FR card
or EN card will be read into a buffer and are executed for each frequency.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-112

14.39 GF card

With this card the Green’s function may be selected. The Green’s function relates the fields in
space to the sources present. The propagation space is usually free space, and FEKO uses the
free space Green’s function by default. For a few specific interaction environments, some of the
complexities of the environments can be taken into account using special Green’s functions. The
Green’s functions that FEKO supports are:

• Homogeneous medium: The dielectric properties for the entire problem space can be set.
This is useful for modelling in a homogeneous medium that differs from free space.

• Layered dielectric sphere: A layered dielectric sphere located at the origin is taken into
account with the Green’s function.

• Planar multilayer substrate: A multilayer dielectric substrate in the x y-plane is taken into
account. The substrate can have metallic ground planes above and below it.

Using Green’s functions to model the presence of these dielectric regions means that their in-
fluence is taken into account implicitly, using less computational resources than modelling them
using either the volume- or surface equivalence principles.
The dialogs for each of the three cases above are discussed separately below.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-113

Homogeneous medium

When this Green’s function is selected, the EM problem under investigation is modelled inside
an infinite space of the homogeneous medium. This is the standard “free space” Green’s function
similar to when the GF card is not used. The medium is normally free space, defined by material
label “0”, but different parameters can be set with the DI card.
Parameters:

Material label: The label of the homogeneous medium to be used, as defined in the DI card.

Metallic structures can have an arbitrary orientation (horizontal, vertical, and also diagonal).
They can lie at an arbitrary position in one or more layers and can lie directly on the border of
two layers (e.g. on the surface of a substrate). The only restriction is that no metallic segment or
triangle may cross a boundary between layers, i.e. it must lie completely within one layer or at
the boundary between layers.
If, for example, a metallic wire penetrates a multilayer substrate, the segmentation must be so
that there is a node on each interface between layers. (See example_31 in the Script Examples.)
The following is not possible with this Green’s function:

• dielectric ground

• hybrid MoM/PO method

• hybrid MoM/UTD method

• dielectric bodies with the surface or volume equivalence principle

July 2011 FEKO User’s Manual

CONTROL CARDS 14-114

Layered dielectric sphere

When this Green’s function is selected, the EM interaction of a layered dielectric sphere located
at the coordinate system centre is taken into account. With this option it is, for example, possible
to analyse a cellphone in front of a spherical shell model of the human head very efficiently.
Parameters:

Configuration list: The drop-down list allows selecting between Single dielectric sphere and a
core with a number of layers. Note that whether metal structures are allowed,
influences the number of layers that are allowed.

Allow metal structures inside sphere: When this item is checked metallic structures can be present in
the inner parts of the sphere.

Use interpolation: When this item is checked interpolation (*.gfe and *.gfh files) is used to
accelerate the computations.

Convergence criterion: Convergence criteria for the summation of the rows of Green’s functions. If
this field is 0 or undefined, a sensible standard criterion is used.

Radius: Radius of the sphere / layer in m (is scaled by the SF card). For the layers,
this is the total radius of the core plus layers up to that point. The highest
numbered layer is outside.

Material label: Label of the material (as defined in the DI card) to be used for the core / layer.

The scaling factor that is entered by the SF card is applied to the radius. Note that the surrounding
medium is defined by material label “0”. By default the values of free space is used, but these
parameters can be redefined in the DI card.
The Green’s function for a homogeneous or layered dielectric sphere can be used with metallic
structures (treated with the MoM) either inside or outside the sphere (but not for example a
wire from inside to outside). It can be used with dielectric bodies treated with the volume
equivalence principle (e.g. the hand of a user around a mobile phone), but the dielectric bodies
must be outside the sphere.
An example of the use of the GF card for a sphere is given in example_15 (Script Examples).

July 2011 FEKO User’s Manual

CONTROL CARDS 14-115

x
4
3
2
1

Figure 14-25: Example of a sphere consisting of 4 media (core and 3 layers) indicating the layer number-
ing.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-116

Planar multilayer substrate

When this Green’s function is selected, the EM interaction of a layered dielectric substrate located
in the x y-plane is taken into account. The Green’s function can also take the effect of a top and/or
bottom ground plane into account. This formulation has been popularised by its application to
planar (microstrip) circuits and antennas, but it is applicable to a larger class of EM problems
e.g. buried antennas.
Parameters:

Number of layers: Number of layers in the substrate (layer 0 — the upper half-space) is not in-
cluded in the number.

Thickness: Thickness of the layer (is scaled by the SF card).

Material label: Label of the material to be used for the layer (as defined in the DI card).

Bottom ground plane: By checking “Bottom ground plane” at the respective layer, a user can choose
to have a ground plane at the bottom of the layer.

Z-value at the top of layer 1: The value of the z coordinate at the transition between layer 0 and layer
1.

Number of additional conducting ground planes: Number of additional conducting ground planes to be
added at arbitrary z-values.

Advanced ground planes at arbitrary z-values: The z-value at which the ground plane is to be added.

As indicated in the figure of the card, layer 0 is the upper half-space, layer 1 is just beneath this,
etc.
Metallic structures can have an arbitrary orientation (horizontal, vertical, and also diagonal).
They can lie at an arbitrary position in one or more layers and can lie directly on the border of
two layers (e.g. on the surface of a substrate). The only restriction is that no metallic segment or

July 2011 FEKO User’s Manual

CONTROL CARDS 14-117

triangle may cross a boundary between layers, i.e. it must lie completely within one layer or at
the boundary between layers.
If, for example, a metallic wire penetrates a multilayer substrate, the segmentation must be so
that there is a node on each interface between layers. (See example_31 in the Examples Guide.)
The following is not possible with this Green’s function:

• dielectric ground (BO Card)

• hybrid MoM/PO method

• hybrid MoM/UTD method

• hybrid MoM/FEM method

• dielectric bodies with the surface or volume equivalence principle

July 2011 FEKO User’s Manual

CONTROL CARDS 14-118

14.40 L2 card

The L2 card allows a load (complex impedance) to be placed on a node in the model.

Parameters:

Select segment: When this item is selected, then the Segment label field becomes active. Here
one specifies the label of the segment which shall be loaded (either start or
end point as determined by the corresponding check box). The load has to be
located at a node, either between two segments, or between a segment and a
triangle, ground plane or polygonal plate. Only one segment with this label
should be declared. If there is more than one segment with this label then all
the corresponding segment vertices will be loaded.

Set load position: When this check box is activated, then the load node is determined by speci-
fying its Cartesian coordinates in the Coordinates of node group. These values
are in m and may be scaled by the SF card.

Load at start of segment: This option is only available when selecting the feed segment by label. If
set, it indicates that the load location is at the start of the wire segment with a
matching label.

Source at end of segment: This option is only available when selecting the feed segment by label. If
set, it indicates that the load location is at the end of the wire segment with a
matching label.

Real part of impedance: The real part of the complex impedance in Ω.

Imaginary part of impedance: The imaginary part of the complex impedance in Ω.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-119

14.41 L4 card

This card can be used to add a load between a metallic triangle and the ground plane for the
planar multilayer Green’s function without having the requirement to model a vertical current
element (analogous to the A4 excitation card).

Parameters:

Define a load at a coaxial attachment point: Define a load with the following parameters.

Remove all L4 type loads previously defined: This L4 card does not define a load, but rather all previ-
ously defined L4 loads are deleted. All the other input parameters of this card
are ignored.

Select element: The label of the triangle to load. If there is more than one triangle with this
label, the one with the highest element number is loaded.

Set source position: Alternatively, the user may specify the Cartesian coordinates x, y and z of the
load point in the Select source coordinates group. The triangle with the centre
point closest to this point is loaded.

Transform impedance to ground plane: The specified impedance refers to the metallic ground plane,
and a transformation must be done to get the correct load impedance at the
triangle.

Real part of impedance: Real part of the load impedance (in Ω).

Imaginary part: Imaginary part of the load impedance (in Ω).

Radius of the connection pin: The radius of the load pin in m.

The source coordinates (if entered) and the radius of the connection pin (if entered) is optionally
scaled by the SF card.
If an L4 card is processed any existing L4 load on that triangle is replaced. For example, if two L4
cards are used directly below each other, the first specifying a 50 Ω and the second a 20 Ω load,
the segment will be loaded with 20 Ω, not 70 Ω.
It must also be noted that if the L4 card is used in conjunction with an A4 card (impressed current
source), the load impedance of the L4 card is placed parallel to the input impedance of the A4
source (it has no effect if it is in series with the current source), i.e. the resulting input admittance
is the sum of input admittance without the L4 load, and the admittance added by the L4 card.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-120

14.42 LC card

With the LC card, complex, series and parallel circuits can be applied between connector pins
and also between a connector pin and ground can be defined.

Parameters:

Define/replace a load at a cable connector: Define/replace a load at a cable connector with the follow-
ing parameters.

Remove a load at a cable connector: A load can be removed between two connector pins by defining
the Connector label and Pin number. A load may also be removed between
ground and a pin by checking the PEC checkbox.

Remove all cable loads previously defined: All previously defined LC type loads are removed.

Define/replace load between: A load can be placed between two connector pins by defining the Con-
nector label and Pin number. A load may also be placed between ground and
a pin by checking the PEC checkbox.

Complex impedance: The real and imaginary part of the complex impedance in Ω.

Series circuit: The resistor value in Ω, inductor value in Henry and the capacitor value in
Farad to be added as a series circuit.

Parallel circuit: The resistor value in Ω, inductor value in Henry and the capacitor value in
Farad to be added as a parallel circuit.

For detailed conductor to cable connector pin relation, see the AK card (section 14.16).

July 2011 FEKO User’s Manual

CONTROL CARDS 14-121

14.43 LD card

With this card it is possible to specify a distributed resistive, capacitive or inductive loading or
even a series combination of these values for a segment.

Parameters:

Label of segments . . .: All segments with this label are subjected to distributed loading.
Ω
Resistance: The distributed resistance in m
.
H
Inductance: The distributed inductance in m
.
F
Capacitance: The distributed capacitance in m
.

The combined impedance of the segment with length l is then

1
 
Zs = R + jωL +
0 0
· l. (14-66)
jωC 0

It should be noted that if the Capacitance (F/m) is left empty, it is treated as infinite, so that it
does not contribute to the impedance.
The LD card may be combined with the LP, LS, LZ and the SK cards, but only one LD card may
be used per label. If a second LD card is used, it replaces the values entered by the first one. This
card has no significance for surface elements, even when these are assigned the same label.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-122

14.44 LE card

With the LE card an edge between surface triangles can be loaded with an impedance Z = R+ jX ,
as shown in Figure 14-26.

Negative side

Positive side

Load impedance
Z = R + jX

Figure 14-26: Application of the LE card

Parameters:

Load edge between regions with multiple labels: The edge between different regions is loaded with a
complex impedance. The following parameters apply when this item is checked:

• Maximum number of labels on a side: This indicates the maximum num-

ber of different labels used on either side of the load. Increasing this
value will add rows to the table for the labels.
• Negative side: The labels of triangles on the one side of the load.
• Positive side: The labels of triangles on the other side of the load.

Load an edge connected to ground/UTD: Load the triangles with specified labels that are connected to
a UTD surface or to a PEC ground plane (as specified with a BO or GF card).
The following parameters apply when this item is checked:

• Maximum number of labels on a side: This indicates the maximum num-

ber of different labels used on the negative side of the load. Increasing
this value will add rows to the table for the labels.
• Negative side: The labels of the triangles that are connected to the ground.

Load microstrip edge between two points: This is a special microstrip port load. The load is placed on
all edges on the line between two points (previously specified with DP cards)
entered into the dialog below. A GF card with a conducting ground plane must
be present. The following parameters apply when this item is checked:

• Start point of edge: The start point (not label) of the edge.
• End point of edge: The end point (not label) of the edge.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-123

Real part of impedance: Real part of the complex load impedance.

Imaginary part of impedance: Imaginary part of the complex load impedance.

See also the AE card for the excitation of such an edge. As shown in Figure 14-26 the edge can
consist of several single edges between regions specified by labels (AE card (see section 14.13)
provides a discussion of the allowed configurations.) Alternatively the edge can be along a con-
nection between triangles and a polygonal plate or a PEC ground plane, or it can be a microstrip
feed line port. The impedance Z applies to the complete edge (i.e. all the single edges in paral-
lel). The LE card can be combined with the AE card to specify both an impedance and a voltage
source over the edge.
Note that the edge between the triangles does not need to be straight. One may, for example,
specify a resistive connection between two half cylinders.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-124

14.45 LF card

This card can be used to impress a complex impedance between two points inside a FEM mesh.

Parameters:

Real part of impedance: The real part of the complex impedance in Ω.

Imaginary part: The imaginary part of the complex impedance in Ω.

Load position: The Cartesian coordinates of the start and end points of the load.

The complex impedance is impressed between two points inside the FEM mesh. The line may be
positioned arbitrarily inside the FEM mesh, i.e. it does not have to be coincident with tetrahedral
edges, but the length between the points should be small compared to the shortest wavelength
in the band of interest to obtain reasonable accuracy.
The LF type load should not be used to model a short-circuit load. A short can be modelled by a
metallic strip (meshed into triangles) inside a FEM region.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-125

14.46 LN card

This card can be used to add a complex load to any non-radiating network port that is not
connected to geometry (i.e. any non-radiating network of the type Internal).

Parameters:

Define a load at a network port: Define a network load with the following parameters.

Remove all LN type loads previously defined: All previously defined LN type loads are removed. This re-
places all network loads with open circuits. Note that setting the load impedance
to zero creates a short circuit between the network terminals.

Network name: The network or transmission line name, with the network port number, uniquely
identifies the connection terminals.

Network port number: The network port number, with the network or transmission line name, uniquely
identifies the connection terminals.

Real part of impedance: The real part of the complex impedance in Ω.

Imaginary part: The imaginary part of the complex impedance in Ω.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-126

14.47 LP card

This card can be used to assign discrete circuit elements (in parallel) to a segment. Figure 14-27
shows the parallel circuit that can be assigned to a segment.

Figure 14-27: Sketch of the parallel circuit

Parameters:

Label of segments . . .: All segments with this label are assigned the parallel circuit values specified
below.

Resistor value: Value of the resistor in Ω.

Inductor value: Value of the inductor in H.

Capacitor value : Value of the capacitor in F.

The impedance is then given by

1
Zp = 1 1
. (14-67)
Rp
+ jωL p
+ jωC p

If the resistance is set to zero, then the resistance is interpreted as infinite, i.e. in the parallel case
it will not change the impedance. The same applies to the inductance.
The LP card may be combined with the LD, LS, LZ and the SK cards, but only one LP card may
be used per label. If a second LP card is used, it replaces the values entered by the first one. This
card has no significance for surface elements, even when these are assigned the same label.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-127

14.48 LS card

This card can be used to assign discrete circuit elements (in series) to a segment.

Figure 14-28 shows this circuit that can be assigned to a segment.

Figure 14-28: Sketch for the serial combination

Parameters:

Label of segments . . .: All segments with this label are assigned the series circuit values specified be-
low.

Resistor value: Value of the resistor in Ω.

Inductor value: Value of the inductor in H.

Capacitor value: Value of the capacitor in F.

The impedance is given by

1
Zs = Rs + jωLs + . (14-68)
jωCs

If a capacitance of zero is selected, it is interpreted as infinite capacitance, i.e. in the case of the
series combination it is zero.
The LS card may be combined with the LD, LP, LZ and the SK cards, but only one LS card may be
used per label. If a second LS card is used, it replaces the values entered by the first one. This
card has no significance for surface elements, even when these are assigned the same label.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-128

14.49 LZ card

This card can be used to assign a complex impedance to a segment.

Parameters:

Label of segments . . .: All segments with this label are assigned the specified impedance.

Real part of impedance: The real part of the complex impedance in Ω.

Imaginary part: The imaginary part of the complex impedance in Ω.

The complex impedance value is a constant with respect to frequency. Frequency dependent
impedances can be realised using the LS or the LP cards.
The LZ card may be combined with the LD, LP, LS and the SK cards, but only one LZ card may be
used per label. If a second LZ card is used, it replaces the values entered by the first one. This
card has no significance for surface elements, even when these are assigned the same label.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-129

14.50 NW card

This card can be used to add a linear non-radiating network.

The non-radiating general network allows the user to do combined analysis of electromagnetics
with linear circuits (e.g. amplifiers, filters, matching networks). It is therefore possible to reduce
computation time by breaking large problems into smaller element blocks. Cascading the solution
of these blocks (represented by S-, Z- or Y-parameters), direct modelling of passive circuits using
SPICE and combining with FEKO geometry, the complete problem solution can be found. The
individual element solutions are without field coupling.
Parameters:

Remove all existing networks: All previously defined non-radiating networks are removed.

New network: A new non-radiating network is created after removing all previously defined
networks.

Add to existing network: A non-radiating network is created and added to any previously defined net-
works.

Network name: The name of the network.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-130

Number of ports: A network can consist of any number of ports, but is required to have at least
one port.

Port n: Each port of a network can be connected to other network ports or geometry.
Note that the orientation of a network port connection can easily be reversed
for all connections except if connected to internal ports.

Wire segment label The label of the segment to which the network port must
be connected. If more than one segment has this label, the network port
is connected to the last segment with this label.
Wire segment position The segment is determined by specifying the Carte-
sian coordinates of the segment centre. These values are in metre and are
scaled by the SF card if Modify all dimension related values is checked.
Internal port The network name and the network port number to connect to.
Edge between regions with multiple labels The positive and negative labels
that define the edge where the network port has to be connected.
Edge connected to ground/UTD The positive or negative labels that define
the edge where the network port has to be connected.
Edge of microstrip between two points The points that define the edge of
the microstrip line where the network port has to be connected.
Vertex by segment label The vertex is determined by specifying a segment
label. Also select whether the start or end point of that segment should
be used.
Vertex by position The vertex is determined by specifying the Cartesian coor-
dinates of the vertex.

Data type: The network data can be specified S-parameters, Z-parameters, Y-parameters
or SPICE *.cir file.

Load data from Touchstone file: The network data can be loaded from a Touchstone file (in v1.0 for-
mat). The data in a Touchstone file is always defined in increasing order and
at specific frequencies only. These may of course not directly coincide with
the frequencies at which the FEKO kernel is run. The solution is to interpolate
both the magnitude and phase data by using cubic spline interpolation. The
FEKO frequency is considered out of bounds when it is more than 0.1% away
from the lowest/highest frequency defined in the Touchstone file. In such an
instance an error will be given and FEKO will terminate. If the FEKO frequency
is within bounds, but not between points, no interpolation will be performed.

Load data from a SPICE file: A passive circuit network can be loaded from a SPICE file. A *.cir file is to
be supplied containing the required SPICE circuit description. This description
should include a subcircuit definition (..SUBCKT subnam N1 <N2 N3 ..>) with
its name identical to the current NW card name. Its external number of ports
should also agree in number with the number of ports defined for this NW
card.

The data follows in the *.pre file: For small networks with four ports or less, the network matrix can
be inserted directly in the *.pre file. The matrix is entered as real and imagi-
nary components. S-parameters also require a real reference impedance to be
specified at each port.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-131

It should be noted that it is not necessary to specify all possible connections. If, for example,
NWName1.Port1 is connected to NWName2.Port1, it is not necessary to connect NWName2.Port1
to NWName1.Port1. The user should ensure that there is enough information to link all con-
nected ports. If an internal port should be left open, then no connection should be entered.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-132

14.51 OF card

This card specifies an offset for the origin of the coordinate system used for near and far field
calculations. In addition it is possible to use only a part of the structure when calculating the
fields (selected using labels).

Parameters:

Calculate near field on offset axis: Use the offset specified below as the origin of the coordinate system
for field calculations.

Use only some labels for field calculation: Use label selection when calculating near and far fields. Only
the currents on structures with a label in the range specified in the fields Start
at label and End at label are used during field computation. (If a basis function
extends over, for example, two triangles it is included if either triangle of the
triangles lies in the specified range.)

Origin of offset coordinate: In this group the Cartesian coordinates of the transformed origin are spec-
ified. Each of x, y and z is scaled by the SF card if the SF card is used.

Rotation about the axis: The angle of rotation α x around the x axis, the angle of rotation α y around
the y axis and the angle of rotation αz around the z axis in degrees.

A possible application of the OF card is, for example, to calculate the near field on the surface
of a sphere whose centre does not lie on the origin. The OF card transforms the origin of the
coordinate system to the centre of the sphere, so that the near field calculation can be executed
in spherical coordinates.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-133

14.52 OS card

With this card the currents on the surfaces and the segments can be extracted.

Parameters:

No currents: No current output (but does start calculation).

All currents: Output all currents on triangles (metallic and dielectric).

Only the currents on triangles: Only output the currents on surface triangles.

Only the segment currents: Only output the currents on wires.

Currents on structures with specified label(s): Output the currents on segments and triangles with the
specified labels.

Currents on structures with label range: Output all currents on segments and triangles in the label
range specified by the fields Extract currents starting on label and And end-
ing at label.

All segment currents to *.rsd (CableMod/CRIPTE) file: Export the currents on all segments to a *.rsd
file in CableMod/CRIPTE/PCBMod format (see the comment below).

Segment currents (label range) to *.rsd file: Export the currents on all segments with labels in the
range specified by the fields Extract currents starting on label and And ending
at label to a *.rsd CableMod/CRIPTE/PCBMod file.

No averaging of currents at triangle corners: For the output of the magnitude of current densities at the
vertices of triangles, neighbouring triangles with common vertices are identi-
fied and the current densities are then averaged over the neighbours. This
ensures that a graphical representation of the magnitude of the current density
(found in the *.out and *.os files) is a smooth colour representation without
discontinuities. Note that this setting has no effect on the graphical represen-
tation in POSTFEKO, the magnitude of the current density in POSTFEKO is
always averaged. Averaging of the current densities at the vertices could po-
tentially be very time consuming, particularly for structures containing a large
number of triangles.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-134

Multiple OS cards can be used to extract currents on multiple, non-consecutive, labels.

The options where a *.rsd file is written permit the creation of a *.rsd file for use with the
transmission line simulation programs CableMod or CRIPTE or the PCB tool PCBMod. The cur-
rents along all or selected segments are exported to the *.rsd file (the file name without exten-
sion is the same as that of the *.fek file). The *.rsd file is an ASCII file and contains first a
description of the geometry of the line, followed by blocks with the current information for each
frequency. It can be read by CableMod, CRIPTE or PCBMod and can also be imported back into
FEKO to realise an impressed line source (see the AC card).
If the current of dielectric triangles (surface current formulation) must be output by the OS card,
both the equivalent electric and magnetic surface currents of the external problem are written
to the output file. (The currents of the internal problem are different to those of the external
problem only in that their sign is reversed.)
If requested by the DA card, a *.os file will be created in addition to the currents written to the
output file.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-135

14.53 PP card

This card defines the phase shift of the excitation between one unit cell and the next when
doing periodic boundary condition calculations. The unit cell for a periodic boundary condition
calculation is specified using the PE card.

Parameters:

Manually specify phase shift: The phase shift is manually specified.

Determine phase shift from plane-wave excitation: When a plane wave is used as excitation the phase
difference between the cells can not be specified, but is determined by the
excitation.

Determine phase shift from beam pointing (“squint”) angle: The phase shift is determined by specifying
the theta and phi angle of the “squint” angle.

Phase shift in û1 direction: Phase shift in the first direction, û1 .

Phase shift in û2 direction: Phase shift in the second direction, û2 .

One dimension:

Theta angle (degrees): Orientation of the squint angle - theta in degrees is the angle between the
squint angle and the plane defined by the û1 vector.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-136

Two dimensions:

Theta angle (degrees): Orientation of the squint angle - theta in degrees is the angle between the
squint angle and the (û1 = 0) plane.

Phi angle (degrees): Orientation of the squint angle - phi in degrees is the angle between the squint
angle and the plane defined by the û1 vector.

Note that multiple PP cards can be used in a model (only one PE card).

July 2011 FEKO User’s Manual

CONTROL CARDS 14-137

14.54 PS card

This card is a general program control and can be used for instance to store the current expansion
coefficients to a file or load them later again to speed up the solution.
It is important to be familiar with the solution process of the MoM to understand this card. The
solution of electromagnetic problems based on the MoM involves a setting up a system of linear
equations, which by default is solved using an LU-decomposition and a subsequent backwards
substitution. This card can be used to save the matrix of the system of linear equations, its LU-
decomposition, or the solution vector (which also includes PO currents etc.). Such elements can
also be loaded again.

Parameters:

Save / read matrix elements: Select this option to save or read the matrix elements of the system of
linear equations. This is typically not recommended, since the file will be large
and the time to recompute the matrix elements is typically much shorter than
the time for the LU-decomposition of the matrix.

Save / read LU decomposed matrix: Select this option to save or read the LU-decomposition of the
system of linear equations. This option is useful if you want to solve a problem
subsequently with multiple different excitations (i.e. right-hand sides). Note
that if you do this in one FEKO run (i.e. one *.pre file), then FEKO keeps the
LU-decomposition automatically in memory. This option is useful if you want
to solve for multiple excitations in different FEKO runs.

Save / read currents: Select this option to save or read the solution vector of the system of linear
equations. The solution vector corresponds to the currents on the structure
being simulated.

FEKO always uses the most efficient computation when doing multiple solutions in one file.
However, sometimes one might also do a solution, look at the results and then change certain
parameters. Then the option to store the solution in the *.str file and load it again or the
similar option for the LU-decomposition are particularly useful. The option to save the currents
is applicable when the solution stays the same (i.e. same geometry, same material parameters
and loads, same frequency, same excitation), but when one wants to compute the near- or far

July 2011 FEKO User’s Manual

CONTROL CARDS 14-138

fields with different options. The storage of the LU-decomposition is useful when only the right-
hand side of the system of linear equations changes (e.g. the direction of incidence of a plane
wave). FEKO checks all this using checksums and reports a warning if, for example, currents
were exported for one frequency and are later imported again for another frequency.
Note that the *.str file can be used for MoM, MLFMM, PO, and FEM, while the *.mat and
*.lud files are only applicable when the standard MoM is used (and then also only for sequential
in-core solutions).
Note that models built with PREFEKO on different computers may not be identical due to very
small rounding differences of different CPUs. It is therefore advisable to run PREFEKO only on
one computer when using this card, to ensure consistency in the *.fek files. The *.fek files can
then be copied to another computer if required. (For example, a user may calculate and store
the current distribution for a large model on a fast workstation and later load this to calculate
different near fields on a small PC. To ensure that current solution is valid on the PC, the original
*.fek file should be generated on the PC and copied to the workstation.)
If the PS card is used to specify that data should be read from a file, the PS card may occur only
once in the input file. It is advisable to place it right after the EG card.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-139

14.55 PW card

When defining the excitation of an antenna, the source is normally specified as a complex volt-
age. The PW card allows the user to specify the radiated power or the source power instead
(FEKO then just internally scales the whole solution to arrive at this desired power). In addi-
tion, it is possible to consider a mismatch between the antenna input impedance and the internal
impedance of a voltage source or the characteristic impedance of a transmission line feed.

Parameters:

No scaling (use specified voltages): PW card is not activated, i.e. the specified value of the voltage
source is used.

Total source power (no internal impedance): PW card is activated and all the currents are multiplied by
a scaling factor so that the total source power (the sum of the power delivered
by all the individual sources) is P0 — the value specified in the Source power
field. Mismatch is not considered.

Total source power (internal impedance): All voltage sources are assumed to have an input impedance
Zi as specified by the parameters Source impedance, real part and Source
impedance, imag part. The currents are scaled such that the total power sup-
plied by the voltage sources equals P0 as discussed below. The mismatch losses
in the source impedance reduce the antenna gain.

Total source power (transmission line feed): All the antennas are assumed to be fed by transmission
lines with a complex characteristic impedance Z L as specified by the parame-
ters Charact. impedance, real part and Charact. impedance, imag part. If there
is a mismatch between Z L and the antenna input impedance Za , some of the
incident power will be reflected back to the source.

Decouple all sources when calculating power: When this item is not checked and multiple impressed
sources (i.e. elementary dipoles A5/A6, impressed current elements AI/AV,
etc.) are present, the mutual coupling of all these sources, as well as the
coupling of the sources with other structures such as ground (BO card), UTD
surfaces, or MoM elements are taken into account when determining the source
power. This is also the default if the PW card is not present. When this item
is checked, however, this mutual coupling is not considered. This is acceptable
for sources which are relatively far away or when no accurate power values
are required. (Since gain/directivity are based on power, they are then also
possibly not very accurate.)

July 2011 FEKO User’s Manual

CONTROL CARDS 14-140

Source power: The total power P0 in Watt, i.e. the total power supplied by all the voltage
sources, or in the case of transmission lines, the total power of all forward
travelling waves.

Details of the various possibilities with the use of the PW card are shown in Figure 14-29.

no internal impedance: internal impedance:

Zi

~ ~ Pi
P0 Pa P0 Pa
~ Za ~ Za

transmission line feed:

~
Transmission line P0
with characteristic Pa Za
impedance ZL Pr

Figure 14-29: Possible applications of the PW card to determine the total power

The options Total source power (internal impedance) and Total source power (transmission line
feed) are only allowed for voltage sources (the A1, A2, A3, A4, A7, AE and AN cards). For models
containing other sources such as dipoles and impressed currents, the option Total source power
(no internal impedance) should be used. For plane waves No scaling (use specified voltages)
must be used.
The power equations for different cases are discussed below. Consider, in general, that there
are N sources (such as in an array antenna) with open circuit voltages U0,ν (before the scaling
operation) where the parameter ν lies in the range 1. . .N . At each source there is an antenna
input impedance Za,ν (as calculated during the FEKO solution) to which power Pa,ν is transferred.

• Total source power (no internal impedance)

Using this option all the source power is delivered to the respective antennas, i.e.

P̃0,ν = Pa,ν (14-69)

as shown in Figure 14-29. To ensure that the total power is P0 , the power must be scaled
with the factor
P0 P0
s= N = N (14-70)
P P
P̃0,ν Pa,ν
ν=1 ν=1
p
The currents on the structure are consequently scaled with the factor s. There is no power
loss.

• Total source power (internal impedance)

When this option is used the internal impedance Zi of the voltage source is considered
as shown in Figure 14-29. Since the same current flows through the internal source

July 2011 FEKO User’s Manual

CONTROL CARDS 14-141

impedance and the antenna input impedance, the power dissipated in the impedance of
the ν th voltage source is given by the relation
Re Zi
Pi,ν = Pa,ν (14-71)
Re Za,ν

and the scaling factor s (to scale the total power supplied by the sources to P0 ) is
P0 P0 P0
s= N
= N €
= N  ‹ (14-72)
Re Zi
Š
Pa,ν + Pi,ν Pa,ν 1 + Re Z
P P P
P̃0,ν
a,ν
ν=1 ν=1 ν=1

The combined loss caused by the mismatched antennas

N N
X X Re Zi
Ploss = s Pi,ν = s Pa,ν (14-73)
ν=1 ν=1
Re Za,ν

reduces, for example, the antenna gain (but not the directivity).

• Total source power (transmission line feed):

When this option is used each antenna (with input impedance Za,ν ) is considered to be
excited by a transmission line with a complex characteristic impedance Z L as shown in
Figure 14-29. For most practical applications the transmission line will be lossless, resulting
in a real characteristic wave impedance. For this lossless case the reflection factor
Za,ν − Z L
%ν = (14-74)
Za,ν + Z L

is taken into account when calculating the incident power at the feed point.
The total incident power is given by
Pa,ν
P̃0,ν = (14-75)
1 − |%ν |2

and the reflected power by

|%ν |2
Pr,ν = |%ν | P̃0,ν = Pa,ν
2
(14-76)
1 − |%ν |2

To ensure that the total incident power is P0 , the power is scaled with the factor
P0 P0
s= = N (14-77)
N
P P Pa,ν
P̃0,ν
ν=1 ν=1 1 − |%ν |2
p
and the currents with the factor s. As before the total reflected power
N N
X X |%ν |2
Ploss = s Pr,ν = s Pa,ν (14-78)
ν=1 ν=1
1 − |%ν |2

reduces the gain of the antenna.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-142

If we have a lossy transmission line, then forward and backward travelling waves can still
be identified on such a transmission line, but a proper definition of power associated with
such a wave is not possible, since due to the losses the power will constantly change along
the length of the transmission line. It is also questionable whether using the PW card
with a lossy transmission line makes any sense. But it has still been implemented, and
then the forward travelling power P0 is interpreted as the maximum available power at the
end terminals of the transmission line. For a lossless transmission line this formulation is
compatible with the above equations.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-143

14.56 RA card

This card is used to define an ideal receiving antenna which is described by an impressed radia-
tion pattern and which can be positioned anywhere in the model.

