See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/268036302

RF PASSIVE CIRCUIT DESIGN

Conference Paper · September 2014

DOI: 10.13140/2.1.4921.0560

CITATIONS READS

0 9,621

2 authors, including:

Ikponmwosa Aikhionbare
University of Portsmouth
1 PUBLICATION 0 CITATIONS

SEE PROFILE

All content following this page was uploaded by Ikponmwosa Aikhionbare on 10 November 2014.

The user has requested enhancement of the downloaded file.

 RF PASSIVE CIRCUIT DESIGN

by

Ikponmwosa Dickson, Aikhionbare

MSc Electronics Engineering

Supervisor: Manish Malik

2013/2014

School of Engineering

This report is submitted in partial fulfilment of the requirements for the

award of the degree of Master of Science.

 STUDENT ID: UP698209

ABSTRACT

The design of Radio Frequency (RF) circuits especially those that are in the microwave
region require a different approach from the conventional circuit theory principles due
to spurious effects that characterize high frequency circuits. Lumped elements which
would otherwise work well in low frequency circuits could be rather unpredictable and
would most likely fail in a microwave circuit. This is due to the fact that the physical
dimensions of the circuit would fall within the region of the wavelength (λ) of the
transmitted signal. With all these factors in mind, it became necessary to device new
means of designing circuits which would take advantage of these effects in order to
make them suited for microwave devices. This work involves extensive research on
existing RF and microwave circuit design techniques and the design, building and testing
of RF passive devices such as filters, splitters and couplers. These components are
implemented on the FR4 1mm microstrip substrate, and are suited to operate within the
lower spectrum of the fourth generation (4G) Long Term Evolution (LTE) mobile
communication network system. The performance of the produced circuit was
investigated and evaluated.

i
 STUDENT ID: UP698209

ACKNOWLEDGEMENT

I am grateful to God for His grace throughout the period of this school year.
I would like to express my profound gratitude and deep regards to my supervisor, Mr
Manish Malik for his guidance and encouragement throughout the course of this project
and Mr Mark Oliver for his support.
My appreciation also goes to my moderator Dr Khalil Alkhadimi and all the lecturers at
the school of engineering for the time and effort they have put into tutoring me in this
past year.
Most of all, to my family who have supported me at every stage of my life, I say thank
you and may the good God replenish all of your efforts.

ii
 STUDENT ID: UP698209

TABLE OF CONTENTS

ABSTRACT .......................................................................................................................................... i
ACKNOWLEDGEMENT ...................................................................................................................... ii
TABLE OF CONTENTS ....................................................................................................................... iii
TABLE OF FIGURES ............................................................................................................................ v
LIST OF TABLES .............................................................................................................................. viii
LIST OF UNITS .................................................................................................................................. ix
ABBREVIATIONS................................................................................................................................ x
CHAPTER 1 ........................................................................................................................................ 1
INTRODUCTION ............................................................................................................................ 1
1.0 CHAPTER OVERVIEW ........................................................................................................ 1
1.1 BRIEF HISTORY OF WIRELESS COMMUNICATION............................................................. 1
1.2 DEFINITION OF RF AND MICROWAVES ............................................................................ 2
1.3 MODERN APPLICATIONS OF RF TECHNOLOGY ................................................................. 3
1.4 PROJECT AIMS AND OBJECTIVES ...................................................................................... 4
CHAPTER 2 ........................................................................................................................................ 5
BACKGROUND THEORY ................................................................................................................ 5
2.0 CHAPTER OVERVIEW ........................................................................................................ 5
2.1 RF CIRCUIT DESIGN THEORY AND CONCEPTS............................................................... 5
2.1.1 The Maxwell Equations ............................................................................................ 5
2.1.2 The Transmission Line Theory .................................................................................. 6
2.1.3 The Transmission Line Model ................................................................................... 6
2.1.4 Transmission Line Characteristics ............................................................................ 7
2.2 PASSIVE CIRCUITS IN RF .................................................................................................... 9
2.2.1 RF Filters ................................................................................................................... 9
2.2.2 Splitters and Couplers............................................................................................. 16
CHAPTER 3 ...................................................................................................................................... 20
3.0 CHAPTER OVERVIEW ...................................................................................................... 20
3.1 WIRELESS CIRCUITS DESIGN REQUIREMENTS ................................................................ 20
3.2 SYSTEM REQUIREMENTS FOR TRANSMITTERS AND RECEIVERS .................................... 20
3.2.1 Receiver Parameters............................................................................................... 21
3.2.2 Transmitter Parameters ......................................................................................... 22

iii
 STUDENT ID: UP698209

3.2.3 Filter Design Requirements for Modern Wireless Communication ....................... 23

3.2.4 Power Splitter Design Requirements for Modern Wireless Communication......... 25
3.2.5 Coupler Design Requirements for Modern Wireless Communication ................... 26
3.2.6 Coupler and Splitter Figures of Merit ..................................................................... 27
CHAPTER 4 ...................................................................................................................................... 29
4.0 CHAPTER OVERVIEW ...................................................................................................... 29
4.1 MICROSTRIP DESIGN CONSIDERATIONS ........................................................................ 29
4.2 PRACTICAL DESIGN OF RF PASSIVE CIRCUITS ................................................................. 32
4.2.1 Frequency and Bandwidth Selection ...................................................................... 32
4.2.2 Substrate Selection for 4G Wireless Application .................................................... 32
4.2.3 Bandpass Filter Design for 4G Wireless Application............................................... 34
4.2.4 Power Splitter Design for 4G Wireless Application ................................................ 54
4.2.5 Rat Race Coupler Design for 4G Wireless Application............................................ 63
4.2.6 Branch Line Coupler Design for 4G Wireless Application ....................................... 67
4.2.7 Comparison of Design on FR4 and Rogers Substrate ............................................. 73
CHAPTER 5 ...................................................................................................................................... 76
5.0 CHAPTER OVERVIEW ...................................................................................................... 76
5.1 PCB PRODUCERS REQUIREMENT.................................................................................... 76
5.2 BANPASS FILTER TEST ..................................................................................................... 79
CHAPTER 6 ...................................................................................................................................... 85
RECOMMENDATIONS ................................................................................................................. 85
CONCLUSION .................................................................................................................................. 86
REFERENCES ................................................................................................................................... 87
BIBLIOGRAPHY ................................................................................................................................ 91
APPENDIX ....................................................................................................................................... 97

iv
 STUDENT ID: UP698209

TABLE OF FIGURES

Figure 1.1: Estimated Mobile Cellular Subscriptions 2014 .................................................. 2

Figure 1.2: The Electromagnetic Spectrum.......................................................................... 3
Figure 2.2: Two Port Network. ............................................................................................. 9
Figure 2.3: Typical Frequency Response of Lowpass Filter ............................................... 10
Figure 2.4: Lowpass Filter Prototype Circuit ...................................................................... 11
Figure 2.5: Typical Highpass Filter Response ..................................................................... 11
Figure 2.6: Highpass Prototype Circuit .............................................................................. 12
Figure 2.7: Bandpass Filter Response ................................................................................ 12
Figure 2.8: Bandpass Prototype Circuit ............................................................................. 13
Figure 2.9: Bandstop Response.......................................................................................... 13
Figure 2.10: Bandstop Prototype ....................................................................................... 14
Figure 2.11: Butterworth Filter Response.......................................................................... 14
Figure 2.3: Chebyshev Filter Response .............................................................................. 15
Figure 2.4: Wilkinson Power Splitter ................................................................................. 17
Figure 2.5: Branch-line Coupler ......................................................................................... 18
Figure 2.6: Rat Race Coupler .............................................................................................. 19
Table 3.2: Receiver Sensitivity ........................................................................................... 21
Figure 3.1: Ideal Bandpass Filter Response ....................................................................... 23
Figure 3.2: Ideal Response of Equal Split Power Splitter ................................................... 25
Figure 3.3: Rat Race Coupler .............................................................................................. 26
Figure 3.4: Power Split of Rat Race Coupler ...................................................................... 26
Figure 4.1: Microstrip Line ................................................................................................. 29
Table 4.1: Typical Characteristics of Substrates Types ...................................................... 31
Figure 4.2: Substrate Comparison ..................................................................................... 34
Figure 4.3: Element Transformation .................................................................................. 35
Figure 4.4: Nomograph for number of resonant sections for Chebyshev filter. ............... 36
Figure 4.5: 3rd Order Bandpass Filter Configuration ......................................................... 38
Figure 4.6: S Parameters of 3rd Order Chebyshev Bandpass Filter ................................... 38
Figure 4.7: Group Delay of 3rd Order Chebyshev Filter...................................................... 39
Figure 4.8: Butterworth Filter with Real Elements ............................................................ 40
Figure 4.9: S Parameters of 3rd Order Butterworth Bandpass Filter ................................ 41
Figure 4.10: Group Delay 3rd Order Butterworth Bandpass Filter .................................... 41
Figure 4.11: Parallel Coupled Line Schematic .................................................................... 42
Figure 4.12: Frequency Response of Coupled Filter .......................................................... 43
Figure 4.13: PCB View of Coupled Line Filter ..................................................................... 43
Figure 4.14: 3D Artwork Model of Coupled Line Filter ...................................................... 44
Figure 4.15: Hairpin Filter Schematic ................................................................................. 44
Figure 4.16: Frequency Response of Hairpin Filter ............................................................ 45
Figure 4.17: PCB View of Hairpin Filter .............................................................................. 45

v
 STUDENT ID: UP698209

Figure 4.18: 3D Artwork Model of Hairpin Filter ............................................................... 46

Figure 4.17: Shunt Stub Filter Schematic ........................................................................... 46
Figure 4.18: Frequency Response of Shunt Stub Filter ...................................................... 47
Figure 4.19: PCB View of Shunt Stub Filter ........................................................................ 47
Figure 4.20: 3D Artwork Model of Shunt Stub Filter ......................................................... 48
Figure 4.21: Combline Filter Schematic ............................................................................. 48
Figure 4.22: Frequency Response of Combline Filter ........................................................ 49
Figure 4.23: PCB View of Combline Filter .......................................................................... 49
Figure 4.24: 3D Artwork Model of Combline Filter ........................................................... 50
Figure 4.25: Interdigital Filter Schematic ........................................................................... 51
Figure 4.26: Frequency Response of the Optimised Interdigital Filter .............................. 51
Figure 4.27: PCB View of Interdigital Filter ........................................................................ 52
Figure 4.28: 3D Artwork Model of Interdigital Filter ......................................................... 52
Figure 4.29:Comparison of Insertion Loss of Microstrip Filters ........................................ 52
Figure 4.30: Comparison of Return Loss of Microstrip Filters ........................................... 53
Figure 4.31: Group Delay of Microstrip Filters .................................................................. 53
Figure 4.32: Line Width Calculation using TXLine Tool ...................................................... 55
Figure 4.33: Wilkinson Power Splitter Layout.................................................................... 56
Figure 4.34: Schematic for Power Splitter ......................................................................... 56
Figure 4.35: Frequency Response of Power Splitter .......................................................... 57
Figure 4.36: Angular Response of Power Splitter .............................................................. 57
Figure 4.37: Wilkinson Power Splitter in Power Combiner Setup ..................................... 58
Figure 4.38: Wilkinson Power Splitter as Combiner .......................................................... 58
Figure 4.39: PCB View of Power Splitter ............................................................................ 59
Figure 4.40: 3D Artwork Model ......................................................................................... 59
Figure 4.41: Schematic of Wilkinson Splitter in Circular Layout ........................................ 60
Figure 4.42: Frequency Response of Wilkinson Splitter in Circular Layout ....................... 60
Figure 4.43: Angular Response of the Circular Power Splitter .......................................... 61
Figure 4.44: PCB View of the Circular Power Splitter ........................................................ 61
Figure 4.45: 3D Artwork View of Circular Splitter.............................................................. 62
Figure 4.46: Compensated Power Splitter ......................................................................... 62
Figure 4.47: PCB View of Compensated Power Splitter .................................................... 62
Figure 4.48a: Useful bandwidth of Uncompensated Power Splitter ................................. 63
Figure 4.48b: Useful bandwidth of Compensated Power Splitter ..................................... 63
Figure 4.49: Rat Race Coupler Geometry .......................................................................... 64
Figure 4.50: Schematic Layout of Rat Race Coupler .......................................................... 65
Figure 4.51: Frequency Response of Rat Race Coupler ..................................................... 65
Figure 4.52: Angular Response of Rat Race Coupler ......................................................... 66
Figure 4.53: PCB View of Rat Race Coupler ....................................................................... 66
Figure 4.54: 3D Artwork Model of Rat Race Coupler ........................................................ 67
Figure 4.55: Branch Line Coupler ....................................................................................... 67

vi
 STUDENT ID: UP698209

Figure 4.56: Branch Line Coupler Schematic ..................................................................... 68

