Gary Rondeau , AF7NX

 Describe the NEC codes

 How to define and edit the models
 Running the simulations and work flow
 Work some examples
 Using Variables
 Deal with Feed Lines in the models
 Grounds
 Optimization
 More Examples
 Numerical Electromagnetic Code (NEC) were developed at LLNL in
the 1970’s by Gerald Burke (r.i.p.) and Andrew Poggio.
 “Engines” NEC2 freely available – plenty good for 90% of ham work
NEC4 pay to license ($110 from LLNL)- better with
buried radials.
NEC5 new rewrite ($110 from LLNL) but also no longer
supported – doesn’t work w/ WIN 11
 “Wrappers” include EZNEC (was commercial), 4NEC2 (free) and
others. EZNEC now free and unsupported.
 EZNEC - popular with ARRL – most popular because it’s “EZ”.
 4NEC2 – Arie Voors, very powerful features – end-of-development
because Visual Basic coding no longer supported by Microsoft. Uses
NEC2 supports NEC4.
4NEC2 is what I use because of the ability to use variables and the ability
to optimize them. – The native NEC2 version is just fine!
 >10,000 wires/ segments possible – only limited by
memory and running time

 Great graphics and plotting

 Use of Variables

 Optimization and Parameter Sweeps

 4NEC2 Enhancements -- wire loading, fast grounds,

auto-segmentation, convergence test, etc.

 Freely Available https://www.qsl.net/4nec2/

 Define Geometry of Antenna Wires and
Radiating Elements

 Define Voltage or Current Source(s)

 Define Ground Plane (usually)

 Define Transmission lines and R,L,C

components and loads.
 NEC card format harkens to IBM puch-card
days and remains the NEC file format

NEC “cards”:

CM 20 M DIPOLE
CE
Tag #segs x1 y1 z1 x2 y2 z2 r
GW 1 100 -5 0 5 5 0 5 0.001
GE 0
GN -1
EK
EX 0 1 50 0 1 0 0
FR 0 0 0 0 14.1 0
EN
CM 20 M DIPOLE
CE
SY H = 10 'Height in Meters
SY LW1 = 10.0 'Length of Wire (meters)
GW 1 100 -LW1/2 0 H LW1/2 0 H #14
GE 0
GN -1
EK
EX 0 1 50 0 1 0 0
FR 0 0 0 0 14.1 0
EN

SY cards define variables that can be used to specify model

parameters. But, use the more intuitive GUI editor!
The SY cards are the glue that makes 4NEC2 so powerful.
Modern GUI makes
entering data easier
and with less
mistakes.

NEC files saved with

“card” format.
Use the “Calculate” tab
to launch the
simulation.

“Far Field Pattern”

most useful. You also
get SWR for your feed
geometry, radiated
efficiency, RDF, etc.

For multi-band
antennas, the
“Frequency Sweep” is
very useful to see how
well you can load the
antenna on various
bands.
 Wires have Tag numbers and are divided into Segments
Typically use 20 to 300 Segments per half-wave.

More Segments take longer to run. Use 4NEC2 “AutoSegment” utility.

 Wires with the same endpoints are joined together

 Sources and Loads are placed on specific Segments
 Transmission Lines and Ground Planes (“perfect” and
“real” grounds) are supported

 NEC determines the self consistent current in all of the wire

segments due to the Sources. Then determines the fields
produced by the currents. Magic as far as we are concerned!
 Enter Geometry
 Validate geometry, feed,
loading, etc. (Look at graphics
and run Segment Checks.)
 Run Simulation
 Look at results
 Can we make it better?
Yes - Modify geometry, feed,
loading, etc. – optimize.
No – Then time to build and test
the antenna and verify the
simulation – or try something
new.
 Following examples illustrate a variety of
4NEC2 functions and abilities.

 Enough input data is shown to allow

duplication so you could run it yourself.

 Wide variety of complexity from very simple to

complicated “cut-to” recipe.
 How long?
 Where is the feed point?
 Use a Balun? What ratio?
 Feed line? - The “Windom” designs include
the vertical feed line section. Does it work?

 I want good performance on 20m & 30m.

 I have a pile of 75 Ω co-ax… can I use it?
Geometry includes the feed coax which you can connect to one side of
the antenna (direct coupling) or not (using a balun).

Place the driving source on antenna wire.

Include the co-ax feed line as antenna element going to ground and the
internal transmission line to the TX.

Driving source at the end of a transmission line.

 Height H = 15m
 Overall Length LW = 41m
 Fractional length at feed F = 0.67

Express various lengths in terms of parameters:

LW1 = LW * F Length of long section
LW2 = LW – LW1 Length of short section

Say all this with SYMBOLS & EQUATIONS

before you define the geometry.
Calculated Far-Field Pattern 14.05 MHz
The impedance graph
shows that the
antenna would do
better driving 200Ω
rather than 50Ω. Plan
on a 4:1 Balun.