Parameters:
Read pattern data from: In this group the user must select one of four options:

• a *.ffe file: Read the radiation pattern from an *.ffe format file (which
may be created with the DA and FF cards).
• an external data file: Read the radiation pattern from an ASCII file (the
format of this file is described at the AR card (see section 14.19).
• after this line in the *.pre file: The radiation pattern is specified in the
*.pre file following the RA card (the format is described at the AR card
(see section 14.19)). With this option one can make use of the FOR loops
to generate patterns from known functions.
• use last pattern defined at previous RA card: When using multiple RA
cards (i.e. different receiving antennas in the same model) then it is quite
common that the shape of the pattern is identical. If this is the case, then
one can load the pattern just once and at subsequent RA cards check this
option here. Then the last defined pattern will be used and memory can
be saved (no need to store it again). Note that one is still able to set the
antenna position and orientation individually.

Include only the scattered part of the field: When this item is checked, only the scattered part of the
field is considered for calculation. Otherwise the total field, the sum of the
scattered and source contributions, are considered for calculation.

Position: In this group the x, y and z coordinates of the receiving point (i.e. the position
where the antenna is placed) are entered in m. This value is affected by the
scale factor of the SF card if used.

Rotation about the axes: In this group the angles with which the imported pattern is rotated around
the x, y and z axes are entered. We will refer to these fields as α x , α y and αz
in the rest of this discussion.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-144

File name: The name of the *.ffe or ASCII input file.

Start from point number: This parameter is only relevant when the data is read from an external file,
and gives the line number of the first line to read from the input file. If the data
must be read from the beginning of the file, the value in this field should be set
equal to 1. This parameter is used when the *.ffe file contains more than one
pattern. For example, if the file contains the pattern at various frequencies, the
correct pattern can be selected by setting this field to the appropriate value for
each frequency.
If the *.ffe file is of a newer format and contains header data in addition to
the data blocks, the point number refers to the actual point number. This is the
same as the line number if all blank lines, comment lines and header lines are
stripped from the file.

Number of ϑ points: The number of ϑ angles in the pattern.

Number of ϕ points: The number of ϕ angles in the pattern.

The definition of the radiation pattern and the various different input formats supported here
at the RA card are identical to those of an impressed transmitting antenna (AR card (see sec-
tion 14.19)). See also the example antenna_coupling in the FEKO Examples Guide.
The ideal receiving antenna is considered to be decoupled from the model (i.e. no change in the
currents and charges by including this antenna in the model) and should ideally be in the far
field of radiating structures (so that incident plane wave approximations apply and the radial
field component is small). Checks are available inside of FEKO to ensure these assumptions.
FEKO computes the power received by this ideal antenna assuming a perfect match, i.e. conjugate
complex load.
The relative phase of the received signal is also printed to the *.out file, which indicates the
relative phase of either the voltage or the current through the termination impedance (which is
an ideal conjugate complex load). When using arrays of identical receiving antennas, then this
relative phase information can be used to obtain the signal phase offsets of the various array
elements (required for instance in the design of the feed network).

July 2011 FEKO User’s Manual

CONTROL CARDS 14-145

14.57 SA card

This card is used to control calculations of the specific absorption rate (SAR) in a dielectric
medium.

Parameters:

Select Calculation: One of three SAR values which could be of interest should be selected in this
group.

Specify the search region: This group can be used to control, either by medium number or by label
number or by layer number (for the special Green’s functions), which dielectric
bodies are used for the specified calculation. It is also possible to specify a user
defined position here for the spatial average SAR computations.

Average/Peak SAR in all media/labels/layers: Select this option if the selected SAR calculation should
be done on a ‘by label’ or ‘by medium’ or ‘by layer’ basis. The whole body
average SAR is also calculated. Selecting the volume by label is only valid for
the FEM analysis.

Average/Peak SAR in a single medium/label/layer: The selected SAR calculation is obtained for the
medium/label specified in the Include medium/label dialog.

Average/Peak SAR in a medium/label/layer range: The selected SAR calculation is performed on the
label range as specified below in the input fields for Include medium/label and
up to medium/label.

Centre of SAR cube: For the spatial average SAR computations using a specified position, the X , Y
and Z coordinates of the cube centre must be specified here.

The required SAR calculation is performed, and the result saved in the *.out file.
If the options Calculate volume-average SAR and Entire region are selected, the SAR, averaged
over all media, is returned. If the options Calculate volume-average SAR and By medium are
selected, the average SAR is calculated per medium and tabulated in the *.out file. If a medi-
um/label/layer range is specified, the SAR is averaged over the volume defined by the medi-
um/label/layer range.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-146

If a spatial-peak SAR calculation is requested, then spatial-peak SAR is computed, averaged over
a mass of either 1 g or 10 g of tissue in the shape of a cube. By default, the search for the spatial-
peak SAR in the entire domain is returned, otherwise the spatial-peak SAR can be requested for
regions in a specified medium/label/layer range, or also at a user specified position.
When a special spherical or multilayer planar Green’s function is used in FEKO, then also spatial-
peak average SAR values can be computed (not volume average SAR). A selection is possible by
a single layer number, a range of layer numbers, or by including the whole dielectric volume in
the search.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-147

14.58 SH card

This card can be used to define solid or braided cable shields.

Solid (Schelkunoff)
Parameters:

PEC: Select this option to set the shield of the filaments to PEC.

Material label for metallic material: The label of the metallic (as defined in the DI card) to be used for
the shield.

Material label for coating: The label of the dielectric (as defined in the DI card) to be used as the
coating for the cable.

Thickness of layer for coating: Define the thickness of the dielectric coating.

Shield thickness: Define the thickness of the shield material.

Braided (Kley)

Parameters:

Number of carriers (m): The number of the carriers in the braided shield.

Weave angle (degrees): The weave angle of the shield.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-148

Inside braid-fixing material label: The label of the inside shield braid-fixing (protective) coating mate-
rial.

Number of filaments(n): The number of filaments in each carrier.

Diameter: The diameter of the filament.

PEC: Select this option to set the shield of the filaments to PEC.

Material label for metallic material: The label of the material (as defined in the DI card) to be used for
the shield.

Material label for coating: The label of the dielectric (as defined in the DI card) to be used as the
coating for the cable.

Thickness of layer for coating: Define the thickness of the dielectric coating.

Define shield properties

Parameters:

Define in the *.pre file: The frequency dependent shield parameters based on measured transfer
impedance and admittance data can be specified in the *.pre file.

Load from file: The file format used when importing the shield properties from file will be
described in the section Load shield properties from file below.

Frequency: The frequency at which the transfer impedance and admittance are specified
in the *.pre file.

Magnitude of transfer impedance: The magnitude of the transfer impedance defined in the *.pre file.

Phase of transfer impedance: The phase of the transfer impedance defined in the *.pre file.

Magnitude of transfer admittance: The magnitude of the transfer admittance defined in the *.pre file.

Phase of transfer admittance: The phase of the transfer admittance in the *.pre file.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-149

PEC: Select this option to set the shield of the filaments to PEC.

Material label of metallic material: The label of the material (as defined in the DI card) to be used for
the shield.

Material label for coating: The label of the dielectric (as defined in the DI card) to be used as the
coating for the cable.

Thickness of layer for coating: Define the thickness of the dielectric coating.

Shield thickness: The thickness of the shield

Load shield properties from file Below is given an XML example containing fictitious measured
data to show the file format for when importing measured cable data.
<?xml version="1.0" encoding="UTF-8"?>
<cableDB creator="name" date="2011-07-30" version="1.0">
<shielding name="shielding1">
<dataPoint freq="100e6" trans_imp_abs="5" trans_imp_phase="0" trans_adm_abs="0" trans_adm_phase="2" />
<dataPoint freq="300e6" trans_imp_abs="6" trans_imp_phase="2" trans_adm_abs="4" trans_adm_phase="1" />
<dataPoint freq="500e6" trans_imp_abs="4" trans_imp_phase="3" trans_adm_abs="3" trans_adm_phase="2" />
<dataPoint freq="700e6" trans_imp_abs="1" trans_imp_phase="5" trans_adm_abs="2" trans_adm_phase="5" />
</shielding>
</cableDB>

July 2011 FEKO User’s Manual

CONTROL CARDS 14-150

14.59 SK card

This card is used to consider the Skin effect or ohmic losses or an arbitrary user defined impedance
boundary condition on wire segments and surface elements. In addition, it can switch from
metallic triangles to thin dielectric layers which may consist of multiple layers and which may be
anisotropic. If no SK card has been defined, FEKO assumes ideal conductivity without any losses.

Skin effect and ohmic losses

Parameters:

Affect all structures with label: Affect all segments and triangles with this label.

Assume ideal conductivity: Ideal conductivity is assumed (also the default when there is no SK card
for a given label). All other parameters are ignored.

High frequency approximation skin effect: High frequency approximation.

Static approximation skin effect: Static (ohmic losses) approximation.

Exact expression for the skin effect: Use the exact expression of the skin effect for wires and/or surfaces
that is valid at arbitrary frequencies.

Triangles as thin isotropic dielectric sheet: Treat metallic triangles as thin isotropic dielectric layers (pos-
sibly consisting of multiple layers). Specify the label of the layered dielectric
(as defined in the DL card) to be used as the medium for the thin isotropic
dielectric layers.

Triangles as thin anisotropic dielectric sheet: Treat metallic triangles as thin anisotropic dielectric lay-
ers (possibly consisting of multiple layers). Specify the label of the layered
dielectric (as defined in the DL card) to be used as the medium for the thin
anisotropic dielectric layers.

User defined surface impedance: An arbitrary user defined complex surface impedance (as defined in
the DI card) can be used by specifying the label of the surface impedance.

Thickness of elements: The thickness d of the surface elements in m (if an SF card is present, this is
always scaled).

Material label: Label of the material which will be used (as specified in the DI card).

July 2011 FEKO User’s Manual

CONTROL CARDS 14-151

Layered dielectric label: Label of the layered dielectric medium which will be used (as specified in the
DL card).

Having both triangles and segments with the label Affect all structures with label should be
avoided. Separate labels and a distinct SK card for each label should be used. In addition all
wires with the label Affect all structures with label must have the same radius. If this is not the
case a unique label must be introduced for each radius.
q
2
It should be noted that the skin depth is given by δskin = ωµσ
, where the radial frequency
ω = 2π f and the permeability µ = µ r µ0 .

• Assume ideal conductivity:

No further parameters.

• High frequency approximation skin effect:

The required parameters are µ r , tanδµ (defined in the DI card) and σ, and for surfaces also
the thickness d. A good conductivity is required, i.e. σ  ω"0 .

Æ δskin < % where % is the wire radius. The

– For wires a further condition requires that
1 jωµ
surface impedance is given by Zs = 2π%
0
σ
.

– For metallic surfaces the condition δskin < d

2
must be met. The surface impedance is
Æ
jωµ
given by Zs = 12 σ
.

• Static approximation skin effect:

The required parameters are µ r , tanδµ (defined in the DI card) and σ, and for surfaces also
the thickness d. A good conductivity is required, i.e. σ  ω"0 .

– For wires a further condition requires that δskin > % where % is the wire radius. The
surface impedance is given by Zs0 = π%1 2 σ1 .

– For metallic surfaces the condition δskin > d

2
must be met. The surface impedance is
1
given by Zs = σd .

• Exact expression for the skin effect:

The required parameters are µ r , tanδµ (defined in the DI card) and σ, and for surfaces also
the thickness d. A good conductivity is required, i.e. σ  ω"0 .

– For wires with radius % the surface impedance is given by

% 
J0 (1 − j) δ

1 − j
Zs0 = skin
(14-79)
2π%σδskin J1 (1 − j) δ %
 
skin

where J0 and J1 are Bessel functions.

– For metallic surfaces the surface impedance is given by
1− j 1
Zs = (14-80)
2σδskin tan (1 − j) 2δd
 
skin

Examples are given in example_02 and example_33 in the Script Examples.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-152

Triangles as a thin isotropic dielectric sheet

Parameters:
Layered dielectric label: Label of the layered dielectric medium which will be used (as specified in the
DL card).

This option only makes sense for triangular surfaces, not for wires. The required parameters are
d, µ r , tanδµ and " r as well as σ or tanδ, so that µ = µ r µ0 and the complex dielectric constant
σ
" = " r "0 (1 − j tan δ) − j ω . Normally either σ or tanδ is entered as zero, but it is possible
to specify both (for example, to approximate a specific frequency response). FEKO will give a
warning which may be ignored.
The triangles with the label Affect all structures with label exist in a certain environment ("e ,µe ),
which is usually specified with the DI card. The surrounding medium is denoted by label ‘0’ and
by default contains the parameters of free space. It is also possible to place the triangles within
a dielectric body — in this case the environment is specified by modifying the parameters of
material ‘0’ in the DI card. An additional condition is that the triangles should be geometrically
thin, i.e. d must be small relative to the lateral dimensions. The mesh size is determined by the
wavelength in the environment (i.e. in the medium "e ,µe ). Thus the layers must be thin relative
to the wavelength in the environment.
When used with the MoM, the use of Triangles as thin isotropic dielectric sheet requires that
µ = µe and " 6= "e . For a single layer, the card consists of only one line. The surface impedance,
as used by FEKO, is then
β
Zs = (14-81)
2 jω (" − "e ) sin(β d2 )
p
where β = ω "µ is the propagation constant. An example is given in example_32: RCS of
a thin dielectric sheet in the ScriptingExamples provided with the FEKO installation. For
multiple layers the card requires one line per layer with the parameters of the first layer on the
same line as the card name. The approximate surface impedances of the different layers are
added to determine the effective surface impedance.
When used with PO, it is not required that µ = µe or " 6= "e . In this case the order of the layers is
also significant. The layer on the side that the triangle normal vector points to, is specified in the
first line with the remaining layers following in sequence.

Triangles as a thin anisotropic dielectric sheet

This option is very similar to Triangles as thin isotropic dielectric sheet, but the layers are
anisotropic. The principle direction in each layer is defined by the angle α (Angle of princi-

July 2011 FEKO User’s Manual

CONTROL CARDS 14-153

ple direction field) relative to the projection of the vector ζ (Reference direction field) onto the
plane of triangle (in the DI card). Here α is measured in the mathematically positive sense with
respect to the normal vector of the triangle. (POSTFEKO can be used to display the fibre direc-
tion and visually check that the input file is correct.) In this case the card line is followed by an
additional line for each layer. The medium properties in the principle direction is different from
those in the orthogonal direction which lies in the plane of the triangle and orthogonal to the
principle direction.
Parameters:

Reference direction: The x, y and z components of the vector ζ (used to define the principle direc-
tion, see Angle of principal direction).

Layered dielectric label: Label of the layered dielectric medium which will be used (as specified in the
DL card).

User defined surface impedance

With this option an arbitrary user defined complex surface impedance Zƒ can be used in FEKO.
The properties of the complex surface impedance are defined in the DI card and are used in the
SK card by specifying the label of the surface impedance sheet.

Parameters:

July 2011 FEKO User’s Manual

CONTROL CARDS 14-154

Material label: Label of the material which will be used as the defined surface impedance (as
specified in the DI card).

For instance, to model a solid dielectric object with relative permittivity " r and with conductivity
σ at a specific angular frequency ω = 2 π f , one could use the surface impedance expression

µ0
r
Zƒ = σ . (14-82)
ε0 " r − j ω

July 2011 FEKO User’s Manual

CONTROL CARDS 14-155

14.60 SP card

This card is used to calculate the S-parameters for the active sources.

Parameters:

Always add port impedance to existing loads: When S-parameters are computed, each port is automat-
ically loaded by FEKO with the S-parameter reference impedance of the port.
If this option is checked, and the user has manually defined a load at a port,
then the S-parameter load will be added to the existing load at the port. If this
option is not checked, then FEKO is backwards compatible to the behaviour of
the SP card in FEKO Suite 5.1 and earlier: FEKO will automatically add the S-
parameter reference loads at the various ports, but possible user defined loads
of the same load type (see discussion below) will be overwritten (i.e. then not
added) at these ports.

Restore loads after calculation: As discussed above, FEKO automatically adds loads to ports when com-
puting S-parameters. With this option the behaviour can be controlled after the
SP card processing is finished.
When this option is enabled, the automatically added loads will be removed
again, and the load situation (for instance for a subsequent far field request) is
the same as if the SP card was not used. Otherwise, all the loads as set during
the SP card processing will remain in place afterwards. This was the default
behaviour of FEKO Suite 5.1 and earlier. See the discussion below regarding a
potential run-time impact of restoring loads (re-computing the MoM matrix is
required then).

System impedance: The reference impedance in Ohm. This is used for all sources for which no
impedance value is specified when defining the source. If this field is empty, it
defaults to 50 Ω. Note that for waveguide sources (AW card) S-parameters are
always related to the corresponding waveguide impedance.

The N ports must be defined before using the SP card. They are identified simply by defining
excitation cards. Currently only A1, A2, A3, A4, AE, AF, AN, AB and AW sources are supported.
A1, A2 and A3 sources must be selected by label (not with position), and unique labels must be
used (i.e. no other segments or triangles may have a label which is used for a port).
If the amplitude of any port is set to zero, it will be used as a receive port (or sink) but not as
a source. For example, if only S21 and S11 are required for a two port network, one may set the
amplitude of the source defining port 2 exactly to zero. Then S12 and S22 are not calculated — in
some cases this may save considerable computation time.
The S-parameter load impedance for each of the port sources can be specified at the source
itself. If no such impedance was specified, the System impedance (Ohm) value specified with the
SP card will be used (if this value is not specified it defaults to 50 Ω). This S-parameter load
impedance will be added automatically to each port. The only exception here are waveguide

July 2011 FEKO User’s Manual

CONTROL CARDS 14-156

ports (AW card) and modal ports (AB card), where S-parameters are related directly to the
corresponding waveguide impedance.
It must be noted that except for waveguide ports the SP card adds load impedances to all the
ports. For A1, A2 and A3 sources it uses LZ type loads; for A4 sources it uses L4 type loads; for
AN sources it uses LN type loads and for AE sources it uses LE type loads. If any similar loads
were applied to the source position before the SP card these loads will either be added or will be
overwritten, as controlled by the Always add port impedance to existing loads checkbox.
When execution continues after processing the SP card these loads will either be removed or
kept, as controlled by the checkbox Restore loads after calculation. This makes a difference when
for instance after the SP card the far field is computed by means of an FF card. If the loads
are removed, then the result for the far field pattern is the same as if there was no SP card (i.e.
far field with ports not loaded by the S-parameter reference impedances). The disadvantage of
restoring the loads is that after the SP card processing the loads change and for instance for the
MoM this means that the MoM matrix changes, and in order to compute the far field pattern, a
full extra matrix computation and LU decomposition must be done. If the loads are kept, then
further results are readily available (i.e. the LU decomposed matrix can be re-used).
The original amplitudes and phases of the excitations will, always be restored. It should, however,
be noted that unlike near- or far field computations or some other results, the amplitudes and
phases of the excitations at the various ports do not influence the S-parameter results (except for
the special case of setting the amplitude of a port to zero which indicates to FEKO that this is a
passive port only). This is not something special in FEKO, but is how S-parameters are defined
(i.e. results are normalised by the incident port signal). It should in particular also be noted
that setting a phase of 180◦ for the excitation of a port does not change the direction of this
port. One rather physically has to define the port the other way round (e.g. other feed segment
orientation). When viewing a model in POSTFEKO then the arrows always indicate the positive
source direction and the arrows will also not change direction when setting a 180◦ phase on the
excitation.
Note that a request to compute S-parameters is not required for 1-port networks, as the S11
(reflection coefficient for a 1-port network) data will be available since it is always calculated for
voltage and current sources. For current and voltage sources, an additional S-parameter block
will not be computed if the model consist of a 1-port network and an SP card was requested and
the reference impedance remains unchanged. For waveguide and modal ports, S-parameters are
calculated by default in the same way that port impedances are calculated for voltage and current
sources. When a redundant S-parameter request has been made, a warning will be displayed to
indicate to the user that the SP card will be ignored. For n-port networks (with n ≥2) while
processing S-parameter request, source, power and impedance data is not written to *.out and
*.bof files since this was misleading as the effect of port loads were included.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-157

14.61 TL card

This card is used to connect a non-radiating transmission line between FEKO geometry or other
general non-radiating networks or transmission lines. Loads and sources can also be connected
on a transmission line terminal using the LN and AN cards respectively.

Parameters:

Remove all existing transmission lines: If checked, all previously defined transmission lines are deleted.
All the other input parameters are ignored.

New transmission line: Defines a new transmission line, all previously defined transmission lines are
replaced.

Add to existing transmission lines: An additional transmission line is defined.

Network name: The name of the transmission line.

Input port: The input port (start of the transmission line) can be connected to geometry or
other non-radiating ports in a number of ways.

Wire segment label: The label of the segment to which the transmission line
port must be connected. If more than one segment has this label, the
transmission port is connected to the last segment with this label.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-158

Wire segment position: The segment is determined by specifying the Carte-

sian coordinates of the segment centre. These values are in metre and are
scaled by the SF card if Modify all dimension related values is checked.
Internal port: The network name and the network port number of another
network port that has to be connected.
Edge between regions with multiple labels: The positive and negative labels
define the positive and negative terminals of the port connection.
Edge connected to ground/UTD: The positive or negative labels that define
the edge where the network port has to be connected to.
Edge of microstrip between two points: The points that define the edge of
the microstrip line where the network port has to be connected to.
Vertex by segment label: The vertex is determined by specifying a segment
label. Also select whether the start or end point of that segment should
be used.
Vertex by position: The vertex is determined by specifying the Cartesian co-
ordinates of the vertex.

Output port: Same as for Input port, but applies to the end of the transmission line.

Cross input and output ports: The positive port voltage is in the direction of the segment that it is
connected to (from the start to the end point of the segment). Thus the input
and output ports of the transmission line have unique orientations. If this item
is checked the transmission line connecting the ports is crossed.

Calculate length from position: If checked, FEKO determines the length based on the geometrical dis-
tance between the start and end points. Note that this feature is only available
for when both transmission line ports are connected to segments or vertices/n-
odes (the ports do not have to be the same type).

Transmission line length: The length of the transmission line in metres. This value is scaled with the
scaling factor of the SF card.

Losses: Losses of the transmission line in dB/m. Note that since the propagation con-
stant is taken as the propagation constant of the medium in which the start and
end ports are located, the attenuation specified by this parameter is added to
any losses of this medium. This factor is not affected by scaling specified with
the SF card, i.e. if a scaling factor which reduces the length of the transmission
line is added, the total loss through the line will be less. (The length is now
less and the loss per distance remained the same.)

Real part of Zo: Real part of the characteristic impedance of the transmission line in Ohm.

Imaginary part of Zo: Imaginary part of the characteristic impedance of the transmission line in Ohm.
Note that the characteristic impedance only defines the ratio between the volt-
age and current of the two waves propagating along the line. It does not specify
any losses.

Real part of shunt Y at port 1: Real part of the shunt admittance at the input port in Siemens. (This
admittance is across the port, connecting the two wires of the transmission
line.)

Imaginary part of shunt Y at port 1: Imaginary part of the shunt admittance at the input port in Siemens.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-159

Real part of shunt Y at port 2: Real part of the shunt admittance at the output port in Siemens. (This
admittance is across the port, connecting the two wires of the transmission
line.)

Imaginary part of shunt Y at port 2: Imaginary part of the shunt admittance at the output port in
Siemens.

Any load impedance defined over the transmission line port with the LZ, LS, LP, LD, L2, LE, CO
or SK cards are placed in series with the port, parallel admittances can be defined directly at the
TL card. To illustrate this, a transmission line port connected wire segment is discussed here, but
this applies to connections at edges and nodes also. See Figure 14-31 for an illustration of loads
and excitations when doing S-parameter calculations. Figure 14-30 is an illustration of loads and
excitation for calculations that do not form part of a S-parameter calculation.

Figure 14-31: Load placement for S-parameter calcu-

Figure 14-30: Normal load placement. lations.

If a voltage source of type A1 or A3 is applied at one of the port segments, then this voltage source
is assumed to be across the port (i.e. feeding the transmission line directly with an impressed
voltage). Any other sources are in series with the port. If S-parameters are computed with
respect to an excitation on a wire segment to which a TL card is connected, then the LZ load may
not be used at this segment.
The wire segments for the two ports should be located in the same medium, so that the prop-
agation constant of this medium can be taken for the transmission line. If the segments are in
different media, then the medium of the segment at the input port is used. Note that the prop-
agation constant and thus also the propagation loss of the transmission line is the same as that
of the medium surrounding input port unless an additional loss tangent is specified in the Losses
field. If this is free space the transmission line will be lossless. For transmission lines with a
propagation constant that is higher than that of the surrounding medium, such as coaxial cables
filled with dielectric material, one needs to reduce the length of the transmission line.
Losses in the transmission line network (due to the shunt admittances or transmission line losses
directly) are taken into account and will for instance reduce antenna efficiency or gain.
The TL card is used in example_39 (Scripting examples) to create a log periodic antenna.

July 2011 FEKO User’s Manual

CONTROL CARDS 14-160

14.62 TR card

This card is used to calculate the transmission and reflection coefficients for a plane wave inter-
acting with a planar structure.

Parameters:

X, Y, Z coordinate: Cartesian coordinates of the position of the plane wave in m (is scaled by the
SF card).

The transmission coefficient is defined as

Et
τ= . (14-83)
Ei

The reflection coefficient is given as

Er
ρ= . (14-84)
Ei

Figure 14-32 shows the incident, transmitted and reflected electric fields on a planar structure.

incident field Ei reflected field Er

transmitted field E t

Figure 14-32: A plane wave interacting with a planar structure

Note that for a transmission and reflection request, only a single plane wave source is allowed
(i.e. no other sources are allowed in the model). The model must either contain a:

• Planar multilayer substrate without any other geometry/mesh in the model or

• 2D periodic boundary condition (PBC).

July 2011 FEKO User’s Manual

CONTROL CARDS 14-161

14.63 WD card

This card is used to define the dielectric properties of each of the windscreen glass layers. These
layers are placed over the antenna elements by defining the relative position of the top layer w.r.t.
the reference plane (in the direction of the reference plane normal).

Note that the the three cards WR (see section 13.52), WA (see section 13.50) and WD (see
section 14.63) should be used together to create windscreen antenna models. These three cards
respectively define the windscreen reference surface, windscreen active elements (antenna) and
the windscreen layered media definition.
Parameters:

Windscreen name: The name of the windscreen.

Offset of top layer (Offset L): The distance from the reference to the top of layer 1.

Layered dielectric name: The label of the layered medium to be used, as defined in the DL card.

July 2011 FEKO User’s Manual

Part V

Working with SECFEKO

THE PROGRAM SECFEKO 15-1

15 The program SECFEKO

15.1 Introduction to SECFEKO

The FEKO licence manager shows all the licences in the specified secfeko.dat file (for node
locked licences) or connects to the floating licence servers and retrieves licence information. This
allows viewing the active components / modules, limits and expiry dates of all licences in a clear
human readable form. In addition — for floating licences — it allows viewing the available
licences, selecting which licence to use, and managing the licences that are in use. See also the
section on floating licences in the FEKO Installation Guide, in particular if the floating licence
servers must be installed or activated.
The graphical licence manager is only available for MS Windows and Linux. For other platforms
(but also for Windows and Linux) a similar functionality is provided by a command line utility
secfeko. Please just run secfeko to see all the various options available, or refer to the Floating
Licence section of the FEKO Installation Guide where some of the options are discussed.
The licence manager is started by typing secfeko_gui from a console window. Under MS
Windows it may also be started by selecting Licence manager from the FEKO menu under the
Windows Start menu.

15.2 Displaying licence information

Information regarding the licences are indicated by the following colours:

green: Node locked licences are indicated by green. Note that FEKO will always use
the first valid node locked licence. There should generally only be one valid
licence on each machine.

cyan: Floating licences are indicated in cyan.

orange: Excluded licences are shown in orange, see section 15.3.3.

red: Expired licences are indicated in red. If only the Maintenance and Support
expiry date is filled in red, it indicates a valid licence but without maintenance
and support.

Figure 15-1 shows the licence manager with an excluded licence, floating licence, expired main-
tenance and support, expired licence and node locked licence.
The licence manager will, by default, open the file %FEKO_HOME%/license/secfeko.dat
which is also used by all FEKO components. The user may, however, also uncheck Default li-
cence file and browse for an alternative file, for example to check the information in a new
secfeko.dat file before replacing an existing licence file. If Local mode is checked, floating
licences are displayed without connecting to the server.
The main table displays — for each licence — the hostname (if available) where the licence is
valid, the licence expiry date, the maintenance and support expiry date, the highest allowed ver-
sion (older versions can usually be used with newer licence files, but not the other way around),

July 2011 FEKO User’s Manual

THE PROGRAM SECFEKO 15-2

Figure 15-1: The licence manager showing an excluded, floating, expired maintenance and support, ex-
pired and node locked licences

the pricing category, the memory limit if there is one, and the number of parallel processes al-
lowed if applicable.
Each licence also contains the following information, shown in Figure 15-2 (click on the icon to
the left of the licence to expand / contract it):

• Machine code: Expanding this item shows the machine code of the machine where this
licence is valid. (Only shown for node locked licences.)

• Components: This item lists all the components (for example, CADFEKO, FEKO, RUNFEKO,
etc.) which are activated in this licence.

• Modules: The modules and extensions that are active.

July 2011 FEKO User’s Manual

THE PROGRAM SECFEKO 15-3

Figure 15-2: The licence manager displaying the Machine code.

15.3 Managing floating licences

Figure 15-3 shows the licence manager with a typical floating licence. If the secfeko.dat
file contains one or more floating licence servers, then this information (server names and port
numbers) is displayed at the top of the licence manager. This also includes the status of a floating
licence server, whether it is active or not. See the discussion of installing and setting up floating
licences in the FEKO Installation Guide. A quorum of licence servers must be active for the FEKO
floating licence scheme to work (so if you have one server this must be active, but if you are
running three redundant floating licence servers, then at least two of them must be active).
The Update button allows the user to update the state of the licences. (Note that floating licences
are checked in and out as they are being used, thus this information is not static as is the case
with node locked licences.)

July 2011 FEKO User’s Manual

THE PROGRAM SECFEKO 15-4

Figure 15-3: The licence manager with a typical licences

With the exception of the machine code of each licence, all the information displayed for node
locked licences is also applicable to floating licences. For both floating and node-locked licences,
a line with a green background indicates a preferred licence (see section 15.3.1) and a line with
a light blue background indicates a licence that is in use.
For each licence that is in use, the table displays the user and hostname where this licence is
used, and the Parallel column indicates how many solver instances are available. (If a user has
a 16 node licence, he may start two concurrent 8 process parallel jobs, but not an 8 process job
and a 10 process job.)
In addition, expanding the licence shows information on when the licence was checked out and
what components are in use. The user and hostname to which each component has been checked
out to, as well as the process ID (PID), is shown for each component. This allows each component
process to be uniquely identified and located for administrative and tracking purposes. It is
possible to manually check in a licence to make it available for a new user (see section 15.3.2).

15.3.1 Concept of a preferred licence

When a new licence is requested, the floating licence server will check out the first available
licence. Where users have multiple licences they may not necessarily be identical. (Some licences
may, for example, be memory limited GUI licences, or some CAD import module may be activated
only in some licences.) In this case the first available licence may not be the desired one.
The licence server supports the setting of a preferred licence for both floating and node-locked
licences. Right click on any licence and select Mark as preferred from the pop-up menu. The
preferred licence is listed at the top right of the licence manager. The next time a new licence is
requested, this preferred one will be used if it is still available. (The preferred licence can also
be selected by clicking on the licence and selecting Server → Set preferred licence from the main
menu.)

July 2011 FEKO User’s Manual

THE PROGRAM SECFEKO 15-5

Click the Clear button or select Server → Clear preferred licence from the main menu to remove
the preferred licence. When using a floating licence, since the server will just use the first avail-
able licence if the preferred licence is already in use, there is very little need for an individual user
to clear his preferred licence. However, if your company has one full and a number of limited
GUI licences you should not check out the full licence while just setting up your model — then
nobody else can use this and you could possibly have used one of the GUI licences.
Finally, note that all licence requests from the same user and the same machine will use the same
licence. Thus if you set a preferred licence while you have other FEKO components running (the
licence manager does not require a licence), the new licence will only be used once you have
closed all components and opened a new one.