Figure 4.57: Frequency Response of Branch Line Coupler ................................................ 69
Figure 4.58: Angular Response of Branch Line Coupler..................................................... 69
Figure 4.59: PCB View of Branch Line Coupler .................................................................. 70
Figure 4.60: Artwork of Branch Line Coupler .................................................................... 70
Figure 4.61: Two Stage Branch Line Coupler ..................................................................... 71
Figure 4.62: Frequency Response of Two Stage Coupler .................................................. 71
Figure 4.63: Angular Response of Two Stage Branch Line Coupler ................................... 72
Figure 4.64: PCB View of Two Stage Branch Line Coupler ................................................. 72
Figure 4.64: Artwork View of Two Stage Branch Line Coupler .......................................... 73
Figure 4.65: Frequency Response of Hairpin filter built on Rogers Substrate .................. 73
Figure 4.66: Frequency Response of Power Splitter built on Rogers Substrate ................ 74
Figure 4.67 Frequency Response of Rat Race Coupler built on Rogers Substrate ............ 74
Figure 4.68: Frequency Response of Single Stage Branch Line Coupler on Rogers
Substrate ............................................................................................................................ 75
Figure 4.69: Frequency Response of Double Stage Branch Line Coupler on Rogers
Substrate ............................................................................................................................ 75
Figure 5.1: Gerber View of Hairpin Filter ........................................................................... 76
Figure 5.2: Price Quote for Bandpass Filter ....................................................................... 77
Figure 5.3: Gerber View of Circular Splitter ....................................................................... 78
Figure 5.4: Gerber View of Rat Race Coupler .................................................................... 78
Figure 5.5: Price Quote for Coupler and Splitter ............................................................... 79
Figure 5.6: Produced Bandpass Filter ................................................................................ 79
Figure 5.7: Setup for Filter Insertion Loss Test .................................................................. 80
Figure 5.8: Setup for Filter Return Loss Test ...................................................................... 80
Figure 5.9: Frequency Sweep Graph .................................................................................. 81
Figure 5.8: Return Loss of Hairpin Filter ............................................................................ 83
Figure 5.9: Insertion Loss of Hairpin Bandpass Filter ........................................................ 84

vii
 STUDENT ID: UP698209

LIST OF TABLES

Table 3.1: Receiver Sensitivity

Table 4.1: Typical Characteristics of Substrates Types
Table 4.2: design properties for Rogers 4350B Substrate
Table 4.3: Design Properties of the FR4 Substrate
Table 4.4: Element Values for Chebyshev Prototype Filter with 0.5 dB Ripple
Table 4.5: Element Values for Butterworth Filter
Table 4.6: Power Splitter Performance Specification
Table 4.7: Rat Race Coupler Design Specifications
Table 4.8: Rat Race Coupler Design Specifications
Table: 5.1: Frequency Response of Bandpass Filter

viii
 STUDENT ID: UP698209

LIST OF UNITS
 cm Centimetre
 GHz Gigahertz (109 Hz)
 Hz Hertz
 KHz Kilohertz (103 Hz)
 m/s meters per second
 MHz Megahertz (106 Hz)
 mm Millimetre
 ns nano seconds

ix
 STUDENT ID: UP698209

ABBREVIATIONS

 4G Fourth Generation Mobile Communication

 CAD Computer Aided Design
 CDMA Code Division Multiple Access
 DBS Direct Broadcast Satellite
 DC Direct Current
 FCC Federal Communication Commission
 FM Frequency Modulation
 GPS Global Positioning Satellite
 GSM Global System for Mobile Communications
 IEEE Institute of Electrical and Electronics Engineers
 LTE Long Term Evolution
 LTE Long Term Evolution
 MMIC Monolithic Microwave Integrated Circuit
 PCB Printed Circuit Board
 QPSK Quadrature Phase Shift Keying
 Radar Radio Detection and Ranging
 RF Radio Frequency
 SCS Single Crystal Sapphire
 SNR Signal to Noise Ratio
 TEM Transverse Electromagnetic Mode
 UHF Ultra High Frequency
 VHF Very High Frequency
 VSWR Voltage Standing Wave Ratio
 WLAN Wireless Local Area Network

x
 STUDENT ID: UP698209

CHAPTER 1

INTRODUCTION

1.0 CHAPTER OVERVIEW

This chapter gives a brief history and an adequate definition and of wireless
communications. It also describes the important modern applications of RF and
microwave. The aims and objectives of this project are also outlined.

1.1 BRIEF HISTORY OF WIRELESS COMMUNICATION

The origin of wireless communication can be traced back to the work of James Clerk
Maxwell and Heinrich Rudolf Hertz. Maxwell predicted the propagation of
electromagnetic waves in free space at the speed of light and presented his Maxwell’s
equations in 1864. These equations were a combination of works by Lorentz, Faraday,
Ampere and Gauss. Maxwell’s theory gained further acceptance in 1884 after the
validation of the theory of propagation of electromagnetic waves by Hertz (Kai, Inder, &
Vijay, 2002, p. 1). Hertz also demonstrated the generation, propagation and reception of
RF in his laboratory, but it was not used for long distance communication until 1895
when Guglielmo Marconi was able to communicate over long distances using RF, and
then went on to carry out the first transatlantic wireless communication using the Morse
code (Probir, 1995, p. 1-2).
The advent of the Second World War created a drastic need for radio communications
and Radio Detection and Ranging (radar). The use of higher frequency waves became of
utmost importance in radar because the resolution of a radar system is proportional to
its wavelength, hence to detect small objects such as aircrafts, microwave frequencies
had to be used. The war also saw significant increase in the use of telegraphs,
telephones and point-to-point radio links. In the 1960s, satellite systems emerged as the
new means of global communications, utilizing broadband high frequency systems to
simultaneously support thousands of telephone users, television channels and data links
(Kai, Inder, & Vijay, 2002, p. 2). Since the introduction of the cellular phone in the early 1980s,

there has been an immense growth in this industry. According to estimates from The
International Telecommunication Union, by the end of 2014 there will be approximately
7 billion mobile subscriptions worldwide, rising from 2 billion subscriptions in 2005, 3

1
 STUDENT ID: UP698209

billion Internet users and 2.3 billion mobile-broadband subscriptions. (International

Telecommunication Union, 2014, p. 3).

Figure 1.1: Estimated Mobile Cellular Subscriptions 2014

1.2 DEFINITION OF RF AND MICROWAVES

RF can be defined as those frequencies that are above human hearing, which extend
into the microwave spectrum to the edge of the infrared region. This range is between
20 KHz and 30 GHz (Carr, 2002, p. 3). RF and microwaves have also been defined as
alternating current signals with frequencies in the range of 100 MHz to 1000 GHz, with
RF frequencies falling into the 30 – 3000 MHz range (Very High Frequency (VHF) in 30 –
300 MHz and Ultra High Frequency (UHF) 300 – 3000 MHz). Microwave refers to
frequencies between 3 and 300 GHz with an electrical wavelength between 10 cm and 1
mm (Pozar, 2012, p. 1). The electromagnetic spectrum is shown in Fig. 1.2.

2
 STUDENT ID: UP698209

Figure 1.2: The Electromagnetic Spectrum (Pozar, 2012, p. 2)

In order to properly characterize RF frequencies it is better to define them based on

their properties, instead of a frequency demarcation.
Besser and Gilmore suggested that an RF circuit or system is one where;
 The phase shift of the signal is significant over the extent of the component,
because its size is comparable with a wavelength.
 The reactance’s of the various circuit elements must be accounted for,
particularly those associated with the parasitic of active devices
 Circuit losses cause degradation of Q value, reduction of frequency selectivity
and thermal noise.
 Electromagnetic radiation and capacitive coupling among elements cause
unintentional loss and may also significantly alter the performance of the circuit.
 Reflections between elements occur because circuit size is of the order of a
wavelength. Circuit design requires special treatment to ensure reflections do
not cause loss of gain, power, or failure of components.
 Nonlinearity which causes distortion and unwanted frequency components, can
be undesired but at the same time is an essential part of the circuit operation, as
in mixers or local oscillators.
(Besser & Gillmore, 2003, p. 2)

1.3 MODERN APPLICATIONS OF RF TECHNOLOGY

In the modern day, RF and microwave technology is mostly applied in wireless
networking and communication systems, wireless security systems, radar systems,
environmental remote sensing and medical systems (Pozar, 2012, p. 3).

3
 STUDENT ID: UP698209

RF and microwave technology has played an indispensible role in the development of

wireless telephony. Ever since the discovery of the concept of cellular frequency reuse,
by Bell Labs, the cellular phone industry has grown immensely (Jagannathan, 2007, p. 17) The
first generation cellular systems commonly referred to as 1G utilized analogue FM
(frequency modulation) with allocation of frequency bands divided into narrow band
voice channels. As the demand for cellular phones increased, this method of modulation
became relatively inefficient. Further improvement in modulation technology saw the
introduction of the second generation (2G) cellular systems such as the GSM (Global
System for Mobile Communications) and CDMA (Code Division Multiple Access). In
recent years, 2.5G, 3G, 3.5G, 3.75G, 4G and LTE have been introduced to provide a wide
range of services to about 7 billon users globally (Pozar, 2012, p. 3). The ‘4G Long Term
Evolution (4G-LTE)’ standard is the most recent advancement in mobile communications
and fixed internet services. This standard provides high data rate transmission for
mobile phones and fixed wireless devices (Khan, 2011). The Satellite systems which
provide voice, video and data communication worldwide are dependent on RF and
microwave technology. Satellite dependent systems such as Direct Broadcast Satellite
(DBS) and Global Positioning Satellite (GPS) have been immensely advanced over time.
Other applications of RF and microwave technology include WLAN (Wireless Local Area
Network) and Radar systems.

1.4 PROJECT AIMS AND OBJECTIVES

The aims and objectives of this project are as follows;

i. Research extensively on the existing state of the art RF and microwave design
techniques in use
ii. Design passive devices such as filter, splitter and coupler
iii. Implement the design on a chosen substrate
iv. Test, investigate and evaluate the performance of the produced devices
v. Suggest possible improvements to the design and implementation process used.

4
 STUDENT ID: UP698209

CHAPTER 2

BACKGROUND THEORY

2.0 CHAPTER OVERVIEW

In this chapter, a relevant background theory and concepts with respect to RF circuits
design such as the Maxwell equations, the transmission line theory, transmission line
model and transmission line characteristics are discussed. RF passive circuits such as
filters, splitters, and couplers are also discussed.

2.1 RF CIRCUIT DESIGN THEORY AND CONCEPTS

Due to the diverse application of RF systems, the design of RF systems has been steadily
improved over time. The basic concepts mostly applied in RF circuit design are discussed
below.
2.1.1 The Maxwell Equations
The electric and magnetic phenomena that affect the design of RF circuits are described
by Maxwell’s equations. According to Pozar 2012, the general form of these equations is
shown below;

(2.1)

, (2.2)

(2.3)
(2.4)
Where is the electric field, in volts per meter (V/m),
is the magnetic field, in amperes per meter (A/m),
is the electric flux density, in coulombs per meter squared (Coul/m 2),
is the magnetic flux density in webers per meter squared (Wb/ m2),
is the (fictitious) magnetic current density, in volts per meter squared (V/m2),
is the electric current density in amperes per meter squared (A/m2),
is the electric charge density, in coulombs per meter cubed (Coul/m 3).
(Pozar, 2012, p. 10)

5
 STUDENT ID: UP698209

2.1.2 The Transmission Line Theory

Due to the significant difference in electrical size (which is directly proportional to λ),
between low frequency and RF or microwave frequency circuits, the basic circuit theory
is not sufficient to analyze circuits or to design their devices. While basic circuit analysis
assumes that the physical size of the network is much smaller than the electrical
wavelength, transmission line theory assumes that the physical size of the network
could be a significant fraction or many wavelengths in size. Hence, the transmission line
theory deals with a distributed parameter network (where voltages and currents can
vary in magnitude and phase over its length) while the circuit theory deals with lumped
element (where voltages and currents do not vary appreciably over the physical
dimensions of the elements) (Pozar, 2012, p. 48).

2.1.3 The Transmission Line Model

For a TEM (Transverse Electromagnetic Mode) wave propagation, the transmission line
can be modelled as a two wire line as shown in Fig. 2.1 below. This model holds because
transmission lines always have a minimum of two conductors.

Figure 2.1: Transmission Line Model (Pozar, 2012, p. 49)

Where: R = series resistance per unit length, for both conductors, in Ω/m.
L = series inductance per unit length, for both conductors, in H/m.
G = shunt conductance per unit length, in S/m.
C = shunt capacitance per unit length, in F/m.
(Pozar, 2012, p. 49)

6
 STUDENT ID: UP698209

2.1.4 Transmission Line Characteristics

2.1.4.1 Wave Velocity on Transmission Lines
The velocity of TEM waves on transmission lines is sometimes referred to as the phase
velocity. It is less than the free space velocity of light.
The phase velocity is given as; (2.5)
Where: wave or phase velocity
3 m/s (free-space velocity of light)
relative dielectric constant or relative permeability.
(Ferril, 2005, p. 227)

2.1.4.2 Reflection Coefficients

When a wave travels down a transmission line, it will be reflected back to source if the
load impedance is not matched to the input impedance. The reflection coefficient for a
load or termination is given by the equation;

Voltage reflection coefficient, (2.6)

Where: voltage reflection coefficient,

voltage of reflected wave,
voltage of incident wave,
characteristic impedance,
characteristic impedance.
(Ferril, 2005, p.228)

ZL may be equal to RL L hence it is a vector quantity and can be a combination of a

resistance and a reactance. This means that has both magnitude and phase. The
voltage reflection coefficient can be determined by measurement and then used to find
using (1.7) below.

2.1.4.3 Standing – Wave Ratio

When a transmission line is terminated with an unmatched load, there will be reflected
waves and hence standing waves on the line. The direct and reflected wave can either
be in phase where they add up to produce a maximum, or 90 degrees electrically away

7
 STUDENT ID: UP698209

from the point of maximum standing wave where they subtract to produce a minimum.
In this case, the direct and reflected wave will be 180 degrees out of phase.
Standing wave ratio S (2.8)

The magnitude of the reflection coefficient is .

(Ferril, 2005, p.228)

2.1.4.4 S Parameters
S parameters refer to the scattering matrix that quantifies the spread of RF energy
through a multiport network. For an N port network, its S parameters matrix contains N2
coefficients. Due to the unavailability of equipment to measure RF total voltage and
current, difficulty of obtaining perfect open and short circuit, and the unstable nature of
active devices under open or short conditions, it is pretty difficult to measure Y, Z or H
parameters but is more convenient to measure the S parameters. For a 2-port device
having ports named 1 and 2, there are four S parameters namely;
 S11 and S22: these are the forward and reverse coefficients with the opposite port
terminated in the impedance Zo.
 S12 and S21: these are the forward and reverse gains, also assuming the generator
and load impedance to be Zo.
For a network of N ports, the following properties apply.
i. For any matched port i, Sii = 0.
ii. For a reciprocal network, Snm = Smn.
iii. For a passive circuit, |Smn| 1
iv. For a lossless and reciprocal network, the ith port is;

The S parameter for a 2 port network is given by;

8
 STUDENT ID: UP698209

Where represents the incident wave voltages and represents the reflected wave
voltages.
The S parameter properties are also defined as;

(Kai, Inder, & Vijay, 2002, p. 123)

Figure 2.2: Two Port Network (Pozar, 2012).

2.2 PASSIVE CIRCUITS IN RF

Passive devices are those devices in circuits that are not equipped to have gain. These
include resistors, inductors and capacitors (Weinstock & Nisenoff, 1999, p. 198). A passive
circuit is one that comprises only of passive devices as components. RF
telecommunication systems normally comprises of both active and passive components.
However, RF circuits such as filters, couplers and splitters can be implemented solely
using passive components.

2.2.1 RF Filters
The design of RF filters is based on resonating circuits, since resonators basically form
filter networks. A combination of inductive L and capacitive C elements form resonating
circuits which behave as various forms of filters.

The resonant frequency is given as (2.14)

(Kai, Inder, & Vijay, 2002, p. 119)

9
 STUDENT ID: UP698209

The property of frequency selectivity and energy loss of a resonant circuit is specified by
the quality factor Q.
At the resonant frequency the loaded and unloaded Q is given as shown in (2.15) and
(2.16) below.

