DESIGN OF MULTIBEAM ANTENNA FOR WIRELESS LOCAL AREA
NETWORK APPLICATIONS
SITI ZURAIDAH IBRAHIM
UNIVERSITI TEKNOLOGI MALAYSIA
iii
ACKNOWLEDGEMENT
First of all, thanks to our creator, “Allah” for the continuous blessing and for
giving me the strength and chances in completing this project.
Special thanks to my project supervisor, Associate Prof. Dr. Mohamad Kamal
A. Rahim, for his guidance, support and helpful comments in doing this project.
My family deserves special mention for their constant support and for their
role of being the driving force towards the success of my project. My friends
deserve recognition for lending a helping hand when I need them. En. Thelaha and
En. Nazri, the closest colleagues at P18, deserves praises for their assistance in the
fabrication and testing of my project.
My sincere appreciation also goes to everyone whom I may not have
mentioned above who have helped directly or indirectly in the completion of my
project.
iv
ABSTRACT
The current trend in most access point in conventional wireless local area
network (WLAN) is to use omnidirectional antennas, which radiates and receives
power equally in all directions. This attribute however gives result of lower power
efficiency and decrease network performance due to co-channel interference that
arrived from undesired directions. One of the proposed solutions to overcome these
constraints is to use multibeam antenna on WLAN access points. Multibeam
antennas are antenna arrays that make use of beamforming networks to produce
multiple independent beams that directed to different directions. In this project,
multibeam antenna comprises of linear antenna array and beamforming network is
presented. It was designed at 2.4 GHz to suit the application of WLAN at 802.11b/g.
Butler Matrix 4 x 4 is chosen as a beamforming network which was designed to
provide four different progressive phase shifts, -45°, +135°, -135°, +45° that coupled
to antenna array. It is made up from four 90° hybrid coupler, two 0 dB crossover and
two -45° phase shifter. Each component is designed and simulated using Agilent
ADS software and fabricated on FR4 board. This network is then combined with a
linear antenna arrays with the aim to produce four independent beams at four
different directions. Three types of antenna array that having different kind of
radiation patterns have been implemented which are square patch antenna, 4 x 2
planar antenna array and dipole antenna. The obtained result shows that 4 beams are
generated by each design where square patch antenna array produce Half Power
Beamwidth, HPBW for each beams about 30° and manage to cover 120° of coverage
area, 4 x 2 antenna array has HPBW about 7° and cover 30° while dipole antenna
produce two kind of beams, broader and narrower beams. Finally, it can be
concluded that the objectives of this project are achieved.
v
ABSTRAK
Hala tuju kebanyakan titik akses pada Rangkaian Kawasan Tempatan
Wayarles (WLAN) masa kini masih menggunakan antena halaan-omni, di mana ia
menyebarkan dan menerima kuasa daripada semua arah. Keadaan ini bagaimanapun
telah menyebabkan kecekapan kuasa menjadi rendah, malah menyebabkan
penurunan prestasi rangkaian yang disebabkan oleh gangguan saluran utama
daripada arah yang tidak dikehendaki. Salah satu jalan penyelesaian kepada
kelemahan ini ialah dengan menggunakan antena pelbagai sinaran pada titik akses
WLAN. Antena pelbagai sinaran adalah tatasusun antena yang menggunakan
jaringan pembentuk alur untuk menghasilkan beberapa sinaran yang menghala pada
arah yang berlainan. Dalam projek ini, antena pelbagai sinaran yang diperbuat
daripada tatasusun antena dan jaringan pembentuk alur dibentangkan. Ia direka pada
frekuensi 2.4 GHz untuk applikasi WLAN pada 802.11b/g. 4 x 4 Butler Matrix
dipilih sebagai jaringan pembentuk alur dan direka untuk menghasilkan empat nilai
anjakan fasa progesif yang berbeza iaitu -45°, +135°, -135°, +45° yang digandingkan
dengan tatasusun antena. It diperbuat daripada empat komponen gandingan hibrid
90°, dua komponen garis silang 0 dB dan dua penganjak fasa -45°. Setiap komponen
direkabentuk dan disimulasi menggunakan perisian Agilent ADS and difabrikasi ke
atas papan FR4. Jaringan ini kemudiannya dicantumkan bersama antena bertatasusun
lurus dengan matlamat untuk menghasilkan empat sinaran yang berasingan. Tiga
jenis antena yang mempunyai corak radiasi berbeza telah digunakan iaitu antenna
tampalan segi empat sama, 4 x 2 tatasusun antena dan antena dua-kutub. Keputusan
yang diperolehi menunjukkan empat sinaran telah dihasilkan, yang mana antena
tampalan segi empat sama menghasilkan HPBW selebar 30° dan kawasan liputan
seluas 120°, 4 x 2 tatasusun antenna mempunyai HPBW selebar 7° dan kawasan
liputan seluas 30° dan antena dua-kutub menghasilkan dua jenis sinaran yang
berbeza, sempit dan lebar. Akhirnya, dapatlah disimpulkan bahawa objektif projek
telah dicapai.
vi
TABLE OF CONTENTS
CHAPTER
1
2
TITLE
PAGE
DECLARATION
ii
ACKNOWLEDGEMENT
iii
ABSTRACT
iv
ABSTRAK
v
TABLE OF CONTENTS
vi
LIST OF TABLES
ix
LIST OF FIGURES
xi
LIST OF ABBREVIATIONS
xv
LIST OF SYMBOLS
xvii
LIST OF APPENDICES
xix
INTRODUCTION
1
1.1
Background of the problem
1
1.2
Problem Statement
4
1.3
Objective
4
1.4
Scope of the Study
5
1.5
Project Contribution
5
1.6
Organization of the Thesis
6
LITERATURE REVIEW
8
2.1
8
Smart Antenna Technology
2.1.1 Motivation towards Smart Antenna
11
2.1.2 Smart Antenna Applications in WLAN
14
vii
2.1.2.1
Standard of WLAN
2.1.2.2
Co-channel interference on
WLAN
2.1.2.3
2.2
2.4
18
Benefits of beam switching in
WLAN
20
2.1.3 Four Beams Multibeam Antenna
23
Antenna Basic
26
2.2.1 Microstrip Antenna
26
2.2.2 Antenna properties
27
2.2.2.1 Radiation Pattern
28
2.2.2.2 Half Power Beamwidth
32
2.2.2.3 Polarization
32
2.2.2.4 Bandwidth
34
2.2.3 Antenna Array
2.3
15
34
2.2.3.1 Uniform Linear Antenna Array
35
2.2.3.2 Beamswitching
43
Beamforming Network
46
2.3.1 Blass Matrix
48
2.3.2 Butler Matrix
48
2.3.2.1 90° Hybrid Coupler
52
2.3.2.2 0 dB Crossover
54
2.3.2.3 Phase Shifter
56
Previous Work
56
2.4.1 Integration between conventional 4 x 4
Butler Matrix and Antenna Array
2.4.2 Development of 4 x 4 Butler Matrix
2.5
3
Chapter Summary
57
62
68
METHODOLOGY
69
3.1
Project Methodology
69
3.2
Design development and software simulation
70
3.2.1 Development of Antenna Array
71
3.2.1.1 (4 x 1) Square Patch Antenna
71
3.2.1.2 4 x (4 x 2) Antenna Array
75
viii
3.2.1.3 (4 x 1) Dipole Antenna
77
3.2.2 Development of Butler Matrix
3.2.2.1 The design of 90° Hybrid
78
3.2.2.2 The design of 0 dB Crossover
81
3.2.2.3 Phase Shifter
83
3.2.2.4 Construction of Butler Matrix
85
3.3
Prototype Fabrication
88
3.4
Measurement Setup
89
3.5
3.6
4
3.4.1 S-parameter
89
3.4.2 Radiation pattern
90
Comparison of the designed Butler Matrix with
other findings
91
Chapter Summary
95
EXPERIMENTAL RESULTS & DISCUSSION
96
4.1
Result of Return Loss
96
4.2
Result of Butler Matrix
99
4.3
Radiation Pattern
100
4.4
Result analysis
107
4.5
Comparison of the measured radiation pattern with
other findings
4.6
4.7
5
77
117
The comparison between commercially used
antenna with designed multibeam antenna
122
Chapter Summary
124
CONCLUSION & FUTURE WORK
125
5.1
Conclusion
125
5.2
Proposed Future Work
127
REFERENCES
Appendices A – E
128
133-152
ix
LIST OF TABLES
TABLE NO.
TITLE
PAGE
2.1
IEEE WLAN standards
16
2.2
The center frequency defined by 802.11b/g specifications
17
2.3
The operation of multibeam antenna
25
2.4
The effect of varying parameter N
39
2.5
The effect of varying parameter d
40
2.6
The effect of varying parameter β
41
2.7
An example of power divider result in ideal case
44
2.8
Progressive phase difference corresponds to each input port
of Butler
44
2.9
Numerical value for 2 x 2 Butler Matrix
50
2.10
Numerical value for 4 x 4 Butler Matrix
50
2.11
Numerical value for 8 x 8 Butler Matrix
51
2.12
S-parameter for ideal case 90°hybrid coupler
53
2.13
S-parameter for ideal case 0 dB crossover
55
2.14
Previous work on the integration between conventional
4 x 4 Butler Matrix and Antenna Array
58
2.15
Previous work on the development of 4 x 4 Butler Matrix
64
3.1
Specifications for the FR4 board
70
3.2
Simulated result analysis for 4x1 square patch antenna
74
3.3
Radiation pattern of 4 x1 square patch interpretation
75
3.4
Simulated result analysis for 4x2 antenna array
76
3.5
Width value for each impedance value in hybrid coupler
78
3.6
The numerical result of simulated hybrid coupler
80
x
3.7
The numerical result of simulated hybrid coupler
83
3.8
The numerical result of simulated 45° phase shifter
84
3.9
The numerical result of simulated 0° phase shifter
85
3.10
Design Specification of the Butler Matrix
86
3.11
The simulated output phase of Butler Matrix (schematic)
87
3.12
Computed phase error (schematic simulation)
87
3.13
The simulated output phase of Butler Matrix (momentum)
87
3.14
Computed phase error (momentum simulation)
87
3.15
The comparison between designed Butler Matrix and other
Findings
92
4.1
The numerical result of square patch antenna
98
4.2
The numerical result of 4 x 2 antenna array
98
4.3
The numerical result of dipole antenna
99
4.4
The measured output phase of Butler Matrix
100
4.5
Computed phase error (measurement)
100
4.6
Numerical result of measured radiation patterns of using
square patch
4.7
Numerical result of measured radiation patterns of using
4 x 2 antenna array
4.8
103
104
Numerical result of measured radiation patterns of using
dipole antenna
105
4.9
AF equations correspond to each β
108
4.10
The comparison between measured radiation patterns of the
design with other findings
4.11
118
The comparison between commercially used antenna with
designed multibeam antenna
122
xi
LIST OF FIGURES
FIGURE NO.
TITLE
PAGE
2.1
Radiation pattern of smart antenna
9
2.2
The functional block diagram of smart antenna
10
2.3
Radiation pattern of Omnidirectional Antenna (Top view)
12
2.4
Directional Antenna Coverage Pattern
12
2.5
Antenna diversity
13
2.6
The motivation towards smart antenna implementation
14
2.7
WLAN with two APs
15
2.8
3 non-overlap channels in 802.11b/g
18
2.9
Devices that cause interference to WLAN AP
18
2.10
APs with 3 non-overlap channel
19
2.11
WLAN with more than 3 APs
19
2.12
An example of multibeam antenna coverage on WLAN
20
2.13
Comparison of throughput between switched beam and traditional
AP
2.14
Simulation results of BER when utilizing switch-beam antenna
in AP
2.15
21
22
The plot of CIR (carrier to interference ratio) as a function of
the cellular frequency reuse factor, K, and the number of beams, m 23
2.16
Block diagram of 4 ports multibeam antenna
2.17
The generated radiation pattern by exciting current at one port at
24
instant
24
2.18
4 beams radiation pattern
25
2.19
Rectangular patch antenna
26
2.20
Various feeding technique
27
xii
2.21
Coordinate system for radiation pattern measurement
28
2.22
Principle and E/H pattern cuts
29
2.23
2D radiation pattern
30
2.24
Radiation pattern of omnidirectional antenna
31
2.25
E-plane linear polarized
33
2.26
Various antenna array configuration
35
2.27
An example of pattern multiplication theorem
36
2.28
N element along x axis
36
2.29
Uniform Linear Array Configuration
37
2.30
Plots of AF with d =
2.31
Plots of AF with, N = 4, β = 0 and d = 0.25λ, 0.5λ, λ, 1.25λ
40
2.32
Plots of AF with, N = 4, d = 0.25λ
41
2.33
Phase scanning block diagram
42
2.34
A switched line phase shifter
42
2.35
The operation of power divider in terms of S-parameter
43
2.36
4 x 4 Butler Matrix configuration
44
2.37
Progressive phase difference corresponds to each input port of
λ
2
, β = 0 and N = 4, 8 and 12
39
Butler Matrix (block diagram form)
45
2.38
Radiation pattern obtained
46
2.39
Flow chart of the type beamformer
47
2.40
Blass Matrix configuration
48
2.41
AF plot for 2 x 2 Butler Matrix (N = 2, β = ±90°)
49
2.42
Block Diagram of 4 x 4 Butler Matrix
50
2.43
AF plot for 4 x 4 Butler Matrix (N = 4, β = ±45°, ±135°)
50
2.44
Block Diagram of 8 x 8 Butler Matrix
51
2.45
AF plot for 8 x 8 Butler Matrix (N = 8, β = ±22.5°, ±67.5°,
±112.5°, ±157.5°,)
51
2.46
Geometry of 90° hybrid coupler
52
2.47
Geometry of 0 dB crossover
54
2.48
Illustration that represents the function of 0 dB crossover
55
3.1
The flow chart of the operational framework
69
3.2
The block diagram of the complete design configuration
70
3.3
The flow chart of the design development of the project
70
xiii
3.4
Square patch antenna configuration
72
3.5
Simulated Return Loss for 4 x 1 square patch antenna
73
3.6
Radiation pattern of 4 x 1 square patch antenna
74
3.7
Layout of 4 x 2 antenna array
75
3.8
Return Loss of 4 x 2 array patch
76
3.9
E-plane co-polarization radiation pattern of 4 x 2 array patch
76
3.10
The flow chart of the Butler Matrix implementation
78
3.11
Designed hybrid coupler
79
3.12
The simulated result of hybrid coupler
80
3.13
Designed 0 dB crossover
82
3.14
The simulated result of amplitude and phase of 0 dB crossover
83
3.15
Designed 45° phase shifter
84
3.16
Designed 0° phase shifter
85
3.17
The block structure and layout of the Butler Matrix.