15.3.2 Managing floating licence servers

While a licence is checked out, nobody else can use that licence. It may happen that a user has
say CADFEKO open on his machine while he is on leave, or that the licence is urgently needed for
another run and the user is not available to stop his run. The licence manager therefore allows
checking in a licence so that it becomes available again. Right click on the licence and select
Check in licence from the resulting pop-up menu, or click on the licence and select Server →
Check in licence from the main menu. For security reasons, this check in operation is allowed
only for users with administrative privileges (Windows) or the root user (UNIX) or the user who
has initially checked out this particular licence.
If a licence has been checked in manually, all running components using that licence will fail
the next time they check the licence server. In the case of the solver components (RUNFEKO,
PREFEKO, FEKO, ADAPTFEKO, OPTFEKO and TIMEFEKO) the error is fatal and the entire solu-
tion will be lost. The GUI components (CADFEKO, POSTFEKO and EDITFEKO) will close after
giving the user an option to save the current model. Thus it is usually preferred to only use this
manual check-in feature when only GUI components are checked out for a particular licence, and
if these GUI components cannot be closed directly.
In addition, the Server item on the main menu allows Shutdown and Reset. If multiple redun-
dant floating licence servers are used, then both Reset and Shutdown apply to all the running
and active servers. For security reasons, both these operations are restricted to users with ad-
ministrative privileges (Windows) or the root user (UNIX) or the user account under which the
licence server is running. Generally it should only be necessary to Reset the licence server when
modifying the licence file (for example when obtaining a new licence file or when changing the
TCP port). Note that resetting the licence server will check in all licences, so that all active user
jobs will be terminated similar to when checking in individual licences. Users should not need
the Shutdown option unless the floating licence server executable needs to be updated with a
newer one. Note also that once the licence server has been shut down it must be restarted on the
server itself — the licence manager cannot connect to the service if it is not running (for multiple
redundant floating licence servers each must be restarted individually). Please check the FEKO
Installation Guide, for further instructions on the floating licence server installation and startup.

15.3.3 Allow or deny usage of certain floating licences

Upon startup, the floating licence server SECFEKO will in addition to reading the licence file
secfeko.dat, also read a configuration file secfeko.cfg. The configuration file defines the

July 2011 FEKO User’s Manual

THE PROGRAM SECFEKO 15-6

EXCLUDE and the INCLUDE rules for the licences. The same strategy as used for searching for
the licence file will be used to locate the configuration file, namely the %FEKO_HOME%\license,
but if the environment variable FEKO_LICENSE_FILE is set, the floating licence manager will
try to locate the configuration file from the specified directory. If the configuration file is not in
the directory as set by the environment variable, the default setting will apply. By default (when
the configuration file is unavailable or no rule is specified) a floating licence can be checked out
by anybody, from any host.
An INCLUDE rule will result in only specified users or hosts to check out a particular licence. An
EXCLUDE rule can be set that will enable all users or hosts except those specified by the rule to
be able to check out the licence. Note that an INCLUDE and EXCLUDE rule is not allowed for the
same licence, although multiple INCLUDE or multiple EXCLUDE rules are allowed for the same
licence.

Syntax of configuration files

Comment lines are indicated by empty lines or lines starting with a # symbol. The INCLUDE
statement is shown in the following example.
INCLUDE <licence_number> USER <username>
INCLUDE <licence_number> HOST <hostname or IP address>

The use of wildcards are also allowed in the hostname or IP address.

INCLUDE <licence_number> HOST node-*

The same rules apply when creating EXCLUDE rules.

EXCLUDE <licence_number> USER <username>
EXCLUDE <licence_number> HOST <hostname or IP address>

15.4 Determining the machine code and creating licence requests

The machine codes of the current machine can be displayed by selecting Info → Machine code
info from the licence manager main menu. All licences that match any of these machine codes
will be valid on this machine.
By selecting Info → Create request file, a file that contains all of the relevant information for the
creation of a new licence for the machine will be generated. This request file is created in the
license subdirectory of the FEKO installation and can be used to request a new licence file for
the machine as described in the FEKO installation guide.

15.5 SECFEKO command line

The program SECFEKO can also be controlled from a command line. The command line options
available are discussed in the following table.

-m Print the FEKO machine codes for this machine.

-p Print licence status summary in a table. Included in this summary report will be a line
for each floating licence indicating their respective status.

July 2011 FEKO User’s Manual

THE PROGRAM SECFEKO 15-7

-r Reset the floating licence servers (note that this will reset all FEKO licences and then
reread the licence file and start the servers again).

-s Shutdown all floating licence servers (note that this will reset all FEKO licences, and
the licence servers must be started manually again).

-slocal Shutdown the floating licence server running on this machine (note that the licence
server must be started manually again).

-e xxx Set licence no. xxx to be the preferred licence to be checked out by you from this host
(one can reset a previously selected licence by using the licence number 0, i.e. by using
the option -e 0).

-c xxx Manually check in licence no. xxx into licence server (normally not required to be done
only possibly after a power failure or similar).

-f xxx Instead of using the default FEKO licence file secfeko.dat in the default directory, we
use an alternative licence file xxx.

-q Quiet mode, don’t ask user for confirmation.

−−version Print the version information and then exit.

−−delay Special internal delay during version output.

−−run-from-gui Special execution mode for the graphical user interface.

−−local-mode For floating licences no connection will be made to the floating licence server, only the
local licence file will be parsed (option only applicable with -p).

July 2011 FEKO User’s Manual

Part VI

Working with QUEUEFEKO

INTRODUCTION TO QUEUEFEKO 16-1

16 Introduction to QUEUEFEKO

QUEUEFEKO (and the GUI component QUEUEFEKO_GUI) facilitates the creation of a package
which can be transported to remote cluster machines. These packages can be placed in an exe-
cution queue (such as PBS) and executed when time and other resources are available. All the
information required to run FEKO on the cluster is included in the package. The package is ex-
tracted on the remote machine and repackaged once the simulation has been completed. Results
can be viewed by copying the correct package back from the cluster and extracting the contents.
Packages can optionally be encrypted to allow secure transfer of models and results between the
user’s local machine and the remote cluster. Various options are available for the encryption of
packages (see section 16.2.5).

16.1 QUEUEFEKO system overview

QUEUEFEKO_GUI is a graphical interface that allows users to create and extract packages. QUEUE-
FEKO is a console application that adds a package to an execution queue and takes care of all
the management tasks required for the successful execution of the simulation. QUEUEFEKO_GUI
is used on the users’s local machine for package configuration, creation and extraction, while
QUEUEFEKO runs on the remote cluster and oversees the simulation described in the package.
Figure 16-1 is a flowchart describing the general process of using QUEUEFEKO_GUI and QUEUE-
FEKO.

Figure 16-1: Remote execution scenario highlighting QUEUEFEKO’s role.

16.2 Creating and extracting packages

The following steps should be followed in order to generate a new package for remote execution.
Each step in the process is discussed in the sections that follow.

• Create a new configuration file (or edit an existing configuration file) (see section 16.2.1).

• Set the configuration options (see sections 16.2.2 to 16.2.5).

July 2011 FEKO User’s Manual

INTRODUCTION TO QUEUEFEKO 16-2

• Generate the package (see section 16.2.6).

• Add the package to the execution queue (see section 16.2.7).

• Extract the package containing the simulation results (see section 16.2.8).

16.2.1 Package configuration files

The QUEUEFEKO_GUI allows a user to easily create packages with similar settings by storing
these settings in a package configuration file. The package configuration is not the package itself,
but it contains all the information required to create a package. The QUEUEFEKO_GUI allows
the user to access all the settings required to control the simulation and the queueing process.
Once all the settings have been configured, package creation is as simple as choosing a name for
the package.
A new package configuration file can either be created by selecting File→New configuration or
by opening an existing configuration file by selecting File→Open configuration.
When a new configuration is created, a dialog (Figure 16-2) is displayed containing the most
often used settings from all three categories (files, FEKO options, cluster settings).

Figure 16-2: New package configuration dialog.

These settings can be adjusted after a new package configuration file is created by selecting the
FEKO or Cluster tabs on the main window.

16.2.2 Including package files

The files required to perform the simulation are determined by reading the *.pre file. Files that
are not required for the simulation (but still form part of the FEKO project) are not included
in the package by default, but can be added manually. Unnecessary files are not included to
minimise upload times to the remote cluster.
Package file selection is controlled by the items in the Package→File selection menu. The avail-
able options are:
Select a different *.pre file to be the new base file for this package.

Add additional files to the package.

July 2011 FEKO User’s Manual

INTRODUCTION TO QUEUEFEKO 16-3

Remove a file from the package.

Figure 16-3 shows a package where two of the files (startup.pre and startup.cfm) have
been included since they are required for the simulation, but the file network.s2p has been
added manually.

Figure 16-3: QUEUEFEKO_GUI showing the list of files that will be included in the package.

Note that all package files must be located either in the same directory as the base *.pre file,
or in a subdirectory inside the base directory. This is required since the files are packaged with
the directory structure intact. The directory containing the base *.pre file is considered the root
directory of the package.

16.2.3 FEKO options

Start by selecting the FEKO component (Figure 16-4) that should be launched. Each of the
components that form part of the simulation have options that can be set. The Advanced edit box
on each of the component tabs allow additional command line parameters to be specified that
are not available in the graphical interface.
The most important FEKO kernel option is to specify the number of processors that should be
used.

16.2.4 CLUSTER options

The remote cluster options allow a user to define the specific job queue that should be used and
the manner in which the user should be notified when the job starts and completes. Additionally,
resources such as RAM and run time required for the simulation must also be specified. The
wallclock time and RAM fields are required; all other fields are optional. Figure 16-5 shows the
options available for the cluster.

July 2011 FEKO User’s Manual

INTRODUCTION TO QUEUEFEKO 16-4

Figure 16-4: The FEKO component dialog.

Figure 16-5: Options available for the remote cluster in QUEUEFEKO.

16.2.5 Using encryption (optional)

The package can be encrypted, allowing for secure transfer of models and results between the
user’s local machine and the remote cluster. Package encryption is only possible if the user has
an encryption engine installed locally that is compatible with the encryption engine installed on
the remote cluster, and also has the public key of the remote cluster.
The Security dialog shown in Figure 16-6 allows encryption to be enabled or disabled for the
package, and allows selecting users who will be able to decrypt the input and output packages. It
is important that the remote cluster be chosen as one of the users that will be able to decrypt the
input package, otherwise QUEUEFEKO will not be able to extract and simulate the model. Also,
remember to add yourself to the list of users that will be able to decrypt the result package!

July 2011 FEKO User’s Manual

INTRODUCTION TO QUEUEFEKO 16-5

Figure 16-6: Security options available in QUEUEFEKO.

16.2.6 Generate package

A new package can be generated by selecting Package→Generate package.

The user will be prompted to enter the destination path of the newly created package. Packages
are created with a .pkg extention.

16.2.7 Adding the package to the execution queue

The package is transferred to the remote cluster where it is placed in an execution queue. Users
using the CrunchYard website can simply upload the package to the website and mark the pack-
age for execution. For other clusters the file may be copied and added to the PBS queue manually.
On the cluster machine, the package is added to the queue using the following command:
queuefeko mypackage.pkg

Simulation of the package will commence as specified by the user in the package configuration
file.

16.2.8 Extract package

After the simulation has completed, a new output package (*.output.pkg) will be available
for downloading from the remote cluster machine to the local machine. The package can be
extracted using the QUEUEFEKO_GUI.

July 2011 FEKO User’s Manual

INTRODUCTION TO QUEUEFEKO 16-6

A package is extracted by selecting Package→Extract package, selecting the package

to extract and selecting the location on the disk where the contents of the package
should be extracted to. Once the package has been extracted, the entire directory
contents as it was on the remote machine is made available on the local machine.

16.2.9 Decryption of package (if required)

The package will be decrypted using the first available local Private key. If it was protected by a
passphrase, the user will be prompted to enter it. The typed characters are not saved, but sent
directly to the encryption engine for verification.

16.2.10 View results

Results obtained from the remote cluster machine can now be viewed in POSTFEKO on the user’s
local machine as if the simulation was run locally.

16.3 Setting preferences

The Preferences dialog allows the user to specify the PDF viewer to be used when opening the
FEKO user manual, as well as the encryption engine that should be used. Currently only GNU
Privacy Guard is supported.

July 2011 FEKO User’s Manual

Part VII

Preprocessing and the FEKO solver

THE PREPROCESSOR PREFEKO 17-1

17 The preprocessor PREFEKO

17.1 Description

The surface of the structure to be analysed with the program FEKO, has to be subdivided into
elementary surfaces. Wires have to be subdivided into segments. The mesh size is dependent on
the wavelength in the medium surrounding the structure. The program PREFEKO can do all the
meshing. It automatically generates the geometric data, in the form required by FEKO, from the
data given by the user. The mesh density is controlled by a couple of parameters. PREFEKO also
imports mesh geometry, for example constructed in CADFEKO, integrating it with the final FEKO
input file.
This section describes the principle workings of the program. The user first defines the location
of points in space with the DP card. Structures are then defined in terms of these points. For
example, two points may be joined to form a line (BL card), or four points for a parallelogram
(BP card).

17.2 Running PREFEKO

If, for example, a file example.pre has been created using a text editor, PREFEKO is started
using the following command:
prefeko example

After successful execution a file example.fek, is created. This is the input file for FEKO. The
program PREFEKO allows a number of options, which are mainly used for debugging purposes.
Entering prefeko without arguments will give an overview of the syntax and supported options.
The options available for PREFEKO are listed below:

−−version Print the version information and then exit.

−−fek-format x Write the *.fek file in the x th file format.

-#var=value Set a variable #var to the value value.

-d Debug mode with some extra output (can be used to troubleshoot errors)

−−ignore-errors Error messages are non-fatal, i.e. PREFEKO will continue with the processing after
encountering errors. This can result in more errors as a consequence of the first one,
but it could also be useful to see all geometry modelling errors at once, and not always
the first one only.

−−print-variables Print a list of all variables (name, value, comment) to stdout. The output also
includes info whether the variable is set for the first time or whether the value of an
existing variable is changed.

−−print-variables-to-out Print a list of all variables (name, value, comment) to the FEKO output
file (*.out). The output also includes info whether the variable is set for the first time
or whether the value of an existing variable is changed.

July 2011 FEKO User’s Manual

THE PREPROCESSOR PREFEKO 17-2

When defining variables (see the next section) from the command line, for example calling
PREFEKO with
prefeko filename -#variable1=value1 -#variable2=value2 ...

It is a good idea to use the !!print_to_out command to write these variables to the output
file in order to keep a record of their values.
Note - prior to FEKO Suite 5.2, the syntax
prefeko filename #variable1=value1 #variable2=value2 ...

(i.e. without the minus sign) was supported, but this has been superseded by the new format
to avoid shell escaping problems under UNIX where the # must be masked as \#. While the old
syntax is still supported for backwards compatibility, users are advised to use the new syntax with
the leading minus sign.

July 2011 FEKO User’s Manual

THE FEKO SOLUTION KERNEL 18-1

18 The FEKO solution kernel

18.1 Introduction

The program FEKO does the actual calculations involved in the solution. Input and output of the
FEKO solver is based on files. During a solution, the status of the calculation phases is indicated,
on screen.

18.2 Running the FEKO kernel

18.2.1 Running the sequential version

Normally users are advised to run the FEKO kernel directly from the GUI components CADFEKO,
EDITFEKO or POSTFEKO. Once a session or model has been loaded, the sequential FEKO solver
can be started from the Run menu, by selecting FEKO (the shortcut key <Alt><4> can also be
used).
When the FEKO kernel is not executed from within the GUI, it can be started in a command
window (on a Windows PC) or a shell (in UNIX) by executing the command
runfeko example08

where example08.pre must be an existing input file. RUNFEKO executes PREFEKO if the *.fek
file is missing or older than the *.pre file and then executes the appropriate FEKO solver.
RUNFEKO accepts the optional parameters listed below. More information regarding additional
options for launching and controlling the parallel version of the solver can be found in sec-
tion 18.2.2. Additional options for the remote launching of FEKO are found in section 18.2.3. In
the GUI these settings are available through the Run menu → Component parameters → FEKO
dialog (see also Section 6.16.1).

−−execute-prefeko Always execute PREFEKO — even if the existing *.fek file is newer than the
*.pre file.

−−no-execute-prefeko PREFEKO will not be run to generate a new *.fek file before the FEKO
solver is launched, even if the *.fek and/or *.cfm file are older than the existing
*.fek file.

−−execute-cadfeko_bash Always execute CADFEKO_BASH first(by-pass automatic checks based on

file existence and date stamps.)

−−no-execute-cadfeko_bash CADFEKO_BATCH will not be run to create a new *.cfm and *.pre
file.

−−version Print the version information and then exit.

−−priority x The value x specifies the CPU usage priority of the FEKO run: 0 = idle, 1 = below
normal, 2 = normal, 3 = above normal and 4 = high. If not specified, the default is 2.
This option might not be available for specific systems or specific FEKO versions, then
it is just ignored.

July 2011 FEKO User’s Manual

THE FEKO SOLUTION KERNEL 18-2

−−use-gpu Execute FEKO using GPU acceleration.

−−use-job-scheduler Run the parallel FEKO kernel within a queuing system (see the list of sup-
ported systems below) and obtain the number of parallel processes as well as the host
list directly from the respective job scheduler.

−−remote-use-mpi Activates the MPI method on Windows.

−−prefeko-options ... All options following this, up to the next –xxx-options are passed to
PREFEKO.

−−feko-options ... All options following this, up to the next –xxx-options are passed to FEKO.

−−adaptfeko-options ... All options following this, up to the next –xxx-options are passed to
ADAPTFEKO.

-d Debug mode with some extra output (can be used to troubleshoot errors).

Various queuing systems are supported by FEKO under Windows and Linux when using Intel MPI
(which is the default on these systems). The queuing systems that are currently supported are
Windows CCS (compute cluster server) Torque 1.2.0 and higher, Altair PBS Pro 7.1 and higher,
OpenPBS 2.3, Platform LSF 6.1 and higher, Parallelnavi NQS V2.0L10 and higher, Netbatch 6.x
and higher.
The optional command line parameters for FEKO (specified after –feko-options) are listed
below.

−−check-only If this option is used, FEKO processes and checks the model, but does not start a
solution. This is useful to for instance check an input file on a local PC before submitting
it to a parallel computer.

-e ENV=value This has the same effect as starting FEKO with the environment variable ENV set to
value. More than one -e ... argument is allowed.

18.2.2 Running the parallel version

The parallel version of FEKO may be used on any system that is licenced to run multiple FEKO
processes concurrently. If a system has a multiple-core CPU (e.g. a quad-core processor) then
a sequential FEKO licence will allow a parallel solution with up to 4 parallel processes to be
launched. For multiple-CPU systems (e.g. a system with 2 separate dual-core CPUs) a 2-CPU
parallel FEKO licence will be required in order to run parallel solutions involving all 4 of the
available cores.
In order to use the parallel version of FEKO from the GUI, one first has to configure the host-
names and number of processes that will be used for each node. (This will initially be set up
during installation, meaning that reconfiguration is only necessary if changes are made.) To do
this configuration, open the Run menu → Component parameters (see also Section 6.16.1 for
additional detailed information on the settings shown on that dialog). On the FEKO tab, under
Parallel execution click the Configure button. In the dialog (shown in Figure 18-1), the host-
names and number of processes to be started on each host must be entered. Usually one process
per core available on each machine should be chosen. For example 2 processes for a dual-core
machine. One may also use this to implement a crude load balancing system, running more

July 2011 FEKO User’s Manual

THE FEKO SOLUTION KERNEL 18-3

Figure 18-1: Dialog for the parallel host configuration

processes on hosts with faster CPUs or more memory. Nodes may be added or removed from the
current cluster setup using the Add machine and Remove machine buttons respectively.
A special note for parallel Windows PC clusters: FEKO must be installed in the same location on
each host. It is recommended that the parallel job is started from a PC that forms part of the
cluster and that this host is listed first.9
When the user clicks OK on this window, the hosts are saved to a file machines.feko in the
directory specified by the environment variable FEKO_USER_HOME. This file is then used in the
actual parallel process launching.

Figure 18-2: Dialog for setting FEKO solution options

On the Run → Component parameters → FEKO tab (shown in Figure 18-2), further settings
with regard to the FEKO solution can be made. If the Output MFLOPS rate . . . and Network
9
It is possible to launch the job without including the local machine. The *.fek input file must then be located
on the first PC in the list and the *.out and *.bof output files are created on this PC — both in the exact same
directory as the project directory on the local machine. It is the user’s responsibility to transfer the files between the
local machine and the first machine in the list if these are not the same, or one can also use remote parallel launching
(see section 18.2.3) where FEKO does this copying explicitly.

July 2011 FEKO User’s Manual

THE FEKO SOLUTION KERNEL 18-4

latency and bandwidth options are selected, FEKO will also print a table giving the performance
of the various nodes. It is recommended that this is used during setup to ensure an optimum
configuration. These checks are repeated each time FEKO calculates the solution. A significant
amount of time may be required if the test file contains multiple frequencies. These options
should therefore not be kept selected after the initial setup, except for debugging purposes.
The target priority of the FEKO run may also be set on this tab. Setting the priority below normal
will allow the user to continue with other interactive work. However, all machines in the cluster
operate at the speed of the slowest node, so starting other CPU-intensive jobs on one of the nodes
in a cluster is generally not recommended.
After having configured the parallel FEKO version and after having set any possible special op-
tions, the parallel FEKO version can be run. To do this, Run → Parallel FEKO execution should
be selected so that a check-mark is shown next to the menu option. Any FEKO solver runs that
are launched while this option is checked will use the parallel version of FEKO.
From the command line (e.g. on a UNIX workstation), the parallel FEKO version is started with
runfeko example_08 -np x

where the parameter x following -np indicates the required number of processes to be used in
the parallel solution. In addition to the arguments listed in section 18.2.1, the parallel version
accepts the following optional parameters:

-np x Start the parallel FEKO version with x processes. The -np all option is also supported
when all available processors in the machines file should be used.

−−use-openmp-threading Enable shared-memory parallel communication between multiple pro-

cesses residing on the same host.

−−machines-file machname The file machname is the machines file with the node names and the
number of CPUs (see below).

−−mpi-options ... All options following this (if another –xxx-options parameter is used, all ar-
guments before the second –xxx-options parameter) are passed to the MPI launcher
(e.g. mpirun or mpiexec).

−−parallel-authenticate <method> Sets the authentication method to be used for parallel FEKO
runs. The following authentication methods are available:

default Platform dependent default (same as if option not specified).

localonly Run the parallel job on local host only and thus no authentication is required.
sspi Windows Active Directory (SSPI) authentication is used. This option is available
on Windows only. 10
registry Encrypt the credentials (username and password) into the registry. This option
is available on Windows only.
10
Note that maybe additional (one time) configuration settings have to be done (by the domain administrator)
to prepare the Windows domain for this kind of authentication requests (see the Intel MPI (User Authorization —
Active Directory Setup) and/or MPICH2 (Runtime Environment — Security) documentation shipped in the directory
mpi\<mpi-version>\doc of the FEKO installation directory for details).

July 2011 FEKO User’s Manual

THE FEKO SOLUTION KERNEL 18-5

The number of processes to launch on each available host is specified in a machines file with the
following general syntax
Hostname:Number of processes

with names host1 and host2 (this is the output of the UNIX command hostname), and 4 and
8 processors respectively, the machines file will be
host1:4
host2:8

With this machines file, the example with 6 processes given above will run with 4 processes on
host1 and 2 processes on host2.
If only one process is to be started on any host, then instead of the entry host3:1 in the machines
file, the shorter form host3 may be used.
Such a machines file (the file mpi/share/machines.feko under the FEKO installation path
FEKO_HOME) is automatically created during the installation of the parallel version of FEKO. By
default FEKO uses this file. If a different distribution of the processes is required, this file can be
manually edited - this is, however, strongly discouraged. The user should rather create a separate
machines file with the syntax described above. The environment variable FEKO_MACHFILE can
be used to force RUNFEKO to use this file instead of the default. The required commands (for
the sh shell) assuming the desired machines file is, for example machname, are
FEKO_MACHFILE=./machname
export FEKO_MACHFILE
runfeko example_08 -np 6

Alternatively one may pass the name of the machines file as an argument to RUNFEKO on the
command line like

runfeko example_08 -np 6 --machines-file ../../machname

Using RUNFEKO is independent of the respective platforms and MPI implementations (the dis-
cussion of the environment variable FEKO_WHICH_MPI (see section 24.3) contains more informa-
tion). For very special applications or experienced users it may be necessary to pass additional
options to MPI. (In such a case, the appropriate MPI manuals should be considered.) These
are added after the argument –mpi-options. For example on a ScaMPI cluster (assuming
FEKO_WHICH_MPI=6), the call
runfeko example_08 -np 6 --mpi-options -immediate_handling \
threaded -smtrace 5-6

(all on one line) is interpreted internally and FEKO is executed with the command
/opt/scali/bin/mpimon -export env -immediate_handling threaded \
-smtrace 5-6 /opt/feko/bin/feko.csv example_08 -- host1 4 \
host2 2

(Note that host1 and host2 are examples only — the actual information is taken from the
machines file.)
In addition to using the –mpi-options command line option , the MPI environment can be
controlled by setting certain environment variables. For instance, when using the Intel MPI
the environment variable I_MPI_DEVICE is quite important to control which device should be
used (sockets or shared memory, or some RDMA device etc.). Such environment variables should
preferably be set in the FEKO initialisation script initfeko. Check the comments there, and also

July 2011 FEKO User’s Manual

THE FEKO SOLUTION KERNEL 18-6

check the supplied MPI documentation (typically in the subdirectory mpi/<mpi-version>/doc

in the FEKO installation).
If using the −−use-openmp-threading option and the number of parallel processes are speci-
fied as 12, 2 MPI processes with 4 and 8 threads respectively will be used. The machines file will
then be as follows:
host1:4
host2:8

for the example machines used throughout this section.

18.2.3 Running on a remote host

The FEKO kernel can also be started on a remote host with automatic file transfer. For instance
the user can run the FEKO user interface on a Windows PC, but start a sequential or parallel
FEKO job directly from this user interface on some other remote workstation or cluster. There are
two main mechanisms for remote launching: The SSH/RSH based method and the MPI based
method.

The SSH/RSH method This remote launching method is cross platform capable, e.g. one can
launch a remote job from a Windows PC on an UNIX workstation or vice-versa.
In order to use the remote launching facility, SSH must be available with public key au-
thentication.

The MPI method This method is currently only available when doing remote launching from a
Windows host to a Windows host. It is based on pure Windows commands and relies on a
network share for copying files and uses the MPI daemon (as shipped with the FEKO Suite)
for starting the remote process.
Also for this method to work properly, the related option must have been selected during
installation of FEKO on the remote machine. This consists of creating a shared network
directory.

For more information regarding the setup requirements for remote launching using either method,
please see the detailed installation and setup instructions in the FEKO Installation Guide.

General settings and usage

After the setup, using remote launching is simple. On Windows and Linux, this remote launching
facility can be used directly from within the GUI components CADFEKO, EDITFEKO or POST-
FEKO. As described above for the parallel launching, open the Component parameters dialog
(see also Section 6.16.1) from the Run menu. This dialog is shown in Figure 18-2. Enter the
hostname or IP address of the remote host in the Remote host input field under Remote execu-
tion and select the appropriate Remote execution method. In order to use the remote launching
after it has been set up, select Run → Remote FEKO execution so that a check mark is shown next
to the Remote FEKO execution option. Runs of the FEKO solver (either sequential or parallel if
Parallel FEKO execution is also checked) will employ Remote FEKO execution on the remote host
while this option remains checked.

July 2011 FEKO User’s Manual

THE FEKO SOLUTION KERNEL 18-7

In order to use the remote launching facility from the command line, the command
runfeko example_08 --remote-host h

can be used, where the parameter h following –remote-host gives the hostname or the IP
address of the remote host. This will automatically use the SSH based remote launching method.
If one wants to use the MPI based method, then the additional option –remote-use-mpi must
also be given as
runfeko example_08 --remote-host h --remote-use-mpi

This command line option of RUNFEKO may be combined with other options, for instance using
runfeko example_08 --remote-host h -np 4 --machines-file m

would launch a parallel job with 4 processes using the nodes as listed in the machines file m, and
the parallel job is then launched from the remote host h (typically the control node of a cluster).
As mentioned earlier, the remote launching facility has an automatic file transfer included, so
it is not necessary to work on a shared network drive. On the remote host, FEKO will create
a temporary sub-directory in the user’s home directory with the name remote_FEKO_job_xxx
(xxx is a unique number) and all the FEKO files will be placed there for the duration of the FEKO
run. After the completion of the remote run, all files will be copied back to the client and this
temporary subdirectory on the remote machine will be removed.
A final note on remote launching of parallel FEKO jobs:

• If a machines file is specified while launching the job locally, this will also be used on
the remote host (i.e. it will be copied over). So a parallel job can be configured on the
local client on (for example, two hosts node1 and node2) but the FEKO solution can be
launched remotely on another computer - which will then be the control node of the parallel
solution. This makes sense when launching a parallel FEKO job from a Windows PC on a
Linux cluster.

• If no local machines file is specified when launching a remote solution (when launching
from the GUI it will always be there, but from the command line this can be omitted,
the parallel hosts are then found using some default mechanism (environment variable
FEKO_MACHFILE, default location for the file machines.feko etc. More information can
be found in section 18.2.2). It is important to realise that these default options are used as
set on the remote host where the parallel job is launched (this is e.g. the control node of
a parallel cluster). Thus, FEKO will read the machines.feko file on the remote host, and
not on the local host where jobs are launched.

July 2011 FEKO User’s Manual

DESCRIPTION OF THE OUTPUT FILE OF FEKO 19-1

19 Description of the output file of FEKO

The program FEKO writes all the results to an ASCII output file *.out as well as to a binary
output file *.bof. The latter is used by POSTFEKO for the result extractions. The *.out file is
human readable and contains all of the expected results.
In this section the various parts of the output file *.out are described.