Q0= (2.15)

QL= (2.16)

Where F0 is the resonant frequency and is the 3 dB half power bandwidth

(Kai, Inder, & Vijay, 2002, p. 119 &p. 121)

For microstrip circuits, open ended, stub, dielectric and ring resonators are most
commonly used.
A filter is a network with an input and output port which is applied in communication
systems to control the frequency response of the system by transmitting frequencies
within the pass band and attenuating frequencies in the stop band (Pozar, 2012, p. 49). An
ideal filter is characterized by perfect impedance matching, zero pass band insertion loss
and infinite rejection loss (Pozar, 2012, p. 129). The rejection rate of a filter can be improved
increasing the number of resonators. There are four types of filters namely lowpass,
highpass, bandpass, and bandstop. These filters are briefly discussed below.

2.2.1.1 Lowpass Filter

Lowpass filters allow the transmission of frequencies that are below the frequency of
interest and attenuate the frequencies above the cut off frequency.

Figure 2.3: Typical Frequency Response of Lowpass Filter (AWR Corporation)

10
 STUDENT ID: UP698209

A sample prototype circuit to implement a lowpass filter is shown in Fig. 2.4

Figure 2.4: Lowpass Filter Prototype Circuit (AWR Corporation)

2.2.1.2 Highpass Filter

Highpass filters allow the transmission of frequencies that are above the frequency of
interest and attenuate the frequencies below the cut off frequency.

Figure 2.5: Typical Highpass Filter Response (AWR Corporation)

A sample prototype circuit to implement a highpass filter is shown in Fig. 2.6

11
 STUDENT ID: UP698209

Figure 2.6: Highpass Prototype Circuit

2.2.1.3 Bandpass Filter

Bandpass filters are those that pass frequencies within a specific pass band and
attenuate or reject other frequencies. The rejection ratio of the filter is specified by the
Q value.

Figure 2.7: Bandpass Filter Response (AWR Corporation)

12
 STUDENT ID: UP698209

Figure 2.8: Bandpass Prototype Circuit (AWR Corporation)

2.2.1.4 Bandstop Filter

Bandstop filters are those that stop or attenuate frequencies within a stop band and
allow other frequencies.

Figure 2.9: Bandstop Response (AWR Corporation)

13
 STUDENT ID: UP698209

Figure 2.10: Bandstop Prototype (AWR Corporation)

Filters have various pass band response characteristics depending on the design
procedure that was used in the design. The two most common are the maximally flat
(Butterworth) and equal ripple (Chebyshev) response.

2.2.1.5 Butterworth Filter Response

The Butterworth filter is a medium Q filter whose amplitude response is required to be
as flat as possible within the pass band resulting in zero derivatives of amplitude with
frequency at DC (Direct Current) (Kikkert, 2009, p. 2). The characteristics of Butterworth
filters are often see as good and acceptable despite the trade off in attenuation
steepness.

Figure 2.11: Butterworth Filter Response (AWR Corporation)

14
 STUDENT ID: UP698209

The circuit elements for a Butterworth filter can be easily obtained using the equations
(2.17) and (2.18).

(Kikkert, 2009, p. 2)

Where R is the desired load impedance, is the cut off frequency, is the capacitor

value and is the inductor value. Also to be noted here is that the term is in

radians.
This lowpass prototype procedure forms the basis of Butterworth filter design, as other
filter types can be obtained by transformation.

2.2.1.6 Chebyshev Filter Response

The Chebyshev filter is sometimes referred to as an elliptical filter because its pole
positions lie in an ellipse. This filter has a response which is characterised by pass band
ripples which can be determined as part of the filter design. The Chebyshev filter
possesses a much sharper roll of than the Butterworth filter, but with a much larger
group delay which makes it not very suitable for data communications. (Kikkert, 2009, p. 6)

Figure 2.3: Chebyshev Filter Response (AWR Corporation)

15
 STUDENT ID: UP698209

2.2.2 Splitters and Couplers

Splitters are non-directional with non-isolated port passive microwave components with
two or more ports which receive an input signal and split them into two or more output
signals of lesser power. (Microwaves 101. 2010) The output signals are usually in phase with
an equal power division ratio of 3 dB. Splitters can be used in reverse as combiners to
sum up two or more input signals to one output. (Thomas, 2004, p. 184)
Splitters are characterised in the simplest form as a three-port network having one input
and two outputs, which can be expressed by its scattering matrix as:

(2.19)

Assuming that all three ports are perfectly matched and the network is reciprocal and
lossless, (2.19) becomes:

(2.20)

(2.20) above leads to the conditions

(2.21a)
or
(2.21b)
(Pozar, 2012, p. 318)

Directional couplers are passive reciprocal four port networks in which one port is
isolated from the input port. In directional couplers, when power is fed into the input
port, most of the incident signal exits at one port known as the through port, a fraction
of the input signal appears in the coupled port, while the isolated port is usually
terminated. If the signal is fed into the through port, it is coupled to the isolated port
and not to the coupled port. Directional couplers only couple power flowing in one
direction (Annapurna & Sisir, 2009, p. 214).
A reciprocal directional coupler can be expressed in its characteristic scattering matrix
as:

16
 STUDENT ID: UP698209

 0 S 12 S 13 S 14 
 S 12 0 S 23 S 24
  (2.22)
 S 13 S 23 0 S 34
 
 S 14 S 24 S 34 0 
(Pozar, 2012, p. 320)

The performance of directional couplers can be quantified based on the parameters of

coupling(C), directivity (D), Isolation (I) and Insertion loss (L). (Annapurna & Sisir, 2009, p. 214)

(2.23a)

(2.23b)

(2.23c)

(2.23d)

(Pozar, 2012, p. 322)

2.2.2.1 The Wilkinson Power Splitter

The Wilkinson power splitter is a simple form of power splitter which offers broad band
width and equal phase at the output ports. A series resistor with a value of 2 Z0 is used
to terminate both output ports 2 and port 3, with Z0 being the system impedance. The
2 Z0 resistance also ensures that all the ports are matched.

Figure 2.4: Wilkinson Power Splitter (Microwaves 101, 2010)

When a signal enters port 1, it splits into two equal amplitude, equal phase output
signals at port 2 and port 3. Ports 2 and 3 are at the same potential, hence there is no
flow of current between both ports. The Wilkinson splitter can also be used as a
combiner by feeding signals into ports 2 and 3. The sum of which is supplied to port 1. A
matched input is achieved when ports 2 and 3 are terminated by the characteristic
impedance of the quarter wave length line which must be equal to Z0. In situations

17
 STUDENT ID: UP698209

where space is not a hindrance in the circuit surrounding, the performance of this
splitter can be improved by compensation. This is done by adding a quarter wavelength
line between the input port and the split. This compensation results in a better VSWR
(Voltage Standing Wave Ratio) and better performance in wide band operations. (Ferril,
2005, p.248)

An unequal power split can be achieved by modifying the characteristic impedance of

the quarter wave transformer and the other λ/4 sections (Bahl, 2003, p. 371). The scattering
matrix of the Wilkinson power splitter is of the form:

[S]= (2.24)

(Frank, 2012, p. 224)

2.2.2.2 The Branch-Line Coupler

The branch-line coupler or quadrature 90° hybrid is a 3 dB directional coupler with each
of the transmission line being λ/4 wavelength long.

Figure 2.5: Branch-line Coupler. (Microwaves 101, 2012)

With all the ports matched, an input signal fed into port 1 is split into two quadrature
signals see at port 2 and port 3, with a 90° phase difference between them, while port 4
is completely isolated from port 1 at the centre frequency. (Pozar, 2012, p. 343)
The transmission used can be 3/4, 5/4 or 7/4 of a wavelength long, but this would
reduce the bandwidth of operation (Microwaves 101, 2012). The scattering matrix of the
branch-line coupler is in the form shown in (2.25). (Pozar, 2012, p. 343)

0 j 1 0 
 j 0 0 1
  (2.25)
1 0 0 j 
 
0 1 j 0 

18
 STUDENT ID: UP698209

From observing the scattering matrix above, it can be seen that the branch-line coupler
is very symmetric, hence any port can be utilised as input, the output port will always be
the opposite junction from the input and the isolated port will always be on the same
side as the input port.

2.2.2.3 The Rat Race Coupler

The rat race coupler or 180° hybrid is a lossless, matched and reciprocal four-port
network. The rat race coupler has an input port, two output ports and one isolated port.

Figure 2.6: Rat Race Coupler (Microwaves 101, 2012)

The device can be operated both as a signal splitter and as a signal combiner. As a
splitter, when a signal is fed into port 1, it is evenly split into two in-phase signals which
appear at ports 2 and 3, while port 4 is isolated. On the other hand, when a signal is fed
into port 4, it is evenly split into two signals with a 180° phase difference between them.
The 3 dB power split occurs at the centre frequency. The device can be used as a
combiner by feeding two signals into ports 2 and 3. The sum of these two signals is seen
at port 1 (sum port) and the difference is seen at port 4 (difference port).
The scattering matrix is of the form:

0 1 1 0
1 0 0  1
 (2.26)
1 0 0 1
 
0  1 1 0
(Pozar, 2012, p. 343)

19
 STUDENT ID: UP698209

CHAPTER 3

3.0 CHAPTER OVERVIEW

The efficient design of RF telecommunication systems is aimed at ensuring effective,
portable and low cost implementation of RF circuit elements on a suitable platform
which in the case of RF circuits in microstrip. These circuits are to be realised according
to the requirements set laid out by the IEEE (Institute of Electrical and Electronics
Engineers) and other regulating bodies. The design requirements and parameters are
discussed in this chapter.

3.1 WIRELESS CIRCUITS DESIGN REQUIREMENTS

RF and microwave frequencies significantly alter the physical behaviour of circuit
components and this can make the design of RF circuit components quite complicated.
The design of RF circuits has to meet some basic requirements to make them suitable for
wireless application. These requirements are:
 High Quality: The RF components have to operate at a certain specified level of
quality with minimum deviation from ideal requirements.
 Compact Size: In order to facilitate mobility, it is essential for wireless devices to
be as compact as possible, without any negative effects on the performance of
the device.
 Low-Cost: The circuit components need to be low cost in order for the finished
product to be affordable by the final consumer.
 Repeatability: Circuit components should be easily reproduced on the selected
platform such as strip line or microstrip.

Other important requirements are discussed in the sections below.

3.2 SYSTEM REQUIREMENTS FOR TRANSMITTERS AND RECEIVERS

Wireless communication systems all consist of transmitters and receivers. The
transmitter conveys the modulated carrier signal, while the receiver recovers the
information within the carrier signal. There are certain parameters that must be
considered in general when designing any RF wireless communication system. These
parameters differ in receivers and transmitters.

20
 STUDENT ID: UP698209

3.2.1 Receiver Parameters

In wireless systems, the receiver is required to have the capacity to process the carrier
signal into useful information with minimal distortion. The acceptable minimal signal to
noise ratio is largely dependent on the application of the system. For wireless mobile
communication systems, the minimum recommended SNR (Signal to Noise Ratio) is 20
dB. (Geir, 2013)
The parameters for receiver systems are;
 Sensitivity: The receiver sensitivity is a measure of the ability of the receiver to
respond to weak signals. Sensitivity is specified as signal to noise ratio in
analogue receiver systems and bit error rate in digital receiver systems. Table 3.1
shows the sensitivity of a receiver with signal to noise ratio in a 4G system.
Table 3.2: Receiver Sensitivity (Source: Geir, 2013)
SNR System Response
40 dB Excellent Signal; Lightening Fast
25 dB to 40 dB Very good signal; very fast
15dB to 25 dB Low signal; fast
10 dB to 15 dB Very low signal; mostly slow
5 dBto10 dB No signal; no receiver response

The SNR for a system is given by:

(3.1)

 Selectivity: This is the measure of the ability of the receiver to reject unwanted
signals that exist on adjacent channel frequencies. The frequency selectivity for
receivers range from 70 to 90 dB, although this is not usually achievable in
reality. (Kai, Inder, & Vijay, 2002, p. 89)
 Spurious Response Rejection: Due to the non-linearity of the receiver, spurious
signals are generated. A good receiver system should have the ability to
distinguish and reject these generated spurious signals. Good spurious response
rejection can be achieved by choosing the right intermediate frequency at the

21
 STUDENT ID: UP698209

design stage and using the appropriate filters. Rejection of 70 - 100 is usually
achievable. (Rohde, 2005, p. 12)
 Intermodulation Rejection: Apart from generating spurious signals the receiver
also generates its own on-channel interference known as intermodulation
products. Good receivers should have intermodulation rejection greater than 70
dB.
 Frequency Stability: The receiver local oscillator should maintain a good level of
stability for good modulation to take place and to prevent phase noise.
Frequency stability can be achieved by using dielectric resonators at the source.
 Radiation Emission: In receivers, the signal generated by the local oscillator can
leak and radiate through the antenna into free space. This radiation is set at a
minimum by regulatory authorities in order to prevent interference and to
ensure that the communication devices are safe. The FCC (Federal
Communications Commission) specifies that the public exposure for cellular
systems must not exceed a specific absorption rate of 1.6 watts per kilogram (1.6
W/kg) (www.fcc.gov, 2013).

3.2.2 Transmitter Parameters

In wireless communication systems, the transmitter performs the function of generating
and sending out the signal in a form that is detectable by the receiver. The
characteristics of transmitters systems are;
 Power Output and Operating Frequency: The transmitter used by any individual
operator is restricted to a certain output RF power level at the assigned
frequency or frequency range.
 Efficiency: This defines how efficient the process of DC to RF conversion takes
place in the transmitter given as;

 Power Output Variation: This characteristic refers to disparity of the transmitted

power level over the allocated frequency range of operation. This variation ought
to be kept as low as possible.

22
 STUDENT ID: UP698209

 Frequency Tuning Range: This refers to the range over which the transmitter can
be adjusted either by electronic or mechanical means whilst still maintaining its
properties.
 Stability: Transmitter stability is its capacity to return to its original operating
point after being subjected to some form of electrical or mechanical disturbance.
 Circuit Quality Factor: The quality factor, Q of the oscillator’s circuit in resonance.
 Noise: All the unwanted variations due to modulation.
 Spurious Signals: These are any unwanted output signals generated at
frequencies other than the desired transmission frequency.
(Kai, Inder, & Vijay, 2002, p. 99 &p. 100)

3.2.3 Filter Design Requirements for Modern Wireless Communication

The filter design techniques for modern wireless communication applications are aimed
at increasing the efficiency and performance of the filters while keeping the size as small
as possible in order to facilitate mobility. The degree of miniaturization is mostly
prohibited by the associated degradation of the filter performance (Ahmad, Ali, Naseer &
Amjad, 2014, p. 1). Filters for modern 4G – LTE wireless communication for high end
applications in diplexers and quad – band operations are required to be of high quality,
compact in size and also reasonably priced.
In most 4G applications, only bandpass filters are required. An ideal bandpass filter will
have constant group delay and no insertion loss in the pass band frequencies, with
infinite attenuation everywhere else.