86
3.18
Fabricated prototype
89
3.19
The configuration of the project
90
4.1
Measured return loss of square patch antenna correspond to
each port
97
4.2
Measured return loss of each 4 x 2 antenna array
98
4.3
Measured return loss of each dipole antenna
99
4.4
Measured radiation pattern of single antenna
101
4.5
Measured radiation patterns of using square patch
102
4.6
Measured radiation patterns of using 4 x 2 antenna
103
4.7
Measured radiation patterns of using dipole antenna
104
4.8
Overlapped radiation pattern
106
4.9
The computed radiation pattern of AF corresponds to each β
108
4.10
Computed radiation pattern of AF
109
4.11
Conversion of array pattern from linear unit to dB
110
4.12
Pattern multiplication of square patch antenna case
111
4.13
Pattern multiplication of 4 x 2 antenna array case
112
4.14
Pattern multiplication of dipole antenna case
113
4.15
Radiation pattern comparison between computed and measured
result (square patch antenna case)
4.16
Radiation pattern comparison between computed and measured
114
xiv
result (4 x 2 antenna array case)
4.17
115
Radiation pattern comparison between computed and measured
result (dipole antenna case)
116
xv
LIST OF ABBREVIATIONS
2D
-
Two dimensional
3D
-
Three dimensional
3G
-
Third Generation
AF
-
Array Factor
AP
-
Access Point
BER
-
Bit Error Rate
BPSK
-
Binary Phase Shift Keying
CCK
-
Complementary Code Keying
CIR
-
Carrier to Interference Ratio
CPW
-
Co-planar waveguide
DBPSK
-
Differential Binary Phase Shift Keying
DQPSK
-
Differential Quadrature Phase Shift Keying
DSSS
-
Direct Sequence Spread Spectrum
FCC
-
Federal Communications Commission
FR4
-
Fire Retardant Type 4
FHSS
-
Frequency Hoping Spread Spectrum
GFSK
-
Gaussian Frequency Shift Keying
HPBW
-
Half-power beamwidth
IEEE
-
Institution of Electrical and Electronic Engineer
IF
-
Intermediate Frequency
ISM
-
Industrial, Scientific, Medical
LAN
-
Local Area Network
LOS
-
Line of Sight
NLOS
-
Non line of sight
OFDM
-
Orthogonal Frequency Division Multiplexing
xvi
QPSK
-
Quadrature Phase Shift Keying
QAM
-
Quadrature Amplitude Modulation
RF
-
Radio Frequency
SDMA
-
Spatial Division Multiple Access
SINR
-
Signal to Interference and Noise Ratio
SIR
-
Signal to Interference Ratio
SLL
-
Side lobe level
SNR
-
Signal to Noise Ratio
UV
-
Ultra Violet
VoWi-Fi
-
Voice over Wide Fidelity
WLAN
-
Wireless Local Area Network
xvii
LIST OF SYMBOLS
dB
-
decibel
1R
-
First beam on the right side of polar plot
1L
-
First beam on the left side of polar plot
2R
-
Second beam on the right side of polar plot
2L
-
Second beam on the left side of polar plot
W
-
Width of rectangular patch antenna
L
-
Length of rectangular patch antenna
εr
-
Dielectric constant
h
-
Substrate height
λg
-
Guided wavelength
(r,θ,φ) -
Spherical coordinate system
E
-
Electric
H
-
Magnetic
P(θ)n -
Normalized radiated power pattern
P(θ)
θ component of the radiated power as a function of angles θ
-
P(θ)max -
The radiated power maximum value
Eθ
-
E field existing θ direction
Eφ
-
E field existing φ direction
fu
-
Upper cutoff frequency
fl
-
Lower cutoff frequency
N
-
Number of elements
d
-
distance between antenna elements
θ
-
phase
β
-
phase difference between antenna elements
k
-
wave number
xviii
λ0
-
wavelength in free space
l
-
transmission line length
Zo
-
characteristic impedance
w
-
transmission line width
εeff
-
effective dielectric constant
c
-
velocity of light in free space
fr
-
operating frequency
tan δ
-
dissipation factor
Leff
-
Effective length
∆L
-
length extension
BW% -
bandwidth in percentage
xix
LIST OF APPENDICES
APPENDIX
TITLE
PAGE
A.
FR4 general technical specifications
133
B.
Simulation result of Butler Matrix
134
C.
H-Co measured radiation pattern for square patch
136
D.
E-Co measured radiation pattern for square patch
E.
antenna when multiple input activated simultaneously
137
Submitted papers for proceedings
139
CHAPTER 1
INTRODUCTION
This dissertation proposes the development of multibeam antenna that can be
implemented for WLAN application. In this first chapter, the background of the
project is discussed providing the problem statement, objective, scope of the study
and project contribution.
1.1
Project Background
In recent years, wireless networking has become a key solution to various
data communication needs. Wireless LANs are fast, flexible and cheap compared to
conventional wired LANs and they are still improving [1]. While in wireless
communications two most important restricting factors are interference and multipath
fading [2]. Multipath is a condition which arises when transmitted signal undergoes
reflection from various obstacles in the propagation environment which cause the
multiple signals arrive from different directions [3]. The result is degradation in
signal quality when they are combined at the receiver due to the phase mismatch. Cochannel interference is the interference between signals that operate at the same
frequency.
2
Smart antenna is one of the most promising technologies that will enable
higher capacity in wireless networks by effectively reducing multipath and cochannel interference [4]-[6]. This achieved by focusing the radiation only in the
desired direction and adjusting itself to change traffic conditions or signal
environment. The early smart antenna systems were designed for use in military
applications to suppress interfering or jamming signals from the enemy [3]. Since
interference suppression was a feature in this system, this technology was borrowed
to apply to personal wireless communications where interference was limiting the
number of users that a network could handle [3].
It has been studies [7] and tested [8], [9], that applying simple smart antenna
systems and algorithm to WLAN, would improve the performance worthily [1].
Taking into account that IEEE 802.11a WLAN the bit rate rises with an increase in
the Signal to Interference and Noise Ratio (SINR), developing a smart antenna
solution for WLAN application becomes more valuable [1]. Multiple beam antenna
array, a part of smart antenna system is known to be able to provide capacity
enhancement by means of interference reduction though spatial filtering [9]. It
provides a considerable increase in network capacity when compared to traditional
antenna systems or sector based systems [9].
The current trend in most access point in conventional WLAN is to use
omnidirectional antennas, which radiates and receives power equally in all directions
[10]. The implementation of this antenna is simple but it forms some limitations on
the performance of the network. As the direction is not specific, only small
percentage of the overall energy is reaching to the desired user which resulting the
lower power efficiency. It also suffers of co-channel interference as signals that
operate at the same frequency from undesired directions also capable to reach the
antenna. Due to the limited spectrum allocation in WLAN, co-channel interference
will become an issue which is the major problem of omnidirectional antenna
broadcast [11].
3
One of the proposed solutions to overcome these constraints is to use
multibeam antenna on WLAN access points, AP [12]. Multibeam antennas are
antenna array that make use of beamforming network to produce multiple
independent beams that directed to different directions. By offering independent
beams or channels, AP will switch between these channels to select the channel that
has the highest received power. This feature assist the antenna system to maximize
the power received in the desired directions.
The implementation of multibeam antenna is not new. In fact, it has been
implemented in Cellular Radio Systems in a few years back. It has been reported in
[6] that by having a multiple beam, the selected beam can reduce the interference,
increase the system carrier to interference level and then offer an opportunity for the
greater capacity by tighter frequency reuse. The increase in frequency reuse permits a
75% increase in the number of RF channels at the site and doubling the overall
number of the subscriber capacity [6].
As the implementation of multibeam antenna gives a tremendous result in
cellular communications, people interested to apply multiple beam antennas for
WLAN communications. So far, most studies had done on the simulation to observe
the performance in WLAN [1], [9], [12]-[15]. For example in [12], it has been
proved that the multibeam antenna is capable to reduce the value of Bit Error Rate
(BER). Compare to omni directional antenna application, the simulation results show
that utilizing switch-beam antenna in AP the BER performance improve about 2 dB
in light-of-sight (LOS) case, and 6 dB in non-light-of –sight (NLOS) case.
With the motivation gained from the simulation that has been done in
literature [1], [9], [12]-[15], this project aims to produce physical implementation of
multibeam antenna so that the actual performance of multibeam antenna in WLAN
could be observed in future.
4
1.2
Problem Statement
By applying omnidirectional antenna on the WLAN AP, it is suffering from a
number of disadvantages which can be summarized as follows:
I. Lower power efficiency
II. Capacity Limitation due to contribution of co-channel interference
These disadvantages can be overcome by using multibeam antenna on the
WLAN access point. This project will only focus on the development of the antenna
itself to meet the satisfied performance that can be used in WLAN system. Thus far,
a few studies has been done on the multibeam antenna, but most of them were
discussing more about beamforming network [16]-[23] which are in terms of
bandwidth and compactness while put less concentration on constructing the
multibeam antenna itself. The main task here is to design a multibeam antenna
system by integrating an antenna array with beamforming network so that the overall
performance of the multibeam antenna in terms of radiation pattern could be
observed.
1.3
Objective
The main objectives of this study can be divided into two goals;
I. To design an antenna system that capable to produce multiple beam of the
radiation pattern
II. To simulate and fabricate that antenna system design so that its performance
can be observed
This project is aim for WLAN application at 2.4GHz
5
1.4
Scope of Study
The first part of this study is to understand the concept of multibeam antenna.
The needed of multibeam antenna on WLAN, the function of the multibeam antenna
and implementation of the antenna system are studied.
In the second part of the study, the multibeam antenna is designed and
simulated. At this stage, the fundamental of antenna parameters, microstrip antenna
characteristics and the theory about multibeam antenna has been covered. After that,
the antenna array and Butler Matrix are designed and simulated.
The third part is the fabrication and measurement of the design. At this stage,
related equipment such as UV Light Equipment, Network Analyzer, Spectrum
Analyzer, Signal Generator, etc. are expected to be familiarized and well handled.
The last part of the study is the analysis part. It is expected that during the
study, the measured result and the theoretical should be compared and observed.
1.5
Project Contribution
The application of smart antenna is not limited to the WLAN network only,
but also can be implemented in most communication network. As provided in
Chapter 2, Section 2.4, most studies have been done on the improvement of
beamforming network [16]-[23] rather than constructing the multibeam antenna
itself. This dissertation will give a basic idea about the integration between antenna
array and beamforming network and the performance of multibeam antenna using 3
6
different types of antenna (directional antenna, omnidirectional antenna, and broader
beamwidth antenna) are observed. Previous works only show the result of using
omnidirectional antenna [24] and broader beamwidth antenna [25]-[27] which
constructed and discussed independently. The multibeam antenna produced by this
project could be integrated later with RF switch and controller part that constructed
by other parties so that a complete switched beam antenna system could be
constructed. Multibeam antenna also can be implemented in the case of Spatial
Division Multiple Access (SDMA) by injecting different signal to each input ports
[28], [29].
1.6
Organization of Thesis
The thesis is divided into five chapters. The first chapter is Introduction,
which provides information regarding the project background, objectives, scope of
project, project contribution and the layout of the thesis.
The second chapter is Literature Review. In this chapter, the concept of
smart antennas, antenna theory, beamforming network and related previous works
are thoroughly explained.
The third chapter is Methodology, in which the methods employed in this
project will be explained. The design procedures and simulation results for this
project will be presented in detail. The simulation results and subsequent analysis
will be discussed. Prototype fabrication and measurement setup are also presented.
7
Results and analysis of the fabrication and measurement are presented in
Chapter 4. The comparison between simulations, fabrication results, measurement
results, and computation result will be explained in this chapter.
The last chapter is Conclusion and Future Work. This chapter will conclude
the findings of the project and provide recommendations for future work.
CHAPTER 2
LITERATURE REVIEW
This chapter explains about smart antennas technology and the motivation
towards smart antenna. This chapter also presents the basic concept of the antenna
properties, antenna array concept and the operation of beamforming network. The
principal ideas are established before the design of the system is done.
Further
more, a previous work will also be introduced.
2.1
Smart Antenna Technology
Smart antenna systems are an active research topic nowadays due to the
improvements and advantages over omni-directional systems. A smart antenna
system is defined by the IEEE as, an antenna system that has circuit elements
associated with its radiating elements such that one or more of the antenna properties
are controlled by the received signal [30]. There are basically two types of smart
antennas: switched beam systems and adaptive array systems.
9
Switched beam antenna system has a fixed number of beams, which one or
more beams can be selected from the array for transmission or reception. The main
motivation of the switched beam antenna is to increase the antenna gain. For
example, a four-switched beam antenna system that used in 120° sector antenna, the
resultant increased gain can be calculated using the formula as follows [5];
Gain = 10 log (M)
(2.1)
where M is the number of beams per sectors. Thus for a sector containing 4 beams
(M=4), the gain increase is 6dB over the original sector antenna.
Adaptive Array systems provide more intelligent operation where it has
ability to adapt in real time radiation pattern to the RF environment. The comparison
can be seen in Figure 2.1 where switched beam antenna provides many narrow
predefined beams and activate one or more beam at instant while adaptive antenna
array provides a beam according to the targeted and interfering user locations.
(a)
(b)
Figure 2.1: Radiation pattern a) Switched beam antenna, b) Adaptive antenna array
The topic of switched beam antenna as a smart antenna has been discussed
vigorously as the implementation of it is simple and requires less cost as compared to
adaptive antenna array. Unlike the adaptive antenna, switched beam antenna only
constructed by an antenna array, a beamforming network, RF switch and simple
10
controller part [3], [24] while adaptive array systems provide more intelligent
operation and needs more advanced signals processing to function. The comparison
of the block diagram between switched beam and adaptive beam can be seen in
Figure 2.2. It can be seen that adaptive array system is more complex compared to
switched beam system. For this project, discussion will be carried on to the switched
beam antenna only.
(a)
(b)
Figure 2.2: The functional block diagram, a) Switched beam antenna, b) Adaptive
antenna array [3]
By referring to Figure 2.2 (a), a switched beam antenna system can be
realized by breaking the whole system down into four major building blocks for ease
of analysis. Figure 2.2 (a) shows how the system can be broken down into a
beamforming network to form independent beams, an RF switch for switching
between input ports, a power detector to monitor signal strength and a control logic
11
running an algorithm that controls the whole system. The basic operation of the
switched beam can be explained as follows.
First of all, let consider switched beam antenna at a receiver side. The input
of the RF switch is connected to the beamforming network and its outputs connected
to the power detector. At instant, only one switch will be turned on while others will
be turned off. The power detector will measure the signal strengths for that incoming
beam, which then will be connected to control logic. The main function of this
control logic is to samples all the power and then makes comparison between them.
After that, logic control will feedback to RF switch and select the beam which
received the highest signal to noise ratio, SNR by sending certain amount of voltage
to RF switch. Since the power detector operated in analog form while
microcontroller in digital form, analog-to-digital converter may needed for
interfacing part. However, this is just a general idea about how a switched beam
antenna can be implemented as a system but the complete implementation of this
switched beam system of course, is not as simple as what was mentioned above. For
more details technical information on the complete implementation, publication in
[10] could be referred.
2.1.1 The Motivation towards Smart Antenna
Since early days of wireless communication, omnidirectional antenna has
been used in mobile and networks to radiate and receives equally in all directions
(Figure 2.3). At that time, to design a very good signal reception is more focusing on
the antenna radiating element itself; there is no system for it. This attributes however
gives lower power efficiency as most of the radiated power is directed towards the
undesired user location.
12
Figure 2.3: Radiation pattern of Omnidirectional Antenna (Top view)
Later, directional antenna has been introduced to improve the power
efficiency of the antenna [4], [28]. The antenna is constructed to have certain fixed
transmission and reception directions. As corresponding to that, cellular area for
mobile network is subdivided into sectors where each of it can be covered by the
particular directional antenna. By doing this, the antenna provide increased gain as
the power only concentrated in particular direction as compared to omnidirectional
antenna. However, this sectorized antenna still does not overcome the co-channel
interference which the major problems of omnidirectional antenna broadcast [4].
Figure 2.4: Directional Antenna Coverage Pattern
In the next step toward smart antenna, diversity antenna has been introduced.
In this form, two or more radiating antenna element is integrated at the base station,
where each of it separated at a certain space, a few wavelengths apart at different
locations. This is to ensure that each antenna will cover different angle of coverage
area. This condition historically improves the signal reception by counteracting the
negative effects of multipath [28]. There are two types of diversity which are [28]:
13
i.
Switched diversity:
This system comprises of several radiating antenna element but only one which
gives the highest output will be connected to the channel at instant. The selection
may be based on the power of the desired signal, the total power or the signal to
interference ratio, SIR obtained by each antenna [28].
ii.
Diversity combining:
This systems comprises of multiple antennas with the receive signals weighted
and combined to produce an output signal. The maximal ratio combiner applies
weight in proportion to the SNR and combines the weighted signals in phase so
that the ratio of combined received SNR can be maximized [28]. Unlike the
antenna array, the signals is processed in IF or baseband level, thus it will not
affect the radiations pattern characteristics of each antenna, only baseband signal
is improved. Antenna array on the other hand, combine the signal in RF level
which cause the radiation pattern of the antennas change [28].
(a)
(b)
Figure 2.5: Antenna diversity (a) Switched diversity, (b) Diversity combining
14
However, both diversity systems still attain less performance in terms of gain,
capacity and error due to major co-channel interference and multipath problems [28].
This situation boosts an idea among researchers to produce antenna with intelligent
systems which is known nowadays as Smart Antenna Technology. The motivation
towards smart antenna can be summarized as shows in Figure 2.6.
Omnidirectional Antenna
Directional Antenna and Sectorized Systems
Antenna Diversity systems
Smart Antenna
Figure 2.6: The motivation towards smart antenna implementation
2.1.2 Smart Antenna Applications in WLAN
Early smart antenna was designed for governmental use in military
applications, which used directed beams to hide transmission from an enemy. Since
the narrow beams of early smart antenna created less overall interference, researchers
began to explore the possibility of applying smart antenna in cellular communication
network. As the demand of higher data rate increases, particularly in 3G applications,
number of simultaneous users that the network could handle becomes limited.