19.1 Geometric data

First the geometric data is given (if it has not been suppressed at the EG card). For the metallic
triangles the following extract is written:
DATA OF THE METALLIC TRIANGLES

no. label x1 in m y1 in m z1 in m edges

medium x2 in m y2 in m z2 in m
medium x3 in m y3 in m z3 in m
nx ny nz area in m*m
1 0 0.0000E+00 0.0000E+00 0.0000E+00 1
0 0.0000E+00 2.0000E-01 0.0000E+00
0 3.3333E-02 0.0000E+00 0.0000E+00
0.0000E+00 0.0000E+00 -1.0000E+00 3.3333E-03
2 0 3.3333E-02 2.0000E-01 0.0000E+00 -1 2 3
0 3.3333E-02 0.0000E+00 0.0000E+00
0 0.0000E+00 2.0000E-01 0.0000E+00
0.0000E+00 0.0000E+00 -1.0000E+00 3.3333E-03

In the first column the number of the triangle is written. In the second column, the label is given
followed by the medium in which the triangle is situated. A 0 means that it is in free space. The
next three columns are the x, y and z coordinates of the three corner points of the triangles.
In the first row of each triangle follows another list of the numbers of the edges of the adjacent
triangles. A positive sign indicates that the positive current direction is away from the triangle. A
negative sign indicates that the positive current direction is toward the triangle. Below the edge
numbers the area of the triangle is given in m2 .
Following this is an extract of the data for the edges between the triangle. Whenever two triangles
have two common vertices, such an edge is generated.
An additional line (or row) gives the components (nx, ny, nz) of the normal vector of each
triangle.
DATA OF THE METALLIC EDGES (with MoM)

triangle no. points of tr. ...

no. type length/m media KORP KORM POIP POIM
1 1 2.0276E-01 0 -1 1 2 1 1
2 1 2.0000E-01 0 -1 2 3 3 3
3 1 3.3333E-02 0 -1 2 7 2 2

information on symmetry
yz xz xy status
0 0 0 unknown
1 0 0 0
0 0 0 unknown

Each edge is assigned a consecutive number, which appears in the first column. The second
column indicates the type of the edge. The length of the edge is given in the third column and
the medium in which the edge is found appears in the fourth column. On an edge there are
exactly two triangles. In the columns KORP and KORM the numbers of these two triangles are

July 2011 FEKO User’s Manual

DESCRIPTION OF THE OUTPUT FILE OF FEKO 19-2

given and the positive current direction is from the triangle KORP to the triangle KORM. In the
column POIP the number of the corner point of the triangle KORP, which is opposite to the edge,
is given. The same applies to the column POIM.
The next four columns contain information concerning the symmetry. In the column yz the
number of the edge appears, corresponding to the plane x =0 ( yz plane) of symmetry. A positive
sign indicates that the currents are symmetric and a negative sign indicates that the currents are
anti-symmetric. If there is a 0 present in this column, then a symmetric edge does not exist. The
same applies to the next columns xz and x y concerning the planes y =0 and z =0.
The last column with the heading STATUS has the following meaning: If unknown appears in it,
the edge has an unknown status. The applicable coefficient of the current basis function cannot
be determined from the symmetry, but has to be determined form the solution of the matrix
equation. If 0 is present in the STATUS column, then the coefficient of the current basis function
is 0 due to electric or magnetic symmetry and does not have to be determined.
If there is any other number in the STATUS column then this number indicates another edge
whose coefficient is equal to (positive sign in the column STATUS) or the negative of (negative
sign in the column STATUS) of the coefficient of the current basis functions. From symmetry the
coefficient of the current triangle does not have to be determined.
The data of the dielectric triangles (surface current method) differ only slightly.
DATA OF THE DIELECTRIC TRIANGLES

no. label x1 in m y1 in m z1 in m edges

medium x2 in m y2 in m z2 in m
medium x3 in m y3 in m z3 in m
nx ny nz area in m*m
1 0 7.1978E-01 0.0000E+00 7.1978E-01 1 2 3
1 9.4044E-01 0.0000E+00 3.8954E-01
0 8.6886E-01 3.5989E-01 3.8954E-01
8.2033E-01 1.6317E-01 5.4812E-01 7.2441E-02
2 0 9.4044E-01 0.0000E+00 3.8954E-01 4 5 6
1 1.0179E+00 0.0000E+00 0.0000E+00
0 9.4044E-01 3.8954E-01 0.0000E+00
9.6264E-01 1.9148E-01 1.9148E-01 7.8817E-02

Two additional columns, POIA and POIE are provided, but they are not relevant anymore and
should be ignored. The symmetry information is shown for the basis functions of both the equiv-
alent electric or magnetic current densities.
For the edges the extract is

DATA OF THE DIELECTRIC EDGES (with MoM)

triangle no. points of the triangle ...

no. type length/m media KORP KORM POIP POIM POIA POIE
1 3 3.6694E-01 0 1 1 3 1 3 3 2
2 3 5.1069E-01 0 1 1 4 2 1 1 3
3 3 3.9718E-01 0 1 1 45 3 2 1 2

electr. info of symmetry magnet. info of symmetry

yz xz xy status yz xz xy status
40 75 141 unknown 40 75 141 unknown
41 76 142 unknown 41 76 142 unknown
42 -3 143 0 42 -3 143 unknown

In addition to the data that is given for the metallic triangles, the following columns are provided
POIA, POIE, KNP and KNM. The column POIA contains the number of the corner point of the
triangle in KORP, where the basis function for magnetic surface current begins and the column

July 2011 FEKO User’s Manual

DESCRIPTION OF THE OUTPUT FILE OF FEKO 19-3

POIM contains the number of the end point of the triangle where the basis function ends. The
sizes KNP and KNM are the lengths when the vertices are connected to the middle of the opposite
edge in the triangles KORP and KORM. The symmetry information is shown for the basis functions
of both the equivalent electric or magnetic current densities.
The data for the segments follows the data for the triangles.
DATA OF THE SEGMENTS

No. label x1 in m y1 in m z1 in m nodes

medium x2 in m y2 in m z2 in m length in m radius in m
1 0 0.0000E+00 0.0000E+00 0.0000E+00 1
0 0.0000E+00 0.0000E+00 1.4286E-01 1.4286E-01 2.0000E-02
2 0 0.0000E+00 0.0000E+00 1.4286E-01 -1 2
0 0.0000E+00 0.0000E+00 2.8571E-01 1.4286E-01 2.0000E-02
3 0 0.0000E+00 0.0000E+00 2.8571E-01 -2 3
0 0.0000E+00 0.0000E+00 4.2857E-01 1.4286E-01 2.0000E-02

Here each segment is assigned a consecutive number. In the second column the label of the
segment appears and below it the number of the medium in which it finds itself. A zero (0)
means free space (vacuum). Then the coordinates of the start and end points of the segment
follow. In the previous row the numbers of the nodes, that are adjacent, appear. A positive
sign for the node number indicates that the positive current direction is defined away from the
segment. When there is a negative number, then the positive direction is directed toward the
segment. In the next row the length of the segment appears, followed by the radius.
For the data of the nodes between the segments a data table is given.
DATA OF THE NODES BETWEEN THE SEGMENTS

no. of segment points of segm. ...

No. ISEGP ISEGM KNOP KNOM
1 1 2 2 1
2 2 3 2 1
3 3 4 2 1
info of symmetry
yz xz xy status
0 0 5 unknown
0 0 6 unknown
0 0 7 unknown

The consecutive numbers of nodes are given in the first column. Then the number ISEGP and
ISEGM of the two connected segments follow. A positive current direction is defined from the seg-
ment ISEGP to the segment ISEGM. The column KNOP indicates whether the starting point(KNOP
= 1) of the segment ISEGP connects to the node or whether it is the end point(KNOP = 2). The
following four columns contain the data about the symmetry and are the same as for the metallic
triangles described above.
If there are any connections between triangles and segments, then the following data is given.
GEOMETRIC DATA OF CONNECTIONS SEGMENTS - TRIANGLES

Data of triang.data of segm. info of symmetry

no. DRENUM DREPOI SEGNUM SEGPOI angle yz xz xy status
1 11 1 360.0000 0 0 0 unknown
15 1 45.0000
33 1 45.0000
55 1 45.0000

Each connection point is assigned a consecutive number which is given in the first column. The
number of the triangle at the connection point is given in the column DRENUM with the connecting
vertex (1 to 3) in the column DREPOI. Likewise the connecting segment’s number is given in the
column SEGNUM and the connecting end in the SEGPOI column. (If the start point of the segment

July 2011 FEKO User’s Manual

DESCRIPTION OF THE OUTPUT FILE OF FEKO 19-4

is connected, SEGPOI = 1; else the end point is connected and SEGPOI = 2.) The column angle
gives the angle that is formed by the triangle at the connection point (in radians). The meaning
of the symmetry information in the last four columns is the same as that of the metallic triangles
given above.
If dielectric volume elements are used, then the following data block is given in the output:
DATA OF THE DIELECTRIC CUBOIDS

No. x1 in m y1 in m z1 in m
label x2 in m y2 in m z2 in m
medium x3 in m y3 in m z3 in m
x4 in m y4 in m z4 in m
1 0.0000E+00 0.0000E+00 0.0000E+00
0 3.3333E-01 0.0000E+00 0.0000E+00
Cube 0.0000E+00 3.3333E-01 0.0000E+00
0.0000E+00 0.0000E+00 3.3333E-01
2 0.0000E+00 0.0000E+00 3.3333E-01
0 3.3333E-01 0.0000E+00 3.3333E-01
Cube 0.0000E+00 3.3333E-01 3.3333E-01
0.0000E+00 0.0000E+00 6.6667E-01
3 0.0000E+00 0.0000E+00 6.6667E-01
0 3.3333E-01 0.0000E+00 6.6667E-01

Each cuboid is given a consecutive number. The x, y and z corner point coordinates are given in
the first three columns. The first row is the reference point. The second row is the corner point
to which from the reference point the first basis function is defined. Further, the third and fourth
rows define the next two basis functions with respect to the reference point.
In each dielectric cuboid there are three basis functions, in each coordinate direction one. The
data of these basis functions is given in the following format:
DATA OF THE BASIS FUNCTIONS FOR DIELECTRIC CUBOIDS

Symmetry information
No. cuboidno. direc. yz xz xy status
1 1 1 28 55 109 unknown
2 2 1 29 56 110 unknown
3 3 1 30 57 111 unknown
4 4 1 31 58 112 unknown
5 5 1 32 59 113 unknown

In the first column the consecutive number of the basis function is given. The next column
indicates the number of the cuboid. The column direction indicates the direction of the basis
function in the respective cuboid. 1 indicates that e.g. the basis function is defined from the
reference point to the second corner point. The last four columns contain information concerning
the symmetry properties of the cuboid, where the structure and the meaning is the same as with
the other basis functions.
For the FEM, the data of the tetrahedral volume elements is printed in a table like this:

July 2011 FEKO User’s Manual

DESCRIPTION OF THE OUTPUT FILE OF FEKO 19-5

DATA OF THE TETRAHEDRAL VOLUME ELEMENTS (FEM)

no. label x1 in m y1 in m z1 in m nodes

medium x2 in m y2 in m z2 in m faces
x3 in m y3 in m z3 in m edges
x4 in m y4 in m z4 in m volume in m*m*m
1 DRA1 0.0000E+00 -1.3337E-02 1.3828E-03 1 2 3 4
air 0.0000E+00 -1.4000E-02 0.0000E+00 1 2 3 4
0.0000E+00 -1.2500E-02 0.0000E+00 1 2 3 4 5 6
1.0157E-03 -1.2403E-02 1.1734E-03 3.5111E-10
2 DRA1 0.0000E+00 -1.4000E-02 0.0000E+00 2 3 4 5
air 0.0000E+00 -1.2500E-02 0.0000E+00 4 5 6 7
1.0157E-03 -1.2403E-02 1.1734E-03 4 5 7 6 8 9
2.7313E-03 -1.3731E-02 0.0000E+00 8.0125E-10
3 DRA2 0.0000E+00 -1.2500E-02 0.0000E+00 3 4 5 6
air 1.0157E-03 -1.2403E-02 1.1734E-03 7 8 9 10
2.7313E-03 -1.3731E-02 0.0000E+00 6 8 10 9 11 12
2.4386E-03 -1.2260E-02 0.0000E+00 7.1540E-10

The consecutive numbers of the elements are given in the first column. The label and dielectric
medium label of the element are given in column 2. Columns 3, 4 and 5 provide the x-, y- and
z-coordinates of the vertices of the element. The numbers of each node, face and edge bounding
the tetrahedral element follow in the last columns. Thereafter information follows regarding the
number of basis functions.
DATA FOR MEMORY USAGE

Number of metallic triangles: 0 max. triangles:MAXNDR = 176

Number of dielectric triangles: 176
Number of diel. GO triangles: 0
Number of FEM surface triangles: 0
Number of metallic segments: 0 max. segments: MAXNSEG = 0
Number dielectr./magnet. cuboids: 0 max. cuboids: MAXNQUA = 0
Number of tetrahedral elements: 0 max. tetrah.: MAXNTETRA= 0
Number of edges in PO region: 0 max. edges: MAXPOKA = 0
Number of wedges in PO region: 0 max. wedges: MAXPOKL = 0
Number of Fock regions: 0 max. Fock reg.:MAXFOGE = 0
Number of polygonal surfaces: 0 max. surfaces: MAXPOLYF = 0
max. corner p.:MAXPOLYP = 0
Number of UTD cylinders: 0

Number of metallic edges (MoM): 0 unknown: 0 max. edges: MAXNKA = 264

0 unknown: 0 (magnet.)
Number of metallic edges (PO): 0 unknown: 0 (electr.) 0 (magnet.)
Number of dielectric edges (MoM): 264 unknown: 66 (electr.)
264 66 (magnet.)
Number of dielectric edges (PO): 0 unknown: 0 (electr.) 0 (magnet.)
Number of edges FEM/MoM surface: 0 unknown: 0 (electr.)
0 0 (magnet.)
Number of nodes between segments: 0 unknown: 0 max. nodes: MAXNKNO = 0
Number of connection points: 0 unknown: 0 max. conn.: MAXNV = 0
Number of dielectric cuboids: 0 unknown: 0 max. cuboids: MAXNQUA = 0
Number of magnetic cuboids: 0 unknown: 0

Number of basis funct. for MoM: 528 unknown: 132 max. basisf. MAXNZEILE = 528
Number of basis funct. for PO: 0 unknown: 0 max. basisf. MAXNKAPO = 0

Here the data, e.g. the number of basis functions on the nodes between segments, can be ex-
tracted. It is also indicated how many have the status unknown, i.e. how many have to be
determined by solving the matrix equation.

19.2 Excitation

The data here is structured depending on the means of excitation. A voltage source on a segment
generates the following data block is generated:

July 2011 FEKO User’s Manual

DESCRIPTION OF THE OUTPUT FILE OF FEKO 19-6

EXCITATION BY VOLTAGE SOURCE AT SEGMENT

Name: VoltageSource1
Number of voltage source: N = 1
Frequency in Hz: FREQ = 7.49481E+07
Wavelength in m: LAMBDA = 4.00000E+00
Open circuit voltage in V: |U0| = 1.00000E+00
Phase in deg.: ARG(U0) = 0.00
Source at segment w. label: ULA = Line1.Wire1.Port1
Absolute number of segment: UNR = 11
Location of the excit. in m: X = 0.00000E+00
Y = 0.00000E+00
Z = 0.00000E+00
Positive feed direction: X = 0.00000E+00
Y = 0.00000E+00
Z = 1.00000E+00

Similar information is provided for other voltage excitations (excitation on an edge, microstrip
excitation, voltage source connected to a general network, etc).
If a waveguide excitation is used, the information in the output file has the following format:
EXCITATION BY IMPRESSED WAVEGUIDE MODE

Name: WaveguideExcitation1
Number of the excitation: N = 1
Port label: Union2.Face44
Frequency in Hz: FREQ = 1.64500E+09
Wavelength in m: LAMBDA = 1.82245E-01
Port type: Rectangular
Port dimensions: width = 1.29600E-01 m
height = 6.48000E-02 m
Port reference location: X = -6.48000E-02 m
Y = -3.24000E-02 m
Z = -3.02000E-01 m
Impressed mode: TE 1 0
Direction of propagation: NX = 0.00000E+00
NY = 0.00000E+00
NZ = 1.00000E+00
Amplitude in A/m: 1.00000E+00
Phase in degrees: 0.00
Transmitted power in W: 1.13772E+00

Information regarding the mode expansion at the waveguide excitation follows the waveguide
excitation information as follows:
MODE EXPANSION DATA OF WAVEGUIDE PORT

Port label: Union2.Face44

Port type: Rectangular
Port width: 1.29600E-01 m
Port height: 6.48000E-02 m
Frequency: 1.64500E+09 Hz

Mode indices Cutoff freq. Transverse wave impedance ...

m n in Hz real part imag. part
TE 0 1 2.31321E+09 0.00000E+00 3.81056E+02
TE 1 0 1.15661E+09 5.29795E+02 0.00000E+00
TE 1 1 2.58625E+09 0.00000E+00 3.10534E+02
TM 1 1 2.58625E+09 0.00000E+00 -4.57038E+02

Propagation factor Description

real part imag. part
3.40852E+01 0.00000E+00 Evanescent
0.00000E+00 2.45159E+01 Propagating Impressed (active)
4.18260E+01 0.00000E+00 Evanescent
4.18260E+01 0.00000E+00 Evanescent

For a FEM excitation (impressed current source) the following information is provided:

July 2011 FEKO User’s Manual

DESCRIPTION OF THE OUTPUT FILE OF FEKO 19-7

EXCITATION BY IMPRESSED CURRENT ELEMENT (FEM)

Name: CurrentSource1

Number of excitation: N = 1
Frequency in Hz: FREQ = 3.00000E+09
Wavelength in m: LAMBDA = 9.99308E-02
Amplitude in A: |I| = 1.00000E+00
Phase in deg.: ARG(I) = 0.00
start point end point
Element location in m: X = 0.00000E+00 0.00000E+00
Y = 6.50000E-03 6.50000E-03
Z = -5.00000E-04 -1.00000E-03
Element length in m: LEN = 5.00000E-04

If an incident plane wave is used then the output file has the following format:
EXCITATION BY INCIDENT PLANE ELECTROMAGNETIC WAVE

Name: PlaneWave1
Number of excitation: N = 1
Frequency in Hz: FREQ = 1.00000E+08
Wavelength in m: LAMBDA = 2.99792E+00
Direction of incidence: THETA = 20.00 PHI = 50.00
Polarisation: LINEAR
Axial ratio: V = 0.0000
Dir. of polarisation: ETA = 60.00
Direction of propag.: BETA0X = -4.60764E-01
BETA0Y = -5.49117E-01
BETA0Z = -1.96945E+00
Field strength in V/m: |E0X| = 9.65425E-01 ARG(E0X) = -180.00
(Phase in deg.) |E0Y| = 1.96747E-01 ARG(E0Y) = 0.00
|E0Z| = 1.71010E-01 ARG(E0Z) = 0.00

~ whose components are given, is the vector which points in the direction of propa-
The vector β,
~ 0 represents the direction of the electric field.
gation. The vector E
For an impressed aperture excitation, the following information is given:
EXCITATION BY AN APERTURE

Number of the aperture: 1

Frequency in Hz: FREQ = 1.25000E+10
Wavelength in m: LAMBDA = 2.39834E-02
No. of dipoles electr.: 1296
magnet.: 1296
Extent of the aperture: X = -1.70096E-02 ... 1.34267E-01 m
Y = -7.56383E-02 ... 7.56383E-02 m
Z = -7.59272E-02 ... 7.59272E-02 m

No specific information regarding the magnitude and phase of the dipole elements that make up
the excitation is given in the output.
Excitation by an impressed radiation pattern point source is shown in the output information as:

July 2011 FEKO User’s Manual

DESCRIPTION OF THE OUTPUT FILE OF FEKO 19-8

EXCITATION BY IMPRESSED RADIATION PATTERN

Name: RadiationPattern1
Number of the excitation: N = 1
Frequency in Hz: FREQ = 1.60000E+09
Wavelength in m: LAMBDA = 1.87370E-01
Max. field strength * dist. in V: 5.90127E+01 * PWFAKTOR (see below)
Radiated power in W: 1.12011E+00 * PWFAKTOR**2 (see below)
Directivity of the antenna in dB: 17.148
Distance for far-field cond. in m: 1.96884E+00
Local co-ordinate system of the source:
Number of grid points NTHETA = 37
NPHI = 73
Angular range THETA in deg.: 0.00 ... 180.00
PHI in deg.: 0.00 ... 360.00
Source position in m: X = -2.16000E-01
Y = 0.00000E+00
Z = 0.00000E+00
Rotation angle in deg. with respect to the global co-ordinate system
around X-axis: 0.00000E+00
around Y-axis: 0.00000E+00
around Z-axis: 0.00000E+00

No specific information regarding the magnitude and phase of the impressed pattern is given.
For an impressed spherical mode excitation, the following information is written to the output:
EXCITATION BY IMPRESSED SPHERICAL MODE

Name: SphericalSource1
Number of the excitation: N = 1
Number of modes: MODES = 880
Frequency in Hz: FREQ = 1.25000E+10
Wavelength in m: LAMBDA = 2.39834E-02
Propagation direction: C = 4 (outwards)
Mode location in m: X = 0.00000E+00
Y = 0.00000E+00
Z = 0.00000E+00
mode indices coefficient in sqrt(Watt) rad. power(Watt)
J S M N magn. phase power
3 1 0 1 9.21068E-02 -88.00 4.24183E-03
4 2 0 1 7.28918E-18 -16.25 2.65661E-35
11 1 0 2 1.49455E-17 -89.63 1.11684E-34
...

Point source type (Hertzian dipole) excitations result in the following information output:
EXCITATION BY HERTZIAN DIPOLE

Name:
Number of excitation: N = 1
Frequency in Hz: FREQ = 2.99792E+08
Wavelength in m: LAMBDA = 1.00000E+00
Amplitude in Am: |IL| = 1.00000E+00
Phase in deg.: ARG(IL) = 0.00
Dipole location in m X = 2.00000E+00
Y = 0.00000E+00
Z = 0.00000E+00
Orientation of dipole: THETA = 0.00
PHI = 0.00

19.3 Currents and charges

The OS card can request the current distribution. The following data is given for each triangle

July 2011 FEKO User’s Manual

DESCRIPTION OF THE OUTPUT FILE OF FEKO 19-9

Current request with name: Currents1

VALUES OF THE CURRENT DENSITY VECTOR ON TRIANGLES in A/m (no averaging)

Triangle centre JX JY JZ ...

number x/m y/m z/m magn. phase magn. phase magn. phase
1 1.70187E+00 7.99633E-02 0.00000E+00 2.674E-02 98.24 2.687E-03 -74.34 0.000E+00 0.00
2 1.81048E+00 1.96490E-01 0.00000E+00 6.015E-03 95.34 4.069E-03 96.46 0.000E+00 0.00
3 3.94881E-01 8.42417E-02 0.00000E+00 1.456E-01 137.29 3.578E-03 80.18 0.000E+00 0.00
4 1.97230E-01 8.42417E-02 0.00000E+00 2.087E-01 141.97 9.018E-02 141.94 0.000E+00 0.00
5 9.92519E-02 2.03608E-01 0.00000E+00 9.135E-02 141.60 1.895E-01 141.57 0.000E+00 0.00

Current magnitude in the

3 corner points
1.984E-02 2.979E-02 3.175E-02
1.912E-02 1.961E-02 8.007E-03
1.503E-01 1.438E-01 1.433E-01
1.433E-01 1.918E-01 1.216E+03
1.216E+03 1.683E-01 1.551E-01

At the position (x, y, z) the current density vector J~ in the complex form is given. The
last three columns indicate the value for the surface current density in the three vertices of the
triangles.(Note that the value of the current written in the *.out file will be affected if averaging
of the currents is de-activated at the OS card. If averaging is requested, the average of the current
at the vertices of all three adjacent triangles is shown). If the current is requested, the charge on
each triangle is also written to the output file. Only the charge is given as the position of each
triangle is the same as written for the currents.
VALUES OF THE SURFACE CHARGE DENSITY ON TRIANGLES in As/m^2

Triangle SIGMA
number magn. phase
1 1.42700E-10 15.66
2 3.34062E-10 5.41
3 1.80384E-10 170.18
4 2.57440E-10 176.74
5 2.43734E-10 174.98

The current on the segments is written as

VALUES OF THE CURRENT IN THE SEGMENTS in A

Segment centre IX ...

number x/m y/m z/m magn. phase
1 0.00000E+00 0.00000E+00 9.60000E-01 0.000E+00 0.00
2 0.00000E+00 0.00000E+00 8.70000E-01 0.000E+00 0.00
3 0.00000E+00 0.00000E+00 7.70000E-01 0.000E+00 0.00
4 0.00000E+00 0.00000E+00 6.70000E-01 0.000E+00 0.00
5 0.00000E+00 0.00000E+00 5.70000E-01 0.000E+00 0.00

IY IZ
magn. phase magn. phase
0.000E+00 0.00 2.202E-02 -38.72
0.000E+00 0.00 6.907E-02 -38.55
0.000E+00 0.00 1.172E-01 -38.30
0.000E+00 0.00 1.614E-01 -38.03
0.000E+00 0.00 2.009E-01 -37.73

With the associated charge

July 2011 FEKO User’s Manual

DESCRIPTION OF THE OUTPUT FILE OF FEKO 19-10

VALUES OF THE LINE CHARGE DENSITY ON SEGMENTS in As/m

Segment Q
number magn. phase
1 1.16913E-09 -128.72
2 1.06276E-09 -128.23
3 9.82679E-10 -127.64
4 8.92312E-10 -126.94
5 7.85966E-10 -126.07

For every voltage source the current at the feed point is determined and thus the impedance. The
following is the result:
DATA OF THE VOLTAGE SOURCE NO. 1

real part imag. part magn. phase

Current in A 2.4933E-01 -1.7696E-01 3.0574E-01 -35.37
Admitt. in A/V 1.5541E-02 -1.1030E-02 1.9058E-02 -35.37
Impedance in Ohm 4.2790E+01 3.0371E+01 5.2473E+01 35.37
Inductance in H 6.4494E-08

Power in Watt: 2.00000E+00

19.4 Finite conductivity

Firstly the block with the set of characteristics for the single labels is displayed:
DATA OF LABELS

Label Cuboid1.Face6: Skin = 3 Load = 0 Coating = 0

Triangle thickness: 5.00000E-03 m
Sigma = 1.000E+02 S/m Mue_r = 1.000E+00 tan(delta_mue) = 0.000E+00
Penetration depth of the skin effect: 5.81353E-03 m

All segments and triangles without a listed label are perfectly conducting

After the calculation of the currents the losses that result from finite conductivity are displayed.
POWER LOSS METAL (in Watt)

| in the segments | in the

Label | skineffect conc.load distr.load coating | triangles
Cuboid1.Face6| 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 | 1.3348E-06
Cuboid1.Face1| 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 | 3.7427E-07
Cuboid1.Face3| 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 | 1.0371E-06
Cuboid1.Face4| 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 | 8.1218E-06
Cuboid1.Face5| 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 | 1.0371E-06
Cuboid1.Face2| 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 | 1.3348E-06
total | 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00 | 1.3240E-05

Total loss in the segments: 0.0000E+00 W

Total loss in the triangles: 1.3240E-05 W

Loss metal (total): 1.3240E-05 W

SUMMARY OF LOSSES

Metallic elements: 1.3240E-05 W

Dielectric (surface equiv. princ.): 0.0000E+00 W
Dielectric (volume equiv. princ.): 0.0000E+00 W
Dielectric (FEM region): 0.0000E+00 W
Mismatch at feed: 0.0000E+00 W
Non-radiating networks: 0.0000E+00 W
Backward power at passive waveguide ports: 0.0000E+00 W
-------------
Sum of all losses: 1.3240E-05 W

Efficiency of the antenna: 99.7266 %

(based on a total active power: 4.8418E-03 W)

July 2011 FEKO User’s Manual

DESCRIPTION OF THE OUTPUT FILE OF FEKO 19-11

In the first column the label is displayed, the lowest row displays the sum of losses in all labels.

19.5 Near fields and SAR

If near fields are calculated, the extract shown below is given for each solution request.
The position as well as the individual components of the electric and the magnetic field strength
are given. This is — if not otherwise requested at the FE card — the total value of the field, i.e.
the sum of the incident wave and the scattered field.
Near-field request with name: NearField1

VALUES OF THE MAGNETIC FIELD STRENGTH in A/m

(total field, incident and scattered)

The following offset is in effect for the co-ordinates:

X= 0.00000E+00 m Y= 0.00000E+00 m Z= 0.00000E+00 m

LOCATION EX EY ...
medium X/m Y/m Z/m magn. phase magn. phase
0 0.00000E+00 0.00000E+00 -1.00000E+00 6.70088E+01 99.86 7.65636E-01 166.42
0 1.00000E-01 0.00000E+00 -1.00000E+00 6.46235E+01 74.23 1.14589E+00 166.13
0 2.00000E-01 0.00000E+00 -1.00000E+00 6.23014E+01 47.98 1.55289E+00 165.83
0 3.00000E-01 0.00000E+00 -1.00000E+00 5.99908E+01 21.51 1.95743E+00 163.95

... EZ
magn. phase
6.89061E+01 126.74
7.17685E+01 98.76
7.35678E+01 70.70
7.41473E+01 42.30

If the electric field inside dielectric cuboids is determined, then the value for the SAR (specific
absorption rate) and the cuboid number are also given:
VALUES OF THE ELECTRIC FIELD STRENGTH in V/m

inside the dielectric cuboids

LOCATION EX EY EZ ...
X/m Y/m Z/m magn. phase magn. phase magn. phase
0.050 0.050 0.050 5.776E+00 59.89 1.259E+01 -177.82 1.415E+02 -125.12
0.050 0.050 0.150 2.192E+01 33.75 4.114E+00 122.93 1.640E+02 -130.45
0.050 0.050 0.250 2.584E+01 31.18 3.420E+00 19.21 1.679E+02 -137.51
0.050 0.050 0.350 2.625E+01 22.29 8.499E+00 -24.72 1.551E+02 -144.87

... SAR cuboid no.

in W/kg
0.000E+00 1
0.000E+00 2
0.000E+00 3
0.000E+00 4

For specific SAR solution requests, the following output is shown (note that the extract shown
below is representative for a volume-average SAR calculation — the output for other options like
spatial peak SAR calculations will differ):

July 2011 FEKO User’s Manual

DESCRIPTION OF THE OUTPUT FILE OF FEKO 19-12

AVERAGE SPECIFIC ABSORPTION RATE in W/kg

Average SAR for the entire domain: 6.47236E+00 W/kg

AVERAGE SPECIFIC ABSORPTION RATE in W/kg

medium relative conductivity relative ...

permittivity in S/m permeability
air 1.00000 0.000000E+00 1.00000
dome 9.50000 1.585526E-03 1.00000

magnetic density volume SAR

conductivity in kg/m^3 in m^3 in W/kg
0.000000E+00 1.000000E+03 1.567165E-06 0.000000E+00
0.000000E+00 1.000000E+03 4.036295E-06 6.472361E+00

19.6 Far fields and receiving antenna

If the far field is calculated, the following block in this form is displayed:
VALUES OF THE SCATTERED ELECTRIC FIELD STRENGTH IN THE FAR FIELD in V
Factor e^(-j*BETA*R)/R not considered

The following offset is in effect for the co-ordinates:

X= 0.00000E+00 m Y= 0.00000E+00 m Z= 0.00000E+00 m

LOCATION ETHETA EPHI directivity in dB ...

THETA PHI magn. phase magn. phase vert. horiz. total
90.00 0.00 1.235E+00 168.98 0.000E+00 0.00 7.1722 -999.9999 7.1722
90.00 2.00 1.233E+00 168.90 0.000E+00 0.00 7.1583 -999.9999 7.1583
90.00 4.00 1.227E+00 168.65 0.000E+00 0.00 7.1166 -999.9999 7.1166
POLARISATION
axial r. angle direction
0.0000 180.00 LINEAR
0.0000 180.00 LINEAR
0.0000 180.00 LINEAR

Gain is a factor of 1.00000E+00 ( 0.00 dB) larger than directivity

The directivity/gain is based on an active power of 4.88015E-03 W

and on a power loss of 0.00000E+00 W

For incident plane waves, the values that are displayed here are the values of the scattered field,
i.e. the incident field is not taken into account. However, for any other sources (e.g. elementary
Hertzian dipoles or impressed radiation pattern or transmission line), the fields radiated by the
excitation are considered.
~ far is defined using the relation
In the far field a complex field strength E

e− jβ0 R
~ (~r ) =
lim E ~ far
E (19-1)
R→∞ R

with a large distance R = |~r | which tends to infinity (and which in the FEKO calculations is
~ far is voltage due to this extra distance
identical to infinity). Please note that the dimension of E
factor R.
In the *.out file the ϑ (vertical) and ϕ (horizontal) components of E ~ far are tabulated by magni-
~ far,ϑ and E
tude and phase, i.e. E ~ far,ϕ . With POSTFEKO also results for other polarisations can be
extracted. The corresponding formulas used are then

July 2011 FEKO User’s Manual

DESCRIPTION OF THE OUTPUT FILE OF FEKO 19-13

1 €
~ far,S = p ~ far,ϕ − E
~ far,ϑ
Š
E E (19-2)
2

for S-polarisation,

1 €
~ far,Z = p ~ far,ϕ + E
~ far,ϑ
Š
E E (19-3)
2

for Z-polarisation,

1 €
~ far,LHC = p ~ far,ϕ + j E
~ far,ϑ
Š
E E (19-4)
2

for left-hand circular polarisation, and

1 €
~ far,RHC = p ~ far,ϕ − j E
~ far,ϑ
Š
E E (19-5)
2

for right-hand circular polarisation.

If the excitation is an incident wave, the results include the radar cross section. In the case of
voltage sources, the gain or directivity is included (see the parameters of the FF card).
~ 0 carries a power
For the radar cross section, the incident plane wave with complex amplitude E
density of

~ 0 |2
1 |E
Si = · (19-6)
2 ZF 0

(Z F 0 denotes the wave impedance of the surrounding medium) which gets scattered on the object
and a wave is reflected with the scattered power density

1 |Eϑ |2 + |Eϕ |2
Ss = · . (19-7)
2 ZF 0

The radar cross section (RCS) σ is then defined as

Ss |R Eϑ |2 + |R Eϕ |2 |Efar,ϑ |2 + |Efar,ϕ |2
σ Tot al = lim 4πR2 = lim 4π = 4π ; (19-8)
R→∞ Si R→∞ ~ 0 |2
|E ~ 0 |2
|E
|R Eϕ |2 |Efar,ϕ |2
σH or izont al = lim 4π = 4π ; (19-9)
R→∞ ~ 0 |2
|E ~ 0 |2
|E
|R Eϑ |2 |Efar,ϑ |2
σVer t ical = lim 4π = 4π . (19-10)
R→∞ ~ 0 |2
|E ~ 0 |2
|E

For antenna and general radiation problems, as mentioned above, FEKO is computing either the
gain or the directivity depending on the FF card setting (this applies to the values tabulated in the

July 2011 FEKO User’s Manual

DESCRIPTION OF THE OUTPUT FILE OF FEKO 19-14

*.out file only, in POSTFEKO any quantity can be selected). Let us assume that Pt is the source
power and Pv are some losses in the structure (e.g. dielectric losses), then a power Pr =Pt -Pv will
be radiated. The directivity (as compared to an isotropic point source) is then defined as

Ss 2π |Efar,ϑ |2 + |Efar,ϕ |2
D Tot al = 4π R2 = · ; (19-11)
Pr ZF 0 Pr
2π |Efar,ϕ |2
DH or izont al = · ; (19-12)
ZF 0 Pr
2π |Efar,ϑ |2
DVer t ical = · . (19-13)
ZF 0 Pr

For the gain a similar definition is used, just now the source power Pt and not the radiated power
Pr is acting as reference:

2π |Efar,ϑ |2 + |Efar,ϕ |2
Ss
G Tot al = 4π R =2
· ; (19-14)
Pt ZF 0 Pt
2π |Efar,ϕ |2
GH or izont al = · ; (19-15)
ZF 0 Pt
2π |Efar,ϑ |2
GVer t ical = · . (19-16)
ZF 0 Pt

Between gain and directivity there is the relation

G Pr Pt − Pv
= = = η, (19-17)
D Pt Pt

where η represents the antenna efficiency.