Figure 3.1: Ideal Bandpass Filter Response. (Poole, 2013)

23
 STUDENT ID: UP698209

In practice, filters will significantly deviate from the ideal response shown above but
these deviations are bounded by tolerance limits. The performance of a typical bandpass
filter used in practice should have -3 dB attenuation within the pass band, transition
band steepness of -40 dB and input reflection should be less than or equal to -6 dB
(Ahmad, Ali, Naseer & Amjad, 2014, p. 3). The physical size of the filter can be adjusted to suit
the desired purpose as long as the filter remains within the performance specifications.
The various design techniques of high quality RF filters often involves cumbersome
mathematical calculations. The most commonly utilized method is the low-pass
prototype method. Filter design using this method involves the following steps:
 The first step is to design the prototype lowpass filter with the desired pass band
frequency response.
 This prototype design is then transformed to the desired type of filter (low-pass,
high-pass, band-pass or band-stop).
 The transformed design is then realized in hardware either in lumped elements
or in distributed elements. This is mostly determined by the frequency range the
filter is required to operate in.

3.2.3.1 Definition of Filter Parameters

The power that is fed into a filter can be viewed as three distinct components which are:
The incident power form the generator (Pin), the power reflected back toward the
generator (PR) and the power absorbed by the filter (PA) (Kai, Inder, & Vijay, 2002, p. 171).
In accordance with the law of energy conservation,

Filters also have important parameters such as:

 Insertion loss: This is the loss of signal power as a result of the insertion of a
device on a transmission line. It is the difference between the power supplied by
the generator and the power that actually gets to the input of the load.
(Annapurna & Sisir 2009, p. 507). The insertion loss IL in decibels is given as:

(3.4)

 Return Loss: The return loss is the ratio of the input power to the reflected
power. The return loss RL is given as:
(3.5)

24
 STUDENT ID: UP698209

 Group Delay: The group delay shows the phase characteristics of the filter. It is a
measure of the time taken for a signal to propagate through the filter. The group
delay τD is given as:

Where and is in radians per seconds.

Modern filter designs are rarely done using manual calculations. There are quite a lot of
CAD (Computer Aided Design) tools based on numerical methods to assist in very
accurate filter design.

3.2.4 Power Splitter Design Requirements for Modern Wireless Communication

Ideally, for a Wilkinson power splitter, there ought to be a 3 dB power and perfect
isolation between the two output ports at the centre frequency. There is also supposed
to be a symmetric power split in both amplitude and phase. In the worst case, the
difference in power should not exceed 0.5 dB and the phase difference should not
exceed 8° (Lourandakis, Weigel, Mextorf, & Knoechel, 2012).

Figure 3.2: Ideal Response of Equal Split Power Splitter (Foley, 2013)

25
 STUDENT ID: UP698209

3.2.5 Coupler Design Requirements for Modern Wireless Communication

In digital cellular wireless communications, couplers and hybrids are used for
determining the phase error of transmitters using the quarter wavelength (λ/4) shift
QPSK (Quadrature Phase Shift Keying) modulation. This requires the coupler or hybrid to
have an equal power split quadrature phase difference between the output ports. In
order for couplers and hybrids to be acceptable for use in communication systems, the
maximum allowable deviations are 0.4 dB for amplitude of the power output (maximum
loss of 3.4 dB) at the output ports and ±5° phase shift (Lourandakis, Weigel, Mextorf, &
Knoechel, 2012). Other requirements include a return loss (S11) below 15 dB and isolation
below 20 dB (Mocanu, 2014, p. 5-6).

Figure 3.3: Rat Race Coupler (AWR Corporation)

Figure 3.4: Power Split of Rat Race Coupler (AWR Corporation)

Fig. 3.4 shows the power split of a rat race coupler. For an ideal equal split rat race
coupler, there is a -3 dB power drop at the output ports, in this case ports 2 and 3.

26
 STUDENT ID: UP698209

3.2.6 Coupler and Splitter Figures of Merit

Similar parameters are used to specify the quality and performance characteristics of
couplers and splitters although these figures differ slightly. These figures are specified in
the datasheets. They are briefly discussed below.

 Splitting Ratio: The splitting ratio or coupling ration is the ratio of the power at
the output to the power at the input. A power splitter that possess a 2:1 power
split ratio will supply 50% (-3 dB) of the input power to each of the output ports,
while a splitter with 3:1 splitter will supply 33% (-4.8 dB) of its input power t the
output port. An unequal power split can also be achieved but this is normally
done using directional couplers due to the impracticality of producing
asymmetrical Wilkinson power splitters (Marki Microwave, Inc. 2012).
 Relative Phase Shift: The relative phase shift is the phase difference between the
various output signals. Power splitter and coupler circuits can be designed to give
0 °, 90° or 180° phase shift.
 Amplitude Balance: The measure of how equally the output power is split at the
output port is referred to as the amplitude balance. It is the difference between
the amplitude of the output signals. Ideally, this value should be zero. Good
splitters and couplers have below 0.25 dB and 0.4 dB for splitters and couplers
respectively (Marki Microwave, Inc. 2012).
 Amplitude Ripple: In ideal situations, couplers and splitters ought to maintain a 0
dB ripple over the operable bandwidth. However in real devices, there is almost
always some form of rippling within the pass band. There tends to be less
rippling in tight band devices as compared to wide band devices, hence there
exists a trade off between broad band operation and amplitude rippling. In
general, power splitters have better amplitude flatness than couplers (Marki
Microwave, Inc. 2012).

 Phase Balance: Phase balance is only relevant to equal power split devices. This
parameter measures the difference in phase shift between the output arms of
the device. Ideally, this difference should be zero, but in reality it is usually in the
region of a few degrees (Lourandakis, Weigel, Mextorf, & Knoechel, 2012).

27
 STUDENT ID: UP698209

 Phase Ripple: Phase ripple usually occurs in devices as a result of output

transmission line asymmetry, poor VSWR matching at the ports and high
operating frequency. Ideal devices have no phase ripple, but well designed and
manufactured devices fluctuate by a few degrees (usually less than 10° (Marki
Microwave, Inc. 2012))

 Output Ports Isolation: Ideal power splitters have both output ports perfectly
isolated, meaning that there are no signal leakages between the output ports.
This also applies for isolated coupler ports. However, real splitters that can
achieve values above 15 dB isolation are considered acceptable. While the
isolation for coupler is often desired to be as much as 30 dB.

28
 STUDENT ID: UP698209

CHAPTER 4
4.0 CHAPTER OVERVIEW
This first section of this chapter will discuss the various design considerations with
respect to the implementation of circuits on microstrip substrates. Subsequent sections
will then address the practical design of RF passive circuits (filter, splitter and coupler) as
well as the available techniques and tools for design and production.

4.1 MICROSTRIP DESIGN CONSIDERATIONS

The selection of appropriate microstrip substrate is anchored on the proper
understanding of the factors that affects its performance.

Figure 4.1: Microstrip Line (Jefferies, 2002)

The physical dimensions of the microstrip transmission line are “W” which represents
the width and “t” which is the thickness of the conductive strip. The “ground plane” is a
conductive material that runs across the entire lower surface. “H” is the height of the
dielectric substrate.
The dielectric can be made from various materials which have their characteristic
dielectric constant “Ɛr”, loss tangent “tan δ” and thermal conductivities. Some common
materials are:

29
 STUDENT ID: UP698209

 Plastic: This offers a low cost, readily adhesive and production friendly option.
The major disadvantage of plastic is its poor dimensional stability, large thermal
coefficient and poor conductivity (Jefferies, 2002).
 Ceramic: Ceramics are mostly used in high end microstrip antennas and optical
filters ( arneiro ilho, ra o, inani, d ssun o, Mendon a, 2009, p. 1-4). Some variants
of ceramic substrate such as Titanium Dioxide (TiO2) have dielectric constant of
up to 100 (Lundstrom, Smith, Rinehart, Pate, Krogh & Huebner, 1999, p. 1489-1491).
Ceramics have good thermal characteristics and dimensional stability. The major
disadvantage of ceramics substrates is the cost, rigidity and difficulty to shape,
cut and drill (Jefferies, 2002).
 Single Crystal Sapphire (SCS): SCS substrate is mostly used for very challenging
applications such as broadband terahertz frequency applications (Shou, 2009, p.4-6).
It is characterised by uniform permittivity along the entire line which makes the
propagating waves not to be subject to changing permittivity at bends or corners
in the line (Owens, Aitken, & Edwards, 1976, p. 499-504). Its high cost is mostly
prohibitive.
 Single Crystal Gallium Arsenide (GaAs): The GaAs is mostly applied in MMICs
(Monolithic Microwave Integrated Circuit). Due to its acoustic mode propagation
property, transmitted wave can be coupled on to the conductor. It piezoelectric
property makes it not very suitable for passive devices, especially if the circuit
will be subjected to any form of mechanical shock (Jefferies, 2002).
 Silicon (Si): Silicon has the advantage of good thermal conductivity, affordability
and established technology due to its long legacy of use. The major setback is the
low resistivity of 10KΩ/cm as compared with 10MΩ/cm for a s (Reyes, El-
Ghazaly, Doml, Dydyk, & Schroder 1994, p. 1759 - 1762).

In general, the power handling capacity of ceramics and single crystals are much higher
than plastics and silicon, hence high Q structures tend to remain more stable despite
higher frequencies. Table 4.1 shows a summary of the typical characteristics of various
substrate types.

30
 STUDENT ID: UP698209

Table 4.1: Typical Characteristics of Substrates Types (Jefferies, 2002)

Single Single Crystal
Plastic Ceramic Crystal Gallium Silicon
Sapphire Arsenide(GaAs)

Ɛr 2.2-10.4 8 – 10 9.4 – 11.6 13 12

Tan δ 0.001 1 10-4 - 5 10-5 6 10-4 5 10-

10-3 3

Surface
Roughness (mm) 0.006 0.05 0.01 0.025 0.025
Thermal
Conductivity
(watts per cm 0.003 0.3 0.4 0.3 0.9
per degree)

To obtain workable microstrip hardware, other important factors such as size, power
handling and higher order modes also need to be considered.

 Size: Wireless circuits for RF and microwaves are required to be as small as

possible, in order to improve mobility of wireless devices. However, the sizes also
have minimal limits to make them convenient for manufacturing.
 Higher Order Modes: The operating frequency must be kept below the cut off
frequency of the first order modes of the selected substrate. This is done in order
to avoid excitation of higher modes. The cut off frequency for the higher order
modes is given in (4.1) (Hong & Lancaster, 2001, p. 84).

(4.1)

Where “W” is the width, “C”, is the speed of light in free space, “h” is the substrate
height and is the dielectric constant.

 Power Handling Capacity: Various microstrip substrates have their distinctive

power handling capacities. This property is proportional to the breakdown

31
 STUDENT ID: UP698209

voltage of the substrate, and is limited by heating (Ramesh, Inder & Maurizio, 2013, p.
82).

4.2 PRACTICAL DESIGN OF RF PASSIVE CIRCUITS

This work will focus on the creation of implementable design of RF passive circuits using
the Microwave Office software from AWR Corporation. The circuits to be designed are
bandpass filter, power splitter and rat race coupler.

4.2.1 Frequency and Bandwidth Selection

In order to make this research work relevant to the modern trends in wireless
communication, the frequency spectrum used by the Fourth Generation (4G) mobile
communication systems is selected as the basis of all the designs. The 4G standard
utilizes the 800 MHz, 1800 MHz, 2600 MHz, 3500 MHz and 3600 MHz frequency spectra
for communication. The various bands have different data carrying capacity and distance
coverage. Lower frequency bands have lower data carrying capabilities but can travel
further than higher bands. The 800 MHz frequency band is mostly utilized in areas that
are not densely populated because it has a lower data carrying capacity than the higher
bands and also travels over much longer distances. However, the 1800MHz band strikes
a balance between coverage and capacity (Rogerson, 2013). This work will be focused on
designs that are applicable in 4G LTE 800 MHz band, with centre frequency at 840 MHz,
and bandwidth of 120 MHz.

4.2.2 Substrate Selection for 4G Wireless Application

Substrate selection is a major concern when implementing RF circuits in a PCB (printed

circuit board) because the performance of the finished product is greatly dependent on
the performance qualities of the selected substrate, despite the design looking good on
paper. Various substrates were considered for the implementation of the design. In the
substrate consideration process, the availability, workability performance and ability to
handle high frequencies were well considered.
The Rogers 4350B substrate was considered for use due to its relatively stable properties
over the frequency of interest (Refer to Appendix A). The major drawback with the Rogers
substrate is the difficulty associated with working with it after production, due to its
fragility.

32
 STUDENT ID: UP698209

The design properties for Rogers 4350B are shown in Table 4.2
Table 4.2: Design properties for Rogers 4350B Substrate (Appendix A)
Typical Value Units Condition
Dielectric Constant (Ɛr) 3.66 - 8 to 40 GHz
10 GHz at 23°C
Loss Tangent (tan δ) 0.0037 - 2.5 GHz at 23°C
Substrate Height 1 mm -
Available in: mm -
0.168, 0.254,
Standard Thickness 0.338, 0.422,
0.508, 0.762,
and 1.524.

The FR-4 substrate was also considered for the design. This substrate is very commonly
used because it is low cost and easy to handle after production. It also has a linear
response over the frequencies of interest.
The design properties for the FR4 are shown in Table 4.3.

Table 4.3: Design Properties of the FR4 Substrate. (Appendix B)

Typical Value Units Condition
Dielectric Constant (Ɛr) 4.5 - 1 GHz
Loss Tangent (tan δ) 0.02 - 1 GHz
Substrate Height 1 mm -
Standard Thickness 1 mm -

To further investigate the effects of these properties on circuit designs, transmission

lines on both substrates were simulated. Both substrates were subjected to the same
frequencies and their S parameters were measured. Fig. 4.2 shows the result of the
simulation.

33
 STUDENT ID: UP698209

Figure 4.2: Substrate Comparison

The simulations in Fig. 4.2 shows that the Rogers 4350B substrate is able maintain a
much broader linear window than the FR4. The FR4 substrate maintains linearity to
3.828 GHz, while the Rogers 4350B substrate is able to stay linear up till 50 GHz.
Despite the obvious difference in performance, the FR4 was still selected for this work
because it was readily available, less fragile than the Rogers 4350B and it also showed
linear response across the frequency of interest.