Although the frequency reuse could increase the capacity of the network, this idea
still restricted due to co-channel interference that arises when reducing the cell size.
By having narrow beams that directed to a certain angle, the co-channel interference
can be reduced, thus increasing the total number of users that the network could
handle.
15
It has been discussed from literature that the smart antenna could improve the
co- channel inference, not only to the cellular communication network, but also to
WLAN as well. As described previously, smart antenna systems may be broadly
categorized into two types: switched beam antenna and adaptive array. As stated in
Chapter 1, there are two major problems in wireless communications that can be
reduced by smart antenna which are multipath and co-channel interference. As
adaptive antenna array involves intelligent processing part, only adaptive antenna
array has a capability to reduce both limitations; multipath effect and co-channel
interference, while switched beam antenna only capable to reduce co-channel
interference. This dissertation is focuses on the application of multibeam antenna as
part of switched beam antenna on WLAN by narrowing the scope to the
improvement of co-channel interference only.
2.1.2.1 Standard of WLAN
A WLAN is a type of Local Area Network (LAN) that uses high frequency
radio waves than wires to communicate and transmit data. A WLAN basically
consists of one or more wireless devices connected through APs, which is then
connected to the backbone network providing wireless connectivity to the covered
area [49]. Figure 2.7 shows a typical layout of a WLAN with two AP.
Figure 2.7: WLAN with two APs [49]
16
The WLAN technology is defined by the IEEE 802.11 family of
specifications. There are currently four specifications in the family that has been
officially released: 802.11, 802.11a, 802.11b, and 802.11g. Table 2.1 below
summarizes the IEEE WLAN standards.
Table 2.1: IEEE WLAN standards [49]
802.11
802.11a
802.11b
802.11g
Standard
approved
1997
1999
1999
2003
Available
bandwidth
83.5 MHz
300 MHz
83.5 MHz
83.5 MHz
Unlicensed
frequencies
of operation
2.4000 – 2.4835
GHz DSSS, FHSS
5.15 – 5.35 GHz OFDM
2.4000 –
2.4835 GHz
DSSS
2.4000 – 2.4835
GHz DSSS,
OFDM
3 (Indoor /
Outdoor)
3 (Indoor /
Outdoor)
11, 5.5, 2, 1
Mbps
54, 36, 33, 24,
22, 12, 11, 9, 6,
5.5, 2, 1 Mbps
5.725 – 5.825 GHz
OFDM
4 Indoor (UNII1)
Number of
nonoverlapping
channels
3 (Indoor / Outdoor)
Data rate per
channel
2, 1 Mbps
4 Indoor / Outdoor
(UNII2)
4 Outdoor (UNII3)
54, 48, 36, 24, 18, 12, 9,
6 Mbps
OFDM/CCK (6,
9, 12, 18, 24, 36,
48, 54)
DQPSK (2Mbps
DSSS)
Modulation
Type
DBPSK (1 Mbps
DSSS)
4GFSK (2 Mbps
FHSS)
BPSK ( 6, 9 Mbps )
QPSK (12, 18 Mbps)
16-QAM (24, 36 Mbps)
64-QAM (48, 54 Mbps)
2GFSK (1 Mbps
FHSS)
DQPSK/CCK
(11, 5.5
Mbps)
OFDM (6, 9, 12,
18, 24, 36, 48,
54)
DQPSK (2
Mbps)
DQPSK/CCK
(22, 33, 11, 5.5
Mbps)
DBPSK (1
Mbps)
DQPSK (2
Mbps)
DBPSK (1
Mbps)
Compatibility
802.11
Wi-Fi5
Wi-Fi
Wi-Fi at 11
Mbps and below
In Malaysia, only the spectrum of 2.4 GHz is allowed while the frequency
spectrum at 5.2 GHz is not allowed because it is reserved for aeronautical navigation
17
and fixed satellite communications [44, 51]. The 802.11b/g standard defines a total
of 14 frequency channels. The FCC (Federal Communications Commission) allows
channel 1 through 11 within the U.S, whereas most of Europe (ETSI) can use
channels 1 through 13. In Japan, only one channel is available; channel 14. Table 2.2
shows the center frequency defined by 802.11b/g specifications.
Table 2.2: The center frequency defined by 802.11b/g specifications [50]
There is only 5 MHz separation between the center frequencies, and an
802.11b/g signal occupies approximately 20 MHz of the frequency spectrum [50].
The signal falls within about 10 MHz of each side of the center frequency. 5 MHz is
needed as a frequency separation between channels to avoid adjacent channel
interference. As a result, an 802.11b/g signals overlaps with several channel
frequencies. This leaves only three channels (Channel 1, 6 and 11) that can be used
without causing interference between access points.
18
Figure 2.8: 3 non-overlap channels in 802.11b/g
2.1.2.2 Co-channel interference on WLAN
A major challenge when deploying WLAN is the channel allocation problem.
The 802.11b/g WLAN operates in the 2.4 GHz unlicensed Industrial, Scientific and
Medical (ISM) band. This introduces interferences from other electronic devices,
such as microwave ovens, cordless phone, Bluetooth enabled devices and garage
door openers [49]. Poor WLAN performance is expected to happen if any of these
devices is used within the same room as WLAN. The interference could be avoided
by switching off any of the devices when accessing WLAN or place them far enough
from AP.
Figure 2.9: Devices that cause interference to WLAN AP
19
The problem of a WLAN is that there could be interference not only caused
by other equipment that using the same frequency band, but also caused by
neighbouring WLAN or the other APs that located in the owner’s network. If the
owner only has a single AP and the interference is caused by a different company
next door, the solution is simpler because the owner can simply assign his AP to
operate on the other channels as 802.11b/g has 3 non-overlap channels.
CH 6
Neighbour
CH 11
CH 1
Neighbour
Owner
Figure 2.10: APs with 3 non-overlap channel
The problem may arises when the owner places his AP more than three on his
network As the number of simultaneous users increases, users may experience a
lower data rate or applications timeout if they being served only by single AP. To
solve these issues, conventional solution is to install more access points at a closer
spacing. As the 2.4 GHz space only supports three non-overlap channels, meaning
that any installation that exceeds three APs will see some co-channel interference.
Co-channel interference may cause serious degradation to the desired signal which is
not suitable for network that targeted a high bandwidth (upload media) or real time
applications (VoWi-Fi) [49].
CH 1
CH 6
CH 11
CH 11
CH 1
Figure 2.11: WLAN with more than 3 APs
20
Since the application of switched beam antenna on cellular communication
system has capability to suppress interference, researcher began to investigate its
implementation on the application of WLAN. By having narrow and directional
beam to desired user and null to interfering user, co- channel interference
theoretically could be reduced. The active beam may selected dependent on SNR,
CIR (carrier to interference ratio), BER or power level of received signal that
calculated by processing part. This project is focused on how to provide an antenna
that capable to produce multiple beams.
Figure 2.12: An example of multibeam antenna coverage on WLAN
2.1.2.3 Benefits of beam switching in WLAN
As mentioned in Chapter 1, traditional WLAN AP only use single
omnidirectional antenna to provide omnidirectional coverage while mutlibeam
antenna make use of antenna array and beamforming network to provide multiple
beam coverage. Although the complete implementation of multibeam antenna on
WLAN still on study by many researchers all over the world, because of its special
features (multiple beams), it could be foreseen that it has several advantages over the
traditional AP’s antenna which can be listed as follows:
21
1.
Higher performance
Beam switching APs provide higher performance where it has consistent RF
coverage. RF coverage more evenly distributed, enabling very consistent
performance throughout an entire building [49]. In the other words, the average
fluctuation in throughput is much smaller with beam switching than a traditional AP,
which makes performance much more predictable [49]. Figure 2.13 shows the
comparison of throughput result between traditional AP and beam switching.
Figure 2.13: Comparison of throughput between switched beam and traditional AP
[49]
2.
Lower BER
According to [12], it has been shown through simulation that the beam switching
antenna also be able to reduce the value of Bit Error Rate (BER). Compare to omni
directional antenna application, the simulation results show that utilizing switchbeam antenna in AP the BER performance improve about 2 dB in light-of-sight
(LOS) case, and 6 dB in non-light-of –sight (NLOS) case (Figure 2.14).
3.
Improve security
By providing directional beam, beam switching antenna can also be used to direct RF
signals directly to a client rather than radiating these signals to every corner of a
wireless network. By focusing the RF energy in a specific direction, it minimizes the
ability to “eavesdrop” on the wireless conversation [49].
22
(a)
(b)
Figure 2.14: Simulation results of BER when utilizing switch-beam antenna in AP
(a) light-of-sight (LOS) case, (b) non-light-of –sight (NLOS) [12]
4.
Improve CIR
Figure 2.15 illustrates the plot of CIR (carrier to interference ratio) as a function of
the cellular frequency reuse factor, K, and the number of beams, m. This graph is
plotted based on cellular network system but it can give a brief idea about the
relation between CIR and number of beams. It can be said that the maximum
frequency reuse for 802.11b/g is 3 (K = 3) as it has 3 non-overlap channels. For
traditional AP, it has one beam (m = 1) and for switched beam antenna, in general it
has 12 beams (m = 12). As is shown, for the reuse factor, K = 3, for m = 1 the CIR is
about 10 dB which is very low, while for m = 12, the CIR is about 22 dB. Care must
be taken to manage the location of 3 AP if traditional AP is used, because 10 dB CIR
is considered too low and poor performance is expected to be happen on the network.
By having multiple beams, CIR of the network could be improved and the owner no
needs to worry much about the location of those APs.
As conclusion, switched beam antenna system has the advantages over
traditional AP – higher performance; improve BER, CIR and tighter security. With
benefits like these, multibeam antenna on switched beam antenna system is sure to
play an important role as wireless network become significant nowadays.
23
Figure 2.15: The plot of CIR (carrier to interference ratio) as a function of the
cellular frequency reuse factor, K, and the number of beams, m [6]
2.1.3 Four Ports Multibeam Antenna
As shown in Figure 2.2 (a), the switched beam antenna consists of antenna
array, beamforming network, RF switch, power detector and controller unit. By
referring to the objective of this project, the main task here is to design an antenna
system that produces multiple beams on radiation pattern’s plot. This multibeam
pattern could be constructed by integrating antenna array and beamforming network
(Figure 2.16). To make the design implementation easier, the scope of the designs is
limited to be as follows:
i. Antenna array – consists of 4 number of antenna elements
ii. Beamforming Network – Having 4 number of inputs and 4 number of outputs
24
Figure 2.16: Block diagram of 4 ports multibeam antenna
This multibeam antenna system can be a receiver or transmitter device but for
the analysis purpose, the multibeam antenna will be considered at the transmitter
side. Figure 2.17 shows the block diagram about how does each beam generated and
the operation can be summarized as shown in Table 2.3.
(a)
(c)
(b)
(d)
Figure 2.17: The generated radiation pattern by exciting current at one port at instant
a) Port 1, b) Port 2, c) Port 3, d) Port 4
25
Table 2.3: The operation of multibeam antenna
Operation
Result
Excite current to Port 1
Beam 1R is generated
Excite current to Port 2
Beam 2L is generated
Excite current to Port 3
Beam 2R is generated
Excite current to Port 4
Beam 1L is generated
This will give result of 4 number of beams will be generated as shown in
Figure 2.18. The notation 1R, 2R, 1L and 2L shows the first beam on the right,
second beam on the right, first beam on the left and second beam on the left
respectively.
Figure 2.18: 4 beams radiation pattern
In general, four-beam multibeam antenna can be used in 120° sector where
each beam has approximate 30° half power beamwidth (HPBW). Adjacent beam
overlapped level is determined to be at 3 dB crossover to ensure that all beams can
serve 120° angular area [5]. In other cases, as proposed in [10] and [31], 6 beams are
generated to cover a 90° space [10], while in [31], only 3 beams are generated to
cover the same space. For this project, only 4 beams will be generated using 4 x 4
beamforming network that integrated with 4 elements antenna array.
26
2.2
Antenna Basic
Antenna is defined as a transducer between a guided wave (propagating in a
transmission line) and an electromagnetic wave (propagating in an unbounded
medium or free space) or vice versa [32]. Antenna is made in various shape and size
and therefore can be categorized into several groups [33]. As only microstrip antenna
design is involved in this project, only this topic will be discussed here. Also, there
are many properties that related to an antenna such as radiation pattern, polarization,
bandwidth, gain and so on, however, only the most relevant to the project will be
discussed here.
2.2.1 Microstrip Antenna
Microstrip patch antennas are widely used due to the fact that they are highly
efficient, structurally compact, and conformal [34]. One of the most common types
of microstrip antenna is the rectangular patch. To conserve space, square patch
antenna could be designed using rectangular patch antenna formula as the model
analysis for both of configuration is almost similar. Figure 2.19 shows a typical
rectangular patch antenna with width W and length L over a grounded dielectric
plane with dielectric constant εr. Ideally, the ground plane on the underside of the
substrate is of infinite extent [34]. Normally, the thickness of the dielectric substrate,
h, is designed to be ≤0.02 λg, where λg is the wavelength in the dielectric [34].
Figure 2.19: Rectangular patch antenna [34]
27
Patch antennas can be fed by many different ways. The four most popular are
the coaxial probe, microstrip line, aperture coupling and proximity coupling [33].
These are displayed in Figure 2.20. The probe feed method is simple but is not
attractive from the fabrication point of view. The microstrip line method has the
advantage that both patch and feed line can be printed together but gives a limited
bandwidth in practical design. Other feed methods, aperture-coupling feed and
proximity coupling involves two substrates. Aperture coupling gives good
polarization purity and no cross-cross polarized radiation in the principle planes
while proximity coupling has potential to give largest bandwidth [33].
(a)
(c)
(b)
(d)
Figure 2.20: Various feeding technique a) Coaxial probe, b) Microstrip line, c)
Proximity coupling, d) Aperture coupling [33]
2.2.2 Antenna Properties
In this project, there are few essential concepts to know when dealing with
the antenna which are radiation pattern, HPBW, polarization and bandwidth. It is
important to know the properties of radiation pattern and polarization as this
information is very useful when doing radiation pattern measurement.
28
2.2.2.1 Radiation pattern
The radiation pattern of an antenna is defined as a mathematical function or
graphical representation of the radiation properties of the antenna as a function of
space coordinates [35]. In simple words, it simply describes about how an antenna
focuses the energy in space, which represents the coverage area of an antenna itself.
The standard (r, θ, φ) spherical coordinate system is typically used to represent field
pattern as shown in Figure 2.21.
Figure 2.21: Coordinate system for radiation pattern measurement [35]
However, three-dimensional (3D) pattern typically provided by sophisticated
simulation software while in practical world, it is hard to measure the radiation
pattern in 3D form. Thus, to describe this 3D pattern with two planar patterns, it is
called as the principal plane cut patterns. This principal plane cut pattern can be
obtained by making two slices through the 3D pattern through the maximum value of
the pattern or by direct measurement [35].
29
A cut in the plane of the theta unit vector is the theta cut, while a pattern cut
in the phi vector direction is the phi cut, as shown in Figure 2.21 [35]. The principal
plane cuts are two dimensional, orthogonal antenna cuts taken in the E field and H
field planes of the power radiation patterns, as illustrated in Figure 2.22 [35]. It is
important to understand the concept of principle plane cuts as it will help people to
visualize the radiation pattern in three-dimensional form.
Figure 2.22: Principle and E/H pattern cuts [35]
The two-dimensional radiation pattern can be plotted in two ways: Cartesian
plot and polar plot (Figure 2.23). The most common used method is polar plot and
this type will be used throughout this report. It can be seen that it is made up of many
different parts pointing at different directions. These different parts of a radiation
pattern may be classified as main, side and back lobes.
30
¾ Main lobe: It is defined as the radiation lobe containing the direction of
major radiation. In this report, it is referred as main beam.
¾ Minor lobe/side lobe/grating lobe: Any lobes other than a main lobe. As
they usually represent radiation in undesired directions, they should always
be minimized.
¾ Back lobe: It is a radiation lobe that is located approximately 180° from the
main lobe. The front-to-back ratio is measure of the power of the main beam
to that of the back lobe.