The last three columns of the far field output give the polarisation information of the scattered
wave. In general the polarisation is elliptical as shown in Figure 19-1. The coordinates are ~e r , ~eϑ
and ~eϕ , and the view is in the direction of the propagation of the wave (~e r ).
In order to evaluate these quantities, let us define the magnitude and phase of the far field
components as
Efar,ϑ = A · e jα Efar,ϕ = B · e jβ . (19-18)
Using the abbreviation τ=ωt-β 0r one finds the temporal field strength vector in space as

A B
~ (τ) =
E · cos (τ + α) · ~eϑ + · cos (τ + β) · ~eϕ . (19-19)
r r
This equation describes then the polarisation ellipse of Figure 19-1. One can find the minimum
and maximum values of the field strength magnitude at these times

1 A2 · sin (2α) + B 2 · sin (2β)

τ1 = − · arctan 2 (19-20)
2 A · cos (2α) + B 2 · cos (2β)

July 2011 FEKO User’s Manual

DESCRIPTION OF THE OUTPUT FILE OF FEKO 19-15

E max

left

right

ej
er

E min
g

eJ

Figure 19-1: Elliptic polarisation in the far field

and
π
τ2 = τ1 + . (19-21)
2
~ (τ1 )| and E2 = | E
Let E1 = | E ~ (τ2 )| and assuming that we have E1 >E2 , then according to Fig-
ure 19-1 we have Ema x =E1 and Emin =E2 .
The axial ratio (Minor/Major) is defined as

Emin E2
v = = (19-22)
Emax E1

The axial ratio (Major/Minor) is defined as

Emax E1
v = = (19-23)
Emin E2

A ratio (Minor/Major) of 0 means that the wave is a linearly polarised wave, but if the ratio
(Minor/Major) has a value of 1 then it is a circularly polarised wave. The direction of rotation is
right hand circular (RHC) for 0<α-β<π and left hand circular (LHC) for π<α-β<2π.
FEKO also computes and prints the polarisation angle γ. It is the angle between the major axis
of the polarisation ellipse and the unit vector ~eϑ and can be computed using

B · cos (τ1 + β)
γ = arctan . (19-24)
A · cos (τ1 + α)

If the FF card is used with NTHETA ≥ 2 and NPHI ≥ 2 the Poynting vector is integrated over the
specified sector, see the detailed discussion given with the FF card (see section 14.37). The result
is the radiated power and is given below the field values.
When analysing an antenna the source power (calculated from the input impedance) should
equal the integral of the radiated power over the surface of a sphere. This may be used as
a partial validation of the result. Note that power losses in dielectrics and finite conductivity
should be taken into account separately.

July 2011 FEKO User’s Manual

DESCRIPTION OF THE OUTPUT FILE OF FEKO 19-16

The user may also elect to integrate the far field power without writing the field values to the
output file (using the FF card with FFREQ = 3). FEKO then produces the output
VALUES OF THE SCATTERED ELECTRIC FIELD STRENGTH IN THE FAR FIELD in V
Factor e^(-j*BETA*R)/R not considered

The following offset is in effect for the co-ordinates:

X= 0.00000E+00 m Y= 0.00000E+00 m Z= 0.00000E+00 m

Integration of the normal component of the Poynting vector in the angular

grid DTHETA = 2.00 deg. and DPHI = 1.00 deg. ( 1810 sample points)
angular range THETA angular range PHI radiated power
-181.00 .. 181.00 deg. -0.50 .. 9.50 deg. 2.82057E-04 Watt
-180.00 .. 180.00 deg. 0.00 .. 9.00 deg. 2.53851E-04 Watt
Polarisation dependent radiated power:
horizontal polarisation: 0.00000E+00 Watt ( 0.00 %)
vertical polarisation: 2.53851E-04 Watt (100.00 %)
S polarisation: 1.26925E-04 Watt ( 50.00 %)
Z polarisation: 1.26925E-04 Watt ( 50.00 %)
left hand circular pol.: 1.26925E-04 Watt ( 50.00 %)
right hand circular pol.: 1.26925E-04 Watt ( 50.00 %)

where the first line of total power is calculated assuming that each specified point lies at the
centre of an incremental integration area. The effective area is therefore slightly larger than the
area defined by the start and end angles. The second line gives the power integrated over an area
defined by the start and end angles. For example, one may request an integration from ϕ=0◦ to
ϕ=350◦ and ϑ=5◦ to ϑ=175◦ both in 10◦ increments in which case the first total will give the
total power through the sphere. One may also request the integration from ϕ=0◦ to ϕ=360◦
and ϑ=0◦ to ϑ=180◦ in which case the second total will give the correct total power through the
sphere.
The polarisation dependent power on the second block is calculated according to the effective
area of the second line.
When using a receiving antenna, the received power and phase of the received signal is given as
follows:
Ideal receiving antenna with name: ReceivingAntenna1

RECEIVED POWER IDEAL RECEIVING ANTENNA

Received power (ideal match assumed): 3.2871E-03 W

Relative phase of received signal: 9.0641E+01 deg.

19.7 S-parameters

If S-parameters have been requested with an SP card, FEKO prints different tables to the output
file. The first lists the impedance at each port (all sources that are active when the SP card is
processed are considered as ports).

S-PARAMETER REFERENCE IMPEDANCES AT PORTS

port impedance in Ohm

1 5.00000E+01
2 1.00000E+02
3 5.00000E+01

Then the S-parameters are listed for each source as shown below. Note that sources whose
amplitude are set to exactly zero are only used as sink ports, i.e. they are not excited and no such
block is created. All the ports are loaded and FEKO therefore also writes this information to the

July 2011 FEKO User’s Manual

DESCRIPTION OF THE OUTPUT FILE OF FEKO 19-17

output file. The second data line below gives S21 or the coupling to port 2 when port 1 is excited.
In the second block here under the first line gives S13 or the coupling into port 1 when port 3 is
excited.
SCATTERING PARAMETERS

ports magnitude phase

sink source real part imag. part linear in dB in deg.
S 1 3 -1.50775E-01 3.89893E-02 1.55735E-01 -16.1523 165.50
S 2 3 7.64835E-01 3.34088E-01 8.34618E-01 -1.5702 23.60
S 3 3 4.46365E-01 -1.47829E-01 4.70207E-01 -6.5542 -18.32
Sum |S|^2 of these S-parameters: 9.41935E-01 -0.2598

...

SCATTERING PARAMETERS

ports magnitude phase

sink source real part imag. part linear in dB in deg.
S 1 1 1.50829E-01 -8.81429E-01 8.94240E-01 -0.9709 -80.29
S 2 1 1.78833E-01 -1.48216E-01 2.32270E-01 -12.6802 -39.65
S 3 1 -1.24083E-01 1.07973E-01 1.64484E-01 -15.6775 138.97
Sum |S|^2 of these S-parameters: 8.80670E-01 -0.5519

19.8 Computation time and peak memory

The final section in the output file gives an overview of the computation time, in seconds, in a
tabular format:
SUMMARY OF REQUIRED TIMES IN SECONDS

CPU-time runtime
Reading and constructing the geometry 0.750 0.750
Checking the geometry 0.500 0.500
Initialisation of the Greens function 0.000 0.000
Calcul. of coupling for PO/Fock 0.000 0.000
Calcul. of the FMM transfer function 0.688 0.687
Fourier transform of FMM basis funct. 3.453 3.453
Calcul. of matrix elements 214.187 214.187
Calcul. of right-hand side vector 0.016 0.016
Preconditioning system of linear eqns. 709.953 710.000
Solution of the system of linear eqns. 58.640 58.641
Determination of surface currents 0.000 0.000
Calcul. of impedances/powers/losses 1.578 1.578
Calcul. of averaged SAR values 0.000 0.000
Calcul. of power ideal receiving ant. 0.000 0.000
Calcul. of cable coupling 0.000 0.000
Calcul. of electric near field 3.688 3.687
Calcul. of magnetic near field 0.000 0.000
Calcul. of far field 0.000 0.000
other 3.688 3.641
----------- -----------
total times: 997.141 997.140
(total times in hours: 0.277 0.277)

This table is followed by an output of the peak memory (main memory, excluding possible out-
of-core files) usage which FEKO encountered during any solution phase:
Peak memory usage during the whole solution: 516.765 MByte

July 2011 FEKO User’s Manual

Part VIII

FEKO Utilities
THE OPTIMISER OPTFEKO 20-1

20 The optimiser OPTFEKO

20.1 Introduction

Optimisation is the process of changing parameters in a model to achieve a specific predefined

response or state. The parameters of the optimisation are usually associated with geometric
dimensions, material properties or excitations/loadings. For example, the gain of a horn antenna
may be maximised by varying the size of the horn aperture.
The component of the FEKO Suite that controls the optimisation process is OPTFEKO. OPTFEKO
requires two components for successful execution. The first is a parametric FEKO model, and the
second is an *.opt file specifying how this model is to be optimised. A parametric FEKO model
may be built up in various ways, and will consist of at least a *.pre file, or a *.cfx file — or
possibly both.
As discussed in Setting up optimisation in CADFEKO (see section 7), all options relating to op-
timisation are specified through the CADFEKO interface and is stored in the *.opt file. The
parametric model may be prepared using CADFEKO (see section 4) and/or EDITFEKO (see sec-
tion 11).
For completeness sake, we will give a brief summary of the concepts involved in setting up a
specific optimisation in CADFEKO.
Optimisation is based on Searches, comprising a number of parts

• Method to be used for the search (including method settings regarding accuracy and stop-
ping criteria)

• Parameters for the search

• Goals of the search

The Method defines how the optimisation search will be performed. The stopping criteria and
accuracy define under what circumstances the search will be terminated. The optimisation Pa-
rameters define the range in which the search will be performed and the Goals specify the desired
result of the optimisation process.
Multiple Parameters and Goals may be defined as part of a single Search. The goals are combined
into a single representative function that is minimised or maximised (the relevant case is chosen
automatically based on the Goal definitions) using the optimisation algorithm indicated by the
Method selection.

20.2 The optimisation method

During optimisation, new values for parameters are chosen based on an algorithm known as
the optimisation method. OPTFEKO provides various optimisation methods, each with different
characteristics that are suitable for a different set of problems. Specifying the most appropriate
method to apply to a given problem is not a trivial task. It is a function of the number and range
of the parameters, the required outcome of the optimisation, the model size and the resources

July 2011 FEKO User’s Manual

THE OPTIMISER OPTFEKO 20-2

available. In this section we will shortly discuss each of the methods available in OPTFEKO based
on the algorithm that is used, the termination criteria as well as other general method-specific
factors. We will also consider the text-log that is produced using each method.

20.2.1 Simplex (Nelder-Mead)

The Simplex Nelder-Mead Algorithm can be categorised as a local or hill-climbing search method,
where the final optimum relies strongly on the specified starting point.
The geometric figure formed by a set of N + 1 points in an N -dimensional space is called a
simplex. The basic idea in the simplex method is to compare the values of the combined goals
at the N + 1 points of a general simplex (where each point represents a single set of parameter
values) and move the simplex gradually toward the optimum point during an iterative process.
The movement of the simplex is achieved using three operations: reflection, contraction and
expansion.
The initial simplex in a 2-dimensional search-space is represented by the points X 1 , X 2 and X 3 in
Figure 20-1

Figure 20-1: Reflection, Expansion and Contraction for the Simplex method

Reflection

Considering the diagram in Figure 20-1, if X h is the point corresponding to the poorest fitness
value among the points of the initial simplex, it may be expected that the point X r obtained by
reflecting the point X h around the axis defined by the other points in the simplex (X 1 and X 2 ) may
(when evaluated according to the optimisation goals) provide a better fitness value. If this is the
case, a new simplex can be constructed by rejecting the point X h from the simplex and including
the new point X r . This process is illustrated in Figure 20-1 where the points X 1 , X 2 and X r form
the new simplex. Since the direction of movement of the simplex is always away from the worst
result, movement will always be in a favourable direction. If the global goal function does not
have steep valleys within the space defined by the parameter ranges, repetitive application of the
reflection process will lead to a zigzag path in the general direction of the optimum.

Expansion

If a reflection process finds a point X r which is a better fitness than any point in the simplex (a
new optimum point), it may be expected that the best fitness value may be improved even further

July 2011 FEKO User’s Manual

THE OPTIMISER OPTFEKO 20-3

by moving along the direction pointing from X 0 to X r . An expansion is therefore performed from
X r to X e .
If the evaluated fitness at X e is better than the fitness at X r , the expansion was successful; X h
is then replaced with X e and the reflection process is restarted. If the evaluated fitness at X e
is poorer, the expansion attempt has failed; X h is replaced by X r (as identified in the previous
reflection operation) and the reflection process is continued.

Contraction

If the reflection process finds a point X r with a better fitness than the second-best point in the
current simplex (X nh), a contraction operation will be performed.
If the contraction process produces a point X c which has a better fitness than a point in the
simplex, the contraction was successful and X h is replaced with X c before continuing with the
reflection process. If the contraction process produces a point X c which has a poorer fitness, the
contraction process has failed and the simplex base is reduced by scaling all the points in the
simplex by an internal factor before restarting with the reflection process.
The Simplex method can be summarised as shown in Table 20-1. (The F (X ) operator represents
the evaluation of the fitness at the point X in the parameter space.)

Table 20-1: A summary of the Simplex operations

Objective Function Operation

F (X r ) < F (X l ) Expansion
F (X l ) ≤ F (X r ) < F (X nh) Reflection
F (X nh) ≤ F (X r ) < F (X h) Positive Contraction
F (X h) ≤ F (X r ) Negative Contraction

Error treatment and termination

Simplex Nelder-Mead terminates naturally when

• The maximum number of FEKO solver runs has been reached

• The standard deviation between the simplex vertices is small enough

• The simplex base is small enough

• The optimisation goal has been reached

Failure during re-evaluation and meshing (in the CADFEKO batch meshing tool or in PREFEKO)
for a specific set of parameters (including violation of boundaries placed on parameter values)
is treated by writing an appropriate error message to the *.log file. A fictitious fitness value
is assigned to such a parameter set and the simplex will continue. If too many consecutive
parameter sets result in evaluation failures, or if failure of a parameter set occurs before the
initial simplex has been established, OPTFEKO will terminate with an error.

July 2011 FEKO User’s Manual

THE OPTIMISER OPTFEKO 20-4

The text log of the Simplex method

During an optimisation, OPTFEKO maintains a text log of the optimisation process in the project
*.log file. The structure of this file is primarily determined by the optimisation method. When
the Simplex (Nelder-Mead) method is used, the log file structure is as shown below.

Section 1: General information regarding the optimisation setup.

========================= L O G - FILE - OPTFEKO =========================

Version: 13.22 of 2007-05-08

Date: 2007-06-06 16:45:43
File: test

OPTIMISATION WITH FEKO

=============== Optimisation variables ===============

No. Name Beg.value Minimum Maximum

1 sigma 3.503500000e+07 1.000000000e+07 5.000000000e+08

=============== Optimisation goals ===============

No. Name Expression

1 search1.goals.nearfieldgoal1 nearfieldgoal1

July 2011 FEKO User’s Manual

THE OPTIMISER OPTFEKO 20-5

Section 2: Information regarding the Simplex method parameters.

=============== Optimisation method: SIMPLEX NELDER-MEAD ===============

Maximum number of iterations: 1000

Base of the simplex: 1.500000000e-01
Reduction factor of the base: 5.000000000e-01
Termination at minimal base: 1.000000000e-03
Termination at standard deviation: 1.000000000e-04
Standard reflection coefficient (R): 1.000000000e+00
Contraction coefficient (-C, +C): 5.000000000e-01
Expansion coefficient (E): 2.000000000e+00

Section 3: Information regarding the parameter values, goal values and Simplex operations at
each iteration.

=============== SIMPLEX NELDER-MEAD: Intermediate results ===============

No. sigma nearfieldgoal1 global goal operation global best aim

1 3.503500000e+07 6.488107157e-02 3.488107157e-02 ----- 3.488107157e-02
2 3.774988237e+07 5.929294284e-02 2.929294284e-02 ----- 2.929294284e-02
3 4.516707895e+07 6.328463622e-02 3.328463622e-02 -----
4 4.788196133e+07 5.795280540e-02 2.795280540e-02 R 2.795280540e-02
5 5.430544199e+07 5.500882669e-02 2.500882669e-02 E success 2.500882669e-02
6 4.153550651e+08 3.036429062e-02 3.642906239e-04 R 3.642906239e-04
7 4.356175335e+08 2.964433899e-02 3.556610122e-04 E success 3.556610122e-04
8 4.457496125e+08 2.929870383e-02 7.012961666e-04 R
9 4.356179559e+08 2.964446921e-02 3.555307926e-04 +C success 3.555307926e-04

Section 4: Information regarding the termination reason and optimisation results. If sufficient
information was available for a sensitivity analysis to be completed, the results of the sen-
sitivity analysis are also given.

=============== SIMPLEX NELDER-MEAD: Finished ===============

Optimisation finished (Standard deviation small enough: 5.020322005e-06)

Optimum found for these parameters:

sigma = 4.356179559e+08

Optimum aim function value (at no. 9): 3.555307926e-04

No. of the last analysis: 9

Sensitivity of optimum value with respect to each optimisation parameter,

i.e. the gradient of the aim function at 1% variation from the optimum:
Parameter Sensitivity
sigma 8.344260771e-01

20.2.2 Particle swarm optimisation (PSO)

Particle Swarm Optimisation (PSO) is a population-based stochastic evolutionary computation

technique based on the movement and intelligence of swarms. As a global search algorithm, the
technique has been shown in certain instances to outperform other methods of optimisation like
Genetic Algorithms (section 20.2.3).
PSO can be best understood through an analogy similar to the one that led to the development
of the PSO. Imagine a swarm of bees in a field. Their goal is to find in the field the location with
the highest density of flowers. Without any a priori knowledge of the field, the bees begin in

July 2011 FEKO User’s Manual

THE OPTIMISER OPTFEKO 20-6

random locations with random velocities (direction and speed) looking for flowers. Each bee can
remember the location at which it found the most flowers, and is aware of the locations at which
each of the other bees has found an abundance of flowers.
Based on its own experience (local best, pbest) and the known best position (global best, gbest)
found so far, each bee in turn adjusts its trajectory (position and velocity) to fly somewhere be-
tween the two points depending on whether nostalgia or social influence dominates its decision.
When each bee is done flying, it communicates its new-found information to the rest of the swarm
who in turn adjust their positions and velocities accordingly.
Along the way, a bee might find a place with a higher concentration of flowers than it had found
previously. It would then be attracted to this new location as well as the location of the most
flowers found by any bee in the whole swarm. Occasionally, one bee may fly over a place with
more flowers than have thus far been encountered by any bee in the swarm. The whole swarm
would then be drawn toward that location in addition to the location of own personal best
discovery. In this way the bees explore the field: overflying locations of greatest concentration,
then being attracted back toward them. Eventually, the bees’ flight leads them to the one place
in the whole field with the highest concentration of flowers.

Population size and Number of iterations

The default swarm/population size is set to 20 and the number of iterations to 50, resulting in
a default maximum allowed number of FEKO solver runs of 1000. While too small a swarm size
prevents the search algorithm from properly traversing the parameter space, larger swarm sizes
require more computational time. Compared to GA (section 20.2.3), the PSO technique tends to
converge more quickly with smaller population sizes.
When the maximum number of solver runs, (C), is specified by the user, this needs to be con-
verted into a population size (A) and number of iterations (B), with A ∗ B ≤ C. A is selected as a
function of the number of parameters (Np ), with an internal upper limit, while the requirement
that B ≥ 5 must be satisfied.

Error treatment and termination

PSO terminates naturally when

• The maximum number of FEKO solver runs has been reached

• The standard deviation between the best positions of the swarm is small enough

• The optimisation goal has been reached

Failure during re-evaluation and meshing (in the CADFEKO batch meshing tool or in PREFEKO)
for a specific set of parameters is treated by writing an appropriate error message to the *.log
file before computing a new parameter set to replace the failed one. If too many consecutive
parameter set failures occur, then the optimisation will terminate with a message indicating this.
The *.log file for the optimisation can be consulted for further information.
Due to the nature of the technique, the parameters naturally adhere to boundaries defined in the
parameter space.

July 2011 FEKO User’s Manual

THE OPTIMISER OPTFEKO 20-7

The text log of the PSO method

During an optimisation, OPTFEKO maintains a text log of the optimisation process in the project
*.log file. The structure of this file is primarily determined by the optimisation method. When
the PSO method is used, the log file structure is as shown below.

Section 1: General information regarding the optimisation setup.

========================= L O G - FILE - OPTFEKO =========================

Version: 13.22 of 2007-05-08

Date: 2007-06-06 16:32:51
File: test

OPTIMISATION WITH FEKO

=============== Optimisation variables ===============

No. Name Beg.value Minimum Maximum

1 zf0 2.000000000e+00 1.000000000e+00 1.000000000e+01

=============== Optimisation goals ===============

No. Name Expression

1 search1.goals.farfieldgoal1 farfieldgoal1

Section 2: Information regarding the PSO method parameters.

=============== Optimisation method: PSO ===============

Maximum number of iterations: 3

Population size: 1
Acceleration constant 1: 2.800000000e+00
Acceleration constant 2: 1.300000000e+00
Termination at standard deviation: 1.000000000e-04
Pseudorandom number generator seed: 1

Section 3: Information regarding the parameter values, goal values and PSO best aims at each
iteration.

=============== PSO: Intermediate results ===============

No. zf0 search1.goals.f global goal local best aim global best aim
1 2.000000000e+00 2.267123373e-01 7.732876627e-01 7.732876627e-01 7.732876627e-01
2 2.000000000e+00 2.267123373e-01 7.732876627e-01
3 2.000000000e+00 2.267123373e-01 7.732876627e-01

Section 4: Information regarding the termination reason and optimisation results. If sufficient
information was available for a sensitivity analysis to be completed, the results of the sen-
sitivity analysis are also given.

July 2011 FEKO User’s Manual

THE OPTIMISER OPTFEKO 20-8

=============== PSO: Finished ===============

Optimisation finished (Maximum number of analyses reached: 3)

Optimum found for these parameters:

zf0 = 2.000000000e+00

Optimum aim function value (at no. 1): 7.732876627e-01

No. of the last analysis: 3

Sensitivity of optimum value with respect to each optimisation parameter,

i.e. the gradient of the aim function at 1% variation from the optimum:
Parameter Sensitivity
zf0 8.344260771e-01

20.2.3 Genetic algorithm (GA)

Genetic algorithm (GA) optimisers are robust, stochastic search methods modelled on the Darvinian
principles and concepts of natural selection and evolution. Like the PSO method (see sec-
tion 20.2.2), GA’s are classified as global optimisers. FEKO employs a real genetic algorithm
(RGA).
GA optimisation borrows from the natural world in a number of ways. Conceptually, during a GA
optimisation, a set of trial solutions (a generation) is chosen. This generation is assigned the role
of ‘parents’, from which a new generation of ‘children’ are derived. In an evolutionary ‘survival-
of-the-fittest’ process, each consecutive generation moves toward an optimal solution under the
selective pressure of the fitness/goal function criteria.

Population size and Number of iterations

As a default, the generation size for the GA method is set to 20 and the maximum number of
iterations to 50, resulting in a maximum allowed number of FEKO solver runs of 1000.
If the user specifies the maximum number of solver runs (C), this needs to be converted into a
generation size (A) and number of iterations (B), with A ∗ B ≥ C. A is selected as a function of
the number of parameters in the optimisation problem (Np ), with an internal upper limit. It is
also internally required that B be chosen such that B ≥ 5.

Error treatment and termination

The Genetic algorithm terminates naturally when

• The maximum number of FEKO solver runs has been reached

• The standard deviation between the current generation chromosomes is small enough

• The optimisation goal has been reached

Failure during re-evaluation and meshing (in the CADFEKO batch meshing tool or in PREFEKO)
for a specific set of parameters is treated by writing an appropriate error message to the *.log
file before computing a new random parameter set to replace the failed one. Due to the nature
of the technique, the parameters naturally adhere to boundaries defined in the parameter space.

July 2011 FEKO User’s Manual

THE OPTIMISER OPTFEKO 20-9

The text log of the GA method

During an optimisation, OPTFEKO maintains a text log of the optimisation process in the project
*.log file. The structure of this file is primarily determined by the optimisation method. When
the GA method is used, the log file structure is as shown below.

Section 1: General information regarding the optimisation setup.

========================= L O G - FILE - OPTFEKO =========================

Version: 13.22 of 2007-05-08

Date: 2007-06-06 16:32:51
File: test

OPTIMISATION WITH FEKO

=============== Optimisation variables ===============

No. Name Beg.value Minimum Maximum

1 zf0 2.000000000e+00 1.000000000e+00 1.000000000e+01

=============== Optimisation goals ===============

No. Name Expression

1 search1.goals.farfieldgoal1 farfieldgoal1

Section 2: Information regarding the GA method parameters.

=============== Optimisation method: RGA ===============

Maximum number of iterations: 3

Population size: 1
Creep mutation with probability: 5.000000000e-01
Elitism, i.e. best individual replicated into next generation
Enforce niching
Uniform crossover with probability: 5.000000000e-01
Termination at standard deviation: 1.000000000e-04
Pseudorandom number generator seed: 1

Section 3: Information regarding the parameter values, goal values and GA aims at each itera-
tion.

=============== RGA: Intermediate results ===============

No. zf0 search1.goals.f global goal global best aim

1 2.000000000e+00 2.267123373e-01 7.732876627e-01 7.732876627e-01
2 1.357202892e+00 2.267123373e-01 7.732876627e-01
3 1.095988516e+00 2.267123373e-01 7.732876627e-01

Section 4: Information regarding the termination reason and optimisation results. If sufficient
information was available for a sensitivity analysis to be completed, the results of the sen-
sitivity analysis are also given.

July 2011 FEKO User’s Manual

THE OPTIMISER OPTFEKO 20-10

=============== RGA: Finished ===============

Optimisation finished (Maximum number of analyses reached: 3)

Optimum found for these parameters:

zf0 = 2.000000000e+00

Optimum aim function value (at no. 1): 7.732876627e-01

No. of the last analysis: 3

Sensitivity of optimum value with respect to each optimisation parameter,

i.e. the gradient of the aim function at 1% variation from the optimum:
Parameter Sensitivity
zf0 8.344260771e-01

20.2.4 Grid search

This method is strictly speaking not an optimisation method. The optimisation parameters are
linearly varied between their minimum and the maximum values in a pre-defined number of
steps.
This can be useful to investigate the parameter space before beginning an optimisation. Due to
the required computational time, it is not generally recommended that this method be applied
for problems containing more than two parameters.
During application of the Grid search method, optimisation goals are evaluated at each of the
specified grid points, and a fitness is assigned to each evaluation. Though this fitness has no
effect on the selection of the ensuing parameter set, an optimum result on the pre-defined grid
will be identified and the solutions at each of the grid points can be compared to evaluate their
performance based on fitness.

Error treatment and termination

Failure during re-evaluation and meshing (in the CADFEKO batch meshing tool or in PREFEKO)
for a specific set of parameters is treated by writing an appropriate error message to the *.log
file before continuing with the next set of parameters in the grid.
The Grid Search Method terminates naturally only when the maximum number of FEKO solver
runs has been reached. The number of solver runs can be computed based on the number of
parameters and the number of steps per parameter as shown in Equation 20-1.

Nparameters
Y
Nsolver runs = Npoints (i) (20-1)
i=1

Internally, a limit of 10 000 is placed on the maximum number of allowed solver runs. For
4 parameters this would mean a maximum of only 10 points per parameter (this indicates how
quickly the algorithm can scale in terms of the number of required solver runs for multi-parameter
problems.)

July 2011 FEKO User’s Manual

THE OPTIMISER OPTFEKO 20-11

The text log of the Grid search method

During an optimisation, OPTFEKO maintains a text log of the optimisation process in the project
*.log file. The structure of this file is primarily determined by the optimisation method. When
the Grid search method is used, the log file structure is as shown below.

Section 1: General information regarding the optimisation setup.

========================= L O G - FILE - OPTFEKO =========================

Version: 13.22 of 2007-05-08

Date: 2007-06-06 16:32:51
File: test

OPTIMISATION WITH FEKO

=============== Optimisation variables ===============

No. Name Beg.value Minimum Maximum

1 zf0 2.000000000e+00 1.000000000e+00 1.000000000e+01

=============== Optimisation goals ===============

No. Name Expression

1 search1.goals.farfieldgoal1 farfieldgoal1

Section 2: Information regarding the grid search parameters.

=============== Optimisation method: GRID SEARCH ===============

No. Name Quantity Minimum Maximum Step

0 zf0 3 1.000000000e+00 1.000000000e+01 4.500000000e+00

July 2011 FEKO User’s Manual

THE OPTIMISER OPTFEKO 20-12

Section 3: Information regarding the parameter values and goal values at each step.

=============== GRID SEARCH: Intermediate results ===============

No. zf0 search1.goals.f global goal

1 1.000000000e+00 2.267123373e-01 7.732876627e-01
2 5.500000000e+00 2.267123373e-01 7.732876627e-01
3 1.000000000e+01 2.267123373e-01 7.732876627e-01

Section 4: Information regarding the termination reason and best results found on the search
grid.

=============== GRID SEARCH: Finished ===============

Optimisation finished (Maximum number of analyses reached: 3)

Optimum found for these parameters:

zf0 = 1.000000000e+01

Optimum aim function value (at no. 3): 7.732876627e-01

No. of the last analysis: 3

20.3 Sensitivity analysis

OPTFEKO computes an indication of the sensitivity of the goal function w.r.t. each parameter
upon termination of an optimisation using the PSO, GA or Simplex methods, if and only if suf-
ficient information is available. The computed sensitivity values are indicated on the screen
output, and stored in the text *.log file. If no sensitivity analysis is performed, the reason is
indicated on the screen output, but no indication is written to the text *.log file.

Figure 20-2: Sensitivity analysis of the goal function f w.r.t the parameter x.

Figure 20-2 shows an example goal function f that varies as a function of the parameter x. The
sensitivity w.r.t. parameter x can be described by the following equation:

∂f
S(x) = (20-2)
∂x x 0 ±∆x

with ∆x equal to 1
∆x = 0.01(x max − x min ). (20-3)

Solving Equation 20-2 directly, however, gives a near zero value when the solution space is
well converged. We therefore rather compute the second derivative from which the sensitivity
parameter can be computed through integration

July 2011 FEKO User’s Manual

THE OPTIMISER OPTFEKO 20-13

∂f ∂2f
S(x) ≈ ≈ · ∆x. (20-4)
∂x x 0 ±∆x ∂ x2 x0

to finally give the sensitivity w.r.t. x as

S(x) ≈ f 00 (x 0 ) · 0.01(x max − x min ). (20-5)

A sensitivity analysis will only be performed if at least 2N +1 samples are available for a problem
with N parameters and these samples should all be within a 5% radius of the optimum. If the
samples under consideration are scattered outside of a 5% radius of the optimum, the stored
data is considered insufficient for proper sensitivity analysis. It should also be realised that as
this computation makes use of already computed samples only, the accuracy of the reported
sensitivity number depends on how well the algorithm has converged.

20.4 Running OPTFEKO

Once a parametric FEKO model and accompanying *.opt file have been created, the optimiser
may be launched. The optimiser may be launched from any of the FEKO GUI components (Run
→OPTFEKO, or using the <Alt><6> shortcut key). Optimisation options may be set in the
Run →Component parameters →Utilities tab. When the optimiser is launched from the GUI,
an execution window will open in which text information regarding the optimisation progress is
displayed.
OPTFEKO may also be launched from the command line using the command:
optfeko dipole

Where dipole is the name of the optimisation project. On the command line the following
parameters may be added:

−−version Output only the version information to the command line and terminate.