4.2.3 Bandpass Filter Design for 4G Wireless Application

In 4G wireless applications, bandpass filters are often used in the RF front end circuits to
select the required band and channel of interest for a given communication system.
Bandpass filters are preferred because lowpass and highpass filters contain frequency
components that are not necessarily of interest, they cost less and they are easy to
manufacture on microstrip (Ahmad, Ali, Naseer, Amjad, 2014, p. 1-15).

4.2.3.1 Bandpass Filter Design Specifications

The specifications for the bandpass filter to be designed are;
 The filter should be able to perform at the centre frequency of 840 MHZ, with a
bandwidth of120 MHz
 There should be a minimum of -3 dB attenuation within the pass band
 A minimum transmission band steepness of -40 dB

34
 STUDENT ID: UP698209

 A minimum input reflection of -6 dB.

4.2.3.2 Bandpass Filter Design Techniques

Bandpass filters can be implemented using a variety of design techniques. These
techniques are all aimed at implementing filters that are small in size and show high
performance. The required filter is implemented using lumped elements, and then
various microstrip formats are also implemented.
In the design of filters for the use of microwave and RF systems, the Chebyshev design
can sometimes be desired over the Butterworth design because for any value of n, “it
provides the maximum possible stop band monotonic insertion loss for a specified
maximum pass band insertion loss ripple” (Kai, Inder, & Vijay, 2002, p. 180). However,
Butterworth filters are utilized in digital wireless communications because of the large
phase delay characteristic of the Chebyshev filter.
The traditional method of designing bandpass filters is to transform from a lowpass
prototype into a bandpass equivalent, and realizing the network in parallel and series LC
resonant circuits as shown in Fig. 4.3.

Figure 4.3: Element Transformation (AWR Corporation)

The capacitors are transformed into series resonant circuits while the inductors are
transformed into parallel resonance circuits.

35
 STUDENT ID: UP698209

4.2.3.3 Lumped Element Bandpass Filter Design

The Microwave Office software from AWR was used to design and simulate a lumped
element bandpass filter which can meet all the specifications. The filter was kept to a
minimum required order in order to keep its size as small as possible.
The design of bandpass filter can be done using the following steps;
 First, the fractional bandwidth (FBW) is calculated using (4.1) (Salima, Nasreddine,
Bendimerad & Bendallah, 2012, p. 207).

Where is the upper pass band frequency, is the lower pass band frequency,
and .

 The number of resonators that will be required is then determined using (4.2)
(Kai, Inder, & Vijay, 2002, p. 183).

Where is the variable frequency and is the bandwidth given by , is

the variable angular frequency and is the 3 dB band edge point frequency. The value
obtained for this relation is then used to select the number of sections using the
nomograph shown in Fig. 4.4.

Figure 4.4: Nomograph for number of resonant sections for Chebyshev filter (Kai, Inder, &
Vijay, 2002, p. 180).

The prototype element values are then obtained from the values in Table 4.4.

36
 STUDENT ID: UP698209

Table 4.4: Element Values for Chebyshev Prototype Filter with 0.5 dB Ripple (Source:
Pozar, 2012, p. 406).

 The lumped element values are then calculated using (4.3), (4.4) for series
resonator and (4.5), (4.6) for parallel resonators.

Where = , and the value is in radians / second.

The bandpass filter was realized using the iFilter Wizard in the Microwave Office
software. For a third order bandpass filter with centre frequency at 840 MHz, with a
bandwidth of 120 MHz and a pass band ripple of 0.5 dB, these values were obtained.

37
 STUDENT ID: UP698209

This software helps to generate lumped filters with realistic response because it makes
use of the parasitic properties of real circuit components. The generated filter is shown
in Fig. 4.5.

Figure 4.5: 3rd Order Bandpass Filter Configuration (AWR Corporation)

The simulation for the insertion loss (S21), the return loss (S11) and the group delay of
the filter schematic are shown in Fig. 4.6 and Fig. 4.7.

Figure 4.6: S Parameters of 3rd Order Chebyshev Bandpass Filter

The response of the filter as seen in Fig.4.6 shows that the centre frequency is at 819
MHz with the 3 dB bandwidth ranging from 765.6 MHz to 900 MHz. This shows that
implementation of this filter using real lumped elements will result in a centre frequency
which is not the same with the desired centre frequency. This filter possesses a good
return loss property of -7.124 dB within the pass band and also a very steep transition
band.

38
 STUDENT ID: UP698209

Figure 4.7: Group Delay of 3rd Order Chebyshev Filter

The simulation of the group delay in Fig.4.7 reveals that the maximum delay within the
bandwidth of interest is 9.163 ns.
The results of these simulations show that a lumped element filter can be utilized for
this frequency. The drawback is physical size and weight and the sheer number of
components that have to be physically connected to realize these filters.
As earlier stated, the Chebyshev filter has a higher phase delay than an equivalent order
Butterworth filter; hence Butterworth filters are most often used in modern digital
communication systems. For the Butterworth filter, its L and C values are computed
using (2.17) and (2.18). This lowpass prototype is then transformed into the bandstop
filter using Table 4.5.
Table 4.5: Element Values for Butterworth Filter (Source: Pozar, 2012, p. 406).

39
 STUDENT ID: UP698209

The Butterworth bandpass filter circuit elements are then calculated using (4.7) for
series components and (4.8) for shunted components (Pozar, 2012, p. 408).

The filter was realized using the same design specifications as the Chebyshev and the
following element values were obtained.

The schematic of the filter which is realized using real elements is shown in Fig. 4.8

Figure 4.8: Butterworth Filter with Real Elements (AWR Corporation)

The S parameters and group delay were simulated and the response shown in Fig. 4.9
and 4.10.

40
 STUDENT ID: UP698209

Figure 4.9: S Parameters of 3rd Order Butterworth Bandpass Filter

The Butterworth filter shows an operation in a narrower bandwidth than the design
specification. Its return loss is much better than the Chebyshev and it also shows a
sharper roll off.

Figure 4.10: Group Delay 3rd Order Butterworth Bandpass Filter

As shown in Fig. 4.10, the maximum group delay in within the band of interest is 8.495
ns. This means that the filter will respond better than the Chebyshev if applied in a
digital communication system.

41
 STUDENT ID: UP698209

4.2.3.4 Microstrip Bandpass Filter Design

Bandpass filters for microwave and high RF frequencies are usually produced on
microstrip substrates because of the significant reduction in size, loss, weight, cost and
their overall better performance than lumped element filters. Microstrip filters are
produced in various configurations which include parallel coupled line, hairpin,
interdigital, combline and shunt stub. The interdigital, combline and hairpin filters are
able to achieve increased compactness through ‘side by side’ placing of the resonators.
The bandpass filter earlier designed in using lumped elements is transformed into the
microstrip equivalent using the various microstrip configurations mentioned.
 Parallel Coupled Line: The parallel coupled line filter is mostly used in bandpass
and bandstop applications, where the frequency bandwidth is narrow (usually
less than 20% of the centre frequency). It is also a low cost option when
compared to other microstrip configurations. (Salima, Nasreddine, Bendimerad &
Bendallah, 2012, p. 207). The parallel coupled line consists of open circuited coupled
microstrip lines that are λ/4 long, thus making them equivalent to shunt resonant
circuits. The mathematical derivation of the parameters for coupled line filters
are very well detailed in (Pozar, 2012, p. 426 - 443).
The design technique utilized here involved the specification of the design parameters
and then using the iFilter CAD tool to generate a schematic. After generation of the
schematic, the variable tuner is then used to tune the filter properties until the best rest
filter response is obtained.

Figure 4.11: Parallel Coupled Line Schematic

42
 STUDENT ID: UP698209

The filter was then tuned to optimise its response as much as possible. The frequency
response of the filter is shown in Fig. 4.12.

Figure 4.12: Frequency Response of Coupled Filter

The filter realized showed a narrower bandwidth than specified. The roll off of the filter
is not steep enough to meet design requirements. Coupled line filters are generally
characterised by a low rejection level hence they are not always used for applications
where a very steep roll off is desired. However, (Lopetegi, Laso & Hernández, 2001, p. 1593 –
1598) showed that the rejection level can be increased by up to 20 dB using “Wiggly-
Lines” instead of the conventional parallel lines.
Despite the low rejection and narrow band characteristic of coupled line filters they are
still in use because of their simplicity and also they form the basis for various other
forms of microstrip filters. The PCB and 3D artwork model generated by the AWR
Microwave Office software of the designed coupled line filter is shown in Fig 4.13 and
Fig. 4.14.

Figure 4.13: PCB View of Coupled Line Filter

The filter has a length of 246 mm.

43
 STUDENT ID: UP698209

Figure 4.14: 3D Artwork Model of Coupled Line Filter

 Hairpin Filter: The hairpin filter is most popularly used in the lower microwave
region. The basic idea behind a hairpin filter is the coupled line filter, with its
resonators folded. This topology makes the filter more compact. The folding of
the resonators results in the reduction of the coupled line lengths, which in turn
results in the reduction of coupling between the resonators. The computation of
the dimensions of the hairpin filter is not an exact science hence adequate
measurements are obtained by simulating different samples until desired result
is achieved (Li, Hu, & Chen, 2008).
A sample hairpin filter schematic was generated using the iFilter CAD tool and its
properties were continuously systematically adjusted until a desired result was
achieved. The hairpin schematic is shown in Fig. 4.15.

Figure 4.15: Hairpin Filter Schematic

44
 STUDENT ID: UP698209

The frequency response of the optimised filter schematic is shown in Fig. 4.16.

Figure 4.16: Frequency Response of Hairpin Filter

The optimised hairpin filter showed a better bandwidth handling and steeper roll off
than the coupled line filter. The PCB and 3D artwork views are shown in Fig. 4.17 and
Fig.4.18.

Figure 4.17: PCB View of Hairpin Filter

The hairpin filter has a length of 47.54 mm. This shows a significant reduction in the
length of the filter as compared with the coupled line design.

45
 STUDENT ID: UP698209

Figure 4.18: 3D Artwork Model of Hairpin Filter

The hairpin filter has the advantages of a sharper roll off and more compact size than
the coupled line filter. It is also easy to manufacture because it does not require critical
grounding (Toledo 2011, p. 2-3).

 Shunt Stub Filter: The shunt stub filters is realized by using a combination of
shorted resonators that are λ/4 long. The filters are relatively easy to design on
microstrip as the characteristic impedances can be computed using (4.9) and
Table 4.5 (Whites, 2013). This characteristic impedance can then be converted into
line thickness using the AWR TXLine tool.

The schematic for the shunt stub filter is shown in Fig. 4.17.

Figure 4.17: Shunt Stub Filter Schematic

46
 STUDENT ID: UP698209

The filter schematic was tuned until a fair response was realized. The frequency
response of the filter is shown in Fig. 4.18.

Figure 4.18: Frequency Response of Shunt Stub Filter

The frequency response of this filter is not as desired. Along with this, the shunt stub
filter requires line widths which are often not practical especially for bandpass filters
with bandwidths less than 50 % (Whites, 2013). In general, the shunt stub filter is not
recommended for bandpass applications. The PCB view and the 3D artwork of the filter
are both shown in Fig. 4.19 and Fig. 4.20.

Figure 4.19: PCB View of Shunt Stub Filter

47
 STUDENT ID: UP698209

The filter has a length of 150.8 mm.

Figure 4.20: 3D Artwork Model of Shunt Stub Filter

 Combline Filter: The combline filter configuration is a variant of the coupled line
filter. It is mostly utilized for the microstrip implementation of narrow and
medium bandwidth microwave bandpass filters. They have been known to show
excellent stopband and selectivity performance. They are also very compact and
easy to integrate (Torregrosa- enalva, L pe -Risueño, & Alonso, 2002, p. 173).
Although useful expressions for the design of combline filters have been shown by
(Rhodes & Hunter, 1982), (Torregrosa-Penalva, L pe -Risueño, & Alonso, 2002) and (Sanchez-Renedo,
Gomez-Garcia, Alonso, Briso-Rodriguez, 2005), the more efficient way to design them is to use
CAD tools to simulate various samples of the combline filter schematic setup by
controlling the resonant section until the filter meets the required frequency response.
This technique was applied making use of the AWR filter design tool. Fig. 4.21 shows the
schematic of the filter.

Figure 4.21: Combline Filter Schematic

48
 STUDENT ID: UP698209

The generated schematic circuit for the filter was tuned adjusting the space between the
coupled lines until the frequency response as shown in Fig. 4.22 was obtained.

Figure 4.22: Frequency Response of Combline Filter

This bandwidth response of the filter is slightly broader than specified although this can
be suitable for production purposes. The PCB and 3D model view of the combline filter is
shown in Fig. 4.23 and Fig. 4.24.

Figure 4.23: PCB View of Combline Filter

49
 STUDENT ID: UP698209

The combline has a length of 29.464 mm. This shows that it is a very compact filter.

Figure 4.24: 3D Artwork Model of Combline Filter

The combline filter shows good bandpass properties. However, the produced piece has
to be critically grounded, and the lumped element capacitors degrade the performance
at higher microwave frequencies.

 Interdigital Filter: The interdigital filter is a derivative of the coupled line filters.
It is widely utilized in microwave systems due to its ease processing, consistency,
reliability and overall good performance. This filter comprises of an array of λ/4
transmission line resonators which are short circuited to ground at one end and
open circuited at the other end with alternative orientation (Ting, Tam, Martins,
2004, p 984-987). Efficient design and implementation of this filter can be achieved
by optimising the traditional schematic (Xu, Yu, Shi & Zhao, 2009). The traditional
filter is obtained by transformation of the lumped bandpass parameters and then
achieving the microstrip parameters as described in (Ziang, 2006).
An initial schematic was created using the iFilter tool and then optimised to obtain an
acceptable frequency response. The gain at the centre frequency was found to increase
with reduction of the resonator length and the bandwidth was found to widen with
reduction in the distance between resonators

50
 STUDENT ID: UP698209

Figure 4.25: Interdigital Filter Schematic

The frequency response is shown in Fig. 4.26.

Figure 4.26: Frequency Response of the Optimised Interdigital Filter

The interdigital filter shows good bandpass frequency response. Its produced piece
however has to be properly grounded to maintain good response. The PCB and 3D
artwork model is shown in Fig 4.27 and Fig 4.28.

51
 STUDENT ID: UP698209

Figure 4.27: PCB View of Interdigital Filter

The filter has a length of 51mm, showing good compactness.

Figure 4.28: 3D Artwork Model of Interdigital Filter

4.2.3.5 Comparison of Microstrip Bandpass Filter Design

The comparison insertion loss of the filters is shown in Fig. 4.29.