(a)
(b)
Figure 2.23: 2D radiation pattern a) Polar plot, b) Cartesian plot
31
It is convenient to normalize the radiation pattern such that its maximum
value is 0 dB. Thus, to normalize the radiated power in dB, the equation 2.2 should
be applied.
Pn (θ )0 = P(θ ) − P(θ )max
(2.2)
where
Pn (θ )0 : The normalized radiated power pattern.
P(θ )
: The θ component of the radiated power as a function of angles θ.
P(θ )max : The radiated power maximum value.
Antenna also can be categorized based on how does its radiation pattern looks
like. For example, dipole antenna can be referred as an omnidirectional antenna,
while directional antenna normally forms by an antenna array. The omnidirectional
antenna radiates and receives equally well in all horizontal directions while
directional antennas focus energy in a particular direction. Figure 2.23 shows an
example of directional pattern while the radiation pattern of omnidirectional antenna
can be illustrated in Figure 2.24.
(a)
(b)
(c)
Figure 2.24: Radiation pattern of omnidirectional antenna a) 3D view, b) E-plane
(2D side view), c) H-plane (2D top view)
32
2.2.2.2 Half-Power Beamwidth
The half-power beamwidth (HPBW) can be defined as the angle between two
directions in the maximum lobe of the power radiation pattern where the directivity
is one half the peak directivity, or 3 dB lower [35]. With reference to Figure 2.23, the
HPBW can be measured at the main beam, by calculating the angle of the gain which
has the value of the maximum beam value minus 3 dB. The beamwidth of the main
lobe is an important parameter in antenna array as there is a trade-off between
beamwidth and side lobe level; i.e. as beamwidth decreases, the side level increases
and vice versa. The effect of this change will be explained briefly in antenna array
section.
2.2.2.3 Polarization
The polarization of an antenna refers to the electric field polarization
properties of the propagating wave received or transmitted by the antenna [35]. In
general, most antennas radiate either linear or circular polarization. As this project
only involves linear polarization, only this topic will be discussed here. A linear
polarized antenna radiates in one plane containing the direction of propagation. The
E-plane and H-plane are reference planes for linearly polarized antennas.
¾ E plane: For a linearly polarized antenna, this is the plane containing the
electric field vector and the direction of maximum radiation. The electric
field or E-plane determines the polarization or orientation of the radio wave.
An antenna is said to be vertically polarized when its E-plane is perpendicular
to the earth surface while horizontally polarized antennas have their E plane
parallel to the earth surface [35].
33
(a)
(b)
Figure 2.25: E-plane linear polarized a) Vertical polarization, b) Horizontal
polarization [35]
¾ H-plane: In the case of the same linearly polarized antenna, this is the plane
containing the magnetic field vector and the direction of maximum radiation.
The magnetic field or H-plane lies perpendicular to the E-plane. For a
vertically-polarized antenna, the H-plane usually concurs with the horizontal
plane. For a horizontally-polarized antenna, the H-plane usually concurs with
the vertical plane.
The radiation pattern is taken both for a co-polarized and cross-polarized
response. Ideally, linear polarization means that the electric field is in only one
direction, but this is seldom the case. For linear polarization, the cross-polarization
level determines the amount of polarization impurity [34]. As an example, for a
vertically polarized antenna, the cross-polarization level is due to the E-field existing
in the horizontal direction; i.e. Eφ is the cross-polarization component of Eθ (Figure
2.21). Normally, cross-polarization level is a measure of decibels below the copolarization level. The polarization quality is expressed by the ratio of these two
responses.
34
2.2.2.4 Bandwidth
Bandwidth can be defined as the range of frequencies within which the
performance of the antenna, with respect to some characteristic, conforms to a
specified standard [33]. In general, the bandwidth is the range of frequencies on both
sides of the centre frequency (resonant frequency) where the antenna characteristics
are within an acceptable value of those at the center frequency. The standard
acceptable return loss value for an antenna is -10 dB which allow 10% power
reflected in an antenna. The bandwidth percentage is calculated as shown in equation
2.3.
BW % =
fu − fl
( f u xf l )
x100%
(2.3)
Where;
f u = upper frequency bandwidth (the second frequency that allow 10% power
reflection)
f l = lower frequency bandwidth (the first frequency that allow 10% power
reflection)
2.2.3 Antenna Array
Antenna arrays consist of many radiating elements, which are fed by signal of
appropriate phase and amplitude. In general, there are several reasons why
researchers interested in having antennas in array; there are common understanding
that antenna arrays could produce higher directivity with a narrow beamwidth and
also a better gain. For some configuration, desired radiation pattern also can be
achieved through an array. It can be designed to control their radiation characteristics
35
by properly selecting the phase and amplitude distribution between elements.
Antenna arrays could be analyzed according to their arrangements that are known as
listed in below. For this dissertation, focus will be carried on to the linear array as
only this topic involved in this project.
a) Linear arrays: antenna elements arranged along a straight line (1Dimension)
b) Planar array: antenna elements arranged over some planar surface (2Dimension)
c) Circular array: antenna elements arranged around a circular ring (2Dimension)
d) Conformal array: antenna elements arranged to conform to some non-planar
surface (3-Dimension)
(a)
(b)
(c)
(d)
Figure 2.26: Various antenna array configuration a) Linear array, b) Planar array, c)
Circular array, d) Conformal array
2.2.3.1 Uniform Linear Antenna Array
For the case of uniform amplitude uniform spacing in linear array, given an
antenna array of identical elements, the radiation pattern of the antenna array may be
found according to the pattern multiplication theorem which is written as equation
2.4 [33]. Pattern multiplication theorem said that the array radiation pattern can be
36
found by multiplying single element pattern with its array factor (AF). Single
element pattern is the pattern of the individual array element while array factor (AF)
is a function dependent only on the geometry of the array and the excitation
(amplitude, phase) of the elements. The example of this theorem is illustrated in
Figure 2.27.
Single element pattern x Array Factor (AF) = Array pattern
x
Single element pattern
(2.4)
=
Array Factor (AF)
Array pattern
Figure 2.27: An example of pattern multiplication theorem
Single element pattern is relied on the type of the radiating element that used
in the array configuration which can be a directional antenna, omnidirectional
antenna or non-directional antenna. The data may obtained from the simulation data
or measured data while for the AF, the own calculation is needed in order to know
how does its radiation pattern looks like. As mentioned previously, AF is dependent
on the geometry of array, since this project only focus on the linear array, only AF
for uniform linear array will be discussed.
Figure 2.28: N element along x axis
37
Before going further, it is important to ensure about the axis of the element
arrangement either arranged along the x, y or z axis because the axis will determine
which formula will be applied for the calculation. For this report, only arrangement
along the x axis is taken into considerations.
Figure 2.29: Uniform Linear Array Configuration
The uniform linear array shown in Figure 2.29 consists of N elements equally
spaced at distance d apart with identical amplitude excitation and has a progressive
phase difference, β between the successive elements. The amplitude excitation and β
are generally controlled by the power distribution network. For standard linear
antenna array, power divider is commonly used as the power distribution network. In
this section, only linear array that having uniform amplitude and uniform spacing
will be discussed while the details about linear array having non-uniform amplitude,
uniform spacing can be found in [33]. The derivation of the AF formula begins by
finding the formula of the phase difference between adjacent elements which is given
as follows:
38
ψ = kd sin θ + β
(2.5)
Where:
Ψ = phase difference between adjacent elements
k=
2π
λ
d = distance between adjacent elements
β = phase difference of excitation current between adjacent elements
θ = angle relative to the normal to the array
Next, the array radiation pattern can be found by summing up the entire received
power of each element. The AF of N-element linear array of isotropic sources is:
AF = e j 0 + e jψ + e j 2ψ ... + e j ( N −1)ψ
(2.6)
The above equation can be simplified and written as:
N
AF = ∑ e + j ( n −1)ψ
(2.7)
n =1
In order to make the pattern analysis more convenient, the above AF relation can be
expressed in a closed form as shown in below. The details about the derivation can
be found in [33].
⎡ ⎡N ⎤⎤
sin ψ
1 ⎢ ⎢⎣ 2 ⎥⎦ ⎥
⎥
AF = ⎢
ψ
N⎢
⎥
⎢
⎥
2
⎣
⎦
Where:
ψ = kd sin θ + β
N = Number of radiating element
(2.8)
39
According to the equation 2.8, it can be observed that only 3 parameters manipulate
the array factor patterns which are:
1) N : number of element
Number of element controls the beamwidth size of main beam pattern. Figure
below shows the effect of this parameter to the radiation pattern of array
factor. In this case, other parameters are set to be constant, while N value is
varies at instant. It can be seen that as the number of antenna element
increasing, the size of main beamwidth becomes narrower.
Figure 2.30: Plots of AF with d =
λ
2
, β = 0 and N = 4, 8 and 12
Table 2.4: The effect of varying parameter N
N
Main lobe beamwidth size
Number of Side Lobes
4
34°
4
8
16°
12
12
10°
20
40
2) d: distance between adjacent element
The distance between elements should be determined carefully with the
intention that the highest gain can be achieved with less appearance of side
lobe. It can be seen that as the distance of the elements increases, the number
of side lobes also increases and the main lobe size decreases.
Figure 2.31: Plots of AF with, N = 4, β = 0 and d = 0.25λ, 0.5λ, λ, 1.25λ
Table 2.5: The effect of varying parameter d
d
Main lobe beamwidth size
Number of Side Lobes
0.25λ
70°
0
0.5λ
34°
4
λ
16°
12
1.25λ
9°
16
3) β : phase difference of excitation current between adjacent elements
Phase difference of the excitation current between the adjacent elements will
control the tilted angle of the main lobe. This effect can be illustrated in
Figure 2.32. When ß have some value, the main lobe can be directed to the
certain direction as shown in Table 2.6.
41
(a)
(b)
(c)
Figure 2.32: Plots of AF with, N = 4, d = 0.25λ, (a) ß = 0°, (b) ß = ±45°, (c) ß = ±90°
Table 2.6: The effect of varying parameter β
ß
Main Lobe Tilt Angle
0°
0°
-45°
-14.5°
+45°
+14.5°
-90°
-30.0°
+90°
+30.0°
From the figure above, it can be concluded that by varying the value of ß, the
main lobe can be tilted to a certain direction. This principle also known as electronic
scanning where is has a capability of changing the direction of the main lobe
electronically through the variation of ß.
There a few method of electronic scanning that commonly known which are:
1) Phase scanning: Main beam of the radiation pattern can be tilted by varying
the current phases of antenna elements; usually phase shifter is used to vary
the current phase.
42
Figure 2.33: Phase scanning block diagram
2) Time delay scanning: Phase changed is achieved by switching in different
length of transmission path; a switched line phase shifter is commonly used
[33].
Figure 2.34: A switched line phase shifter [40]
3) Beamswitching: Phase changed is achieved by using different transmission
path; the most popular one is Butler Matrix.
43
2.2.3.2 Beamswitching (electronic scanning)
In the previous section, the parameters that influence the radiation pattern of
antenna array have been discussed in terms of numerical analyses which are N, the
number of radiating element, d, the distance between elements and ß, the phase
difference of excitation current between adjacent elements. It may easy to understand
the meaning of parameter N and d in practical world, but to understand the concept
of ß; it may need a little knowledge about power divider and s-parameter analysis in
order to understand how this parameter influences the antenna array performance,
mainly in beamswitching scanning.
It is commonly known that for standard antenna array, it is constructed by
integrating a number of antenna elements with a power divider. For a better
understanding, it may easy to analyze power divider network in terms of the S
parameters as illustrated in Figure 2.35. If all ports of power divider are matched, all
outputs will have an equal amplitudes and phases. In this case, 30° is given as an
example of the power divider output phase though is not necessary that all 6 dB
power divider will gives 30° at the output. With reference to Table 2.7, phase
difference between outputs can be obtained by subtracting the phase of S51 with
S41, S41 with S31 and S31 with S21, which then give the result of β = 0°. That is
how the phase difference of the excitation current between adjacent antenna elements
can be obtained.
S21
S31
S41
S51
6 dB Power
Divider
Network
Port 1
Figure 2.35: The operation of power divider in terms of S-parameter
44
Table 2.7: An example of power divider result in ideal case
Coupling parameter
Amplitude (dB)
Phase (°)
S21
10 log (0.25) = -6
30
S31
10 log (0.25) = -6
30
S41
10 log (0.25) = -6
30
S51
10 log (0.25) = -6
30
While in the beamswitching case, antenna array usually integrated with
beamforming network. The function of beamforming network is to control the
amplitude and phase of the power that goes to the antenna array.
Figure 2.36: 4 x 4 Butler Matrix configuration
Table 2.8: Progressive phase difference corresponds to each input port of Butler
Matrix.
Port β
1
-45º
2
+135º
3
-135º
4
+45º
Figure 2.36 shows an example of a Butler Matrix as a beamforming network.
Butler Matrix has a capability to steer a beam electronically to a certain direction by
providing multiple β through different transmission paths. It has four input ports; 1,
45
2, 3 and 4 and four outputs ports; 5, 6, 7 and 8. Four output ports are then used as
inputs to antenna elements. The conventional Butler Matrix can provide four
different value of β which are 45º, -135º, 135º, -45º. Each of input port of the Butler
Matrix will give one value of β at the output ports as shown in Table 2.8. For
example, if port 1 is activated, the phase excitation of the current that coupled to the
antenna will be -45º different between adjacent elements. As explained previously,
the value of β is obtained by subtracting the phase at port 8 with port 7, port 7 with
port 6 and port 6 with port 5. The data in Table 2.8 can be illustrated in terms of
diagram as shown in Figure 2.37. Since Butler Matrix can provide 4 different value
of β, thus four beams at four different directions can be generated. The details about
beamforming network and Butler Matrix will be explained later in the next section.
(a)
(c)
(b)
(d)
Figure 2.37: Progressive phase difference corresponds to each input port of Butler
Matrix (block diagram form) a) Port 1: β =-45º, b) Port 2: β =+135º, c) Port 3: β =135º, d) Port 4: β =+45º
46
2.3
Beamforming Network
Beamforming is the combination of radio signals from a set of small nondirectional antenna to simulate a large directional antenna [4]. By adding
beamforming network to the antenna array, the desired radiation pattern can be
obtained as it has a capability to control the amplitudes and phases of the excitation
current coupled to the radiating elements. In communications, beamforming is used
to reduce interference and improve communications by tilting the main beam to the
desired user without physically moves the antenna.
The example of beamforming function can be illustrated in Figure 2.38. It
shows the comparison of the radiation pattern that obtained by the standard antenna
array and the radiation pattern obtained by the integration of antenna array and
beamforming network. It can be seen that without beamforming network, the main
lobe is directed to the center of the plot, but after adding beamforming network, the
main beam can be tilted to a certain direction.
(a)
(b)
Figure 2.38: Radiation pattern obtained by a) Standard antenna array, b) The
integration between antenna array and beamforming network.
47
As described in previous section about the antenna array, it has been proved
that when β has some value, the main lobe can be tilted to a certain direction. In the
standard antenna array case, power divider produces equal amplitudes and equal
phase at its outputs port, thus no phase differences of the excitation current coupled
to the antenna. In the other words, β = 0° by using standard 6 dB power divider while
in the beamforming network case, β has some value which cause the main lobe to be
tilted to a certain direction.
Thus far, most of the studies have been focused on the beamforming network
[16]-[23] as it is the most important part of the system. The beamforming network
can be classified into 2 different categories which are digital beamforming and RF
beamforming [36]. Digital beamforming usually related to adaptive array system
while RF beamforming usually involved in multibeam system. For this project, only
the topic about RF beamforming will be discussed. There are few known types of the
RF beamforming network for the application of multibeam antenna which are lens
based beamformer and circuit beamformer [36]. The flow chart of the type of
beamformer can be illustrated in the following figure. Lens based beamformer is a
major topic and the details regarding this topic can be found in [36].
Beamformer
Digital beamformer
RF beamformer
Lens based beamformer
Circuit beamformer
Blass Matrix
Butler Matrix
Figure 2.39: Flow chart of the type beamformer
48
2.3.1 Blass Matrix
The Blass matrix is a multiple beam matrix that uses a set of N antenna array
element transmission lines that intersects a set of M beam port lines, with a
directional coupler at each intersection [37]. This matrix is terminated with matched
load as shown in Figure 2.40. It can be seen from the figure that in order to produce
M beams from N radiating elements, it is required to have M x N couplers, radiating
elements and matched load. This type of beamforming network would require a large
amount of couplers and would have a very large size.