-r All interim model files are deleted after each analysis. The optimum results are, how-
ever, not deleted, and are available with the string (_optimum) appended to the name.
This saves disk space during and after the optimisation process.

−−restart x Resumes an optimisation process that has been stopped, provided that all of the interim
optimisation files (*.fek *.bof and *.cfx) are still available (for example, the pre-
vious optimisation has been stopped by pressing <Ctrl><C> or due to a power failure
or a FEKO error etc.).

-np x The number of processors to be used for farming out of the individual optimisation
steps (see below).

−−machines-file machname The file machname is the machines file with the node names and the
number of CPUs to be used for farming of the individual optimisation steps. This ma-
chines file is used for both farming as well as parallel execution when farming and
parallel execution is used simultaneously. Some examples regarding this topic is dis-
cussed below.

July 2011 FEKO User’s Manual

THE OPTIMISER OPTFEKO 20-14

−−eval-aim-only x The value of the goal function is calculated only for one existing file (x) — no
optimisation is done. (This is mostly used for debugging.)

−−runfeko-options After this option one can specify additional options which will be used in the
launcher RUNFEKO for the FEKO kernel. For instance in order to use the parallel FEKO
solver during the optimisation one can use the command
optfeko file --runfeko-options -np 2

or also
optfeko file --runfeko-options -np 2 --machines-file m

where m indicates the machines file. For a remote execution of the FEKO runs during
the optimisation on another host, the suitable command would be
optfeko file --runfeko-options --remote-host hostname

Additional options for ADAPTFEKO and PREFEKO may also be included in the OPT-
FEKO command as part of the RUNFEKO options. These will be passed to the relevant
component by RUNFEKO as needed. This allows for control of all of the FEKO compo-
nents during the optimisation process. See section 18.2 for a list of all such RUNFEKO
options.

Farming and advanced control

Farming out of the steps of an optimisation involves the concurrent solution of various optimi-
sation steps on a number of available processors or hosts. Farming may not be used with the
Simplex method.
When farming out the individual optimisation steps, the number of processes to start on each
available host is specified in a so-called machines file. This machines file has a syntax identical to
that used for parallel runs. The basic syntax is
Hostname:Number of processes

using a new line for each host. For example, if two hosts are available with names host1 and
host2 (this is the output of the UNIX command hostname), and 4 and 8 processors respectively,
the machines file will be
host1:4
host2:8

Launching an optimisation run with 12 processors for farming using this machine file would
cause the first 4 optimisation steps to be launched on host1 and the next 8 steps to be launched
on host2. The example command is running model <x> and farming out to <n_farming>
processors listed in the machines file <m>.
optfeko <x> -np <n_farming> --machines-file <m>

During optimisation new model files are continuously created by adding the string _opt_ and a
sequentially incremented number to the file name of each relevant component file of the para-
metric model.
Some large problems may require that FEKO be run in parallel (simulation spread over more
than one host or processor - not the same as farming.) This is possible by adding the number of
cores that FEKO should use as a RUNFEKO option. The example command is running model <x>

July 2011 FEKO User’s Manual

THE OPTIMISER OPTFEKO 20-15

and each solution is using <n_parallel> processors listed in the machines file <m>. Note the
similarity and the differences between this command and the previous command.
optfeko <x> --runfeko-options -np <n_parallel> --machines-file <m>

It is also possible to use both the power of farming and parallel computing to optimally utilise
resources. The example command shows how to run model <x> by farming to <n_farming>
processors while running every solution on <n_parallel> processors. The available processors
are listed in a single machines file <m>. OPTFEKO automatically creates machines files for every
solution that is farmed out.
optfeko <x> -np <n_farming> --machines-file <m> --runfeko-options -np <n_parallel>

OPTFEKO output

Information on the optimisation process is stored in a log file with the extension .log — in the
example above the file name will be dipole.log. The structure of this file is dependent on
the optimisation method, and it is therefore discussed in the sections dealing with each of the
method algorithms.
It should be noted that when using the remote launching facility (see section 18.2.3) or farm-
ing out of optimisation steps, the actual optimisation is done on the local machine, only the
FEKO kernel runs (which are the time and memory consuming part) are done on the remote
machine(s).
The optimisation process may be interrupted at any time by clicking the Stop button in the GUI
process information window, or <Ctrl><C> in a command line. Confirmation of the interruption
will be requested, and if given, the optimisation process will be stopped. If an optimisation is
interrupted, all of the interim files created during the optimisation will be kept, except if the
Delete all files option was selected when running from the GUI, or if the -r option was added
when running from a command line. If the files were kept, the optimisation can be restarted at a
later stage using the –restart x option.

July 2011 FEKO User’s Manual

THE PROGRAM ADAPTFEKO 21-1

21 The program ADAPTFEKO

21.1 Description

In examples with narrow resonances a fine frequency resolution is required to locate these res-
onances. If the frequency band is large, a very large number of analyses may be required when
simple linear or multiplicative frequency stepping is used. ADAPTFEKO is used to overcome
these problems. It uses adaptive frequency sampling and interpolation methods to automatically
choose smaller frequency steps near resonances and larger steps where the results are relatively
smooth.
For each frequency ADAPTFEKO creates a *.pre file and calls PREFEKO and FEKO. The file
names are derived from the original name plus _fr_n_ada_m where n and m are incremental nu-
merical values (for example, the new files associated with test.pre are test_fr1_ada_1.pre,
test_fr1_ada_2.pre, . . .).

21.2 Running ADAPTFEKO

ADAPTFEKO is started automatically by RUNFEKO (see section 18.2) if a Continuous (interpo-

lated) range is selected in the CADFEKO Solution frequency setting or if the FR card (see sec-
tion 14.38) defining the range of solution frequencies contains the adaptive frequency sampling
option.
The syntax is:
runfeko filename

or
runfeko filename --adaptfeko-options [options]

where the optional argument options in the second line may be

−−version Output only the version information to the command line and terminate.

−−keep-files All solution files (*.pre, *.fek, *.out, etc.) are preserved.

−−restart x Restart an adaptive frequency analysis using results for the frequency points 1. . .(x-1)
obtained in a previous run. (Then the previous run must have used –keepfiles.)

21.3 The *.pre input file for adaptive sampling

The *.pre file is created (manually in EDITFEKO or automatically by CADFEKO) as for linear or
multiplicative stepping.
During solution, the variable #adaptfreq is defined automatically at the start of the single
frequency input files generated by the ADAPTFEKO utility. This variable may be used to allow
for solution frequency-based variation (for example, adaptive meshing). One should not directly
assign this name to a variable inside the *.pre file or in CADFEKO as this will overwrite the
value specified by ADAPTFEKO. If this variable is needed (for example to run PREFEKO during

July 2011 FEKO User’s Manual

THE PROGRAM ADAPTFEKO 21-2

model setup when using adaptive meshing), the DEFINED function should be used in the pre
file (an example of this usage shown below).
** define a frequency variable if it is not already defined by an ADAPTFEKO run
!!if (not(defined(#adaptfreq))) then
#adaptfreq = 250.0E6
!!endif

Note that care must be taken when using adaptive meshing with ADAPTFEKO. Small discontinu-
ities that may result from changes in the mesh can influence the convergence and accuracy of the
adaptive sampled results dramatically.

July 2011 FEKO User’s Manual

THE PROGRAM TIMEFEKO 22-1

22 The program TIMEFEKO

22.1 Description

With the program TIMEFEKO electromagnetic scattering problems can be solved in the time
domain. It is based on the program FEKO (that does the relevant calculations in the frequency
domain) and an FFT algorithm (that transforms the data to the time domain).
For the excitation a number of different pulses have been pre-defined and stored in the function
library (see section 22.3.1).
The program TIMEFEKO is constructed in such a way that all the data in the output file (*.out)
from FEKO is transformed, i.e. in the input file the cards have to be specified in the correct way to
ensure that the correct data is transformed. Information on the correct card definitions is given
in the following section.
The program TIMEFEKO uses two input files,11 that have to be created by the user.

22.2 The *.pre input file for TIMEFEKO

The input file with extension *.pre is the normal input file for PREFEKO, in which the frequency
has to be expressed i.t.o. the variable #freq. While the program TIMEFEKO is running, the
frequency is constantly changed. TIMEFEKO generates new *.pre files in which the actual
numerical value of the frequency is assigned. The user may use the DEFINED function to assign
a numerical value to the frequency variable #freq in the *.pre file — this is useful for running
PREFEKO and viewing the geometry in POSTFEKO (see the TIMEFEKO example in the Examples
Guide).
The FR card must use the frequency variable #freq. The number of frequencies to be examined
must be one (NFREQ=1).
Example:
** Set the frequency
FR 1 #freq

In the input file (*.pre), all the desired output parameters in their respective cards must be set
(for example, FF card, FE card, OS card, . . .). Only the data in the output file (*.out) can be
transformed.
It is also possible to vary the segmentation for certain applications (far field, near field) to save
memory and computation time. To do this one may use the variable #freq (which is constantly
changed by TIMEFEKO) in the definition of the segmentation parameters.
Example:
** Define some constants
#maxfreq = 250.0e+08
#minlambda = #c0/#maxfreq

** Define the edge length - note the use of #freq

#edgelen = (2.0-#freq/#maxfreq)*#minlamdba/4.0

** Set the segmentation parameters

IP #edgelen

11
The files have the same file name, but with *.pre and *.tim extensions.

July 2011 FEKO User’s Manual

THE PROGRAM TIMEFEKO 22-2

22.3 The *.tim input file

In the file *.tim the pulse form, position and the point in time are assigned. This file consists of
a number of sections which are optional at present.

• Set the pulse form with characteristic value and the time shift.
• Set the value of the highest frequency and the sampling points.
• Set the normalisation (normalise time to the speed of light).
• Set whether the output is written to the output file.
• Set the time points.

In the *.tim file, empty or comment lines (starting with **) are allowed. The parameters need
not be entered in any particular column, but they have to be in the correct order.
The keywords used in the *.tim file exist in German and English (for example ANREGUNG and
EXCITATION). Both versions of each keyword will be given in the discussion below.

22.3.1 Defining the pulse form

The assignment of the pulse form is necessary. Each pulse has a predefined name and has par-
ticular characteristic parameters. The parameters must be assigned absolute values and not
normalised values. The amplitude factor u0 is 1. It can be changed by using the AX card (in the
file *.pre) with the appropriate amplitude value.
Time shifting t 0 : indicates the time (in seconds) that the pulses are delayed (see the shift in
Figure 22-1). The time shift should be such that the excitation of the structure at t = 0 is
approximately 0. Since both the time and frequency domain data are continuous this is not
strictly required, but it simplifies working with the results.

July 2011 FEKO User’s Manual

THE PROGRAM TIMEFEKO 22-3

Figure 22-1: Time function u1 (t) shifted by t 0

The following pulses are available:

1. Gaussian pulse GAUSS / GAUSS (see Figure 22-2)

2
(t−t 0 )2
u1 (t) = u0 e−a (22-1)

Figure 22-2: Gaussian pulse

Example:
** Pulse form
GAUSS
** t0 Exponent a
2.0E-08 3.0E+08

2. Triangular pulse DREIECK / TRIANGLE (see Figure 22-3)


 u 1 − |t − t 0 |
 
for |t − t 0 | ≤ T
u1 (t) = 0
T (22-2)
0 for otherwise


Figure 22-3: Triangular pulse

Example:
** Pulse form
TRIANGLE
** t0 Impulse Duration T
2.0E-08 1.0E-08

July 2011 FEKO User’s Manual

THE PROGRAM TIMEFEKO 22-4

3. Double exponential pulse DEXP / DEXP (see Figure 22-4)


0 for t ≤ t0
t−t 0
 ‚ Œ
−

1 − e τ1 t0 ≤ t ≤ T + t0

u1 (t) = u1 for (22-3)


 t−t 0
−
u2 e τ2 for t ≥ T + t0
u0 u0
u1 = T −t 0
u2 = T −t 0
(22-4)
− −
1 − e τ1 e τ2

Figure 22-4: Double exponential pulse

Example:
** Pulse form
DEXP
** t0 T tau1 tau2
0.0 10.0E-09 5.0E-9 10.0E-9

4. Ramp pulse RAMPE / RAMP (see Figure 22-5)


0 for t ≤ t1 + t0

|t − t 2 − t 0 |
  
t1 + t0 ≤ t ≤ t2 + t0

u0 1 − for


τ1



u1 (t) = u0 for t2 + t0 ≤ t ≤ t3 + t0 (22-5)

|t − t 3 − t 0 |

  
u0 1 − for t3 + t0 ≤ t ≤ t4 + t0


τ


 2
0 for t ≥ t4
T + τ1 T − τ1 T − τ2 T + τ2
t1 = − t2 = − t3 = t4 = (22-6)
2 2 2 2

Example:
** Pulse form
RAMP
** t0 Impulse Duration T tau1 tau2
20.0E-09 15.0E-09 5.0E-09 10.0E-09

5. Double exponential impulse (second type) DBLEXP / DBLEXP (see Figure 22-6)

 0 for t ≤ t0
t−t 0 t−t 0
‚ Œ
u1 (t) = − − (22-7)
 u0 e τ1 − e τ2 for t > t0

July 2011 FEKO User’s Manual

THE PROGRAM TIMEFEKO 22-5

Figure 22-5: Ramp pulse

Figure 22-6: Double exponential impulse (second type)

Example:
** Impulse form
DBLEXP
** Time. t0 Parameter tau1 Parameter tau2
20.0E-9 70.0E-9 5.0E-9

22.3.2 Definition of the frequency block

The upper frequency limit f ma x and the number of frequency points N are defined in the fre-
quency block FREQUENZ / FREQUENCY.
Example:
FREQUENCY
** Upper frequency limit Number of frequency points
250.0E+06 34

The maximum frequency f ma x should be large enough such that the whole spectrum of the ex-
citing pulse is covered. For example, for the Gaussian pulse — GAUSS (see section 22.3.1), we
find
ap p
f3d B = ln 2 ≈ 0.187 a (22-8)
π
with a as defined in Equation 22-1. One should select approximately f max = 4 f3dB .

July 2011 FEKO User’s Manual

THE PROGRAM TIMEFEKO 22-6

The number of frequency points N is then selected so that the total time T is long enough for
the exciting pulse and all included currents, radiated fields, etc. to have decayed. Once the total
time has been determined, the number of samples may be determined from the relation
1 f max
∆f = = (22-9)
T (N − 1)
or
N = 1 + T f max (22-10)

Finally, let P be the smallest power of 2 which is larger than or equal to N − 1. (For example:
for N=10, P=16; for N=33, P=64; and for N=50, P=64.) Then the time stepping will be
∆t = 2PT
= 2 Nf −1 P . With the relation P2 < N − 1 ≤ P we get the bounds
max

1 1
< ∆t ≤ (22-11)
4 f max 2 f max

22.3.3 Definitions of the normalisation

Using the keyword NORM / NORM, the time can be normalised to the speed of light in a vacuum
c0 . The normalised time then has a unit of lm (light-metre).

t nor m = t · c0 (22-12)

Example:
** Normalising time with respect to the speed of light
NORM

22.3.4 Definition of the excitation output

The keyword ANREGUNG / EXCITATION indicates whether the time variation of the exciting
pulse is to appear in the output file.
Example:
** Output the excitation pulse
EXCITATION

22.3.5 Definition of a time point

With the keyword ZEITPUNKTE / POINTS_IN_TIME, the near fields, surface currents and line
currents can be calculated at certain points in time. Using time points has no effect on the
calculation of source currents or far fields. If normalisation has been set (NORM), then the time
points must also be given in normalised form.
Example:
** Specifying the time points -- time has been normalised
POINTS_IN_TIME
5.0 10.0 15.0 20.0

July 2011 FEKO User’s Manual

THE PROGRAM TIMEFEKO 22-7

22.4 Running TIMEFEKO

Firstly, the input files *.pre and *.tim have to be created. During execution new input files
(*.pre) are continuously generated by adding the string _tim_ and a sequentially incremented
number to the file name. If, for example, the input files cube.pre and cube.tim have been
created, TIMEFEKO is executed with the command
timefeko cube

On the command line the following parameters may be added:

-a Here an FFT is only performed on the already available data.

-r The *.pre and *.out files are deleted after each iteration. This saves disk space.

−−runfeko-options After this option one can specify additional options which will be used in the
launcher RUNFEKO for the FEKO kernel. For instance in order to use the parallel FEKO
solver during the time domain analysis one can use the command
timefeko file --runfeko-options -np 2

or also
timefeko file --runfeko-options -np 2 --machines-file m

For a remote execution of the FEKO runs during the analysis on another host, the suit-
able command would be
timefeko file --remote-host hostname

See section 18.2 for a list of all such RUNFEKO options.

−−version Output only the version information to the command line and terminate.

It should be noted that when using the remote launching facility (see section 18.2.3)), the actual
TIMEFEKO process is executed on the local machine, only the FEKO kernel runs (which are the
time and memory consuming part) are done on the remote machine.

22.5 TIMEFEKO output

TIMEFEKO generates different *.pre files at the different frequencies. The *.out files for all
these runs are available if TIMEFEKO was called without the -r option. These results are sum-
marised — if they are requested in the *.pre file — in *.ost (triangle currents), *.oss (seg-
ment currents), *.cur (the currents on voltage sources), *.far (far fields), *.nfe (near electric
fields) and *.nfh (near magnetic fields). In each case the results for each frequency is listed one
after the other with frequencies separated by a line containing a single #-character.
The time domain results are given in the *.aus file for all outputs requested in the *.pre file.
Currents are requested by the OS card. Note that requesting the current on a large number of
structures will result in very large *.aus files.

July 2011 FEKO User’s Manual

THE FEKO SOFTWARE UPDATER 23-1

23 The FEKO software updater

23.1 Introduction

With the FEKO GUI update utility or the command-line FEKO_UPDATE utility the latest FEKO
updates can automatically be downloaded and installed. Users are given the option to download
the latest FEKO update from a master or a local repository. The master repository is the FEKO
website and requires internet access. The command-line updating utility enables scripted updates
whilst the GUI version of the updater is interactive and allows the user to set preferences as well
as checking for new updates.
Note that no updates will be made available for FEKO Lite installations. Other installations will
require a valid M&S licence to download updates. FEKO Suite updates will only apply to the
current FEKO suite installation.
No information is sent to the FEKO website during an update, a list containing the latest software
is retrieved and compared to the currently installed components.

23.2 GUI update utility

The FEKO software updating utility is automatically launched every time one of the GUI com-
ponents is launched, but it only checks for updates if the check interval time has elapsed (the
default check interval is one week). When updates are available the user will be notified and also
has an option to download and install the updates. It is also possible to force a Check for update
from within CADFEKO, EDITFEKO and POSTFEKO. Select Help →Check for updates.
On the Info tab, information regarding the available updates will be displayed. Settings regarding
the update schedule, update source and proxy settings are given on the Settings tab as shown in
Figure 23-1.
The name of the GUI update utility is feko_updchecker_gui and allows the following options
to be set as well as checking for updates and performing updates.

Schedule: The user may set the intervals for when the FEKO updater should check for
updates. The option to automatically check for updates can also be disabled by
unchecking the checkbox Check for updates automatically on the settings tab
of the FEKO update dialog.

Update from: The web or a local repository may be set as the desired location from where the
latest updates are to be downloaded from. The user has the option to download
updates via the web or from a local repository. Using the local repository is
recommended when the computer network or cluster does not have internet
access due to security reasons or only limited bandwidth is available. In this
case, the updates may be downloaded from the FEKO website by the system
administrator and placed at a location accessible for the computer network or
cluster. The Local repository can be set on the Settings tab of the FEKO update
dialog. Select Local repository and enter the location of the updates.
When the Web option is selected, the updates will be downloaded from the
FEKO website.

July 2011 FEKO User’s Manual

THE FEKO SOFTWARE UPDATER 23-2

Proxy: If the web is used as the repository, the user can set the proxy server and
authentication (if required).

Figure 23-1: The FEKO update dialog with the Settings tab selected

23.3 Command-line update utility

FEKO_UPDATE is a command-line application that enables scripted updates. Launching the appli-
cation without any arguments will list the possible argument options. In most cases feko_update
will be launched as shown below to check for updates and then download and install any avail-
able updates.
feko_update --update

The optional arguments for feko_update are:

−−check [ <host>[:port][user [password]]] Check if updates are available from the master
repository. If the computer is behind a proxy server, the proxy server address and
the login details can be supplied as required. If updates are available, the following
information is printed to the screen:

• The date of the available update.

• A list of the updated FEKO Suite components with their version numbers.
• The total download size of the update.
• The release notes for all updates for the Suite version.

If no updates are available, the user is informed and the application exits.

−−update [ <host>[:port][user [password]]] Check if updates are available from the mas-
ter repository, then download and install all available updates. If the computer is behind
a proxy server, the proxy server address and the login details can be optionally supplied.
If updates are available, the following will occur:

July 2011 FEKO User’s Manual

THE FEKO SOFTWARE UPDATER 23-3

• Print the same output as that of the −−check option.

• Print each file which is being downloaded.
• Print each file which is being updated.
• Print a message stating that the update was successful and exit.

If no updates are available, the user is informed and the application exits.

−−check-from <path> Similar to the −−check option, the updater checks if updates are available,
but the update source is not the FEKO website. The update source is the local repository
specified by <path>.

−−update-from <path> Similar to the −−update option, the updater checks if updates are avail-
able and install the updates, but the update source is not the FEKO website. The update
source is the local repository specified by <path>. The path must be an absolute file
path which can point to a unmapped network share (Windows), mapped (mounted)
network share or a directory on a local drive.

−−version Print the version information and then exit.

23.4 Proxy settings - advanced

The FEKO updater (GUI and command line versions) will use the system proxy by default. On
Windows the proxy is the same as is used by Internet Explorer. The proxy can be specified
manually or by using PAC file. On Linux the system proxy is defined by the environment variable
http_proxy. If the the environment variable http_proxy is not defined, then no proxy will be
used.
The parameter −−no-proxy was added to bypass the system settings and use a direct connec-
tion.

23.5 Creating a local update repository

The following describes the steps required for the system administrator to create a local FEKO
update repository. Note that the path for the local FEKO update repository must be an absolute
file path which can point to a unmapped network share (Windows), mapped (mounted) network
share or a directory on a local drive.

• If previous updates are still located at the required destination, first delete the old updates.

• Download the updates for the required platforms.

• Unzip the folders in the required destination.

If multiple platforms are downloaded, ensure that the various platform updates are located at the
same destination. The zip file contains an entire directory structure which must be kept intact.
When updates for multiple platforms are required, the directories must be merged.
If for example WIN64_AMD64 and WIN64_EM64T updates were downloaded to
C:\Updates, unzipped and merged, the following directory tree will result:

July 2011 FEKO User’s Manual

THE FEKO SOFTWARE UPDATER 23-4

C:\
|__Updates
|__FEKO_EMSS
|__SuiteXX.YY
|__Win64_AMD64
|__Win64_EM64T

In the above case set the local repository path to C:\Updates (or a network reachable equivalent
UNC path or mapped drive). Always point the local repository path to the directory that contains
the root item of the *.zip file.

July 2011 FEKO User’s Manual

Part IX

Appendix
ADVANCED MODELLING AND SOLUTION CONTROL 24-1

24 Advanced modelling and solution control

24.1 Utilisation of symmetry

It is possible to reduce the calculation time and memory usage if symmetry is utilised. This can be
done by using the SY card (see section 13.43), or by setting symmetry for the model in CADFEKO
(see section 6.8).
Three coordinate planes, x = 0 ( yz plane), y = 0 (xz plane) and z = 0 (x y plane) may be
defined as planes of symmetry. There are three different types of symmetry. They are described
below.

24.1.1 Geometric symmetry

With this type of symmetry the geometry of the modelled solid or part of the solid is symmetric
about one or more coordinate planes. The interaction between any two basis functions must be
the same as that between their symmetrical counterparts. Everything which affects this must be
symmetrical, i.e. loading, losses, Green’s functions, etc. The source, however, is not symmetric,
thus a symmetric current distribution does not exist. This asymmetric current distribution leads
to asymmetric electric and magnetic fields.
The body of a truck with an antenna placed at the front left hand side, will be used as an example.
In the input file, half of the body is constructed (either the left or right side). The other half is
then created with the SY command. Finally the antenna is placed in the correct position on one
side.
A rectangular metallic plate, illuminated by an electromagnetic wave from a direction outside
the principle planes, is another example. In this case a quarter of the plate is constructed and
the rest is created using the SY card (see section 13.43) with geometric mirroring around two
coordinate planes.
When defining symmetry in CADFEKO, symmetrical geometry must be created and the relevant
symmetry set before meshing. In the case where a model is to be considered that is not entirely
symmetrical (as in the first example above), the asymmetrical part/s of the model must be added
by editing the pre file and cannot be defined in CADFEKO.
Geometric symmetry does not reduce the number of unknown coefficients in the current basis
functions. Therefore there is no reduction in memory usage. There is, however, a reduction in
computation time when the matrix elements are determined.

24.1.2 Electric symmetry

An electric symmetry plane is a plane which can be replaced by an ideal electrically conducting
wall without changing the field distribution.
~ field’s tangential component
In Figure 24-1 an electric symmetry plane is displayed. The electric E
disappears and the magnetic H ~ field’s normal component disappears. The electric current density
~
J is anti-symmetric and the magnetic current density M ~ is symmetric.

July 2011 FEKO User’s Manual

ADVANCED MODELLING AND SOLUTION CONTROL 24-2

As with geometric symmetry, less computational time is required to calculate the matrix elements.
The number of unknown coefficients of the current basis functions are also reduced, leading to
linear equation system of a lower order. This leads to a further reduction in the computation time
and requires less memory due to the reduction in the number of matrix elements.

Figure 24-1: Electric symmetry plane

24.1.3 Magnetic symmetry

A magnetic symmetry plane is a plane which can be replaced by an ideal magnetically conducting
wall without changing the field distribution.
In Figure 24-2 a magnetic symmetry plane is displayed.

Figure 24-2: Magnetic symmetry plane

The electric E~ field’s normal component and the magnetic H ~ field’s tangential disappears. The
~
electric current density J is symmetric and the magnetic current density M~ is asymmetric.

When using magnetic symmetry there is a reduction in the computational time when determining
the matrix elements. The order of the matrix equation is also reduced, which leads to a further
reduction in the computational time and reduces the amount of memory needed to determine
the matrix elements.

24.1.4 Example of the application of symmetry

In Figure 24-3 a dielectric sphere is shown with a linear polarised incident electromagnetic field.
The full description of the problem is given in example_04 in the Script Examples. Only the use
of symmetry is described here.
The plane z = 0 (x y plane) is a plane of geometric symmetry, because the excitation does not
have any symmetry in this plane. The plane x = 0 ( yz plane) is a plane of electric symmetry,
because the electric field is perpendicular to this plane and the magnetic field has a tangential
component only. Similarly the electric field only has a tangential component in the y = 0 plane

July 2011 FEKO User’s Manual

ADVANCED MODELLING AND SOLUTION CONTROL 24-3

Figure 24-3: Dielectric sphere with incident field

Table 24-1: A summary of the Simplex operations

Symmetry type Number of Memory Solution

unknowns usage time [s]
Symmetry not used 1032 8.38 MByte 3.860
All 3 coordinate planes declared as planes of 1032 8.41 MByte 2.015
geometric symmetry
Plane x = 0 declared as electric plane of symmetry, 258 2.36 MByte 1.094
plane y = 0 as magnetic plane of symmetry and plane
z = 0 as geometric plane of symmetry

and the magnetic field is perpendicular. This is thus a plane of magnetic symmetry. The reduction
in time and resources is tabulated below 12 .
This example has relatively few unknowns. Most of the computational time is therefore used
to determine the matrix elements in comparison to the time taken to solve the matrix equation.
For applications with more unknowns, the reduction of unknowns could make a considerable
difference in the time and memory required.

24.1.5 Special enforcement of symmetry: Even – odd method

The table in the previous section demonstrates the advantage of using electric and/or magnetic
symmetry. For very large structures which have only geometrical symmetry, it may be worthwhile
to consider two separate problems with electrical and magnetic symmetry as described below.
Figure 24-4 (a) shows the original problem. It consists of a dipole antenna with a passive wire
below it. This is admittedly a very simple problem, normally this procedure would only be applied
to much more complex structures.
The structure in Figure 24-4 is symmetric about the plane z = 0, but the excitation is asymmetric
and thus only geometric symmetry can be applied in FEKO. This problem may be separated into
12
The runtimes in this table are dependant on the platform on which the problem is run. These numbers are
therefore only indicative of the relative effects on overall runtime.

July 2011 FEKO User’s Manual

ADVANCED MODELLING AND SOLUTION CONTROL 24-4

Figure 24-4: Separating a problem with geometrical symmetry into two problems with electric and mag-
netic symmetry respectively.

the two sub-problems shown in Figures 24-4 (b) and (c). These two problems can be solved
using respectively electric and magnetic symmetry about the plane z = 0. Each of these problems
require only half the number of unknowns required for case (a). Superposition of the problems
(b) and (c) (this must unfortunately be done externally to FEKO, as such superposition is not
currently catered for) yields a result to the original problem.
The solution of each of the sub-problems requires only half of the storage space required for case
(a). For very large problems the solution time is dominated by the time required to solve the
system of linear equations. In this case the two sub-problems each requires only one eighth of
the time required for direct solution of case (a).

24.2 Dynamic memory management

24.2.1 Telling FEKO how much memory can be used

FEKO has the ability to manage the memory dynamically, i.e. the memory required for the geom-
etry data and matrix elements etc. is determined and allocated at run time. When FEKO tries to
allocate memory, in principle the operating system offers a certain address space, which might
either be physically installed memory (i.e. RAM), but also virtual memory (system swap space
swapped to the hard disk).
If FEKO starts to swap using virtual memory, then the whole solution process can be slowed down
quite significantly, and this is not recommended. FEKO also has an out-of-core solution which
uses the data on disk in a much more efficient way. (The out-of-core technique is also used, of
course, if the problem requires more memory than is available in both RAM and virtual memory.)
For solutions which do not fit into the available RAM, but do fit into the RAM plus virtual memory,
the user should tell FEKO what the actual physically installed memory is, less a certain reserve
for the operating system.
Up to FEKO Suite 4.2 the variables #maxalloc or #maxallocm were used for this purpose:

#maxallocm This sets the maximum allocatable memory in MBytes that can be used together by
all FEKO processes launched as part of one parallel FEKO job per host. For example

July 2011 FEKO User’s Manual

ADVANCED MODELLING AND SOLUTION CONTROL 24-5

the definition #maxallocm = 400 will allow a maximum of 400 MByte of memory
to be allocated. If this is not enough, the matrix will be saved to the hard disk or
the program will be halted. For parallel versions of FEKO the memory limit for each
process is determined by dividing the value of #maxallocm by the number of parallel
processes on the same host. Note that if #maxallocm is used, it will have preference,
and #maxalloc will be ignored if it is used in the same *.pre file.

#maxalloc Similar to #maxalloc but just specifies the memory in Bytes and not in MBytes.

These two variables #maxalloc and #maxallocm are still supported in FEKO Suite 4.2 and later
for backwards compatibility reasons, however, their usage is strongly discouraged when using
newer FKO versions.
In newer FEKO versions (Windows and Linux) the FEKO memory management is fully automatic
and the usage of virtual memory (swap space) is avoided as far as possible. On some UNIX
versions (such as HP-UX or SUN Solaris), an environment variable FEKO_MAXALLOCM is defined
during the installation and set in the initfeko initialisation script (this specifies the physically
installed RAM less an operating system reserve for a specific machine). The advantage then is
that the FEKO *.pre files are machine independent, i.e. if you have two different computers
with different memory, no variable #maxallocm has to be changed in the *.pre file as such,
everything is either fully automatic (Windows or Linux) or defined on a per machine basis (other
UNIX versions).

24.2.2 Variables that are automatically set correctly

Note that normally PREFEKO estimates the following variables correctly and they should only be
declared in cases where there is an explicit error message stating that larger memory blocks are
required. Under normal circumstances these variables should not be set, as they could have a negative
impact on FEKO performance.

#maxaeedges The maximum number of edges between triangles that may be excited with the AE card.

#maxanr The maximum number of sources.

#maxapo The size of the memory block that is used to save the coefficients in the physical optics
approximation. For #maxapo=0 the necessary amount will be dynamically allocated.

#maxarang The maximum number of ϑ or ϕ angles used with the AR card (excitation by a point
source with a specified radiation pattern).

#maxarpat The maximum number of radiation pattern excitations (AR card) allowed simultane-
ously.