Figure 4.29:Comparison of Insertion Loss of Microstrip Filters

52
 STUDENT ID: UP698209

Fig. 4.29 shows that all the microstrip filters designed show good bandpass
characteristics except for the coupled line filter which shows poor rejection of out of
band frequencies. The return loss of the microstrip filters is shown in Fig. 4.30.

Figure 4.30: Comparison of Return Loss of Microstrip Filters

Fig. 4.30 shows that all the microstrip filters designed showed return loss below the
design specification of -6 dB at the centre frequency.
Although the group delay property was not specified in the design, filters used for
modern digital communications are required to have their group delay to be as low as
possible in order to achieve more synchronised communication between transmitter
and receiver. The group delay for the filters is shown in Fig. 4.31.

Figure 4.31: Group Delay of Microstrip Filters

53
 STUDENT ID: UP698209

From Fig. 4.31, it is clear that the microstrip filters have less group delay than the
lumped element filters earlier designed, except for the combline filter which has a delay
of 9.146 ns. The hairpin filter showed the lowest group delay property across the
frequency band of interest hence it is applicable for digital communication.

4.2.4 Power Splitter Design for 4G Wireless Application

As earlier discussed, a power splitter is a three port device for creating two equally split
signals from one input port. Both output signals are of the same phase and power level.
This split is achieved by applying two matched quarter wavelength lines to the input
port. Splitters are applied in communication systems for splitting RF power into load
circuits. They are also used for switches and antenna array feed networks (Choinsk, 1984, p.
613-614).

4.2.4.1 Power Splitter Design Specifications

According to (Kim, Jung & Yun, 2004), at the design level, a power splitter is required to
meet the following requirements;

 The power splitter should have a return loss (S11) of -9.54 dB

 For and equal split splitter, an insertion loss across the splitting arms of not less
than -3.6 dB
 The isolation between the output arms of -15.7 dB.

The datasheet of a Wilkinson power splitter manufactured by Marki Microwave

Incorporated (appendix C1 and C2) shows the following specifications;

Table 4.6: Power Splitter Performance Specification (Appendix C1 and C2)

Parameter Value
Excess Insertion Loss (S12) -3 dB 2
Return Loss (S11) 12.7 - 17.7dB
Isolation (S23) -22 dB
Amplitude Imbalance 0.1
Phase Balance 1

54
 STUDENT ID: UP698209

4.2.4.2 Power Splitter Design Procedure

Section 2.2.2.1 of this work shows the theory behind the Wilkinson type power splitter.
This splitter is basically made up of an input line and two λ/4 lines which act as λ/4
transformers. The λ/4 lines have an impedance of Zo which is terminated by a
Zo to isolate the output ports.
To design the splitter, the value for λ/4 is calculated. The length λ in free-space can be
calculated using (4.10). However for microstrip designs (4.10) does not apply because of
the effect of the dielectric material.

The λ/4 length is computed using (4.11) (Kai, Inder, & Vijay, 2002, p. 130).

Where is the wavelength in free-space, C is is is the speed of light F is the centre

frequency, and is the effective dielectric constant given by (4.12).

(4.12) is only accurate when , where w is the transmission line thickness, is the
dielectric constant and h is the substrate height. Applying (4.11) and (4.12);

, and

The TXLine tool was used to compute the line width that would give the characteristic
impedance Zo) for the λ/4 transformer section. This is shown in ig. 4.32.

Figure 4.32: Line Width Calculation using TXLine Tool

55
 STUDENT ID: UP698209

s shown in ig. 4.32 the required width to give the characteristic impedance of 70.71 Ω
is 0.743059 mm, hence the transformer lines will have this thickness.
The AWR software was then used to implement the design parameters as shown in Fig
4.33.

Figure 4.33: Wilkinson Power Splitter Layout

With reference to Fig. 4.33, W is the width of the input and output ports, W1 is the
width of the λ/4 transformer sections and R is the resistance of the output port isolation.
This layout shown in Fig. 4.33 cannot be synthesized by the software to a microstrip
substrate, so a different layout had to be built. This layout is shown in Fig. 4.34.

Figure 4.34: Schematic for Power Splitter

The frequency response of the power splitter built is shown in Fig. 4.35.

56
 STUDENT ID: UP698209

Figure 4.35: Frequency Response of Power Splitter

The simulation shows the power splitting properties. The input signal which is fed
through the port 1 is split into two equal amplitude signals which are -3.174 dB less than
the input signal, showing that about 42% of the input power is split and fed into the two
output ports. The parameter S11 which is at -32.25 dB shows that the ports are well
matched. The S23 which is the isolation between ports 2 and 3 is -31.79 dB. The response
shows conformity with the design specifications. Another important parameter of the
power divider is the equal phase of the output signals. The angular component of the
design power splitter is shown in Fig 4.36.

Figure 4.36: Angular Response of Power Splitter

57
 STUDENT ID: UP698209

Fig. 4.36 shows that the output signals of the power splitter are the same, which
conforms to the design specifications.
The power splitter can also be used as a power combiner by feeding two signals of equal
magnitude and phase into the ports 2 and 3. This property was also investigated. Fig.
4.37 shows the schematic set up for this investigation.

Figure 4.37: Wilkinson Power Splitter in Power Combiner Setup

For the purpose of this test, the output lines feeding ports 2 and 3 were merged into a
single line, and the then fed with an input signal. The power reflected at port 1 is shown
in Fig. 4.38.

Figure 4.38: Wilkinson Power Splitter as Combiner

58
 STUDENT ID: UP698209

Fig. 4.38 shows a reflection of -0.6338 dB (about 85%) of the input power is reflected at
the output port.
The PCB view and the 3D artwork model were generated using the AWR software. They
are shown in Fig. 4.39 and Fig. 4.40.

Figure 4.39: PCB View of Power Splitter

Figure 4.40: 3D Artwork Model

A known drawback of the Wilkinson design is the relatively large length of the splitter
when it is designed to operate at lower frequencies. A more improved design was done
by replacing the quarter wavelength line with a curve and a reduction in the physical
length of the splitter was achieved, while remaining within the splitter design
specifications. The design was done and the circuit schematic is shown in Fig. 4.41.

59
 STUDENT ID: UP698209

Figure 4.41: Schematic of Wilkinson Splitter in Circular Layout

Figure 4.42: Frequency Response of Wilkinson Splitter in Circular Layout

The circular layout shows a power output of -3.169 dB which is fed to the two ports. The
S11 parameters shows good matching as the S23 shows good isolation between the
output ports.

60
 STUDENT ID: UP698209

Figure 4.43: Angular Response of the Circular Power Splitter

The circular power splitter has an angular response which is within the design
specifications. Both output signals are of equal phase.

The PCB view of this layout is shown in Fig.4.44.

Figure 4.44: PCB View of the Circular Power Splitter

The circular design shows a reduction of 24.8 mm (about 40 %) as compared to the

conventional design.

61
 STUDENT ID: UP698209

Figure 4.45: 3D Artwork View of Circular Splitter

The bandwidth of operation of the Wilkinson power splitter can be improved by

inserting a λ/4 section before the split section. This is shown in ig 4.46 and ig 4.47. The
compensating line has an impedance of 42 Ω and the section lines have an impedance of
59.4 Ω.

Figure 4.46: Compensated Power Splitter

Figure 4.47: PCB View of Compensated Power Splitter

62
 STUDENT ID: UP698209

The compensated splitter shows a 95% increase in length.

Fig 4.48a shows the useful bandwidth of an uncompensated power splitter while Fig.
4.48b shows the useful bandwidth of a compensated power splitter.

Figure 4.48a: Useful bandwidth of Uncompensated Power Splitter

Figure 4.48b: Useful bandwidth of Compensated Power Splitter

From Fig.4.48a and Fig. 4.48b, it is clear that the bandwidth of operation which is
restrained by the isolation between ports 2 and 3 is improved by compensation.

4.2.5 Rat Race Coupler Design for 4G Wireless Application

The rat race coupler is a four port device network which is used as a 180° hybrid. This
network is circular, with its circumference equal to 1.5λ at the operating frequency, with
the λ/4 sections having an impedance of Z0 (Tu, & Chao-Hsiung, 2010, p 1224 - 1227). Rat
race couplers are applied in the development of various mixers, amplifiers, antenna

63
 STUDENT ID: UP698209

arrays and they can also be used in feed networks to replace power splitters (Wu, Mao, &
Yin, 2012).

4.2.5.1 Rat Race Coupler Design Specifications

A 180° rat race coupler manufactured by RF Lambda gives the performance as shown in
Table 4.7.
Table 4.7: Rat Race Coupler Design Specifications (Source: Appendix D)
Frequency Range 0.75 - 1.5 GHz
Insertion Loss -3 dB 0.3
Isolation 22 dB
Phase Imbalance 10°
VSWR 1.25:1
Return Loss 19.085 dB

The design to be undertaken in this work will be based on the specifications listed in
Table 4.7
4.2.5.2 Rat Race Coupler Design Procedure
The length of the λ/4 sections of the rat race coupler is computed using (4.11) and
(4.12), and the impedance of the line is also computed as;
=70.71 Ω, giving a width of = 0.743 mm

Figure 4.49: Rat Race Coupler Geometry

Fig. 4.49 shows the geometry of the rat race coupler. The schematic layout of the rat
race coupler is shown in Fig. 4.50.

64
 STUDENT ID: UP698209

Figure 4.50: Schematic Layout of Rat Race Coupler

The radius of the curved section was computed using the relation;

The frequency response of the rat race coupler is shown in Fig. 4.51

Figure 4.51: Frequency Response of Rat Race Coupler

65
 STUDENT ID: UP698209

Fig. 4.48 shows the frequency response of the rat race coupler. It shows an output signal
of -3.86 dB across ports 2 and 3 (S12 and S13), while port 4 is isolated from the input
signal by -30.27 dB. The S11 parameter (-29.43 dB) shows that all the ports are properly
matched. The design specification requires the output signals at the ports 2 and 3 are
180° 10°apart. Fig. 4.52 shows that this specification is met.

Figure 4.52: Angular Response of Rat Race Coupler

Fig 4.49 shows that between ports 2 and 3, a phase difference of 179.73° was achieved.

The PCB view of the rat race coupler is shown in Fig. 4.53.

Figure 4.53: PCB View of Rat Race Coupler

66
 STUDENT ID: UP698209

The 3D artwork of the rat race coupler is shown in Fig. 4.54.

Figure 4.54: 3D Artwork Model of Rat Race Coupler

The rat race coupler designed is able to operate at 21% of the bandwidth of the centre
frequency.

4.2.6 Branch Line Coupler Design for 4G Wireless Application

The branch line coupler is a 4 port coupler with each section separated by a transmission
line which is λ/4 long. The branch line coupler splits the input signal into two quadrature
signals which are equal in magnitude but 90 degrees out of phase. It also has an isolated
port. 3/4 λ, 5/4 λ, and 7/4 λ sections can also be used but these would reduce the
bandwidth of operation.

Figure 4.55: Branch Line Coupler (AWR Corporation)

67
 STUDENT ID: UP698209

With port 1 as the input port, ports 3 and 4 acts as the output ports while port 2 is the
isolated port.

4.2.6.1 Branch Line Coupler Design Specifications

A branch line manufactured by Marki Microwave gives the performance as shown in
Table 4.8.

Table 4.8: Rat Race Coupler Design Specifications (Source: Appendix E)

Frequency Range 650-1300 MHz
Insertion Loss -3 dB 0.5
Isolation 13 dB
Phase Imbalance 3°
VSWR 1.3:1
Return Loss 17.7 dB

The design in this work will be implemented with reference to Table 4.8.

4.2.5.2 Branch Line Coupler Design Procedure

The length of the λ/4 sections of the branch line coupler is computed using (4.11) and
(4.12). The impedance of these λ/4 sections are as shown in the layout on Fig. 2.5. The
AWR schematic is shown in Fig. 4.56.

Figure 4.56: Branch Line Coupler Schematic

68
 STUDENT ID: UP698209

The λ/4 sections connecting ports 1 and 4; and ports 2 and 3 have an impedance of 35.4
Ω while the lines connecting ports 1 and 2; and ports 3 and 4 have an impedance of 50 Ω
(Ferril, 2005). The S parameters and the angular response are shown in Fig. 4.57 and Fig.
4.55.

Figure 4.57: Frequency Response of Branch Line Coupler

As shown in Fig. 4.57, the branch line coupler shows a coupling of -3.652 dB across the
output ports and a bandwidth of operation between 709.51 MHz and 872.52 MHz. This
response shows that the coupler can only be used across a narrow bandwidth.

Figure 4.58: Angular Response of Branch Line Coupler

The angular response of the coupler shows a difference of 90.6° at the output, which is
well within the design a specification.

69
 STUDENT ID: UP698209

Figure 4.59: PCB View of Branch Line Coupler

Figure 4.60: Artwork of Branch Line Coupler

The bandwidth can be improved by adding λ/4 sections (Ferril, 2005). The two stage
coupler is obtained by inserting λ/4 sections between ports 1, 3 and ports 2, 4. The
impedance of the line connecting ports 1, 2 and ports 4, 3 now becomes 121 Ω. This
applies for a 50 Ω system (Ferril, 2005). The schematic for the two stage branch line
coupler is shown in Fig 4.61.

70
 STUDENT ID: UP698209

Figure 4.61: Two Stage Branch Line Coupler

The frequency and angular response are shown in Fig. 4.59 and Fig 4.60.

Figure 4.62: Frequency Response of Two Stage Coupler

As expected, the frequency response of the two stage coupler has a wider bandwidth
than the single stage coupler. It shows a 91.6% increase in the bandwidth.

71
 STUDENT ID: UP698209

Figure 4.63: Angular Response of Two Stage Branch Line Coupler

Fig. 4.63 shows a phase difference between the output signals of 90°. This also shows
good response.
Although the two stage branch line coupler meets the design specifications, it cannot be
implemented on the R4 substrate because the 121 Ω line is too thin to be physically
produced (www.tecbridgecircuits.co.uk).
The PCB and the artwork views are shown in Fig. 4.64 and Fig 4.64.

Figure 4.64: PCB View of Two Stage Branch Line Coupler

72
 STUDENT ID: UP698209

Figure 4.64: Artwork View of Two Stage Branch Line Coupler

4.2.7 Comparison of Design on FR4 and Rogers Substrate

The hairpin filter, power splitter, branch line couplr and rat race circuits were
implemented on the Rogers substrate, using the design parameters shown in Table 4.2.
The hairpin filter seemed to show better bandwidth response and out of band rejection
than the filter implemented on the FR4 substrate but had a higher group delay. The
better bandwidth and rejection is due to the lower loss tangent property of the Rogers
substrate, while the increased group delay is due to the increase in size. The response of
the filter is shown in Fig. 4.65.