Figure 2.40: Blass Matrix configuration [37]
2.3.2 Butler Matrix
The most popular beamforming network based on circuit type is Butler
Matrix. It is easier to implement using microstrip transmission lines as its
implementation is less complex as compared to Blass Matrix. The conventional
Butler Matrix is a 2n x 2n network which has 2n inputs, 2n outputs, 2n-1 log2 2n hybrids
49
and several phase shifters [16]. Typically, it has the N number of input, and N
number of output to produce N number of orthogonal beam. Nevertheless, it may not
necessary that it will have same number of input and output ports. This feature may
change according to application of Butler matrix. For example, in [38], the Butler
Matrix is modified to have only 3 inputs but maintaining 4 outputs ports with the aim
is to produce 3 beams, in which, one of them is broadside beams. With N input and
M output port, Butler Matrix can produce N orthogonal beams. N number of input
will determine the number of beam that produced by the Butler Matrix while the M
number of output will determine the number of radiating elements, which will affect
the beamwidth size of the main lobe.
The most famous configuration of the Butler Matrix is 4 x 4, which has 4
input and 4 output ports to generate four independent beams at four different
directions. As discussed previously, the Butler Matrix has capability to steer a beam
by providing progressive phase difference between adjacent output ports, β. By
knowing the value of β, the patterns of AF that generated by Butler Matrix can be
plotted. The block diagram and AF plot for the conventional 2 x 2, 4 x 4 and 8 x 8
Butler Matrix is shown in following figure. The value of phase differences of each
Butler Matrix is taken from [3]. The input ports of Butler Matrix are named
according to the beam position in polar plot that correspond to each β.
Figure 2.41: AF plot for 2 x 2 Butler Matrix (N = 2, β = ±90°)
50
Table 2.9: Numerical value for 2 x 2 Butler Matrix
Input Port
β
Beam position in polar plot 3 dB beamwidth
R1
-90°
+30°
53.9°
L1
+90°
-30°
53.9°
Figure 2.42: Block Diagram of 4 x 4 Butler Matrix [16]
Table 2.10: Numerical value for 4 x 4 Butler Matrix
Input Port
β
Beam position in polar plot 3 dB beamwidth
R1
-45°
+14.5°
33.4°
L2
+135°
-48.6°
44.3°
R2
-135°
+48.6°
44.3°
L1
+45°
-14.5°
33.4°
Figure 2.43: AF plot for 4 x 4 Butler Matrix (N = 4, β = ±45°, ±135°)
51
Figure 2.44: Block Diagram of 8 x 8 Butler Matrix [39]
Table 2.11: Numerical value for 8 x 8 Butler Matrix
Input Port
β
Beam Position in polar plot 3 dB beamwidth
1R
-22.5°
+7.2°
20°
4L
+157.5°
-61°
42°
3R
-112.5°
+38.7°
24.8°
2L
+67.5°
-22°
21.2°
2R
-67.5°
+22°
21.2°
3L
+112.5°
-38.7°
24.8°
4R
-157.5°
+61°
42°
1L
+22.5°
-7.2°
20°
Figure 2.45: AF plot for 8 x 8 Butler Matrix (N = 8, β = ±22.5°, ±67.5°, ±112.5°,
±157.5°,)
52
As shown in Figure 2.42, typical 4 x 4 Butler Matrix is made of three major
components which are:
1) 90° hybrid coupler
2) 0 dB crossover
3) Phase shifter
2.3.2.1 90° Hybrid Coupler
90° hybrid couplers are 3 dB directional coupler with a 90° phase difference
in the outputs of the through and coupled arms [40]. The 90° hybrid couplers also
known as quadrature hybrids or branch line hybrid. This type of hybrid is usually
implemented using microstrip or stripline technique and the configuration is shown
in Figure 2.46.
Figure 2.46: Geometry of 90° hybrid coupler
The function of hybrid coupler is almost similar to the T-junction power
divider which is to divide the input power equally to both outputs, except in power
divider, both power at the outputs having the same amplitude and phase while in
hybrid coupler, the power amplitudes are same but phases are shifted about 90°
between those outputs. This coupler is designed with the intention that when power
53
entering to port 1, 50% power will go to port 2 and 3, no power is coupled to port 4
and the phase difference between output ports are 90°. The percentage of power can
be analyzed in terms of S-parameter as summarized in Table 2.12.
Table 2.12: S-parameter for ideal case 90°hybrid coupler
From
To
Port 1
Port 2
Port 1
Port 3
Port 4
Characteristics (Ideal Case)
0% Reflected power
Return Loss
50% Transmitted power
Coupling value
50% Transmitted power
Coupling value
0% Transmitted power
Isolation value
S-parameter
S11 = 10 log (0) = - infinity
S21 = 10 log (0.5) = - 3 dB
S31 = 10 log (0.5) = - 3 dB
S41 = 10 log (0) = - infinity
By referring to Table 2.12, if the design for all ports is matched, the result for
S11 and S41 at the operating frequency will show the lowest value while the
obtained coupling value will be approximately about -3 dB. As the properties of the
hybrid coupler itself are symmetrical and reciprocal, the parameters those shown in
Table 2.12 are good enough to analyze the performance of the hybrid coupler. By
knowing the symmetrical properties of the hybrid coupler, port 1 or port 4 can be the
input port and the result will not be much different from those as shown in Table
2.10. By knowing the reciprocal properties of the hybrid coupler, an input can be the
output or vice versa as the result will be same. For example, the value of S12 and
S21 will be same for a reciprocal device. The characteristics of hybrid coupler is
remain same where the output ports will always be on the opposite side of the
junction from the input port and the isolated port will be the remaining port on the
same side of the input port.
54
2.3.2.2 0 dB Crossover
0 dB crossover can be described as a two crossing transmission lines on the
microstrip element. Port 1 to port 3 can be considered as one transmission lines while
port 4 to port 2 is another transmission lines but both of them are crossing between
each other. The basic operation of the crossover is such that with all ports matched,
power entering to port 1 is 100% transmitted to port 3 and no power is coupled to
port 2 and 4 and the phase from port 1 to port 3 should not be changed. This will give
a result of the coupling value, S31 = 0 dB and the phase of S31 = 0°. The crossover
is designed with the intention that good matching and good isolation can be obtained
so that all input power will be transmitted to the desired port. This matching can be
achieved by cascading 2 hybrid coupler into a device as shown in Figure 2.47. This is
the conventional method of designing crossover layout. The other techniques of
designing a crossover could be found in [16]-[17], [20]-[23].
Figure 2.47: Geometry of 0 dB crossover
For a better understanding about the operation of the crossover circuit, the
layout of the crossover in Figure 2.47 can be illustrated as shown in Figure 2.48
(only for explanation purpose, not the actual circuit representation). When the power
entering to port 1, 100% power will be transmitted to the port 3 which is located on
the opposite side of the P1 but it has to cross another transmission line which is port
4 to port 2. Crossover also has the same properties as hybrid coupler where it is a
55
symmetrical and reciprocal device. Table 2.13 shows the s-parameter of crossover
when port 1 becomes the input port. If the design for all ports is matched, the result
for S11, S41 and S21 at the operating frequency will show the lowest value while the
obtained coupling value, S31 will be approximately about 0 dB and phase will be
equal to 0°.
Figure 2.48: Illustration that represents the function of 0 dB crossover
Table 2.13: S-parameter for ideal case 0 dB crossover
From
To
Port 1
Port 2
Port 1
Port 3
Port 4
Characteristics (Ideal Case)
0% Reflected power
Return Loss
0% Transmitted power
Isolation value
100% Transmitted power
Coupling value
0% Transmitted power
Isolation value
S-parameter
S11 = 10 log (0) = - infinity
S21 = 10 log (0) = - infinity
S31 = 10 log (1) = 0 dB
S41 = 10 log (0) = - infinity
56
2.3.2.3 Phase shifter
The phase shifters is implemented using transmission lines which lengths
introduce the required phase shift. The phase shift θ associated to a transmission line
of length l is given by the equation [27]:
θ=
360°l
λg
(2.9)
Where l is in meter, θ is in degree and λg is one wavelength in the transmission line
medium. For microstrip transmission lines, the wavelength is given by [40]:
λg =
λ0
ε eff
(2.10)
Where λ0 is the wavelength at free space, and εeff is the effective dielectric constant of
the microstrip substrate and free space.
The phase shift values are proportional to the length of the transmission line.
The longer the transmission line between the input and output ports, the longer the
delay. A longer delay causes a bigger phase shift. The electrical lengths of the delay
lines are in proportion to the operating frequency of the circuit.
2.4
Previous Work
Previous work that related to this project can be divided into 2 sections which
are the integration between conventional 4 x 4 Butler Matrix and antenna array and
57
the development of 4 x 4 Butler Matrix. For the first section, the intention is to
compare the obtained radiation pattern from literature that corresponds to each input
port of Butler Matrix. In this section, only integration between antenna and
conventional Butler Matrix is considered, so that the performance of using
conventional Butler Matrix with antenna array can be observed. In the second
section, the development of Butler Matrix is highlighted. The development that
brought by other researchers in literature is very helpful to understand about the
properties of Butler Matrix itself.
2.4.1 Integration between conventional 4 x 4 Butler Matrix and Antenna
Array
The application of switched beam antenna system has become more popular
nowadays, as it is capable to increase the SNR ratio, reduce interference and increase
the channel capacity of the system [6]. In recent years, the topic of multibeam
antenna constructed using Butler Matrix as a beamforming network has received
much attention due to its simplicity and low cost of implementation.
Four ports conventional Butler Matrix incorporated with 4 linear antenna
arrays has been presented in several papers such as in [25], [26] and [27]. In paper
[25], 4x2 bowtie antennas have been used as a radiating antenna. It was verified that
four beams have been created with beam angles at 12º, -40º, 40º and -12º. The
design configuration constructed in [26] can generate four beams; 1R, 2L, 2R and 1L
with the beam angles at 16º, -39º, 38º and -15º. A similar design configuration also
presented in [27], but they brought some changes on the feeding techniques where
coaxial probe is used instead of inset feeding. It was shown in [27] that only 1L and
1R appeared as directional patterns and the others showing unidentified direction.
The directions of beam patterns are -19º and 19º, which represent about 13% errors
from the theoretical calculation.
58
All these papers were discussing about switched beam antenna that uses
microstrip antenna array, which only accomplished to serve 120º of angular coverage
area. In [24], the authors present about using monopole antenna as the antenna array.
The result is quite different from the previous mentioned papers where the radiation
patterns of the system are having different types, broader and narrower. The work
done in all these papers can be summarized as shown in Table 2.14.
Table 2.14: Previous work on the integration between conventional 4 x 4 Butler
Matrix and Antenna Array
1. Researchers details
Dau-Chyrh Chang and Shin-Huei Jou,
Da Yeh University [25]
Date of published
June 2003
Designed Frequency
1.7 GHz ~ 2.2 GHz
Bowtie antenna
Antenna Type
Beam angle: 12º, -40º, 40º and -12º
Measured
radiation pattern
59
2. Researchers
details
Date of published
Siti Rohaini Ahmad and Fauziahanim Che Seman,
Kolej Universiti Teknologi Tun Hussein Onn,
Malaysia [27]
December 2005
Layout of
Butler Matrix
Designed Frequency 900 MHz
ƒ
Rectangular Microstrip Antenna
ƒ
Feeding type – coaxial probe
ƒ
Port 1 (-19º)
ƒ
Port 2 (unidentified)
ƒ
Port 3 (unidentified)
ƒ
Port 4 (19º)
Antenna Type
Measured
radiation pattern
60
3. Researchers
Nhi T. Pham1, Gye-An Lee2 and Franco De Flaviis1
details
1
Date of published
July 2005
University of California, 2Skyworks Solution Inc. [26]
Layout of
Butler Matrix
and antenna array
Designed Frequency 5.2 GHz
Antenna Type
ƒ
Rectangular Microstrip Antenna
ƒ
Feeding type – inset feed
ƒ
Port 1R (16º)
ƒ
Port 2L (-39º)
ƒ
Port 2R (38º)
ƒ
Port 1L (-15º)
Measured
radiation pattern
61
4.
Researchers Ying Jung Chang and Ruey Bing Hwang
details
National Chiao-Tung University, Taiwan [24]
Date of published
December 2001
Designed
Frequency
2.4 GHz
ƒ
CPW-fed uniplanar monopole antenna
ƒ
Port 1
ƒ
Port 2
ƒ
Port 3
ƒ
Port 4
Antenna Type
Computed and
measured
radiation pattern
62
2.4.2 Development of 4 x 4 Butler Matrix
A conventional Butler matrix is realized using 3 dB hybrid couplers,
crossover and phase shifters and usually implemented using single layer microstrip
line. However, this conventional Butler matrix has its own disadvantages as it has
narrow bandwidth caused by both the hybrids and phase shifters. Because of its
narrow bandwidth, any fabrication errors will shift the center frequency significantly
[18].
There are few papers discussing on how to improve the bandwidth of Butler
Matrix and each of them define the meaning of ‘wideband’ on their own
perspectives. For example, in [16], the authors define the wideband in terms of
having good return loss over the frequency range and put less concentration of the
coupling values and phases. Wide band Butler Matrix is achieved by designing 2
section hybrids instead of one section hybrid in the conventional case. Simulation
results show that 13% bandwidth can be achieved by the Butler Matrix.
A novel design of Butler Matrix is presented in [17] which enhance the
bandwidth by maintaining the coupling ratio of the hybrids coupler to be equal over a
wide band frequency. This is different with [16] where it only focused to have wide
band return loss, rather than expecting the coupling value to be good over the wide
band. The author adopted the suggestion in [42] to apply matching network for
improving the coupling ratio. One half-wavelength series transformer and half
wavelength stub were attached to all the four ends of the conventional hybrid coupler
and folded backwards to conserve space. The simulation performance shows the flat
coupling for the frequency range of 9.75 GHz to 10.25 GHz.
However, the proposed ideas in both papers give a result of larger in size due
to the multisection hybrids in [16] and half-wavelength stub in [17] which making
them more difficult to fabricate when integrating with crossover and phase shifter
63
[18]. Hence, to achieve compact broadband Butler Matrix, disk hybrid is proposed in
[18] and half wavelength open stubs is used as a phase shifter as it can gives a flat
phase difference over the frequency range. Compact broadband with equal
amplitude, constant phase difference and good return loss over a wider frequency
range is achieved in [18].
In fact, the idea presented in [18] is adopted from [19] where in [19], four
element planar Butler Matrix using hybrid coupler and half-wavelength delay lines
without any crossing is proposed.
The purpose is to improve relative phase
characteristics between output ports over a wide frequency range. Constant phase
difference, about 2° over the frequency range of 0.85 to 0.9 GHz is achieved.
Another concern regarding the implementation of the Butler Matrix is its
large size and difficult to implement crossovers, especially when trying to implement
it for large arrays. Thus, in [20], [21], a new technique of fabricating Butler Matrix is
proposed where the authors used coplanar waveguide technique (CPW) instead of
microstrip technique. By having CPW hybrid couplers, the used of crossover can be
avoided but this technique is hard to be implemented as it involves multi layer
substrate.
While in [22], the size of Butler Matrix is reduced by reducing the size of
hybrid coupler as the conventional hybrid coupler based on quarter wavelength
requires larger area. The design is realized on a low cost substrate (FR4 board) and
the implementation is simple which only use single layer microstrip technique. The
smaller hybrid is realized by introducing 12 additional capacitive plates to the
internal structures of the hybrid couplers. The experimental circuit was reduced by
80% in area and a good frequency characteristic is achieved.
Paper in [23] shows the design of a dual band Butler Matrix with the aim to
cover two different frequencies in WLAN systems, which are 2.4 GHz and 5 GHz.
64
Planar dual band hybrid is designed using 3 branch line 3 dB couplers in which the
length of the coupling lines in one pair of the parallel arms is twice longer (λ/2) than
in the other (λ/4). The measured result of Butler Matrix shows the acceptable result
over the proposed frequencies.
As a conclusion, there are 3 majors area that has been received much
attention in the implementation of Butler Matrix which are bandwidth, size reduction
and dual band operation. The work done in all these papers can be summarized as
shown in the following table.