#maxbsobnr To accelerate the ray path search with PO the area under consideration is divided into
boxes. Information pertaining to which box contains which object must be stored. A
field of size #maxbsobnr is used in this case.

#maxcolayer The maximum number of layers on a CO card which implements thin dielectric sheets.

#maxdrnv The maximum number of triangle elements that can be connected to a segment at an
attachment point.

July 2011 FEKO User’s Manual

ADVANCED MODELLING AND SOLUTION CONTROL 24-6

#maxfepkts The maximum number of points considered for the near field computation with the FE
card when using Specified points.

#maxfoge The maximum number of areas that are described by using the Fock theory.

#maxgfmsia The maximum number of entries in the interpolation tables for the Green’s function of
a planar substrate.

#maxhacards The maximum number of HA cards (used internally to set up microstrip ports) that may
be present in the *.fek file.

#maxkanr The maximum number of “internal” edges (also the number of basis functions) per
triangle. It may be larger than 3 if more than two plates share an edge.

#maxknonr The maximum number of nodes that may lie against a segment.

#maxl4cards The maximum number of L4 type loads.

#maxlab The maximum number of labels in a model.

#maxlecards The maximum number of LE cards (which specify a load on an edge between triangles).
#maxleedges The maximum number of edges between two surface triangles that can be loaded with
a single LE card.

#maxlengz Dimension of the interpolation table used for the planar multilayer Green’s functions.
This variable determines the maximum number of sample points in the z direction.

#maxmedia The maximum number of different media used for the treatment of dielectric bodies
in the surface equivalence principle. The surrounding free space (medium 0) is not
counted (i.e. with #maxmedia=1 one dielectric body can be treated).

#maxnp The maximum number of columns and rows which a block in the matrix consists of in
the Block-Gauss algorithm which solves the matrix equation.
Dynamically 3*#maxnp*#maxnp*16 Bytes are allocated for 3 blocks in the matrix.

#maxnv The maximum number of connection points between wires and surface triangles.

#maxndr The maximum number of triangles.

#maxnka The maximum number of edges between two triangles.

#maxnkapo The maximum number of edges in the physical optics approximation.

#maxnkno The maximum number of nodes between segments.

#maxnlayer The maximum number of layers for the special Green’s function of a planar substrate.

#maxnqua The maximum number of dielectric cuboids.

#maxnseg The maximum number of segments.

#maxntetra The maximum number of tetrahedral volume elements for a FEM solution.

#maxnzeile The maximum number of basis functions in the moment method area.

#maxpoka The maximum number of bordering edges to the PO area.

#maxpokl The maximum number of wedges in the PO area.

July 2011 FEKO User’s Manual

ADVANCED MODELLING AND SOLUTION CONTROL 24-7

#maxpolyf The maximum number of polygonal plates that can be used to represent a body in the
UTD region.

#maxpolyp The maximum number of corner points allowed for a polygonal plate.

#maxpovs The maximum number of label to label visibility specifications set by VS cards (a card
with a range sets a number of entries equal to the size of the range).

#maxsklayer The maximum number of layers at an SK card.

#maxtlcards The maximum number of TL cards.

#maxutdzyl The maximum number of cylinders in the UTD region.

#nmat The memory size that may be allocated for the matrix of the system of linear equations.
For #nmat=0, the necessary amount will be allocated dynamically. The allocation is
not specified in Bytes, but in terms of the number of type DOUBLE COMPLEX num-
bers. (These require 16 Bytes each.) For example, 400 MByte is specified by setting
#nmat = 400*1024*1024/16. The same effect can be achieved by setting the vari-
able #maxalloc so that it is unusual to assign a value to #nmat.

#npuf The maximum number of control cards that may occur in a loop (for example in a
frequency sweep).

24.3 Environment variables and registry keys

This section lists the environment variables that may be used to control the execution of FEKO.
From Suite 5.5 the default location for these settings under Windows are in the registry and not
as environment variables. The preferred way of changing any of these settings under Windows is
by using initfeko.bat. The syntax for initfeko.bat is as follows:

-setreg Write the environment values (as set in initfeko.bat) to the registry. Default values
are defined in initfeko.bt directly by the initial installer and can be modified to
match the users’ needs. Note that the current environment/shell will not be changed.

-queryreg Query the registry and set the values to the current environment if they are not already
set.

-console Opens a FEKO Console Window for running commands directly in the console. All re-
quired environment settings will be loaded (see above option “-queryreg”) and applied
to the current shell.

-d Shows additional debug output while setting the environment.

These variables are also discussed in the FEKO Installation Guide, with respect to the script
initfeko or the batch file initfeko.bat automatically created by the FEKO installation pro-
gram. During the installation process, the environment variables are set correctly and the user
does not generally need to set environment variables manually.
The environment variables can also be specified with registry keys that can be found in
HKLM\SOFTWARE\FEKO\major.minor\Environment (global)

July 2011 FEKO User’s Manual

ADVANCED MODELLING AND SOLUTION CONTROL 24-8

HKCU\SOFTWARE\FEKO\major.minor\Environment (current user)

where major.minor is the FEKO Suite (e.g. 5.5). The variables set in the environment have
precedence over variables in the registry. Also note that the ones in HKCU have precedence over
ones in the HKLM. Please note that the preferred way to change values in the registry is to use
initfeko.bat. Changing the environment variables directly is strongly discouraged.
The following environment variables may be set:

FEKO_CMDINFO If this environment variable is set to the value 1, FEKO writes additional data concern-
ing the number and the value of the received command line parameters to the screen.
This can be useful to trace errors in the parallel version of FEKO used in connection
with some implementations of mpirun, mpiopt, mpprun etc.

FEKO_HOME This variable is set to the FEKO installation path which contains the subdirectories such
as bin, doc, license and, for the parallel version, mpi.
Note that it is not recommended to modify this environment variable.

FEKO_LICENSE_FILE This variable is used to specify the location of the FEKO licence file if it is not
located in the default directory with the default name. The default name and location
is secfeko.dat and %FEKO_HOME%\license (see section 15.1).

FEKO_LITE If this environment variable is set to 1, then FEKO runs for 30 days in the unregistered
FEKO LITE mode. FEKO is then restricted regarding the element sizes etc., but does not
need a licence (see the FEKO Installation Guide for more details).

FEKO_MACHFILE The parallel version of FEKO is started by running RUNFEKO with options -np x.
When FEKO is installed on a parallel computer or a computer cluster, the configuration
of the cluster and the number of processes that should be run on each computer is
specified during the installation. This can be overwritten for any FEKO run by creating
a so-called machines file and setting the environment variable FEKO_MACHFILE to point
to this file. (More detail can be found in section 18.2.)

FEKO_MACHINFO If this parameter is set, FEKO will write information about the machine precision to
both the screen and the output file.

FEKO_MAXALLOCM This environment variable is used to limit memory (in MByte) that FEKO is allowed
to use on this computer. This environment variable is not needed or recommended for
computers running Windows or Linux operating systems. On others the variable is set
at installation time. The value of this variable should usually be set equal to the physical
memory minus 70 MBytes for the operating system. In a few cases a lower limit may be
required, and should be set here.

FEKO_MPISTATISTICS This environment variable provides additional information about the perfor-
mance of the parallel version of FEKO. There are three options:

1: Give a detailed report of the CPU and run times for the individual processes. It is,
for example, possible to determine how much time each process required during
the computation of the array elements.
2: Give as additional output the MFLOPS rate of each process (without network com-
munication time). This is useful to determine the relative performance of nodes
in a heterogeneous cluster.
4: Give information about the network performance (latency and bandwidth). This is
very useful when configuring parallel clusters.

July 2011 FEKO User’s Manual

ADVANCED MODELLING AND SOLUTION CONTROL 24-9

The options can be added in a binary fashion, for example setting the environment
variable equal to 5 will print both the run times and network performance.

FEKO_PARALLEL_DEBUG For parallel runs of FEKO under UNIX, this environment variable can be set
to 1 in order to see all the details and commands used in the parallel launching and
machines file parsing etc. This is helpful for troubleshooting errors.

FEKO_RSH When installing the parallel FEKO version on a UNIX cluster, then communication be-
tween the nodes is required both at installation time (checks on the remote nodes,
remote copying of files, remote execution of utilities etc.), but also when using FEKO
(remote launching of parallel FEKO processes). By default both the installation script
and the parallel launcher will use the remote shell for this purpose (remsh under HP-
UX and rsh for all other UNIX platforms). A typical setup is then to use a /̃.rhosts
file (see detailed comments given during the installation). But this is not quite se-
cure, and one might prefer to rather use the secure shell ssh in connection with public
key authentication (avoids having to type passwords all the time). The actual remote
shell executable (e.g. rsh or remsh or ssh will be determined during the installation
procedure, and the environment variable FEKO_RSH will be set to point to this exe-
cutable. This can always be changed later (e.g. using rsh for the installation as root,
but then ssh for the users using the parallel FEKO version or vice versa) by changing the
value of the environment variable FEKO_RSH in the FEKO initialisation file initfeko
or initfeko.csh accordingly. One can also set this on a user per user basis, then di-
rectly for instance in ∼/.profile after having executed initfeko. This environment
variable will not be used on Windows systems.

FEKO_TMPDIR This variable specifies the directory where FEKO will write paging files, when using the
out-of-core solution. In the past it was required that the definition ended in a backslash
(Windows) or a slash (UNIX). This is no longer required. For example, in UNIX it may
be set with
set FEKO_TMPDIR=/tmp
export FEKO_TMPDIR
FEKO_USER_HOME This directory is used to write user specific initialisation files. This variable replaced
FEKO_WRITE. It is provided to allow different users to save unique configurations, and
for situations where the user does not have write access to the FEKO directory. For
Windows systems this is normally %APPDATA%\feko\xx.yy and on UNIX systems it is
usually set to $HOME/.feko/xx.yy during the installation. Here xx.yy represent the
major and minor suite version numbers.

FEKO_WHICH_MPI FEKO uses different MPI implementations for the different platforms and thus the
different platforms require different command syntax to start FEKO. RUNFEKO pro-
vides an interface that remains the same on all platforms. However, it must know
which MPI implementation is used. This is done by setting the environment variable
FEKO_WHICH_MPI (it is automatically set by the installation script) to one of the fol-
lowing options:

1: MPICH2 (all supported UNIX platforms, currently Linux on Intel, Alpha and SUN.)
2: HP-MPI for HP-UX
3: NEC MPI on NEC SX/5
4: SGI MPI on SGI13
5: CRAY/SGI MPI on CRAY T3E
13
Support for SGI MIPS IRIX discontinued from Suite 5.4

July 2011 FEKO User’s Manual

ADVANCED MODELLING AND SOLUTION CONTROL 24-10

6: ScaMPI (Linux on Intel/AMD and Alpha with SCI/Myrinet/Infiniband interfaces)

7: Compaq MPI
8: SCore (Linux on Intel with Myrinet interfaces)
9: GM (Linux on Intel with Myrinet interfaces)
10: ParaStation MPI
11: Intel MPI (Linux on Intel IA32, IA64, EM64T and on AMD64)
12: IBM Parallel Environment (PE)
13: MS MPI (Microsoft MPI)

FEKO_WHICH_SPICE_EXECUTABLE This variable specifies which SPICE solver is to be used. When left
unset (default), FEKO will use the NGSPICE executable (included in FEKO installation).
A user can set this variable to point to the full path and executable name of the preferred
SPICE solver.

FEKO_WRITE_RHS If this environment variable is set (value arbitrary), FEKO writes the right side of the
set of linear equations to a *.rhs file. This is only useful for test purposes, such as
when one wants to analyse this vector with another program.

MKL_SERIAL If this is set to YES, FEKO (and all other codes using the Intel MKL libraries) will run as
a single threaded application (i.e. it will not utilise multiple CPUs) irrespective of the
value of OMP_NUM_THREADS (see below).

OMP_NUM_THREADS Sequential FEKO versions for Windows or Linux PCs (Intel or compatible) can use
multiple CPUs on one board for the LU-decomposition of the MoM matrix (the parallel
FEKO version will have all phases of the solution in parallel). To use this, the environ-
ment variable OMP_NUM_THREADS must be set to the number of CPUs to use. (See also
MKL_SERIAL above.)

July 2011 FEKO User’s Manual

SUMMARY OF FILES 25-1

25 Summary of files

The table below gives an overview of the different files and their respective functions. (STDOUT
is the standard output - usually the screen; a *-symbol is used to symbolically indicate the file
name.)

STDOUT Usually the screen. This is where comments such as progress, warnings and errors are
sent.

*._14 Page (temporary storage) file for the matrix in the sequential version of FEKO with
out-of-core solution.

*._15 Page (temporary storage) file for the matrix in the sequential version of FEKO with
out-of-core solution.

*._16 Page (temporary storage) file for the matrix in the sequential version of FEKO with
out-of-core solution.

*._20 Page (temporary storage) file for the calculated coupling coefficients of the MoM/PO
hybrid method during an out-of-core solution.

*._21 Page (temporary storage) file for the calculated coupling coefficients of the MoM/PO
hybrid method during an out-of-core solution.

*.afo Continuous frequency results created by ADAPTFEKO.

*.aus Output file of TIMEFEKO.

*.bof Binary version of the output file which is used for post processing.

*.cdb ANSYS mesh file which can be imported with the IN card.

*.cfm CADFEKO mesh file (exported file containing CADFEKO mesh and variables to be im-
ported by PREFEKO).

*.cfs Session file of CADFEKO.

*.cfx Native CADFEKO model file (contains geometry, mesh, solution settings, optimisation
settings etc.)

*.chr FEST3D file containing the generalised S-parameter matrix and may be exported using
the DA card.

*.cir SPICE file which describes a circuit and may be included as a non-radiating general
network by means of the NW card.

*.cgm Contains the size of the residue that results from the iterative algorithm which solves
the matrix equation and the number of iterations. This file is only generated on request
by a DA card (see section 14.31).

*.dbg When using the UTD, it is possible to request an optional output file containing a large
amount of additional data (and may therefore be very large), see the UT card.

*.dxf AutoCAD geometry file which may be imported using the IN card. (Arbitrary surfaces
from meshed *.dxf files may be imported. Lines and polyline surfaces can also be
import and meshed - see IN card (see section 13.22).)

July 2011 FEKO User’s Manual

SUMMARY OF FILES 25-2

*.efe File containing the electric field strengths. Contains both the position and the complex
components of the electric field strength vectors. This file is only generated on request
by a DA card (see section 14.31).

*.fek Output file from PREFEKO — serves as the input file for FEKO.

*.fim FEST3D file containing the waveguide modes which may be imported using the AW
card.

*.ffe File containing the far field data. This file is only generated on request by a DA card
(see section 14.31).

*.gfe Interpolation table of the electric field strengths for the Green’s function of a layered
sphere.

*.gfh Interpolation table of the magnetic field strengths for the Green’s function of a layered
sphere.

*.hfe File containing the magnetic field strengths. Contains both the position and the complex
components of the magnetic field strength vectors. This file is only generated on request
by a DA card (see section 14.31).

*.inc Include file for PREFEKO.

*.inp ABAQUS mesh file.

*.isd Data file containing the field distribution calculated by FEKO for coupling with Cable-
Mod, CRIPTE or PCBMod.

*.log Log file created by OPTFEKO.

*.lud File in which the elements of the LU-decomposed MoM matrix are stored (only gener-
ated on request of a PS card (see section 14.54).

*.mat File in which the matrix elements of the linear equation system, are stored (only gener-
ated on request of a PS card (see section 14.54).

*.nas NASTRAN geometry file that may be imported using the IN card.

*.neu Geometrical data file which is exported by the program FEMAP.

*.opt Input file for the program OPTFEKO.

*.ofc Paging files for the array elements used with sequential and parallel out-of-core solu-
tion. (To avoid the 2 GByte file size limit; or on parallel systems with a distributed file
system, several files may be used. These are distinguished by adding numbers to the
file name.)

*.ol File containing the surface charges and the charges in the segments. The data includes
the physical position and the complex charge density. This file is only generated on
request by a DA card (see section 14.31).

*.os File containing the surface currents and the currents in the segments. The data includes
the physical position and the complex components of the current density vectors. This
file is only generated on request by a DA card (see section 14.31).

July 2011 FEKO User’s Manual

SUMMARY OF FILES 25-3

*.out The ASCII output file generated by the FEKO solver, in which the results of all the
calculations and messages can be found in a human-readable format.

*.pcr Exported ILU preconditioner for the FEM.

*.pfg POSTFEKO graph file. (The *.pfg file is also used to store optimisation process infor-
mation used for graphing in POSTFEKO after/during an optimisation run.)

*.pfs POSTFEKO session file.

*.pre Input file for PREFEKO.

*.ray When using UTD or GO an optional ray file can be requested. This file is not required
when visualising rays in POSTFEKO.

*.rhs File containing the right hand side vector in the system of linear equations.

*.rsd File for coupling of FEKO with CableMod, CRIPTE or PCBMod. It is usually created by
CableMod, CRIPTE or PCBMod, but can also be created by FEKO if requested with the
OS card (field calculation along lines).

*.sha File storing shadowing information for the PO.

*.snp Touchstone format (v1.0) S-parameter file as created by the DA card (see section 14.31).
The n refers to the number of ports.

*.sph Spherical wave expansion (SWE) as used by the reflector antenna code GRASP from
TICRA (may be exported during a FEKO solution using the DA card (see section 14.31)
or used to define a spherical mode excitation in a FEKO solution as part of an AS card).

*.str File in which the coefficients of the basis functions are stored for reuse (generated on
request from a PS card (see section 14.54).

*.tim Input file for the program TIMEFEKO.

*.vis When multiple reflections are used with the PO formulation FEKO determines which
basis functions have line of sight visibility. Since this calculation may require significant
run time, this information can be saved to a *.vis file for reuse.

*.wfg Figure created and saved with GraphFEKO.

*.x_t/*.x_b Parasolid model file.

The files *.efe, *.hfe, *.ffe, *.os, and *.snp are redundant. All the information in these
files is also available in human-readable format in the *.out file (if not explicitly switched off
at the DA card). The format of these redundant files may lend itself more readily to further
processing outside of POSTFEKO.

July 2011 FEKO User’s Manual

SPICE3F5 GENERAL STRUCTURE, CONVENTIONS AND SYNTAX 26-1

26 SPICE3f5 general structure, conventions and syntax

A SPICE circuit to be analysed is described by a set of element lines, which define the circuit
topology and element values, and a set of control lines, which define the model parameters and
the run controls.
Since the FEKO user is only presenting the circuit description (subset of a standard SPICE circuit),
no run controls may be specified. The first line in the input file must be the title and the last line
must be ‘.END’.
Each element in the circuit is specified by an element line that contains the element name, the
circuit nodes to which the element is connected and the values of the parameters that determine
the electrical characteristics of the element. An example SPICE *.cir file is shown below:
Matching circuit

* Note that the subcircuit name should correspond to the name

* of the general network in CADFEKO (or in the *.pre file).
.SUBCKT MatchingNetwork n1 n2
* The next two lines describe a capacitor and an inductor
c1 n1 0 2.17pF
l1 n1 n2 42.29n
.ENDS MatchingNetwork

.END

Extract from the Berkeley SPICE manual regarding SPICE syntax will be presented in the sections
that follow. Only a small subsection of the commands is presented and the descriptions are
incomplete, please refer to the Berkeley SPICE manuals for a more detailed description of the
required syntax.

Title line

The title line must be the first line in the input file and is required. Its contents will be printed
verbatim as the heading for each section of output.

.END line

The ‘.END’ line must always be the last in the input file and is required.

Comments

The asterisk in the first column indicates that the line is a comment line. Comment lines may be
placed anywhere in the circuit description.

.SUBCKT line

General form:
.SUBCKT SUBNAM N1 <N2 N3...>

July 2011 FEKO User’s Manual

SPICE3F5 GENERAL STRUCTURE, CONVENTIONS AND SYNTAX 26-2

A circuit definition begins with a ‘.SUBCKT’ line. ‘SUBNAM’ is the subcircuit name (needs to be
the same name as the general network) and ‘N1, N2...’ are the external nodes, which cannot
be zero. The group of element lines which immediately follows the ‘.SUBCKT’ line define the
subcircuit. The last line in a subcircuit definition is the ‘.ENDS’ line. Control lines may not appear
within a subcircuit definition, subciruit definitions may contain anything else, including other
subcircuit definitions, device models and subcircuit calls. Any device models or subcircuit defini-
tions included as part of a subcircuit definition are strictly local (i.e. such models and definitions
are not known outside the subcircuit definition). Also, any element nodes not included on the
‘.SUBCKT’ line are strictly local, with the exception of 0 (ground) which is always global.

.ENDS line

General form:
.ENDS <SUBNAM>

The ‘.ENDS’ line must be the last one for any subcircuit definition. The subcircuit name, if
included, indicates which subcircuit definition is being terminated; if omitted, all subcircuits
being defined are terminated. The name is needed only when nested subcircuit definitions are
being made.

.INCLUDE lines

General form:
.INCLUDE filename

Portions of circuit descriptions will often be reused in several input files, particularly with com-
mon models and subcircuits. In any SPICE input file, the ‘.INCLUDE’ line may be used to copy
some other file as if that second file appeared instead of the ‘.INCLUDE’ line in the original file.
There is no restriction on the file name imposed by SPICE beyond those imposed by the local
operating system.

Resistors

General form:
RXXXXXXX N1 N2 VALUE

‘N1’ and ‘N2’ are the two element nodes. ‘VALUE’ is the resistance (in ohms) and may be positive
or negative but not zero.

Capacitors

General form:
CXXXXXXX N+ N- VALUE

‘N+’ and ‘N-’ are the positive and negative element nodes, respectively. ‘VALUE’ is the capacitance
in Farads.

July 2011 FEKO User’s Manual

SPICE3F5 GENERAL STRUCTURE, CONVENTIONS AND SYNTAX 26-3

Inductors

General form:
LXXXXXXX N+ N- VALUE

‘N+’ and ‘N-’ are the positive and negative element nodes, respectively. VALUE is the inductance
in Henries.

Coupled (Mutual) Inductors

General form:
KXXXXXXX LYYYYYYY LZZZZZZZ VALUE

‘LYYYYYYY’ and ‘LZZZZZZZ’ are the names of the two coupled inductors. ‘VALUE’ is the coefficient
of coupling, K, which must be greater than zero and less than or equal to one.

Lossless transmission lines

General form:
TXXXXXXX N1 N2 N3 N4 Z0=VALUE <TD=VALUE> <F=FREQ <NL=NRMLEN>

‘N1’ and ‘N2’ are the nodes at port one; ‘N3’ and ‘N4’ are the nodes at port two. ‘Z0’ is the charac-
teristic impedance. ‘N2’ and ‘N4’ are usually used as the ground connections of the transmission
line.
The length of the line may be expressed in either of two forms. The transmission delay, ‘TD’,
may be specified directly (as TD=10 ns, for example). Alternatively, a frequency F may be given,
together with ‘NL’, the normalised electrical length of the transmission line with respect to the
wavelength in the line at the frequency ‘F’. If a frequency is specified but ‘NL’ is omitted, 0.25 is
assumed (i.e. the frequency is assumed to be the quarter-wave frequency). Although both forms
for expressing the line length are indicated as optional, one of the two must be specified.

July 2011 FEKO User’s Manual

CADFEKO GEOMETRY FAULTS 27-1

27 CADFEKO geometry faults

A list of possible geometry faults reported by CADFEKO is listed in the section that follows. A
short description with a possible solution is also given for each fault.

Corrupt data structures: The solid modeller is in an inconsistent state. If this model was im-
ported, either the translation failed or the original model contained errors. Please send the
original model file (in the original format) to the FEKO support team.

Invalid or duplicate identifiers: This situation usually arises when old Parasolid models are im-
ported.

Missing geometry: This situation only arises with imported models. The best solution is to
select the faulty entity in the details tree and remove it. Note that this entity will not be
visible in the 3D view. The geometry has to be recreated.

Invalid geometry: The fault is usually caused by scaled importing. The best solution is to select
the faulty entity in the details tree and remove it. Note that this entity will not be visible in
the 3D view.

Self-intersecting geometry or degenerate geometry: The fault is usually caused by importing

and more specifically scaled importing. The best solution is to select the faulty entity in the
details tree and remove it. The entity can then be recreated.

Geometry not G1-continuous: The fault is usually caused by importing and more specifically
scaled importing. The best solution is to select the faulty entity in the details tree and
remove it. The entity can then be recreated.

Open or non-periodic curve attached to ring edge: This fault can occur when a model is im-
ported or the stitching tool has been used on the model. Please send the original model file
to the FEKO support team.

Open or non-periodic nominal geometry attached to ring edge: This fault can occur when a
model is imported or the stitching tool has been used on the model. Please send the original
model file to the FEKO support team.

Vertex not on curve of edge: The fault is usually caused by importing and more specifically
scaled importing. The best solution is to select the faulty entity and remove it. The entity
can then be recreated.

Vertex not on nominal geometry: The fault is usually caused by importing and more specifi-
cally scaled importing. The best solution is to select the faulty entity and remove it. The
entity can then be recreated.

Edge reversed: This fault only arises with imported models.

Nominal geometry in wrong direction: This situation arises with imported models.

SP-curves of edge not within tolerance: This fault should not arise. Please contact your FEKO
support team.

July 2011 FEKO User’s Manual

CADFEKO GEOMETRY FAULTS 27-2

SP-curves not within edge’s tolerance of nominal geometry: This fault should not arise. Please
contact your FEKO support team.

Vertices of edge touch: The fault is usually caused by importing and more specifically scaled
importing. The best solution is to select the faulty entity and remove it. The entity can
then be recreated. Note that it could be that the face has to be deleted to remove the faulty
edge.

Faces incorrectly ordered at edge: The fault is usually caused by importing. It may be possible
to explode the part and try to combine the entities again.

Vertex not on surface of face: The fault is usually caused by importing and more specifically
scaled importing. The best solution is to select the faulty entity and remove it. The entity
can then be recreated.

Edge not on surface of face: The fault is usually caused by importing. It may be possible to
explode the part and try to combine the entities again.

Self-intersecting face (i.e. edge/edge inconsistency): The fault is usually caused by import-
ing. It may be possible to explode the part and try to combine the entities again. See self
intersecting geometry.

Edges incorrectly ordered at vertex: The fault is usually caused by importing. It may be possi-
ble to explode the part and try to combine the entities again.

Loops inconsistent: The fault is usually caused by importing and more specifically scaled im-
porting. The best solution is to select the faulty entity and remove it. The entity can then
be recreated.

Missing vertex at surface singularity: The fault is usually caused by importing and more specif-
ically scaled importing. The best solution is to select the faulty entity and remove it. The
entity can then be recreated.

Wire-frame edge/face inconsistency: This could be caused by unions involving wires and faced
bodies. Please send the original model file to the FEKO support team.

Wire-frame edge/wire-frame edge inconsistency: This could be caused by unions involving

wires and faced bodies. Please send the original model file to the FEKO support team.

Size-box violation: This fault should not arise. Please contact your FEKO support team. Please
send the original model file to the FEKO support team.

Face-face inconsistency: The fault is usually caused by importing and more specifically scaled
importing. The best solution is to select the faulty entity and remove it. The entity can then
be recreated.

Body is inside out: This situation only arises with imported models.

Shells of region are inconsistent: This fault should not arise. Please contact your FEKO support
team. Please send the original model file to the FEKO support team.

Regions of body are inconsistent: This fault should not arise. Please contact your FEKO sup-
port team. Please send the original model file to the FEKO support team.

July 2011 FEKO User’s Manual

CADFEKO GEOMETRY FAULTS 27-3

Geometry/topology inconsistency in shell: This fault should not arise. Please contact your
FEKO support team. Please send the original model file to the FEKO support team.

Acorn shell/shell inconsistency: This fault should not arise. Please contact your FEKO support
team. Please send the original model file to the FEKO support team.

Unspecified checker failure or checker failure during face-face check: It should be possible
to continue working, but the part may contain faults that can not be detected. Please
contact your FEKO support team.Please send the original model file to the FEKO support
team.

Non-printing character used in name of attribute definition: This situation only arises with
imported models and it should be possible to continue working.

B-geometry has knots closer than the allowed precision: The fault is most probably caused
by importing and more specifically scaled importing. The best solution is to select the
faulty entity and remove it. The entity can then be recreated.

July 2011 FEKO User’s Manual

COPYRIGHT NOTICES AND ACKNOWLEDGEMENTS 28-1

28 Copyright notices and acknowledgements

The copyright for FEKO lies with

Copyright 1998 – 2011: EM Software & Systems-S.A. (Pty) Ltd
However, FEKO uses also third-party libraries and components (both commercial and freely avail-
able ones). For some of the libraries the usage condition is that the corresponding copyright
notices must be reproduced in the accompanying documentation. This shall be done here.

28.1 Copyright to Voronoi

The preprocessor PREFEKO uses part of the program voronoi to execute a Delaunay triangula-
tion for some of the geometric cards. The copyright declaration for voronoi is as follows:
The author of this software is Steven Fortune. Copyright (c) 1994 by
AT&T Bell Laboratories.
Permission to use, copy, modify, and distribute this software for any
purpose without fee is hereby granted, provided that this entire notice
is included in all copies of any software which is or includes a copy
or modification of this software and in all copies of the supporting
documentation for such software.
THIS SOFTWARE IS BEING PROVIDED "AS IS", WITHOUT ANY EXPRESS OR IMPLIED
WARRANTY. IN PARTICULAR, NEITHER THE AUTHORS NOR AT&T MAKE ANY
REPRESENTATION OR WARRANTY OF ANY KIND CONCERNING THE MERCHANTABILITY
OF THIS SOFTWARE OR ITS FITNESS FOR ANY PARTICULAR PURPOSE.

28.2 Copyright to Berkeley Spice 3f5

The FEKO kernel uses Berkeley Spice 3f5 to enable the integration of general SPICE networks
within FEKO.

Copyright (c) 1985-1991 The Regents of the University of California.

All rights reserved.

Permission is hereby granted, without written agreement and without license

or royalty fees, to use, copy, modify and distribute this software and its
documentation for any purpose, provided that the above copyright notice and
the following two paragraphs appear in all copies of this software.

IN NO EVENT SHALL THE UNIVERSITY OF CALIFORNIA BE LIABLE TO ANY PARTY FOR

DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES ARISING OUT
OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN IF THE UNIVERSITY
OF CALIFORNIA HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

THE UNIVERSITY OF CALIFORNIA SPECIFICALLY DISCLAIMS ANY WARRANTIES,

INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
FITNESS FOR A PARTICULAR PURPOSE. THE SOFTWARE PROVIDED HEREUNDER IS ON AN
"AS IS" BASIS, AND THE UNIVERSITY OF CALIFORNIA HAS NO OLBLIGATION TO PROVIDE
MAINTENANCE, SUPPORT, UPDATES, ENCHANCEMENTS, OR MODIFICATIONS.

July 2011 FEKO User’s Manual

COPYRIGHT NOTICES AND ACKNOWLEDGEMENTS 28-2

28.3 Copyright to SuperLU

The FEKO kernel is using for some parts the SuperLU library. The copyright declaration for
SuperLU is as follows:
Copyright (c) 2003, The Regents of the University of California, through
Lawrence Berkeley National Laboratory (subject to receipt of any required
approvals from U.S. Dept. of Energy)

All rights reserved.

Redistribution and use in source and binary forms, with or without

modification, are permitted provided that the following conditions are met:

(1) Redistributions of source code must retain the above copyright notice,
this list of conditions and the following disclaimer.
(2) Redistributions in binary form must reproduce the above copyright notice,
this list of conditions and the following disclaimer in the documentation
and/or other materials provided with the distribution.
(3) Neither the name of Lawrence Berkeley National Laboratory, U.S. Dept. of
Energy nor the names of its contributors may be used to endorse or promote
products derived from this software without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS

IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO,
THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

28.4 Copyright of libcurl

The FEKO GUI makes use of libcurl to connect to the EMSS web-site to check for updates. The
copyright declaration for libcurl is as follows:

COPYRIGHT AND PERMISSION NOTICE

Copyright (c) 1996 - 2005, Daniel Stenberg, <[email protected]>.

All rights reserved.

Permission to use, copy, modify, and distribute this software for any purpose
with or without fee is hereby granted, provided that the above copyright
notice and this permission notice appear in all copies.

July 2011 FEKO User’s Manual

COPYRIGHT NOTICES AND ACKNOWLEDGEMENTS 28-3

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT OF THIRD PARTY RIGHTS. IN
NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,
DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE
OR OTHER DEALINGS IN THE SOFTWARE.

Except as contained in this notice, the name of a copyright holder shall not
be used in advertising or otherwise to promote the sale, use or other dealings
in this Software without prior written authorization of the copyright holder.