Figure 4.65: Frequency Response of Hairpin filter built on Rogers Substrate

The power splitter showed acceptable operation in a narrower bandwidth than the
splitter implemented on the FR4, but the rat race coupler was able to operate within a
slightly broader bandwidth. This implies that implementation of the splitter on a

73
 STUDENT ID: UP698209

substrate with a higher dielectric constant will result in a broader bandwidth of

operation; while a lower dielectric constant will give a broader bandwidth on the rat
race coupler.
Fig. 4.66 and Fig. 4.67 show the frequency response of the power splitter and rat race
coupler.

Figure 4.66: Frequency Response of Power Splitter built on Rogers Substrate

Figure 4.67 Frequency Response of Rat Race Coupler built on Rogers Substrate

74
 STUDENT ID: UP698209

The bandwidth of operation of the power splitter is restricted by the isolation between
the output ports which are to be kept below -20 dB, while the rat race coupler is
restricted by the amplitude imbalance of ±0.5 dB.

The both branch line couplers also displayed better coupling and bandwidth properties
than the couplers implemented on the FR4 substrate. Both coupler layouts showed a
10% increase in the bandwidth when compared to the response of the FR4 circuits.
These responses are shown in Fig. 4.68 and Fig 4.69.

Figure 4.68: Frequency Response of Single Stage Branch Line Coupler on Rogers
Substrate

Figure 4.69: Frequency Response of Double Stage Branch Line Coupler on Rogers
Substrate

75
 STUDENT ID: UP698209

CHAPTER 5
5.0 CHAPTER OVERVIEW
This first section of this chapter will discuss the requirements made by PCB producers for
the manufacture of circuits while the second section will discuss the test of a hairpin
bandpass filter produced on the FR4 substrate.

5.1 PCB PRODUCERS REQUIREMENT

PCB orders sent to producers are required to have the following files available
(Eurocircuits, 2013);

 Gerber files for the copper layers, soldermask and legend layers, mechanical
layers and paste layers as may be needed.
 Excellon drill files, if drilling would be needed.

The Gerber files are generated from the AWR software. Proper care must be taken to
ensure that the files are generated using the correct layout features. The generated files
can be viewed for any errors using the Gerber Logix viewer, available from
www.easylogix.com. Fig. 5.1 shows the Gerber view of the hairpin bandpass filter.

Figure 5.1: Gerber View of Hairpin Filter

76
 STUDENT ID: UP698209

It is also important to ensure that the producers design rules are adhered to. Tecbridge
Circuits requires that (Techbridge Circuits. (2014);
 Minimum track thickness 5 mil (0.13mm)
 Minimum track clearance 5 mil (0.13mm) (For single and double sided boards)
 Minimum track thickness 6 mil (0.15mm)
 Minimum track clearance 6 mil (0.15mm)
A price quotation for the manufacture of a bandpass filter was obtained from Tecbridge
Circuits. The quote is shown in Fig. 5.2.

Figure 5.2: Price Quote for Bandpass Filter

77
 STUDENT ID: UP698209

Figure 5.3: Gerber View of Circular Splitter

Figure 5.4: Gerber View of Rat Race Coupler

The price quotes for the splitter and rat race coupler are shown in Fig. 5.5.

78
 STUDENT ID: UP698209

Figure 5.5: Price Quote for Coupler and Splitter

5.2 BANPASS FILTER TEST

The hairpin bandpass filter with the frequency response shown in Fig. 4.16 was built. Fig.
5.6 shows the photograph of the produced hairpin PCB piece with the ports soldered.

Figure 5.6: Produced Bandpass Filter

79
 STUDENT ID: UP698209

The filter was tested with the use of a signal generator, spectrum analyzer and a
directional coupler. The signal attenuation across the cable was measured to make it
possible to determine the attenuation which is due to the filter alone. After the cable
loss was determined, the insertion loss and return loss of the filter was then measured
using the setup shown in Fig 5.7 and Fig. 5.8.

Figure 5.7: Setup for Filter Insertion Loss Test

Figure 5.8: Setup for Filter Return Loss Test

The insertion loss was measured by connecting the input port of the filter to a signal
generator and the output port to the spectrum analyzer. A frequency sweep from 445
MHz to 1135 MHz was done and traced. The centre frequency was at 843 MHz, with a
gain of -3.33 dB. The trace is shown in Fig. 5.9.

80
 STUDENT ID: UP698209

Figure 5.9: Frequency Sweep Graph

The return loss was measured with the use of a directional coupler. The input port of the
coupler was connected to the signal generator while its output was connected to the
input port of the filter and the “coupled port” connected to the spectrum analy er. The
signals reflected from the input of the filter are coupled to the “coupled” port, and are
then read at the spectrum analyzer. (Setup is shown in Fig. 5.8).

The result of the frequency sweep is recorded as shown in Table 5.1. See Appendix E for
full table.
Table: 5.1: Frequency Response of Bandpass Filter

Frequency (MHz) Gain (dBm) Return Loss(dBm)

445 -30.88 -1.88
455 -30.79 -1.72
465 -30.7 -1.58
475 -30.56 -1.4
485 -30.23 -1.41
495 -29.73 -2
505 -28.97 -2.53
515 -28.2 -2.76
525 -27.75 -2.58
535 -27.66 -2.33
545 -27.82 -2.38
555 -27.9 -2.69
565 -27.61 -3.14
575 -26.96 -3.28

81
 STUDENT ID: UP698209

585 -26.04 -3.16

595 -25 -2.93
605 -24.18 -3.11
615 -23.7 -3.31
625 -23.38 -3.44
635 -22.95 -3.15
645 -22.29 -2.86
655 -21.31 -2.6
665 -20.33 -2.68
675 -19.45 -2.7
685 -18.76 -2.53
695 -18.08 -2.07
705 -17.13 -1.63
715 -15.77 -1.62
725 -14.23 -1.93
735 -12.41 -2.11
745 -10.93 -1.83
755 -9.9 -1.81
765 -8.99 -7.53
775 -7.8 -4.5
780 -7.06 -5.83
783 -6.45 -6.8
785 -6.32 -7.45
795 -4.88 -11.28
805 -3.96 -8.78
815 -3.63 -6
825 -3.49 -4.6
835 -3.46 -4.44
840 -3.45 -4.6
843 -3.33 -4.98
845 -3.49 -5.11
855 -3.57 -6.13
865 -3.67 -7.07
875 -3.78 -8.34
885 -3.9 -10.03
895 -4.28 -9.3
900 -4.68 -8.05
905 -5.08 -7.08
915 -5.9 -5.88
925 -6.49 -5.3
935 -7.03 -4.79
945 -7.86 -4.1
955 -9.01 -3.46
965 -9.9 -3.33
975 -10.17 -3.46

82
 STUDENT ID: UP698209

985 -9.72 -3.88

995 -8.93 -5.14
1005 -8.97 -8.77
1015 -11.28 -5.03
1025 -16.02 -3.24
1045 -26.42 -2.42
1055 -29.29 -1.85
1065 -32.73 -1.57
1075 -36.08 -1.84
1085 -39.3 -2.37
1095 -42.48 -2.55
1105 -45.49 -2.45
1115 -48.29 -2.09
1125 -49.48 -1.97
1135 -50.48 -2.52

The results in Table 5.1 show that the centre frequency of the filter is at 843 MHz which
shows very good performance. The bandwidth ranges from 785 MHz to 925 MHz
(bandwidth of 140 MHz) the obtained bandwidth is 20 MHz wider than the design
specification. Fig. 5.8 shows the plot of the actual return loss of the filter.

Figure 5.8: Return Loss of Hairpin Filter

83
 STUDENT ID: UP698209

Figure 5.9: Insertion Loss of Hairpin Bandpass Filter

The graph showed in Fig. 5.8 and Fig. 5.9 shows that the manufactured filter shows good
bandpass properties when compared with the simulated response, even showing better
rejection rate than the simulated filter. The filter however showed a higher return loss
within the passband than expected. This is most likely due to a slight impedance
mismatch caused by the soldering of the ports.
This test method can be applied in testing the performance of other passive circuits such
as splitters and couplers.

84
 STUDENT ID: UP698209

CHAPTER 6

RECOMMENDATIONS
The manufactured filter showed a slightly wider bandwidth than what was simulated.
Based on this, in order to achieve response that is as close to the design specification as
possible, the designer should target a slightly wider bandwidth on the simulation than
the design specification in order to compensate for the expected increase in bandwidth
in the manufactured filter. This discrepancy is likely due to errors that may occur in the
manufacturing process, as well as the losses in the substrate which cannot be
adequately modelled mathematically. However, this behaviour can be better
understood by producing and testing multiple batches. An efficient and more improved
way of designing microstrip filters is by generating sample schematics using CAD tools
such as the iFilter by AWR Corporation, and then adjusting the resonator spacing and
length until the desired response is achieved.

The narrow band property shown by the splitter and coupler can be improved by using
Composite Right Left Handed (CRLH) and Dual Composite Right Left Handed (D-CRLH)
(Mocanu, 2014) transmission lines and adding additional λ/4 sections before the junction
of the splitter. The size of the Wilkinson splitter was reduced by converting the straight
λ/4 lines to curves. The overall circuit performance can also be improved by utilising
substrates that are more stable and have low loss such as the Rogers R435OB substrate.

The designer can greatly improve and quicken the manufacturing process by studying
the design file parameters set by various PCB producers, and ensuring that the design
meets them. It would also be helpful for the designer to install a Gerber file viewer in
order to facilitate adequate checking of the Gerber files before they are sent to the PCB
producer. The cost of production can be reduced by minimizing the total area of
substrate required for the circuit. Circuit costs can also be reduced by producing as many
circuits as possible on the same wafer and preventing customization by adhering to the
default parameters set by the producer.

85
 STUDENT ID: UP698209

CONCLUSION

This project was focused on the design of RF passive circuits which can function in the
4G-LTE 800 MHz lower spectrum. RF passive circuits in this category are the filter,
splitter and rat race coupler which were implemented as microstrip circuits on the 1mm
FR4 substrate. Various filter configurations such as the edge coupled, hairpin, shunt
stub, combline and interdigital were all designed and simulated using the Microwave
Office software from AWR Corporation. The hairpin filter was built and tested, and the
test results showed close correlation with the design target. The measured bandwidth of
the hairpin filter was found to be 20 MHz more than the design target, and 56 MHz more
than the simulated response.

A Wilkinson power splitter in conventional straight layout and circular layout, single and
double stage branch line coupler as well as a rat race coupler were also designed to
industry standards and were made ready for production but were not built due to
financial cap on the project.

However, sufficient knowledge was gained on the design and production of RF passive
circuits on microstrip substrates.

86
 STUDENT ID: UP698209

REFERENCES

Ahmad, N; Ali, M; Naseer, N; Amjad, N. (2014). Diplexer design for 4G Mobile

Communication using Lumped Component Filters. Journal of Basic and Applied Scientific
Research. 4 (2), p. 1-15.

Annapurna, D; Sisir, D (2009). Microwave Engineering. 2nd ed. New Delhi: Tata McGraw
Hill Education Private Limited. p. 214.

Bahl, J (2003). Lumped Elements for RF and Microwave Circuits. Massachusetts: Artech
House. p. 371.

Besser, L; Gilmore, R. (2003). Practical RF Circuit Design for Modern Wireless Systems.
Massachusetts: Artech House. p. 2.

arneiro ilho, R; ra o, . H; inani,M. ; d ssun o, . ; Mendon a, L. M. (2009).

Experimental and simulation analysis of microstrip patch antennas on BiNbO4 ceramic
substrates. IEEE AP-S Int Symp, Charleston, SC. p. 1-4.

Choinsk, T.C. (1984). Composite Coupler Design. IEEE Transactions on Microwave Theory
and Techniques. 32 (6), p. 613-614.

Eurocircuits. (2013). Data to be Uploaded with PCB order. Available:

http://www.eurocircuits.com/index.php/technology-guidelines/pcb-layout-data/114-
input-data-required. Last accessed 31st August 2014.

fcc.gov. (2013). http://www.fcc.gov/encyclopedia/specific-absorption-rate-sar-cellular-

telephones. Available: http://www.fcc.gov/encyclopedia/specific-absorption-rate-sar-
cellular-telephones. Last accessed 1st August 2014.

Foley, A. (2013). Wilkinson Power Divider Simulation. Available:

http://www.comsol.com/blogs/wilkinson-power-divider-simulation/. Last accessed 15
August 2014.

87
 STUDENT ID: UP698209

Geier, J. (2013). How to: Define Minimum SNR Values for Signal Coverage. Available:
http://www.wireless-nets.com/resources/tutorials/define_SNR_values.html. Last
accessed 1st August 2014.

Hong, J. S; Lancaster, M. J (2001).Microstrip Filters for RF/Microwave Applications, John

Wiley & Sons, Inc., p. 84.

International Telecommunication Union. (2014). ICT Data and Statistics. ICT Facts and
Figures. p. 3.

Jefferies, D. (2002). Microstrip. Available:

http://personal.ee.surrey.ac.uk/Personal/D.Jefferies/mstrip.html. Last accessed 17
August 2014.

Kai, C; Inder, B; Vijay, N. (2002). RF and Microwave Circuit and Component Design for
Wireless Systems. New York: John Wiley & Sons.

Khan, A. (2011). LTE Encyclopedia. Available:

https://sites.google.com/site/lteencyclopedia/home. Last accessed 13th June 2014.

Kikkert, C. J. (2009). RF Electronics (Computer Simulation). California: AWR Corporation.

p 2.

Kimuli, P. (2003). Introduction to GSM and GSM Mobile RF Transceiver Derivation.

Available: http://mobiledevdesign.com/learning-resources/introduction-gsm-and-gsm-
mobile-rf-transceiver-derivation. Last accessed 1st August 2014.

Li Y; Hu, L; Chen J. (2008). A Ku-Band Hairpin Filter Based on LTCC Technology.

Microwave Conference, 2008 China-Japan Joint, Shanghai. p 478-480.

Lourandakis, E. Weigel, R. Mextorf, H. Knoechel, R. (2012). Circuit Agility. IEEE

Microwave Magazine. 13 (1), p. 111- 121.

Lundstrom, J.M; Smith, T.L; Rinehart, L.F; Pate, R.C; Krogh, M.L; Huebner, W. (1999).
Measurement of the Dielectric Strength of Titanium Dioxide Ceramics. Pulsed Power
Conference, 1999. Digest of Technical Papers. 12th IEEE International. 2, p. 1489-1491.

88
 STUDENT ID: UP698209

Marki Microwave, Inc. (2012). Microwave Power Dividers and Couplers Primer2.
Available: http://www.markimicrowave.com/3337/Application_Notes.aspx. Last
accessed 20 August 2014.