Treat baluns as ideal

transformers. Change
source impedance
from 50 Ω to 200 Ω
with 4:1 balun.
Previous design with F=0.67 does pretty
well with the 80, 40, 20, 17, 12, and 10
meter bands.
But I want 30m…

Sweep F from 0.55 to .9 and look at SWR

for 10.1MHz
Unfortunately we lost 20m but improved 30m, 15m & 12m

Use FREQ.TXT list to

emphasize the ham bands

The antenna will work for a particular frequency when:

1) The structure is cut to have a natural resonance at that frequency
2) The structure is fed in a manner that can excite the frequency.
Need both!
Gets more of 20
and some of 30…

Pushing F has
raised the
impedance for
20m and 30m.
Maybe try 300Ω
feed, include
coax transmission
line,
etc…

Pretty soon you

have your own
unique design…
This is the
previous
OCFD on
40m.

Verify
expected
pattern,
gain, and RF
currents for
a particular
band.
 Specify wire size AWG, e.g. #12, or radius.

 Loading – Wire material Copper, Aluminum, etc.

 Loading – Insulated wires. PVC, Teflon, Nylon,

etc. Pitfall - be sure to get the radius correct in the
appropriate units. (feet!?)

 Should get about 4% length correction for THHN

house wire. Check it!
Don’t neglect the effects of the
feed wire itself as an antenna
element!

Transformers are not modeled

directly. Instead change the
impedance, seen at the antenna,
as reflected through the
transformer.

From the antenna’s POV see

feed line impedance with
unknown termination at tuner.
Approximate
Lots of options for
Ground. In the drop-
down menus.

“Fast Ground” allows

for radial array ground
screen.

Main Ground provides

specific soil or
environment type
options for relevant
conductivity and
dielectric constant.

NEC4 handles ground

more realistically than I use “Average Ground” all the time so I can compare
NEC2, especially buried antenna performance rather than ground properties.
radials.
 Look at 15° elevation gain. Sweep NR with 10m
radials.
 Look at 15° elevation Gain and sweep L radial with
16 radials.
  Lwire = 500’ = 152m = NR  Lradial
 Now we can optimize NR

Best is 34 radial about

4.5m long.

How much gain for the

next 500’ spool?
 This time use “Optimize”
rather than “Sweep”

Gone from 0.6 dB to 1.3 dB at 15°,

now with 47 radials about 6.5m
long.

500’ of #14 copper ~$50 so about

$70/dB (plus a bit of shovel work)

Do it again for 1500’ of wire?

 Simple idea uses several modeling features:
 Lumped impedance for feedline choke
 Coated wires – THHN and Coax-Jacket
 Transmission line feed
 Variables to optimize:
 Length of two sections of dipole – use 4NEC2
 Length of transmission line feed – use 5/4 λ
 Impedance of choke, L and C - want trap for 20m
 Trying to use 4NEC2 to optimize everything at
once is a fools errand – make educated guesses!
Modeling inductors properly is tricky because:
• Inductors are often distributed in space (not small
compared to λ) so a lumped circuit model is often not
appropriate.
• Segmentation length limit of λ/1000 prevents direct
modeling for tight helixes.
• Useful inductors are often used near resonance where
properties are rapidly changing.
• Capacitance associate with the inductor is a non-ignorable
nebulous quantity.
My self resonant coil calculator after David Knight, G3YNH
• One piece of 75 Ω coax
• One piece or #14 THHN
house wire
• Single pull-up point.
• No radials
• Good low elevation
pattern.
• SWR ~1
 Large subject…
 A Yagi is a multi-element phased array
 Phase two, three or four verticals to steer the pattern.
 Curtain arrays – VOA did this in a big way.
 Collinear arrays, V-beams, many more…
 Example - consider two low dipoles.
 Low dipoles are sky warmers
 Place two side by side separated by < λ/2
 Worthless sky wave is same on both so subtract signals!
Note impedance seen at antenna is very
close to 75Ω. SWR 1.5 for 50Ω driver.
General representation of
two dipoles, separated by
2* Xo and 2* Yo and at
Angle Ao.
A(0) –B (0) A (0) – B (delay = spacing)
suppresses sky wave emphasizes one direction
Change direction by switching which
cable has the extra phasing length
Sweep PH from
0 to 20 m and
look at Gain,
F/B and SWR.

Best gain has

high SWR.
More work to
do…

Consider series
feed or 4:1 at
radio…
 NEC for VHF & UHF too.
 Can a 20m collinear design I did work on 2m?

 Use 75Ω coax.

 Traps for 145 MHz – use my coil calculator:
 1.6 ft., 2.5 turns at ~2.4 in. dia. of RG-6
 0.54 μH, 2.3 pF
 Optimize design for LQ, TL2, RT for gain at 10°
elevation and SWR.
 This example generates a “recipe” so include
wire coating loads.
Does it work? Need to
build it and find out!
 I wouldn’t… verify!
 Check convergence of model.
Calculate-Optimize-Conv-test will run the setup
with more segments – should get similar answer!
 Measure wires lengths, heights, etc. when you
build. Verify what you actually built!
 Include effects of gutters, Al siding, other
wires, fences if any are near the antenna.
 Measure SWR sweep with antenna analyzer!
 Simple models provide insight even if not a
finished design.
 Running a dozen models will give you more
understanding than building those antennas.
 Models are good enough to “cut to” if you
include everything. Plan to fine tune!
 Models let you exclude bad ideas quickly.

 Be creative!