Table 2.15: Previous work on the development of 4 x 4 Butler Matrix
1. Researchers details
Tayeb. A. Denidni and Taro Eric Libar
University of Quebec, Canada [16]
Date of published
September 2003
Operating Frequency
1.9 GHz ~ 2.2 GHz
Design configuration
¾ Hybrid Coupler Configuration
(wideband)
¾ Crossover configuration
¾ 4 x 4 Butler Matrix configuration
65
2. Researchers details
Manish A. Hiranandani and Ahmed A. Kishk
University of Mississippi [17]
Date of published
July 2005
Operating Frequency
10 GHz
Design configuration
¾ 4 x 4 Butler Matrix
(wideband)
3. Researchers details
S. Zheng, W.S. Chan, S.H. Leung and Q. Xue [18]
Date of published
10th May 2007
Operating Frequency
4.2 GHz
Design configuration
¾ 4 x 4 Butler Matrix
(broadband with flat
coupling)
4. Researchers details
Hitoshi Hayashi, Donald A. Hitko, Charles G. Sodini [19]
Date of published
March 2002
Operating Frequency
0.9 GHz
Design configuration
(constant phase over
wide band)
¾ Schematic block diagram
66
¾ Realized Butler Matrix
5. Researchers details
Mourad Nedil, Tayeb A. Denidni and Larbi Talbi
University of Quebec, Canada [20][21]
Date of published
July 2005 [20], January 2006 [21]
Operating Frequency
5.8 GHz
Design configuration
¾ Hybrid Coupler Configuration
(size reduction)
¾ Crossover configuration
¾ 4 x 4 Butler Matrix configuration
67
6. Researchers details
Gary Kwang T.K, Peter Gardner
University of Birmingham [22]
Date of published
October 2001
Operating Frequency
1.8 GHz
¾ Hybrid Coupler Configuration
Design configuration
(size reduction)
¾ 4 x4 Butler Matrix Configuration
7. Researchers details
Carlos Collado1, Alfred Grau2 and Franco De Flavis2
1
Technical University of Catalonia, 2University of California [23]
Date of published
June 2005
Operating Frequency
2.4 GHz, 5 GHz
Design configuration
(dual band)
68
2.5
Chapter Summary
Therefore this chapter has explained the theory and motivation towards smart
antennas in wireless communications. After that, the basic theory of antenna
including microstrip antenna concept, its properties and antenna array are presented.
Each component that needs to be designed to obtain the Butler Matrix has been
shown to give an overview of the design steps in the next chapter. The works done in
past until present related to this project also presented.
CHAPTER 3
METHODOLOGY
This chapter presents the details methodology of the project. This chapter will
describe about the design procedure such as design specification, numerical
calculation and design configuration. The simulation, prototype fabrication and
measurement setup of the project is also presented. The comparison between
designed Butler Matrix with related published papers is presented as well.
3.1
Project methodology
Operational framework of this project can be divided into 3 major parts which is
shown in Figure 3.1.
Design development and simulation
Fabrication
Measurement and analysis
Figure 3.1: The flow chart of the operational framework
70
3.2
Design Development and Software Simulation
As shown in Figure 3.2, this project is comprises of two major blocks which
are antenna array and Butler Matrix. In order to make the design implementation
easier, the development of the project has been divided into a two stages as shown in
the Figure 3.3. The whole project is implemented using FR4 board and the important
specifications of the board can be found in the Table 3.1 while the details data sheet
can be found in Appendix A.
Figure 3.2: The block diagram of the complete design configuration
Figure 3.3: The flow chart of the design development of the project
Table 3.1: Specifications for the FR4 board
Operating Frequency
2.4 GHz
Dielectric Constant, εr
4.5
Substrate height, h
Dissipation factor, tan δ
1.6 mm
0.019
71
3.2.1 Design and Implementation of the Antenna Array
Three different types of antenna have been used in this project which are:
i)
(4 x 1) Square Patch Antenna – Broader beamwidth antenna
ii)
4 x (4 x 2) Antenna Array – Directional Antenna
iii)
(4 x 1) Commercial Dipole Antenna – Omnidirectional Antenna
These antennas were chosen as each of them has different characteristics, thus the
radiation pattern of the multibeam antenna by using these kinds of antenna could be
observed.
3.2.1.1 (4 x 1) Square Patch Antenna – Broader beamwidth antenna
The first type of the antenna is square patch microstrip antenna with
proximity coupling as the feeding technique. As mentioned in Chapter 2, Section
2.2.1, this technique offer a bigger bandwidth, thus, it is chosen as the feeding
technique. Four identical square patch antennas were designed at 2.4 GHz, arranged
in a linear form and spaced at half-wavelength apart. This half wave spacing ensures
that the array will have the largest gain and directivity that does not have grating
lobes [33].This antenna array is constructed on two layer substrates. Square patch
antenna are printed on the upper substrate while the feeding line are printed on the
lower substrate but both of the substrate have the same specifications as stated in
Table 3.1. The layout of the 4 x 1 square patches is shown in Figure 3.4 (a) and the
3D view can be seen in Figure 3.4 (b).
72
(a)
(b)
Figure 3.4: Square patch antenna configuration a) Top view (2D), b) 3D view
The effective width of the patch antennas is calculated using the theory of
microstrip patch antenna which can be computed by using following equation.
The width of the microstrip patch antenna is given as [34]:
c
W=
2 fr
(3.1)
ε r +1
2
Equation (3.2) gives the effective dielectric constant, ε eff as [34]:
ε eff =
ε r +1 ε r −1 ⎡
+
2
2
1
h ⎤2
⎢⎣1 + 12 W ⎥⎦
(3.2)
Equation (3.3) gives the effective length, Leff as [34]:
Leff =
c
2 f r ε eff
(3.3)
73
Equation (3.4) gives the length extension, ∆L as [34]:
⎛W
⎞
(ε eff + 0.3)⎜ + 0.264 ⎟
⎝h
⎠
∆L = 0.412h
(ε eff − 0.258)⎛⎜ W + 0.8 ⎞⎟
⎝h
⎠
(3.4)
The length of the microstrip patch antenna is given as [34]:
L = Leff − 2∆L
By substituting c = 3x10 8 m , f r =2.4GHz, ε r =4.5 and h = 1.6 mm
(3.5)
to the above
equations, the following results are obtained:
W=37.7mm, ε eff = 4.174 , Leff =30.59mm, ∆L = 0.73655, L=28.6mm
After going through the optimization process using Microwave Office
software, the final optimum parameter for the single square patch antenna is
obtained, where the width is found to be 28 mm x 28 mm. Figure 3.5 shows the
simulated return loss values correspond to the each patch antenna. It is shown that
good impedance matching is obtained at the operating frequency as S11 values are
lower than -10 dB. Table 3.2 shows the details about the simulated result that shown
in Figure 3.5. Bandwidth is calculated using equation 2.2 that written in Chapter 2,
Section 2.2.2.4.
Figure 3.5: Simulated Return Loss for 4x1 square patch antenna
74
Table 3.2: Simulated result analysis for 4x1 square patch antenna
Port1
Port2
Port3
Port4
Return Loss at 2.4GHz (dB)
-12.38
-16.74
16.74
-13.9
fl (GHz)
2.39
2.37
2.37
2.38
fu (GHz)
2.45
2.45
2.45
2.45
BW (%)
2.50
3.35
3.35
2.73
Figure 3.6: Radiation pattern of 4x1 square patch antenna
Figure 3.6 shows the simulated radiation pattern obtained by the 4 x 1 square
patch antenna. Since each port connected to antenna separately, each antenna is
considered as a single antenna, thus the radiation pattern obtained by E-plane has a
broader beamdwidth with HPBW approximately about 106°. As explained in Chapter
2, Section 2.2.2.1 and 2.2.2.3, radiation pattern is obtained by co-polarization and
cross-polarization of E-plane and H-plane. The radiation pattern characteristics of
Figure 3.6 correspond to each plane can be written as shown in Table 3.3. Copolarization shows the maximum radiation value for both plane, E and H-plane,
while cross-polarization shows the other ways.
75
Table 3.3: Radiation pattern of 4 x 1 square patch interpretation
Label
Type of polarization
Plane
PPC_EPhi(0,1)
Co-polarization at φ = 0°
E-plane
PPC_ETheta(0,1)
Cross-polarization at φ = 0°
E-plane
PPC_EPhi(90,1)
Cross-polarization at φ = 90°
H-plane
PPC_ETheta(90,1)
Co-polarization at φ = 90°
H-plane
3.2.1.2 4 x (4 x 2) Antenna Array – Directional Antenna
The directional antenna that has been used in this project is 4 x (4 x 2)
microstrip patch antenna array. It was designed by Hassan [41] and the details about
the design procedure could be found in [41]. It was verified through simulation that
the antenna has the capability to produce a directional beam pattern with the HPBW
of approximately 23°. Coaxial probe was used as a feeding technique and placed at
the centre of the array as this technique produces less radiation effect that caused by
the transmission line. The 4 x 2 element antenna array was simulated and the board
has the same specifications as stated in Table 3.1. The layout of 1 unit of 4 x 2
antenna array can be seen in Figure 3.7 and its simulated return loss and radiation
pattern are shown in Figure 3.8 and 3.9 respectively. No simulation is done for 4 x (4
x 2) antenna array as the feeding line of each 4 x 2 array is coaxial probe which is
difficult to connect by simulation.
Figure 3.7: Layout of 4 x 2 antenna array [41]
76
Table 3.4: Simulated result analysis for 4x2 antenna array
Return Loss at 2.4GHz
fl
Result
-12.30 dB
2.26 GHz
fu
2.47 GHz
BW
9.26 %
Figure 3.8: Return Loss of 4x2 array patch [41]
Figure 3.9: E-plane co-polarization radiation pattern of 4x2 array patch [41]
77
3.2.1.3 (4 x 1) Commercial Dipole Antenna – Omnidirectional Antenna
The radiating elements that have been used in this project are commercial
dipole antennas manufactured by D-Link Company. The antenna is designed to
operate at 2.4 GHz frequency band, and is aimed for WLAN applications. As this
antenna is bought from the market, no simulation is done for this antenna.
3.2.2 Development of the Butler Matrix
In this project, the designed 4 x 4 Butler Matrix consists of four 90° hybrid
coupler, two 0 dB crossover and two -45° phase shifter. Each component is designed
at the operating frequency of 2.4 GHz and simulated using Agilent ADS software.
The width and length of the each component is calculated using the theory of
microstrip patch antenna which can be computed by using the following equation.
Equation 3.6 and 3.7 gives the width of transmission line while equation 3.8 and 3.9
gives the length of transmission lines.
⎛ eH
1
− H
W = ⎜⎜
⎝ 8 4e
−1
⎞
⎟⎟ .h
⎠
(3.6)
⎛ Z 2(ε r + 1) ⎞ 1 ⎛ ε r − 1 ⎞⎛ ⎛ π ⎞
⎛ 4 ⎞⎞
⎟+ ⎜
⎟⎟⎜⎜ ln⎜ ⎟ + ln⎜ ⎟ ⎟⎟
H =⎜ 0
⎜
⎜
⎟ 2 ⎝ ε r + 1 ⎠⎝ ⎝ 2 ⎠
119.9
⎝ π ⎠⎠
⎝
⎠
λg =
ε eff
(3.7)
c
(3.8)
f ε eff
ε + 1 ⎛⎜
1 ⎛ ε r − 1 ⎞⎛ ⎛ π ⎞ 1 ⎛ 4 ⎞ ⎞ ⎞⎟
⎜
⎟⎜ ln⎜ ⎟ + ln⎜ ⎟ ⎟
1
= r
−
2 ⎜⎝ 2 H ⎜⎝ ε r + 1 ⎟⎠⎜⎝ ⎝ 2 ⎠ ε r ⎝ π ⎠ ⎟⎠ ⎟⎠
−2
(3.9)
78
Since there are few components need to be designed to construct the Butler
Matrix, the implementation of the Butler Matrix is divided into a few sections as
illustrated in Figure 3.10.
Design and simulate the hybrid coupler
Design and simulate 0 dB crossover
Design and simulate phase shifter
Integrate the whole block and simulate
Figure 3.10: The flow chart of the Butler Matrix implementation
3.2.2.1 The design of 90° Hybrid Coupler
As described in Chapter 2, Section 2.3.2.1, hybrid coupler has two variations
of impedance value which are Z0 at the shunt arm and Z0/ 2 at the series arm. By
using the equation 3.6 and 3.7 in previous section, the width dimension of the hybrid
coupler are calculated and shown in Table 3.5.
Table 3.5: Width value for each impedance value in hybrid coupler
Z0
2
Z (Ω)
Width (mm)
Z0 = 50
3.008
=
50
2
= 35.35
5.158
79
The guided wavelength of the transmission line is dependent on the operating
frequency, and after calculates using equation 3.8 and 3.9; the obtained value is
67.53 mm. The quarter wavelength of the hybrid coupler can be simply calculated by
dividing λg/4 which then gives a result of 16.88 mm. Figure 3.11 shows the
schematic and layout of the designed hybrid coupler.
(a)
(b)
Figure 3.11: Designed hybrid coupler a) Schematic, b) Layout
80
(a)
(b)
Figure 3.12: The simulated result of hybrid coupler a) Schematic b) Momentum
Figure 3.12 shows the simulated result of the amplitude and phase for the
schematic and momentum simulation. They show a good agreement with the theory
that described in Chapter 2, Section 2.3.2.1. The comparison of the result between
schematic and momentum is summarized and written in Table 3.6. From table below,
it can be concluded that schematic and momentum does not give a big different. The
phase differences between port 2 and 3 also almost 90°.
Table 3.6: The numerical result of simulated hybrid coupler
Input port 1
Schematic
Momentum
to other ports
Magnitude(dB)
Phase (°)
Magnitude(dB)
Phase(°)
Return Loss, S11
-28.7
Not relevant
-25.6
Not relevant
S21
-3.1
-135.5
-2.9
-138.3
S31
-3.5
134.8
-3.9
132.5
-49.0
Not relevant
-37.5
Not relevant
Coupling
Isolation, S41
81
3.2.2.2 The Design of 0 dB Crossover
As described in Chapter 2, Section 2.3.2.2, 0 dB crossover has three
variations of impedance value which are Z0 at the shunt arm, Z0/2 at the middle arm
and Z0/ 2 at the series arm. As the value of Z0 and Z0/ 2 already calculated in the
previous section, only the width of Z0/2 need to be calculated in this section. By
using equation 3.6 and 3.7, the width dimension of the middle arm is found to be
8.578 mm. The quarter wavelength of the 0 dB crossover remain same as the one
that calculated in previous section which is found to be 16.88 mm. Figure 3.13 shows
the schematic and layout of 0 dB crossover.
Figure 3.14 shows the simulated result of the amplitude and phase for the
schematic and momentum simulation. Same case as hybrid coupler, they also show a
good agreement with the theory that described in Chapter 2. The comparison of the
result between schematic and momentum is summarized and written in Table 3.7.
From table below, it can be seen that the simulation result for both method almost
identical except the value of S11 and S21. In the case of momentum simulation, the
graph for S21 is shifted to the right which causes the S21 value become higher about
10 dB at 2.4GHz. On the other hand, S11 performance is improved as the value is
decreases about 10 dB as well. Nevertheless, the value of S21 is in acceptable range
which determined to be lower than -10 dB.
82
(a)
(b)
Figure 3.13: Designed 0 dB crossover a) Schematic, b) Layout
83
(a)
(b)
(c)
(d)
Figure 3.14: The simulated result of amplitude and phase of 0 dB crossover
a) Schematic b) Momentum
Table 3.7: The numerical result of simulated hybrid coupler
Input Port1
Schematic
Momentum
Magnitude(dB)
Phase (°)
Magnitude(dB)
Phase(°)
Return Loss, S11
-26.0
Not relevant
-35.5
Not relevant
S21
-31.9
Not relevant
-21.2
Not relevant
S31
-0.6
-0.5
-0.8
0.4
-40.4
Not relevant
-39.2
Not relevant
Coupling
Isolation, S41
3.2.2.3 Phase Shifter
Phase shifter is implemented using microstrip transmission line delay, and the
length is calculated using equation 2.8 that described in chapter 2, section 2.3.2.3.
84
Figure 3.15 shows the schematic and layout of 45° phase shifter. The comparison of
the result between schematic and momentum is summarized and written in Table 3.8.
From table below, it can be seen that the simulation result for both method almost
identical.