Mocanu, A. I (2014). Dual band rat race coupler for 4G applications using CRLH and D-
CRLH transmission lines. Latest Trends on Communications. p 5-6.

Owens, P; Aitken, J; Edwards, T. (1976). Quasi-Static Characteristics of Microstrip on an

Anisotropic Sapphire Substrate. IEEE Transactions on Microwave Theory and Techniques.
MTT 24 (N/A), p 499-504.

Poole, I. (2013). Quartz crystal filter. Available: http://www.radio-

electronics.com/info/data/crystals/crystal_filter.php. Last accessed 5th August 2014.

Pozar, D. (2012). Microwave Engineering. 4th ed. Massachusetts: John Wiley & Sons.

Probir, K. (1995). Guglielmo Marconi - The Father of Long Distance Radio

Communication - n engineer’s tribute. IEEE Microwave Conference. 2, p1-2.

Ramesh G; Inder B; Maurizio B. (2013). Microstrip Lines and Slotlines. 3rd ed. London:
Artech House. p. 82.

Reyes, C; El-Ghazaly, M; Doml, S; Dydyk, M; Schroder K. (1994). Silicon as a Microwave

Substrate. IEEE MTT-S International. 3, p 1759 - 1762.

Rhodes, J; Hunter I. (1982). Electronically Tuneable Microwave Bandpass Filters. IEEE

Transactions on Microwave Theory and Techniques. 30 (9), p 1354-1360.

Rogerson, J. (2013). 4G and LTE: Everything you need to know. Available:

http://www.techradar.com/news/phone-and-communications/mobile-phones/4g-and-
lte-everything-you-need-to-know-926835. Last accessed 13th June 2014.

Salima S; Nasreddine, B; Bendimerad, F; Bendalah, N. (2012). Design of Parallel Coupled

Microstrip Bandpass Filter for FM Wireless Applications. 6th International Conference on

89
 STUDENT ID: UP698209

Sciences of Electronics, Technologies of Information and Telecommunications (SETIT). p

207-211.

Sanchez-Renedo, M; Gomez-Garcia, R; Alonso, J.I; Briso-Rodriguez, C. (2005). Tunable

Combline filter with continuous control of center frequency and bandwidth. IEEE
Transactions on Microwave Theory and Techniques. 53 (1), p 191-199.

Shou, X. (2009). Braodband Teraheartz Microstrip Waveguide. Department of Electrical

and Computer Engineering, University of Utah. p 4-6.

Thomas, L (2004). Planer Microwave Engineering. A practical Guide to Theory,

Measurements and Circuits. Cambridge: Cambridge University Press. p184.

Ting, S.; Tam, K.; Martins, R. (2004). Novel interdigital microstrip bandpass filter with
improved spurious response. Proceedings of the 2004 International Symposium on
Circuits and Systems. 1, p 984-987.

Toledo, N. (2011). Practical Techniques for Designing Microstrip Tapped Hairpin

Resonator Filters on FR4 Laminates. 2nd National ECE Conference, Manila, Philippines. p
2-3.

Torregrosa- enalva, ; L pe -Risueño, G; Alonso, J. (2002). A simple method to design

wide-band electronically tunable combline filters. IEEE Transactions on Microwave
Theory and Techniques. 50 (1), p 172-177.

Tu, P; Chao-Hsiung, T. (2010). Design of a compact rat-race coupler using dual

transmission lines. Microwave Conference Proceedings (APMC), 2010 Asia-Pacific. p 1224
- 1227.

Whites, K (2013). Resonant Stub Filters. Available:

whites.sdsmt.edu/classes/ee481/notes/481Lecture32.pdf. Last accessed 30th August
2014.

Ziang, X. (2006). Miniaturization of Microwave band-pass filter for research and fractal
structure of the new filter design. Shanghai: East China Normal University.

90
 STUDENT ID: UP698209

BIBLIOGRAPHY

Ahmad, N; Ali, M; Naseer, N; Amjad, N. (2014). Diplexer design for 4G Mobile

Communication using Lumped Component Filters. Journal of Basic and Applied Scientific
Research. 4 (2), p. 1-15.

Annapurna, D; Sisir, D (2009). Microwave Engineering. 2nd ed. New Delhi: Tata McGraw
Hill Education Private Limited. p. 214.

Bahl, J (2003). Lumped Elements for RF and Microwave Circuits. Massachusetts: Artech
House. p. 371.

Besser, L; Gilmore, R. (2003). Practical RF Circuit Design for Modern Wireless Systems.
Massachusetts: Artech House. p. 2.

arneiro ilho, R; ra o, . H; inani,M. ; d ssun o, . ; Mendon a, L. M. (2009).

Experimental and simulation analysis of microstrip patch antennas on BiNbO4 ceramic
substrates. IEEE AP-S Int Symp, Charleston, SC. p. 1-4.

Carr, J. (2002). RF Components and Circuits. Oxford: Newnes.

Choinsk, T.C. (1984). Composite Coupler Design. IEEE Transactions on Microwave Theory
and Techniques. 32 (6), p. 613-614.

Eurocircuits. (2013). Data to be Uploaded with PCB order. Available:

http://www.eurocircuits.com/index.php/technology-guidelines/pcb-layout-data/114-
input-data-required. Last accessed 31st August 2014.

fcc.gov. (2013). http://www.fcc.gov/encyclopedia/specific-absorption-rate-sar-cellular-

telephones. Available: http://www.fcc.gov/encyclopedia/specific-absorption-rate-sar-
cellular-telephones. Last accessed 1st August 2014.

Ferril, A. (2005). RF Systems, Components and Circuits Handbook. 2nd ed. London:
Artech House.

91
 STUDENT ID: UP698209

Foley, A. (2013). Wilkinson Power Divider Simulation. Available:

http://www.comsol.com/blogs/wilkinson-power-divider-simulation/. Last accessed 15
August 2014.

Frank, G (2012). RF and Microwave Engineering: Fundamentals of Wireless

Communications. Chichester: John Wiley & Sons Ltd.

Geier, J. (2013). How to: Define Minimum SNR Values for Signal Coverage. Available:
http://www.wireless-nets.com/resources/tutorials/define_SNR_values.html. Last
accessed 1st August 2014.

Harty, D. (2010). Novel Design of a Wideband Ribcage-Dipole Array and its Feeding
Network. Thesis: Submitted to Worchester Polytechnic Institute.

Hong, J. S; Lancaster, M. J (2001).Microstrip Filters for RF/Microwave Applications, John

Wiley & Sons, Inc., p. 84.

International Telecommunication Union. (2014). ICT Data and Statistics. ICT Facts and
Figures. p. 3.

Jagannathan, S (2007). Wireless Ad hoc and Sensor Networks: Protocols, Performance,

and Control. Florida: Taylor & Francis Group.

Jefferies, D. (2002). Microstrip. Available:

http://personal.ee.surrey.ac.uk/Personal/D.Jefferies/mstrip.html. Last accessed 17
August 2014.

Kai, C; Inder, B; Vijay, N. (2002). RF and Microwave Circuit and Component Design for
Wireless Systems. New York: John Wiley & Sons.

Khan, A. (2011). LTE Encyclopedia. Available:

https://sites.google.com/site/lteencyclopedia/home. Last accessed 13th June 2014.

Kikert, C. J. (2008). A Design Technique for Microstrip Filters. 2nd International

Conference on Signal Processing and Communication Systems, 2008. ICSPCS 2008.

92
 STUDENT ID: UP698209

Kikkert, C. J. (2009). RF Electronics (Computer Simulation). California: AWR Corporation.

p 2.

Kim, D.I; Jung, W.S; Yun, Y. (2004). A High Performance Transformer-Type Wilkinson
Power Splitter with Compensating Circuit for CATV Transmission System and Optimal
Design Method. IEEE Transactions on Consumer Electronics. 50 (3).

Kimuli, P. (2003). Introduction to GSM and GSM Mobile RF Transceiver Derivation.

Available: http://mobiledevdesign.com/learning-resources/introduction-gsm-and-gsm-
mobile-rf-transceiver-derivation. Last accessed 1st August 2014.

Li Y; Hu, L; Chen J. (2008). A Ku-Band Hairpin Filter Based on LTCC Technology.

Microwave Conference, 2008 China-Japan Joint, Shanghai. p 478-480.

Lopetegi, T; Laso, M; Hernánde , . (2001). New microstrip “Wiggly-Line” filters with

spurious passband suppression. IEEE Transactions on Microwave Theory and Techniques.
49 (9).

Lourandakis, E. Weigel, R. Mextorf, H. Knoechel, R. (2012). Circuit Agility. IEEE

Microwave Magazine. 13 (1), p. 111- 121.

Lundstrom, J.M; Smith, T.L; Rinehart, L.F; Pate, R.C; Krogh, M.L; Huebner, W. (1999).
Measurement of the Dielectric Strength of Titanium Dioxide Ceramics. Pulsed Power
Conference, 1999. Digest of Technical Papers. 12th IEEE International. 2, p. 1489-1491.

Marki Microwave, Inc. (2012). Microwave Power Dividers and Couplers Primer2.
Available: http://www.markimicrowave.com/3337/Application_Notes.aspx. Last
accessed 20 August 2014.

Mocanu, A. I (2014). Dual band rat race coupler for 4G applications using CRLH and D-
CRLH transmission lines. Latest Trends on Communications. p 5-6.

Ofcom. (2013). 4G Auction Results. Available: http://consumers.ofcom.org.uk/4g-

auction/. Last accessed 19th June 2014.

93
 STUDENT ID: UP698209

Owens, P; Aitken, J; Edwards, T. (1976). Quasi-Static Characteristics of Microstrip on an

Anisotropic Sapphire Substrate. IEEE Transactions on Microwave Theory and Techniques.
MTT 24 (N/A), p 499-504.

Poole, I. (2013). Quartz crystal filter. Available: http://www.radio-

electronics.com/info/data/crystals/crystal_filter.php. Last accessed 5th August 2014.

Pozar, D. (2012). Microwave Engineering. 4th ed. Massachusetts: John Wiley & Sons.

Probir, K. (1995). Guglielmo Marconi - The Father of Long Distance Radio

Communication - n engineer’s tribute. IEEE Microwave Conference. 2, p1-2.

Ramesh G; Inder B; Maurizio B. (2013). Microstrip Lines and Slotlines. 3rd ed. London:
Artech House. p. 82.

Reyes, C; El-Ghazaly, M; Doml, S; Dydyk, M; Schroder K. (1994). Silicon as a Microwave

Substrate. IEEE MTT-S International. 3, p 1759 - 1762.

Rhodes, J; Hunter I. (1982). Electronically Tuneable Microwave Bandpass Filters. IEEE

Transactions on Microwave Theory and Techniques. 30 (9), p 1354-1360.

Rogerson, J. (2013). 4G and LTE: Everything you need to know. Available:

http://www.techradar.com/news/phone-and-communications/mobile-phones/4g-and-
lte-everything-you-need-to-know-926835. Last accessed 13th June 2014.

Rohde, U. (2005). Receiver Measurements; How to Evaluate Receivers. Published in

Proceedings of USKA Meeting.

Salima S; Nasreddine, B; Bendimerad, F; Bendalah, N. (2012). Design of Parallel Coupled

Microstrip Bandpass Filter for FM Wireless Applications. 6th International Conference on
Sciences of Electronics, Technologies of Information and Telecommunications (SETIT). p
207-211.

94
 STUDENT ID: UP698209

Sanchez-Renedo, M; Gomez-Garcia, R; Alonso, J.I; Briso-Rodriguez, C. (2005). Tunable

Combline filter with continuous control of center frequency and bandwidth. IEEE
Transactions on Microwave Theory and Techniques. 53 (1), p 191-199.

Shou, X. (2009). Braodband Teraheartz Microstrip Waveguide. Department of Electrical

and Computer Engineering, University of Utah. p 4-6.

Techbridge Circuits. (2014). Readme File for Printed Circuit Board. Available:
http://www.tecbridgecircuits.co.uk/resources/readme.txt. Last accessed 23rd August
2014.

Thomas, L (2004). Planer Microwave Engineering. A practical Guide to Theory,

Measurements and Circuits. Cambridge: Cambridge University Press. p184.

Ting, S.; Tam, K.; Martins, R. (2004). Novel interdigital microstrip bandpass filter with
improved spurious response. Proceedings of the 2004 International Symposium on
Circuits and Systems. 1, p 984-987.

Toledo, N. (2011). Practical Techniques for Designing Microstrip Tapped Hairpin

Resonator Filters on FR4 Laminates. 2nd National ECE Conference, Manila, Philippines. p
2-3.

Torregrosa- enalva, ; L pe -Risueño, G; Alonso, J. (2002). A simple method to design

wide-band electronically tunable combline filters. IEEE Transactions on Microwave
Theory and Techniques. 50 (1), p 172-177.

Tu, P; Chao-Hsiung, T. (2010). Design of a compact rat-race coupler using dual

transmission lines. Microwave Conference Proceedings (APMC), 2010 Asia-Pacific. p 1224
- 1227.

Whites, K (2013). Resonant Stub Filters. Available:

whites.sdsmt.edu/classes/ee481/notes/481Lecture32.pdf. Last accessed 30th August
2014.

95
 STUDENT ID: UP698209

Wu, L.S; Mao, J; Yin, W.Y. (2012). Miniaturization of Rat-Race Coupler with Dual-Band
Arbitrary Power Divisions Based on Stepped-Impedance Double-Sided Parallel-Strip Line.
IEEE Transactions on Components, Packaging and Manufacturing Technology. 2 (12).

Xu, J; Yu, N; Shi, B; Zhao, Q. (2009). The miniaturization design of microstrip interdigital
bandpass filter. 2nd International Conference on Power Electronics and Intelligent
Transport Systems (PEITS).

Ziang, X. (2006). Miniaturization of Microwave band-pass filter for research and fractal
structure of the new filter design. Shanghai: East China Normal University.

96
 STUDENT ID: UP698209

APPENDIX
APPENDIX A: ROGERS SUBSTRATE DATASHEET

97
 STUDENT ID: UP698209

APPENDIX B: FR4 DATASHEET

APPENDIX C1: WILKINSON POWER DIVIDER DATASHEET

98
 STUDENT ID: UP698209

APPENDIX C2: WILKINSON POWER DIVIDER DATASHEET

99
 STUDENT ID: UP698209

APPENDIX D: 180° HYBRID DATASHEET

100
 STUDENT ID: UP698209

APPENDIX E: BRANCH LINE COUPLER DATASHEET

101
 STUDENT ID: UP698209

APPENDIX F: FILTER TEST RESULT

102
 STUDENT ID: UP698209

103

View publication stats