28.5 Qwt project usage

For CADFEKO the figures in the mesh information dialogs are based, in part, on the work of the
Qwt project (see http://qwt.sf.net).

28.6 HOOPS and Parasolid

POSTFEKO and CADFEKO use the HOOPS R 3D Application Framework from Tech Soft 3D for
the 3D display and rendering. CADFEKO also uses the Parasolid R Kernel Modeller from UGS to
represent the geometry.

28.7 MeshSim

CADFEKO uses MeshSimTM from Simmetrix Inc. to mesh the geometry.

28.8 FFmpeg (export video files)

POSTFEKO utilises the FFmpeg application for the creation of animation videos (see
http://ffmpeg.org) and is included in the FEKO distributions except FEKO LITE. EMSS-SA
has obtained (as “video provider”) from MPEG-LA a MPEG-4 Visual Patent Portfolio License for
commercial use. The included FFmpeg has been compiled to include only a limited subset of the
video codecs.

28.9 Microsoft

The 2007 Microsoft Office Fluent User Interface is subject to protection under U.S. and interna-
tional intellectual property laws and is used by EM Software & Systems - S.A. (Pty) Ltd under
license from Microsoft.

July 2011 FEKO User’s Manual

COPYRIGHT NOTICES AND ACKNOWLEDGEMENTS 28-4

28.10 Copyright of libXML2

The FEKO kernel is using for some parts the libXML2 toolkit. The copyright declaration is as
follows:

Permission is hereby granted, free of charge, to any person obtaining a copy

of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in
all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
THE SOFTWARE.

July 2011 FEKO User’s Manual

Part X

Index
Index
** card, 13-3, 14-5 archive, 3-22
*.out file array sizes, 24-4, 24-5
basis functions, 19-5 AS card, 14-46
excitation, 19-5 assemblies, 4-27
geometry, 19-1 AutoCAD file, 13-44
*.pre file AV card, 14-52
geometry, 12-1 AW card, 14-54
solution control, 12-1 axes
structure, 12-1 local, 4-7
3D view, 3-1, 3-7 U, V, N, 4-6, 4-7
point entry, 3-8, 3-10
batch
A0 card, 14-9 mesh, 3-6
A1 card, 14-12 beam angle, 6-50
A2 card, 14-13 Bézier, 4-13
A3 card, 14-15 BL card, 13-4
A4 card, 14-17 BO card, 14-61
A5 card, 14-19 *.bof files, 12-1
A6 card, 14-20 Boolean operations, 4-20
A7 card, 14-21 border
AB card, 14-22 PO correction edge, 13-56
ABAQUS, 13-53 PO correction wedge, 13-61
absorption, 6-65 BP card, 13-6
AC card, 14-23 BQ card, 13-8
ACA, 13-27 BT card, 13-10
setting, 6-53
ADAPTFEKO, 21-1 CA card, 14-62
*.pre file for, 21-1 cable coupling, 14-62, 14-67, 14-80, 14-147
running, 21-1 cable cross sections
adaptive cross-approximation, 13-27 bundle, 4-32
adaptive frequency sampling, 6-17, 21-1 coaxial, 4-32
advanced solution options, 6-70 predefined, 4-32
AE card, 14-25 ribbon, 4-32
AF card, 14-28 single conductor, 4-32
AI card, 14-29 cable harness, 4-36
AK card, 14-31 cable irradiation, 14-62, 14-67, 14-80, 14-147
align, 4-29 cable path, 4-34
AN card, 14-34 cable shield
analytical curve, 4-13 braided (Kley), 4-33
angle, 3-21 define shield properties, 4-34
angle between points solid (Schelkunhoff), 4-34
CADFEKO, 3-21 CableMod, 14-23, 14-77
angles cables, 4-31
Kardan, 13-97 CADFEKO, 3-1
anisotropic layers, 9-12, 14-150 geometry and meshing, 4-1
ANSYS, 13-52 Optimisation, 7-1
AP card, 14-35 overview, 3-1
aperture, 14-35 setting solution, 6-1
AR card, 14-41 setting up optimisation, 7-1

I-1
calculating icon indicators, 6-1
antenna reception, 6-67 medium icons, 3-20
error estimates, 6-68 command line
far fields, 6-61 mesh, 3-6
near fields, 6-64 parameters, 6-70
S-parameters, 6-59 comments, 3-21, 13-3, 14-5
SAR, 6-65 complexity, reducing, 4-23
transmission/reflection coefficients, 6-60 component parameters, 6-70
variables, 3-21 authentication, 6-72
capacitance parallel execution, 6-72
loading, 14-121, 14-126, 14-127 remote execution, 6-71
card editing, 11-6 SSPI, 6-72
cascade CONCEPT file, 13-48
Network, 14-129 conducting faces
CB card, 13-12 on dielectric, 6-12
CD card, 14-67 conducting losses, 6-12
CDB file, 13-52 conductivity, 6-1, 6-5
CF card, 14-71 cone primitives, 4-7
CFIE, 14-71 cones, 13-57
*.cfm files, 6-69, 12-1 connection points
CG card, 14-73 definition, 2-2
check for updates, 3-5, 11-5 connectivity
circular arcs, 4-13 rules for, 2-4
circular cone, 13-57 constants, 4-2
circular dependencies, 4-2 constructing solids, 6-11
circular disc, 13-62 contact us, 1-8
circular hole, 13-76 continuous frequency, 6-17
CL card, 13-14 control cards, 12-1, 14-1
clash, 4-37 **, 14-5
clear model, 3-4 A0, 14-9
click mouse, 3-7 A1, 14-12
cluster, 16-1 A2, 14-13
CM card, 14-77 A3, 14-15
CN card, 13-16 A4, 14-17
CO card, 6-13, 14-78 A5, 14-19
coarse segmentation, 13-55 A6, 14-20
coating of wires, 14-78 A7, 14-21
Coatings AC, 14-23
viewing, 6-57 AE, 14-25
coatings, 2-7, 6-6 AF, 14-28
faces, 6-12 AI, 14-29
wires, 6-14 AK, 14-31
coaxial attachment approximation, 14-17 AN, 14-34
coaxial cable, 14-62 AP, 14-35
coil, 13-32 AR, 14-41
collapse, 4-28 AS, 14-46
collapse tree items, 3-15 AV, 14-52
colour AW, 14-54
by coating medium, 3-12 BO, 14-61
by face medium, 3-12 CA, 14-62
by region medium, 3-12 CD, 14-67
dielectrics, 6-1 CF, 14-71

I-2
 CG, 14-73 definition, 2-1
CM, 14-77 current sources, 6-43
CO, 14-78 currents
CS, 14-80 calculation request, 14-133
DA, 14-82 line sources, 14-23, 14-28, 14-29, 14-52
DI, 14-93 written to output files, 6-58
DL, 14-97 curves, 4-22
EE, 14-99 arc, 13-14
EN, 14-100 creating surfaces, 4-24
FE, 14-101 cutplanes, 3-11
FF, 14-106 in POSTFEKO, 9-10
FR, 14-109 through solids, 3-11
GF, 14-112 cylinder
L2, 14-118 UTD, 6-11
L4, 14-119 cylinder primitives, 4-7
LC, 14-120 cylinders, 13-114
LD, 14-121 dielectric, 13-21
LE, 14-122 UTD region, 6-11, 13-107
LF, 14-124
LN, 14-125 DA card, 14-82
LP, 14-126 DD card, 13-17
LS, 14-127 decouple
LZ, 14-128 solutions, 6-54
NW, 14-129 decryption, 16-4
OF, 14-132 default face type, 6-12
OS, 14-133 delete
PS, 14-137 elements, 4-46
PW, 14-139 faces and edges, 4-23, 4-24
SH, 14-147 from archive, 3-22
SK, 14-150 not while in use, 3-9
SP, 14-155 details tree, 3-15
TL, 14-157 media display, 3-20
TR, 14-160 DI card, 14-93
WD, 14-161 dielectric, 2-6, 6-1, 13-68, 14-93, 14-97
convergence accuracy, 6-18 cuboid cylinder, 13-21
copy, 4-27, 4-29 cuboids, 13-18, 13-86, 13-87
geometry, 13-96 for mesh elements, 6-15
multiple, 4-31 losses, 6-1
original part, 4-28 media, 6-1, 6-76
copyright, 28-1 sphere, 14-112
coupling, 14-155 thin sheets, 14-150
transmission line, 14-23, 14-77 Dielectric sheets
create viewing, 6-57
geometry, 4-1, 4-4, 12-1 diffraction theory, 13-103
solid regions, 6-11 dimension, scaling, 13-93
creation history dipole array aperture, 14-35
removing, 4-28 dipoles
CRIPTE, 14-23, 14-77 point electric, 6-24
CS card, 14-80 point magnetic, 6-24
cuboid primitives, 4-7 direction
cuboidal volume elements, 13-18, 13-21, 13-86, 13- normal vectors, 4-10, 4-23, 4-25, 4-26
87 triangles, 4-43

I-3
disabled solution, 6-75 free edges, 4-49
disc, 13-62 ellipse, 4-10
discrete elements, 14-124–14-128 ellipsoid, 13-25
display, see 3D display elliptical arcs, 4-13
items in CADFEKO, 3-12 elliptical hole, 13-76
media, 6-17 ELSE statement, 12-13
distance between points, 3-2 EN card, 14-100
CADFEKO, 3-20 encryption, 16-4
distorted mesh elements, 4-48 end of geometry, 13-23
distributed load, 14-121 end of input file, 14-100
DK card, 13-18 environment of model, 6-46
DL card, 14-97 environment variables, 24-7
domain decomposition, 13-17 equivalent aperture, 14-35
double precision, 6-52 error estimation, 14-99
DP card, 12-15, 13-20 error messages, 3-5
dragging mouse, 3-7 excitations, see sources
duplicate elements, removing, 4-47 current sources, 6-43
DXF file, 13-44 FEM modal mode, 6-45
dynamic memory management, 24-4 Hertzian electric dipole, 6-24
DZ card, 13-21 impressed aperture excitation, 6-29
impressed current, 6-30
edges, 3-15, 6-10 impressed line current, 6-39
definition, 2-2 impressed spherical mode, 6-27
delete, 4-23 incident plane waves, 6-23
length, 4-38, 4-48 magnetic dipole, 6-24
ports on, 6-34 radiation patterns, 6-25
properties, 4-41 total power, 6-19
edit geometry, 4-19 voltage sources, 6-42
EDITFEKO, 11-1 waveguide mode, 6-44
card editor, 11-6 EXIT command, 12-14
edit menu, 11-3 expand tree items, 3-15
file menu, 11-3 explode, 4-22
keystrokes, 11-9 export, 6-69
options, 11-5 POSTFEKO animations, 8-10
preferences, 11-5 POSTFEKO images, 8-10
superuser mode, 11-6 calculated current and charges, 6-58
variables, 11-8 calculated far fields, 6-61
with CADFEKO models, 6-74 calculated near fields, 6-64
EDITFEKO,program flow when using, 12-1 calculated S-parameters, 6-59
editor for model comments, 3-21 geometry, 5-2
EE card, 14-99 Parasolid, 4-1, 5-2
efficiency, 14-139 expressions, 4-2
EFIE, 14-71 functions, 4-2
EG card, 13-23 extents, 4-1
EL card, 13-25 extrude parts, 4-24
electric dipoles, 6-24
electric fields face displacement, 3-4
calculating, 14-101 faces, 3-15, 4-10, 4-22, 6-10, see surfaces
elements, 13-10 coatings, 6-12
count, 4-48 conducting, 6-12
creation, see geometry cards default type, 6-12
deleting, 4-46 delete, 4-23

I-4
 properties, 4-41, 6-12 free edges, 4-37, 4-49
reverse normals, 4-26 free mesh edges, 4-48
solution method, 6-57 free segment nodes, 4-48
solution settings, 6-56 free space, 6-11
far fields, 6-61 frequency, 6-17, 14-109
calculating, 14-106 adaptive sampling, 21-1
FE card, 14-101 functions
feed, see sources in expressions, 4-2
FEK file format, 17-1
*.fek files, 5-5, 12-1 general comments, 2-1
FEKO, 18-1 general networks, 6-20
FEKO_USER_HOME applying excitation, 6-22
environment variable, 24-7 applying loads, 6-22
FEKO_WRITE connecting, 6-22
environment variable, 24-7 creating, 6-21
FEM, 2-6, 13-30 General non-radiating networks
FEM modal port, 6-41 feeding, 14-34
impressed currents, 6-39 geometrical optics, 13-103
Line port, 6-39 geometry
reducing memory requirement, 2-3 align, 4-29
setting, 6-54, 6-55 CAD translation, 5-3
FEM line port, 6-39 creation, 4-1, 4-4, 12-1
FEM modal entering, 13-36
excitation, 6-45 extents, 4-1
FEM modal mode, 6-45 importing, 3-1, 5-2
FEM modal port, 6-41 imprinting points on, 4-21
FEMAP, 13-38 in tree, 3-14
FEST3D file, 14-82 mirror, 4-29
FF card, 14-106 modify, 4-19
file format, 17-1 point entry, 3-10
files, 14-82 point import, 4-18
*.cfm, 6-69 ports on, 6-31
input, 12-1 re-evaluate tree, 4-28
output, 19-1 rotate, 4-29
Parasolid, 5-2 scale, 4-29
*.pre, 6-69 show / hide parts, 3-12
summary of, 25-1 solids, 4-7
used by components, 2-1, 12-1 surfaces, 4-10
finding translate, 4-29
distorted elements, 4-48 validation, 4-37
free mesh edges, 4-48 wires, 4-13
free segment nodes, 4-48 geometry cards, 12-1, 13-1
intersecting mesh elements, 4-48 **, 13-3
flare primitives, 4-7 BL, 13-4
FM card, 13-27 BP, 13-6
FO card, 13-29 BQ, 13-8
Fock area, 13-29 BT, 13-10
FOR/NEXT loops, 12-12 CB, 13-12
formulations CL, 13-14
in FEKO, 1-1 CN, 13-16
FP card, 13-30 DK, 13-18
FR card, 14-109, 21-1 DP, 13-20

I-5
 DZ, 13-21 HE card, 13-32
EG, 13-23 helices, 4-13, 13-32
EL, 13-25 help, 9-28
FM, 13-27 Hertzian electric dipole, 14-19
FO, 13-29 hide
FP, 13-30 items behind cutplane, 3-11
HC, 13-31 single item, 3-12, 3-15
HE, 13-32 history
HP, 13-34 removing, 4-28
HY, 13-35 hot-keys, 3-22
IN, 13-36 HP card, 13-34
IP, 13-55 HY card, 13-35
KA, 13-56 hyperbolic, 13-31, 13-34, 13-35
KK, 13-57
icons
KL, 13-61
in tree, 3-15
KR, 13-62
ideal receiving antenna, 14-143
KU, 13-64
ideal transmission lines, 6-20
LA, 13-66
applying excitation, 6-22
MB, 13-67
applying loads, 6-22
ME, 13-68
connecting, 6-22
NU, 13-71
creating, 6-20
PB, 13-73
IF statement, 12-13
PH, 13-76
impedance, 14-124, 14-125, 14-128
PM, 13-78
loading, 6-45, 14-122, 14-124, 14-125
PO, 13-80
microstrip fed, 14-122
PY, 13-84
sheets, 6-7
QT, 13-86
impedance boundary condition, 14-150
QU, 13-87
import
RM, 13-89
CAD formats, 5-3
SF, 13-93
geometry, 3-1, 5-2, 6-11, 13-36
SY, 13-95
meshes, 3-1, 5-5, 12-1, 13-36
TG, 13-96
points, 4-18
TO, 13-99
impressed aperture excitation, 6-29
TP, 13-101
impressed current, 6-30
UT, 13-103
impressed line current, 14-23, 14-28, 14-29, 14-52
UZ, 13-107
in FEM region, 6-39
VS, 13-108
impressed spherical mode, 6-27
WA, 13-111
imprinting points, 4-21
WG, 13-112
IN card, 13-36
WR, 13-113
in use
ZY, 13-114
deleting items, 3-9
GF card, 14-112
incident plane wave, 14-9
GiD, 13-54
include files, 13-36
GO, 13-103
inductance
setting, 6-54
loading, 14-121, 14-126, 14-127
viewing, 6-57
infinite planes, 6-46
GPU acceleration, 6-71
instances of ports, 6-31
Green’s functions, 2-7, 6-46, 14-112
intersect, 4-20
grid, 4-6
intersecting mesh elements, 4-48
ground planes, 6-46, 14-61
IP card, 13-55
HC card, 13-31
KA card, 13-56

I-6
Kardan angles, 13-97 losses, 14-139
KK card, 13-57 conducting, 6-12
KL card, 13-61 on faces, 6-12
KR card, 13-62 on wires, 6-14
KU card, 13-64 LP card, 14-126
LS card, 14-127
L2 card, 14-118 LZ card, 14-128
L4 card, 14-119
LA card, 13-66 machine codes, 15-6
label selected calculation, 14-132 magnetic cuboids, 13-18
labels, 4-27, 12-4, 13-12, 13-66, 13-67 magnetic dipoles, 6-24, 14-20
changing, 4-45 magnetic fields
using, 4-45, 6-75 calculating, 14-101
large element PO magnetic ring current, 14-15
setting, 6-56 magnitude
large models, 12-4 peak convention, 6-19
layered dielectrics, 6-6 make primitive, 4-28
layered dielectrics (anisotropic), 6-7 manual solution control, 24-1
layered media, 14-160 maxalloc(m), 24-4, 24-5
LC card, 14-120 maximum constants, 24-4, 24-5
LD card, 14-121 MB card, 13-67
LE card, 14-122 ME card, 13-68
length, 3-20 media
lens, 13-31, 13-34, 13-35 colour, 3-12, 6-1
LF card, 14-124 dielectric, 6-1, 13-68, 13-86, 13-87
licence, 15-1 display, 6-17
check in, 15-5 impedance sheets, 6-7
configuration file, 15-6 in details tree, 3-20
floating, 15-3 layered dielectrics, 6-6
include/exclude, 15-5 layered dielectrics (anisotropic), 6-7
information, 15-1 lists of, 6-1
preferred, 15-4 magnetic, 13-86, 13-87
server, 15-4 media library, 6-6
line, 4-13 metallic, 6-1, 6-5
line segments, see segments on imported meshes, 6-15
linear set of equations, 14-73 on imported models, 6-10
LN card, 14-125 properties, 6-76
loading, 14-62, 14-157 setting, 6-10
an edge, 14-122 using, 6-1
attachment point, 14-119 memory
cable, 14-120 allocation, 24-4
distributed, 14-121 reducing FEM requirement, 2-3
FEM, 14-124 menu
impedance, 6-45, 14-128 pop-up, 3-7, 3-15
microstrip line, 14-122 merge
Network, 14-125 mesh, 4-47
parallel circuit, 6-45, 14-126 merging
series circuit, 6-45, 14-127 vertices, 4-46
local mesh parameters, 4-41 mesh
lock point entry, 3-10 batch, 3-6
loft, 4-26 command line, 3-6
log files, 3-5 merge, 4-47

I-7
 union, 4-47 NASTRAN file, 13-42
mesh refinement, 4-44 near fields, 6-64
meshes, 4-38 calculating, 14-101
adaptive refinement, 6-68 calculation offset, 14-132
adding triangles, 4-43 NEC file, 13-47
edit vertex, 4-42 Network
importing, 3-1, 5-5, 12-1 cascade, 14-129
information, 4-48 loading, 14-125
local settings, 4-41 networks, 14-62, 14-157
point refinement, 4-44 Neutral files, 13-38
polyline refinement, 4-44 new
ports, 6-31 model, 3-4
properties, 6-15 nodes
refining, 13-89 defining, 13-20
setting media, 6-15 definition, 2-2
show / hide parts, 3-12 variable names, 12-15, 13-20
small features, 4-40 non-uniform mesh, 13-6, 13-8
smoothing, 4-40 normal vectors, 4-23, 4-25, 13-16
tetrahedra, 4-40 reverse, 4-26, 4-43
meshing, 2-1 notes, 3-21
non-uniform, 13-6, 13-8, 13-10 NU card, 13-71
rules, 2-2, 2-4, 12-3 NURBS, 4-10
message window, 3-1 NURBS surfaces, 13-71
messages, 3-5 NW card, 14-129
metallic media, 6-5
microstrip OF card, 14-132
port, 6-36 offset for near field calculation, 14-132
microstrip port, 6-36 ohmic losses, 6-5, 6-12, 14-150
mirror, 4-29 old models
mismatch loss, 6-19 re-evaluate, 4-28
MLFMM, 13-27 open model, 3-4
setting, 6-53, 6-57 OPTFEKO, 20-1
modal port, 13-67 Optimisation method, 20-1
model Optimisation
creating, 3-4 CADFEKO setup, 7-1
unit, 3-2, 4-1 far field goal, 7-14
validation, 6-69 goal combination tool, 7-19
modelling goal weighting, 7-19
guidelines, 2-1 goals, 7-7
rules, 2-1 impedance goal, 7-11
modify geometry, 4-19 masks, 7-5
mouse interaction, 3-7 method, 7-2
point entry, 3-10 near field goal, 7-12
multilayer substrates parameters, 7-3
defining, 14-112 s-parameter goal, 7-15
multiple reflections, 13-80 SAR goal, 7-17
multiple selection, 3-9 search, 7-1
stopping criteria, 7-2
N axis, 4-6, 4-7 the global goal, 7-19
named points, 4-4 optimisation
in *.pre file, 6-76 farming, 20-13
names, 4-27, 6-75 Genetic algorithm (GA), 20-8

I-8
 Grid search, 20-10 accelerated ray tracing, 6-56
Particle swarm (PSO), 20-5 edge border, 13-56
running, 20-13 plane wave excitation, 6-56
sensitivity analysis, 20-12 setting, 6-54, 6-56
Simplex, 20-2 viewing, 6-57
Optimisation search visibility, 13-108
Defining, 7-1 wedge border, 13-61
Multiple, 7-1 PO card, 13-80
options point dipoles, 6-24
EDITFEKO, 11-5 point entry, 3-10, 4-6
rendering, 3-4 snap mode, 3-8
solution components, 6-70 point source with pattern, 6-25
OS card, 14-133 points
*.out files, 12-1 angle, 3-21
out-of-core, 24-4 distance, 3-20
output file, 19-1 geometry, 3-10
overlap, 4-37 imprinting, 4-21
overview, 1-1 in *.pre file, 6-76
named, 4-4
package, 16-1 names, see nodes
pan polarisation
model view, 3-7 incident waves, 6-23
paraboloid, 4-10, 13-73 polygons, 4-10
parallel execution, 11-4 definition, 2-1
parallelogram, 13-6 meshed, 13-78
parameters normals, 4-43
points, 4-4 polygonal plates, 13-84
segmentation, 13-55 UTD formulation, 13-103
parametric models, 3-1, 4-2 polyline, 4-13
Parasolid, 5-2 ports, 6-31
parents, 3-14 edge, 6-34
parts, 3-14 instances, 6-31
show / hide, 3-12 licence server, 15-3
PATRAN file, 13-51 microstrip, 6-36
patterned point source, 6-25 on wires, 6-33
PB card, 13-73 waveguide, 6-37
PCBMod, 14-23, 14-77 POSTFEKO, 8-1
PDF viewer, 3-4 power input, 6-19, 14-139
PE card, 13-75 PP card, 14-135
peak magnitude, 6-19 *.pre files, 6-69, 12-1
Periodic boundary conditions, 6-50 named points, 6-76
permeability, 6-1 variables, 6-76
permittivity, 6-1 precision, 6-52
PH card, 13-76 preconditioners, 6-53, 6-54
physical optics, 13-80 PREFEKO, 17-1
planar, 6-56 preferences, 3-4
planar substrate, 2-7, 14-112 EDITFEKO, 11-5
plane wave, 14-160 preferred licence, 15-4
plane wave excitation, 6-23, 14-9 preview mode, 3-10
plate with hole, 13-76 primitive
PM card, 13-78 collapsing part tree, 4-28
PO priority

I-9
 setting, 18-1 view operations, 3-13
program execution control, 14-137 referencing elements, 6-75
program flow, 2-1 reflectors, 4-10
project control, 3-22 regions, 3-15, 6-10
project notes, 3-21 properties, 4-41, 6-11
projection, 4-21 registry keys, 24-7
properties relative
medium, 6-11 permeability, 6-1
mesh, 4-41, 6-15 permittivity, 6-1
modify, 4-19 remeshing, 13-89
PS card, 14-137 CADFEKO models, 12-4
PW card, 14-139 remote, 16-1
PY card, 13-84 remote execution, 11-4
remove
QT card, 13-86 duplicates, 4-47
QU card, 13-87 slivers, 4-46
quadrangle, 13-8 renaming
quantity palette, 8-3 geometry, 4-19
QUEUEFEKO mesh elements, 4-45
CLUSTER options, 16-3 rendering options, 3-4
configuration files, 16-2 request
creating packages, 16-1 cable harness, 4-36
Decryption of package, 16-6 antenna reception, 6-67
Execution queue, 16-5 current output, 6-58
Extract packages, 16-5 error estimations, 6-68
extracting packages, 16-1 far fields, 6-61
FEKO options, 16-3 near fields, 6-64
Generate package, 16-5 S-parameters, 6-59
Including package files, 16-2 SAR, 6-65
Setting preferences, 16-6 transmission/reflection coefficients, 6-60
using encryption, 16-4 resistance
queuefeko, 16-1 loading, 14-124–14-126
resistance loading, 14-121, 14-122, 14-127
RA card, 14-143
restore
radiation
deleted faces / edges, 4-28
patterns, 14-106
results, 6-57
patterns as source, 6-25
checking validity, 2-7
patterns as sources, 14-41
continuous, 21-1
radius
time-domain, 22-1
segments, 4-38, 4-41
reuse objects, 4-27
ray file, 6-54
RM card, 13-89
RCS
rotate
radar cross section, 6-61
geometry, 4-29, 13-96
re-evaluate
model view, 3-7
the geometry, 4-28
parts, 4-24
re-use
point, 13-101
solution, 6-52
ruled surface, 4-26
real ground, 14-61, 14-112
run
receiving antenna, 6-67
components, 6-69
rectangle, 4-10
from GUI, 11-4
redo
selection, 3-9 S-parameters, 6-59, 14-155

I-10
SA card, 14-145 slivers
SAR removing, 4-46
specific absorption rate, 6-65, 14-145 small features
save remove, 4-40
as UNIX, 11-3 snap mode, 3-2, 3-8
model, 3-4 solid bodies, 4-7, 6-11
SBR, 13-103 versus shell bodies, 6-11, 6-76
scale, 4-29 solution, 11-4
geometry, 13-93, 13-96 numerical Green’s function, 6-51
Parasolid, 5-2 advanced control, 24-1
units, 4-1, 6-75 component parameters, 6-70
scattering, 6-61 disabled, 6-75
scripting, 2-1, 10-1 face settings, 6-56
secfeko, 15-1 NGF, 6-51
security, 16-4 viewing, 6-57
segmentation solver
parameters, 13-55 settings, 6-52
rules for, 2-2, 2-4, 12-3 solving, 6-69
segments, 13-4 sources, 14-6
arc, 13-14 current sources, 6-43
coating, 14-78 FEM modal, 6-41
creation, see geometry cards Hertzian electric dipole, 6-24, 14-19
definition, 2-1 impressed aperture excitation, 6-29
helix, 13-32 impressed current, 6-30
ports on, 6-33 impressed line current, 6-39, 14-23, 14-28, 14-
radius, 4-38, 4-41 29, 14-52
selection impressed spherical mode, 6-27
in 3D view, 3-8 incident plane wave, 14-9
items, 3-8 incident plane waves, 6-23
multiple, 3-9 magnetic dipole, 6-24, 14-20
toolbar, 3-8 magnetic ring current, 14-15
SEMCAD, 14-83 microstrip line, 14-25
server patch feed pin, 14-17
licence, 15-3 radiation patterns, 6-25, 14-41
setting media, 6-10 spherical modes, 14-46
SF card, 13-93 total power, 6-19
SH card, 14-147 voltage on a node, 14-13
shells, 6-11 voltage on a non-radiating network port, 14-34
shooting and bouncing rays, 13-103 voltage on a radiating cable, 14-31
short-cut keys, 3-22 voltage on a segment, 14-12
EDITFEKO, 11-9 voltage on an edge, 14-21, 14-25
show voltage sources, 6-42
single item, 3-12, 3-15 waveguide modes, 14-54
sides SP card, 14-155
of triangles, 4-43 specified vertex, 4-41
simplify, 4-23 sphere primitives, 4-7
single precision, 6-52 spheres, 13-64
single selection, 3-8 dielectric/magnetic, 13-18
size spherical modes, 14-46
meshing rules, 2-2 SPICE3f5
SK card, 6-13, 14-150 general structure, conventions and syntax, 26-1
skin depth, 2-2, 14-150 spinning parts, 4-24

I-11
spirals, 4-13, 13-32 selection, 3-8
splines, 4-13 toroidal segment, 13-99
split parts, 4-22 TP card, 13-101
squint angle, 6-50, 14-135 TR card, 14-160
SSPI, 6-72 transform
start page, 8-3 view, 3-13
status bar transformation, 4-29, 13-96
cut plane, 3-11 of meshes, 4-31
distance, 3-2 of point, 13-101
model unit, 3-2 translation, 4-29, 13-96, 13-101
snap mode, 3-2 transmission line theory, 14-62
stitch, 4-24 transmission lines, 14-157
STL file, 13-49 coupling, 14-23, 14-77
stripline transmission/reflection coefficients, 6-60, 14-160
feeding, 14-25 tree view, 3-1, 3-14
loading, 14-119 icons, 3-15
substrate, 14-112 triangles, see elements
subtract, 4-20 back side, 4-43
superuser mode creating, 4-43
in EDITFEKO, 11-6 normal side, 4-43
surface currents normals, 4-43
distribution, 14-133 remove duplicates, 4-47
surface equivalence, 2-6
surface resistance, 6-7 U axis, 4-6, 4-7
surface triangles, see elements unconnected edges, 4-37, 4-49
surfaces, 4-10, 4-22 undo
conducting, 6-12 list depth, 3-4
creating solids, 4-24 selection, 3-9
suspect, 4-19 view operations, 3-13
sweeping parts, 4-24 union, 4-20
SY card, 13-95 mesh, 4-47
symbolic variables, 12-6 units, 4-1, 6-75
symmetry, 6-49, 13-95, 24-1 updates, 3-5, 11-5
electric, 24-1 UT card, 13-103
geometric, 24-1 UTD, 13-103
magnetic, 24-2 cylinder, 6-11, 13-107
polygonal plate, 13-84
tab setting, 6-54
Display options, 3-10 viewing, 6-57
tetrahedral volume elements, 13-86 UZ card, 13-107
definition, 2-1
TG card, 13-96 V axis, 4-6, 4-7
thin dielectric sheets, 2-7, 6-6, 6-12, 14-150 validation, 2-7, 6-69
TIMEFEKO, 22-1 geometry, 4-37
*.pre file for, 22-1 variables, 3-21, 4-2
*.tim file for, 22-2 environment, 24-7
output of, 22-7 in *.pre file, 6-76
running, 22-7 in point names, 12-15
TL card, 14-157 memory allocation, 24-4
TO card, 13-99 predefined, 4-2, 24-4
tolerance, 4-46 symbolic, 12-6
toolbars, 3-1 vertex

I-12
 loading, 14-118 workplane, 4-6
vertices coordinates, 4-7
definition, 2-1 WR card, 13-113
editing, 4-42
merging, 4-46 zoom
ports on, 6-33 model view, 3-7
specified location, 4-41 ZY card, 13-114
view
3D default, 3-10
manipulation, 3-7
transform, 3-13
wire-frame, 3-10
visibility
PO region, 13-108
view items, 3-13
voltage sources, 6-42
on a node, 14-13
on a non-radiating network port, 14-34
on a port, 6-31
on a radiating cable, 14-31
on a segment, 14-12
on an edge, 14-21, 14-25
volume equivalence, 2-6
VS card, 13-108

WA card, 13-111
waveguide
excitation, 6-44
port, 6-37
waveguide mode, 6-44
waveguide port, 6-37, 14-54
WD card, 14-161
wedge as PO border, 13-61
WG card, 13-112
windscreen
antenna, 6-8
dielectric, 6-8
reference, 6-8
WA, 13-111
WD, 14-161
WR, 13-113
wire-frame view, 3-10
wires, 6-10
as objects, 4-13
coatings, 6-14
curved, 13-14
grid, 13-112
lossy, 6-14
mesh into segments, 4-38
ports on, 6-33
properties, 6-14
segments, see segments

I-13