(a)
(b)
Figure 3.15: Designed 45° phase shifter a) Schematic, b) Layout
Table 3.8: The numerical result of simulated 45° phase shifter
Input Port1
Coupling, S21
Schematic
Magnitude(dB)
Phase (°)
-0.5
-45.8
Momentum
Magnitude(dB)
Phase(°)
-0.6
-46.3
Figure 3.16 shows the schematic and layout of 0° phase shifter. 0° phase
shifter is needed to be placed at the output ports of Butler Matrix because of the
existence of the crossover between port 6 and 7. The comparison of the result
between schematic and momentum is summarized and written in Table 3.9. It can be
85
seen that the phase obtained from momentum simulation is slightly higher than
schematic simulation. This may be due to transmission line discontinuity at the
bending part.
(a)
(b)
Figure 3.16: Designed 0° phase shifter a) Schematic, b) Layout
Table 3.9: The numerical result of simulated 0° phase shifter
Input Port1
Coupling, S21
Schematic
Magnitude(dB)
Phase (°)
-0.4
0.8
Momentum
Magnitude(dB)
Phase(°)
-0.5
2.8
3.2.2.4 Construction of Butler Matrix
Figure 3.17 shows the block structure and layout of the Butler Matrix. The
Butler Matrix has four inputs 1R, 2L, 2R and 1L, and four outputs A1, A2, A3 and
86
A4. The four outputs are used as inputs to antenna elements to produce four beams.
Four hybrid couplers, two crossovers, two 0° phase shifter and two -45° phase shifter
were combined to produce the Butler Matrix. At the output ports, additional
transmission lines were placed at the input and output ports so that it follows the
design specification that is shown in Table 3.10. It was designed in such a way so
that when current excited to any input ports, the phase different between adjacent
output ports will only has one constant β. In fact, there are many parameters to
analyze the performance of Butler Matrix such as return loss at output and input
ports, isolation, coupling and input-output phase shift. As the main function for
Butler Matrix is to provide four different value of β which are -45º, +135º, -135º,
+45º, thus, only the simulated result of output phase of the Butler Matrix is taken into
consideration. The simulation graph about other parameters can be found in
Appendix B.
(a)
(b)
Figure 3.17: Butler Matrix configuration a) The block structure, b) layout
Table 3.10: Design Specification of the Butler Matrix
Ports
1R
2L
2R
1L
A1
0º
-90º
-45º
-135º
A2
-45º
45º
-180º
-90º
A3
-90º
-180º
45º
-45º
A4
-135º
-45º
-90º
0º
β
-45º
+135º
-135º
+45º
87
Table 3.11: The simulated output phase of Butler Matrix (schematic)
1R
2L
2R
1L
A1
1º
-90º
-44.4º
-134.1º
A2
-44.8º
42.2º
-179.7º
-93º
A3
-93º
-179.7º
42.2º
-44.8º
A4
-134.1º
-44.4º
-90º
1º
Target β
-45º
+135º
-135º
+45º
Table 3.12: Computed phase error (schematic simulation)
Port
1R
| Error|
2L
| Error|
2R
| Error|
1L
| Error|
β1 (A2-A1)
-45.8º
0.8º
132.2º
2.8º
-135.3º
0.3º
41.1º
3.9º
β2 (A3-A2)
-48.2º
3.2º
138.1º
3.1º
138.1º
3.1º
48.2º
3.2º
β3 (A4-A3)
-41.1º
3.9º
135.3º
0.3º
-132.3º
2.8º
45.8º
0.8º
Target
-45º
+135º
-135º
+45º
Table 3.13: The simulated output phase of Butler Matrix (momentum)
1R
2L
2R
1L
A1
-5.1º
-90.4º
-46.6º
-137.5º
A2
-45.4º
35.1º
-170.6º
-92º
A3
-92º
-170.6º
35.1º
-45.4º
A4
-137.5º
-46.6º
-90.4º
-5.1º
Table 3.14: Computed phase error (momentum simulation)
Port
1R
| Error|
2L
| Error|
2R
| Error|
1L
| Error|
β1 (A2-A1)
-40.3º
4.7º
125.5º
9.5º
-142º
7º
45.5º
0.5º
β2 (A3-A2)
-46.6º
1.6º
135.5º
0.5º
135.5º
0.5º
46.6º
1.6º
β3 (A4-A3)
-45.5º
0.5º
142º
7º
-125.5º
9.5º
40.3º
4.7º
Target
-45º
+135º
-135º
+45º
88
Table 3.11 and 3.13 show the simulated results (schematic and momentum
simulation) of the corresponding phase shifts between the inputs and the outputs of
the Butler Matrix. However, the phase values at the output ports are not significant.
The main concern about the result of the Butler Matrix is the phase differences
between the output ports. Thus, the phase differences between output ports for each
schematic and momentum simulation are calculated and the difference between the
obtained value and the target also being compared as shown in Table 3.12 and 3.14.
It can be seen that in the case of schematic simulation, the phase error occurs up to
3.9° while in momentum simulation case, the error occurs up to 9.5°. However,
according to the obtained result in [23], this error will not give a much impact to the
scanning angle of the radiation pattern as the error still less than 13.5°. The error in
scanning angle is approximately about 1° for progressive phase errors up to 13.5°
[23].
3.3
Prototype Fabrication
After designing and determining the dimensions and parameters of the each
component that described in previous section, it was then fabricated using wet
etching technique. The detail about the fabrication process can be found in [43]. For
square patch antenna type, a single antenna is also fabricated so that a radiation
pattern of a single square patch can be measured, which the data can be used later in
pattern multiplication theorem. For 4 x 2 antenna array, each of them was fabricated
onto 4 individual identical FR4 board. Figure 3.18 shows the fabricated prototype of
this project.
89
(a)
(b)
(c)
(d)
Figure 3.18: Fabricated prototype a) 4 x 1 square patch antenna, b) Single square
patch antenna, c) 4 x 2 antenna array, d) Butler Matrix
3.4
Measurement setup
3.4.1 S-parameter measurement
The return loss of each antenna and output phase of Butler Matrix is
measured using vector network analyzer. Measurement of the Butler Matrix is as
follows. The network analyzer is connected to a pair of ports of the Butler Matrix,
while other ports not connected to the network analyzer are terminated by matched
loads.
90
3.4.2 Radiation Pattern Measurement Method
The radiation characteristics of the beams are measured using far-field
method in the anechoic chamber. For this project, only radiation in E-plane is taken
into consideration which is the major radiation of multibeam antenna. Further more,
pattern multiplication that explained in Chapter 2, Section 2.2.3.1 only applied for Eplane. The details of chamber setup can be referred to [44]. The patterns of an
antenna can be measured in transmitting or receiving mode as the properties of
multibeam antenna is reciprocal. For this project, transmitting mode is selected for
measurement as this mode is easier to handle.
(a)
(b)
(c)
Figure 3.19: The configuration of the project a) 4x1 square patch antenna, b) 4 x
(4x2) antenna array, c) 4 x 1 dipole antenna.
91
All input ports are fed with the same signal but only one port is activated at
one time while the other ports are terminated with 50Ω. The configuration of the
Butler Matrix together with the antenna array is shown in Figure 3.19. All antennas
are arranged in a linear form and spaced at half-wavelength apart. During
measurement, antenna array is placed vertically and face towards the transmit
antenna while Butler Matrix is placed horizontally so that the radiation that caused
by the transmission lines of Butler Matrix will not affect the overall radiation pattern
performance.
3.5
Comparison of the designed Butler Matrix with other findings
As provided in Chapter 2, Section 2.4.2, there are 3 major areas that
highlighted by other researchers which are bandwidth, size reduction and dual band
operation. Table 3.15 shows the comparison of Butler Matrix between the design
with other findings (Table 2.15).
92
Table 3.15: The comparison between designed Butler Matrix and other findings
1. Comparison between obtained results with finding in [16]. The highlighted issue
was about the bandwidth of hybrid coupler in terms of return loss and isolation is
wider compared to conventional configuration by cascading two section of hybrid
coupler.
Designed Butler Matrix
Tayeb. A. Denidni and Taro Eric Libar
University of Quebec, Canada [16]
Hybrid Coupler configuration
Hybrid Coupler configuration
Simulated result of Hybrid coupler
Simulated result of Hybrid coupler
With reference level – 15 dB, BW = 7 %
With reference level -15 dB, BW = 13 %
2. Comparison between obtained results with finding in [17]. The highlighted issue
was about the bandwidth of hybrid coupler in terms of coupling ratio is wider by
applying matching network to all four ends of hybrid coupler.
Designed Butler Matrix
Manish A. Hiranandani, Ahmed A. Kishk
University of Mississippi [17]
Hybrid coupler configuration
Hybrid coupler configuration
Coupling ratio of hybrid coupler
Coupling ratio of hybrid coupler
Coupling value varies as frequency
Flat coupling ratio for the frequency range
varies
of 9.75 GHz to 10.25 GHz
93
3. Comparison between obtained results with finding in [18]. The highlighted issue
was about the bandwidth of hybrid coupler in terms of coupling ratio is wider
compared to conventional configuration by applying disk hybrid.
Designed Butler Matrix
S. Zheng, W.S. Chan, S.H. Leung and Q.
Xue [18]
Hybrid coupler configuration
Hybrid coupler configuration
Coupling ratio of hybrid coupler
Coupling ratio of hybrid coupler
For amplitude differences 0.5 dB,
For amplitude differences 0.5 dB,
BW = 5 %
BW = 25 %
4. Comparison between obtained results with finding in [19]. The highlighted issue
was about the constant phase over wide band was achieved by the proposed Butler
Matrix.
Designed Butler Matrix
Butler Matrix configuration
Hitoshi Hayashi, Donald A. Hitko,
Charles G. Sodini [19]
Butler Matrix configuration
Phase error correspond to Port 1
Phase error correspond to Port 1
Phase difference is varies as frequencies
Constant phase difference, from 0.85 to
varies.
0.9 GHz
94
5. Comparison between obtained results with finding in [20], [21]. The highlighted
issue was about size reduction by avoiding the used of crossover in the proposed
Butler Matrix.
Mourad Nedil, Tayeb A. Denidni and
Designed Butler Matrix
Larbi Talbi
University of Quebec, Canada [20][21]
Butler Matrix configuration
Butler Matrix configuration
6. Comparison between obtained results with finding in [22]. The highlighted issue
was about size reduction by reducing the size of hybrid coupler. The smaller hybrid
is realized by introducing 12 additional capacitive plates to the internal structures of
the hybrid couplers.
Designed Butler Matrix
Gary Kwang T.K, Peter Gardner
University of Birmingham [22]
Hybrid coupler configuration
Hybrid Coupler Configuration
Butler Matrix configuration
Butler Matrix configuration
7. Comparison between obtained results with finding in [23]. The highlighted issue
was about dual band Butler Matrix with the aim to cover 2.4 GHz and 5 GHz.
Carlos Collado1, Alfred Grau2 and Franco
Designed Butler Matrix
Butler Matrix configuration
De Flavis2
1
Technical University of Catalonia,
2
University of California [23]
Butler Matrix configuration
95
3.6
Chapter Summary
This chapter discusses the design of antenna array and Butler Matrix. The
schematic simulation and momentum simulation show almost similar result which
shows a good agreement between each others. At the end of the fabrication, the
multibeam antenna is tested to study the performance of the prototype antenna. The
results of the fabricated multibeam antennas are presented in next chapter. The
comparison between the designed Butler Matrix and the other findings (Table 2.15)
is presented as well.
CHAPTER 4
EXPERIMENTAL RESULT & DISCUSSION
In this chapter, the fabrication results such as return loss and radiation pattern
will be shown and analyzed. The measured radiation patterns were compared to the
theoretical calculation, to the result of related published paper and to the commercial
used antenna on WLAN AP.
4.1
Result of return loss of each type of antenna
As explained in previous chapter, there are three types of antenna that has been
used in this project which are:
¾ Square patch antenna
¾ 4 x 2 Antenna array
¾ Dipole antenna
In this section, the return loss of each antenna will be shown; the calculated
bandwidth and the comparison between simulation and measurement result are also
being observed.
97
As can be seen in Figure 4.1, all ports almost have same characteristics. If
compared to the one that obtained from simulation, the graph of measured return loss
is shifted to the right. This may happen due to the air that may trap between two
substrates, as the integration between them only made by tape. Although the
measured return loss of each port is slightly higher than -10 dB, in reality, this value
is still acceptable and reliable to obtain a good radiation pattern characteristic. The
obtained bandwidth is approximately about 4% which can be considered as moderate
bandwidth, this may be due to the selection of the substrate. The bandwidth can be
increased by using substrates that having different dielectric constant and height [45].
For 4 x 2 antenna array, the design of 4 x 2 antenna array has been fabricated
onto 4 identical FR4 board but the measured return loss of each of them slightly
different between each other. This may happen due to the imperfect fabrication
process that may happen during the process as it is not made by machine. However,
this change will not give much effect to the radiation pattern characteristics, and
these antennas still can be considered as identical of each other. For commercial
dipole antennas, the measured result shows good return loss and bandwidth for each
of them.
Square Patch Return Loss
Return Loss
0
-10
-20
-30
2
2.2
2.4
2.6
2.8
3
Frequency (GHz)
S11
S22
S33
S44
Figure 4.1: Measured return loss of square patch antenna correspond to each port
98
Table 4.1: The numerical result of square patch antenna
Port1
Port2
Port3
Port4
Return Loss at 2.4GHz (dB)
-11.5
-12.687
-9.603
-9.613
fl (GHz)
2.39
2.38
2.41
2.41
fu (GHz)
2.49
2.48
2.50
2.51
BW (%)
4.10
4.12
3.67
4.07
Antenna Array Return Loss
0
S11 (dB)
-10
-20
-30
-40
-50
2
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
3
Frequency (GHz)
Array1
Array2
Array3
Array4
Simulation
Figure 4.2: Measured return loss of each 4 x 2 antenna array
Table 4.2: The numerical result of 4 x 2 antenna array
Simulation
Array1
Array2
Array3
Array4
Return Loss at 2.4GHz (dB)
-23.30
-13.35
-11.91
-11.69
-11.80
fl (GHz)
2.21
2.33
2.30
2.33
2.29
fu (GHz)
2.44
2.43
2.56
2.42
2.44
BW (%)
9.90
4.20
10.71
3.79
6.35
99
Return Loss of Dipole Antenna
0
S11 (dB)
-10
-20
-30
-40
-50
2
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
3
Frequency (GHz)
Dipole1
Dipole2
Dipole3
Dipole4
Figure 4.3: Measured return loss of each dipole antenna
Table 4.3: The numerical result of dipole antenna
4.2
Antenna1
Antenna2
Antenna3
Antenna4
Return Loss at 2.4GHz (dB)
-15.45
-21.35
-13.9
-11.9
fl (GHz)
2.35
2.31
2.37
2.39
fu (GHz)
2.61
2.59
2.49
2.53
BW (%)
10.50
11.45
4.94
5.70
Result of Butler Matrix
Table 4.4 shows the measured results of the corresponding phase shifts
between the inputs and the outputs of the Butler Matrix. As described previously in
Chapter 3, Section 3.2.2.4, the main concern about the result of the Butler Matrix is
the phase differences between the output ports, not the value of phase at the output
ports. Thus, the phase differences between output ports are calculated and the
difference between the obtained value and the target also being compared as shown
in Table 4.5. It can be observed that the error occurs up to 23°. If compared to the
100
ones that obtained from the simulation (refer to Table 3. 14), the highest phase error
also happen when port 2L and 2R become an input port. Imperfect fabrication
process and incorrect soldering technique may contribute to the increment of the
phase errors of the measurement result.
Table 4.4: The measured output phase of Butler Matrix
Port
A1
A2
A3
A4
1R
-82º
-130º
-169º
145º
2L
-166º
-53º
84º
-125º
2R
-128º
82º
-54º
-166º
1L
141º
-172º
-131º
-82º
Table 4.5: Computed phase error (measurement)
Port
1R
| Error|
2L
| Error|
2R
| Error|
1L
| Error|
4.3
β1 (A2-A1)
-48º
3º
113º
22º
-150º
15º
48º
3º
β2 (A3-A2)
-39º
6º
137º
2º
-136º
1º
41º
4º
β3 (A4-A3)
-46º
1º
151º
16º
-112º
23º
48º
3º
Target
-45º
+135º
-135º
+45º
Radiation pattern
At first, the radiation pattern of the single antenna is measured. The obtained
radiation pattern of the single antenna can be used later to calculate the predicted
radiation pattern of the integrated project. The measured radiation pattern of the
single antenna for each type of antenna is shown in Figure 4.4.
101
(a)
(b)
(c)
Figure 4.4: Measured radiation pattern of single antenna a) Square patch antenna, b)
4 x 2 antenna array, c) Dipole antenna
By referring to Figure 4.4, it can be seen that the radiation pattern of square
patch antenna has a broader HPBW which is about 89° while 4 x 2 antenna array
provide a directional pattern with HPBW approximately about 27°. On the other
hand, dipole antenna shows an omni direction pattern which receives power almost
equal to all direction. Each of 4 x 2 antenna array and dipole antennas has an
identical measured radiation pattern.
102
For 4 x 2 antenna array case, it can be observed that there is presently a slight
difference between simulated and measured radiation patterns. The beamwidth of the
main lobes are similar for both simulated and measured result. The side lobes levels
and beamwidth size for measured results are slightly higher than those simulated
results. This is may be due to the fabrication process that cause the occurrence of a
mutual coupling and mismatch between feeding network and antenna.
Figure 4.5, 4.6 and 4.7 illustrates the measured radiation patterns of port 1R,
2L, 2R and 1L of each type of antenna while Table 4.6, 4.7 and 4.8 shows the
numerical value corresponds to each radiation pattern.
(a)
(b)
(c)
(d)
Figure 4.5: Measured radiation patterns of using square patch a) Port 1R, b) Port
2L, c) Port 2R, d) Port 1L
103
Table 4.6: Numerical result of measured radiation patterns of using square patch
Port -3 dB beamwidth Beam Angle Maximum SLL (dB)
1R
32.5°
+14.5°
-8.0
2L
10.3°
-40.6°
-10.0
2R
18.0°
+43.0°
-9.0
1L
31.6°
-14.0°
-7.9
(a)
(c)
(b)
(d)
Figure 4.6: Measured radiation patterns of using 4 x 2 antenna array a) Port 1R, b)
Port 2L, c) Port 2R, d) Port 1L
104
Table 4.7: Numerical result of measured radiation patterns of using 4 x 2 antenna
array
Port -3 dB beamwidth Beam Angle Maximum SLL (dB)
1R
6.2°
+4.4°
-9.1
2L
6.6°
-10.0°
-3.7
2R
5.1°
+11.8°
-5.6
1L
6.8°
-2.7°
-9
(a)
(b)
(c)
(d)
Figure 4.7: Measured radiation patterns of using dipole antenna a) Port 1R, b) Port
2L, c) Port 2R, d) Port 1L
105
Table 4.8: Numerical result of measured radiation patterns of using dipole antenna
Port -3 dB beamwidth
Beam Angle
Maximum SLL (dB)
1R
20°, 40°
+18°, +157°
-6.0
2L
29.4°
-90°
-3.0
2R
67.5°
+90°
-5.0
1L
26.0°, 40°
-16°, -155°
-4.3
With reference to Figure 4.6, four different beams at four different directions
(+14.5°, -40.6°, +43.0°, -14.0°) have been generated by exciting signal at one port at
instant. Because of the variation of 3 dB levels, the beamwidth size of beam 2L and
2R seems narrower compare to beam 1R and 1L, but if the beamwidth size is not
taken at exactly – 3 dB levels, by observation, all beams has a beamwidth size
approximately about 30°.
For the case of using 4 x 2 antenna array, four different beams at four
different directions were also generated but the tilted angle of each beams only a few
degrees different between adjacent beams (+4.4°, -10.0°, +11.8°, -2.7°). The
beamwidth size of each beams also very narrow which approximately about 7°.
Some more, a few side lobe with a higher amplitude level also exists particularly at
beam 2L and 2R.
For the case of using dipole antennas, it can be observed that when port 1R is
excited, two main beams appeared and directed to the upper, +18º and bottom part,
+157º of the polar plot. Similar thing happen to port 1L where two main beams also
appeared at upper part, -16º, and bottom part, -155º of the polar plot. The beamwidth
size of the main beams is narrower while for port 2L and 2R, the beamwidth size of
the main lobe is broader.
Figure 4.8 shows the overlapping radiation pattern of the Port 1R, 2L, 2R and
1L on the same plot for each type of antenna. It can be shown that, in the case of
106
square patch antenna, the system capable to cover up to 120º of coverage area, in the
case of 4 x 2 antenna array, it covers up to 30º of coverage area while in the case of
dipole antenna, it covers up to 360º of coverage area. It can be concluded that, by
applying different kind of antenna array to the 4 x 4 Butler Matrix, the obtained
radiation patterns has different characteristics mainly in terms of beamwidth and
coverage area, but the directions of main beam still appear 2 at the right side and 2 at
the left side of polar plot for all cases.
(a)
(b)
(c)
Figure 4.8: Overlapped radiation pattern a) Square patch antenna, b) 4 x 2 Antenna
array, c) Dipole antenna
107
The entire radiation patterns above are taken based on E-plane as this field
will determine the polarization or orientation of the radio wave. For the case of Hplane, only square patch antenna is taken into consideration that can be found in
Appendix C. The radiation pattern when multiple input ports are activated also can
be found in Appendix D.
4.4
Result Analysis
The radiation pattern of the project can be predicted using pattern
multiplication theorem, equation 2.4, as written in the Chapter 2, Section 2.2.3.1. The
data of a single array unit pattern is obtained from the measured result of the single
antenna (Figure 4.4) while the data of the array factors can be calculated using
equation 2.8, also written in Chapter 2 as well, which then expanded as follows:
⎛ N (kd sin θ + β ⎞
sin ⎜
⎟
2
⎠
⎝
AF =
N (kd sin θ + β )
2
(4.1)
By substituting N = 4, k = 2π/λ and d = λ/2 the above equation, it can be simplified
as:
AF =
sin (2(π sin θ + β ))
2(π sin θ + β )
(4.2)
Since there are four different values of ß, four different equation will be obtained
when substituting β = -45º, +135º, -135º, +45º to the equation 4.2 and the result is
written in Table 4.9.
108
Table 4.9: AF equations correspond to each β
β
AF
β
AF
-45º
AF =
sin (2π (sin θ − 0.25))
2π (sin θ − 0.25)
-135
AF =
sin (2π (sin θ − 0.75))
2π (sin θ − 0.75)
+45º
AF =
sin (2π (sin θ + 0.25))
2π (sin θ + 0.25)
+135º
AF =
sin (2π (sin θ + 0.75))
2π (sin θ + 0.75)
The computed radiation pattern of AF can be plotted by substituting θ from 0º
to 360º to the equation above. The computed radiation pattern of AF corresponds to
each β can be shown in Figure 4.9.
(a)
(c)
(b)
(d)
Figure 4.9: The computed radiation pattern of AF corresponds to each β (a) β = -45º,
(b) β = +135º, (c) β = -135º, (d) β = +45º
109
However, this plotted AF is in linear value, which varies from 0 to 1. Since
the measured radiation pattern is taken in dB, so it is necessary to convert the
computed AF into dB unit. This can be done by computing through the following
equation:
AF (dB ) = 10 log( AF (linear ) )
(4.3)
Figure 4.10 shows the plotted computed radiation pattern of AF in dB unit.
(a)
(c)
(b)
(d)
Figure 4.10: Computed radiation pattern of AF, (a) β = -45º, (b) β = +135º, (c) β = -
135º, (d) β = +45º
Next, the radiation patterns of the built project can be predicted using
equation 2.4. Since there are three types of antenna that has been used in this project,
thus the analysis also divided into 3 major parts:
110
1) The analysis of using 4x1 square patch antenna
2) The analysis of using 4 x 4x2 Antenna Array
3) The analysis of using omnidirectional antenna
In order to compute the predicted radiation pattern, the radiation pattern of
the single element array is needed. In this case, the data for the single radiation
pattern is obtained from the measurement. Since the plotted AF and the measured
data already converted in dB unit, thus the multiplication operation should be
changed to plus (+) operation as well. This can be clearly seen if this term is written
in the mathematical form as shown in the following figure.
Array pattern (linear) = Single Unit pattern (linear) x Array Factor
Array Pattern (dB) = Single Unit pattern (dB) + Array Factor (dB)
Figure 4.11: Conversion of array pattern from linear unit to dB
The multiplication of the project of each type of antenna is illustrated in
Figure 4.12, 4.13 and 4.14. The comparison between computed and measured
radiation pattern is shown in Figure 4.15, 4.16 and 4.17.
111
Radiation pattern of
single element
Plotted AF
Multiplication Result
¾ β = -45º
¾ β = +135º
¾ β = -135º
¾ β = +45º
Figure 4.12: Pattern multiplication of square patch antenna case
112
Radiation pattern of
single element
Plotted AF
Multiplication Result
¾ β = -45º
¾ β = +135º
¾ β = -135º
¾ β = +45º
Figure 4.13: Pattern multiplication of 4 x 2 antenna array case
113
Radiation pattern of single
element
Plotted AF
Multiplication Result
¾ β = -45º
¾ β = +135º
¾ β = -135º
¾ β = +45º
Figure 4.14: Pattern multiplication of dipole antenna case
114
(a)
(b)
(c)
(d)
Figure 4.15: Radiation pattern comparison between computed and measured result
(square patch antenna case) a) Port 1R, b) Port 2L, c) Port 2R, d) Port 1L
115
(a)
(c)
(b)
(d)
Figure 4.16: Radiation pattern comparison between computed and measured result
(4 x 2 antenna array case) a) Port 1R, b) Port 2L, c) Port 2R, d) Port 1L
116
(a)
(b)
(c)
(d)
Figure 4.17: Radiation pattern comparison between computed and measured result
(dipole antenna case) a) Port 1R, b) Port 2L, c) Port 2R, d) Port 1L
For the case of square patch antenna, it can be observed that the measured
radiation has a similar pattern to the computed pattern. This is verified that the
experiment is reliable as it has a good agreement with the theoretical calculation.
117
For the case of 4 x 2 antenna array, it can be observed that the calculated
patterns have a bigger beamwidth compared to measured result. The direction of
beams of 1R and 1L are similar to the calculated patterns which are directed to 5º and
-2º respectively. The main beam of port 2R and 2L are directed to 13º and -9º
respectively, which are a little bit different from the computed patterns. Computed
patterns show that the main beam appears at the centre of the polar plot with lower
magnitude as the beam position of the array factor, 2R and 2L are directed to 48.6º
and -48.6º. The measured result shows a narrower beamwidth and more side lobes.
This may be due to non-uniform surface of the antenna holder which is then caused
the distance between 4 x 2 antenna array is higher than λ/2. As described in Chapter
2, Section 2.2.3.1, side lobes will be appeared when distances between elements is
increases.
For the case of dipole antenna, it can be observed that the measured radiation
has a similar pattern to the computed pattern. The little differences may be due to the
misalignment of the rotator inside the chamber.
As a conclusion, the entire measured result shows a good agreement with the
theoretical calculation.
4.5
Comparison of the measured radiation pattern with other findings
As provided in Chapter 2, Section 2.4.1, the discussion of four ports
conventional Butler Matrix incorporated with 4 linear antenna arrays has been
presented in several papers which are [24], [25], [26] and [27]. In this section, the
obtained radiation pattern of the design with dipole antenna is compared to [24] as
the author used the same type of antenna, omnidirectional antenna as an antenna
array. The result of using square patch antenna is compared to [25], [26] and [27] as
118
the author used the microstrip antenna as well. Table 4.10 shows the comparison of
measured radiation pattern between the design with other finding (Table 2.14). It can
be seen that the measured radiation pattern has not much different compared to the
obtained result by other researchers.
Table 4.10: The comparison between measured radiation patterns of the design with
other findings
1. Comparison between obtained results using dipole antenna array with finding in
[24]. The used antenna type is CPW-fed uniplanar monopole antenna.
Ying Jung Chang and Ruey Bing Hwang
Measured radiation patterns
National Chiao-Tung University, Taiwan
[24]
Port 1
Port 2
Port 3
119
Port 4
2. Comparison between obtained results using square patch antenna array with
finding in [25]. The used antenna type is bow tie antenna.
Measured radiation patterns
Dau-Chyrh Chang and Shin-Huei Jou,
Da Yeh University [25]
Angle: -40º, -12º, 12º, 40º
Angle : -40.6°, -14.0°, 14.5°, 43.0°
3. Comparison between obtained results using square patch antenna array with
finding in [26]. The used antenna type is rectangular patch antenna with coaxial
probe as a feeding technique.
Siti Rohaini Ahmad, Fauziahanim Che
Seman,
Measured radiation patterns
Kolej Universiti Teknologi Tun Hussein
Onn,
Malaysia [27]
Port 1 (1R)
120
19°
14.5°
Port 2 (2L)
unidentified
-40.6°
Port 3 (2R)
unidentified
43.0°
Port 4 (1L)
-19°
-14.0°
4. Comparison between obtained results using square patch antenna array with
finding in [27]. The used antenna type is rectangular patch antenna with inset feed as
a feeding technique.
Measured radiation patterns
Nhi T. Pham1, Gye-An Lee2 and Franco De
Flaviis1
121
1
University of California, 2Skyworks
Solution Inc. [26]
Port 1 (1R)
16º
14.5°
Port 2 (2L)
-39º
-40.6°
Port 3 (2R)
38º
43.0°
Port 4 (1L)
-14.0°
-15º
122
4.6
The comparison between commercially used antenna with designed
multibeam antenna
The benefit of implementing multibeam antenna on WLAN such as
improvement of CIR and BER has been discussed in Chapter 2, Section 2.1.2.3
However, these benefits over traditional AP could not be demonstrated in this project
as multibeam antenna is not a stand alone device, it needs a processing part to select
which is the best beam to operate at a given time. Thus, the comparison of designed
antenna with commercially used antenna for WLAN AP can be made based on its
radiation pattern characteristics only.
Table 4.11: The comparison between commercially used antenna with designed
multibeam antenna
Antenna
Radiation pattern
Commercial
antenna for
WLAN AP
Use Single antenna omnidirectional antenna
Max received power: -26 dB
Coverage : 360º
Design with
square
patch
antenna
array
Max received power: -26 dB
Coverage : 120º
123
Advantage over traditional AP:
¾ It has multiple directional beams, which may reduce co-
channel interference as the beam only focus to the desired user
location and null to undesired user.
Disadvantage:
¾ It only cover 120º angular area, thus to cover 360º, it needs
total 3 (360º / 120º) multibeam antenna
Other issue:
¾ Although multibeam antenna comprises of multiple antennas,
the maximum received power is almost similar to the
traditional AP due to the loss that contributed by Butler Matrix
Max received power: -20 dB
Coverage : 30º
Design with
4x2
antenna
array
Advantages over traditional AP:
¾ It has multiple directional beams, which may reduce co-
channel interference as the beam only focus to the desired user
location and null to undesired user.
¾ It has a higher gain as the maximum received power is higher
than single omnidirectional antenna
Disadvantages:
¾ It only cover 30º angular area, thus to cover 360º, it needs total
12 (360º / 30º) multibeam antenna
¾ If no improvement is done on the Butler Matrix, side lobes will
contribute to the increment of co-channel interference level
124
Max received power: -27 dB
Design with
dipole
antenna
array
Coverage : 360º
Advantages over traditional AP:
¾ It has multiple beams that directed to different angle, which
may reduce co-channel interference as the beam only focus to
the desired user location and null to undesired user.
¾ It has same coverage area as provided by traditional AP (360º)
Disadvantage:
¾ It has broader beamwidth for beam 2R and 2L and 2 main
beams for port 1L and 1R, thus less co-channel interference
could be reduced compared to square patch antenna and 4 x 2
antenna array.
4.7
Chapter Summary
This chapter has presented the result of fabricated multibeam antenna in
terms of return loss for each part and radiation pattern of the integration of the
multibeam antenna. The theoretical calculation of array pattern also presented and it
correlates well with the measured result. The comparison between the designed
multbeam with other findings, and the comparison between designed antenna and
commercially used antenna are presented as well.
CHAPTER 5
CONCLUSION AND FUTURE WORK
5.1
Conclusion
The design of multibeam antenna constructed by antenna array and 4 x 4
Butler Matrix at 2.4 GHz has been presented. 3 types of antenna array have been
used in this project which are square patch antenna, 4 x 2 planar antenna array and
dipole antenna. They have been proved that through measurement each of them
produces different kind of radiation patterns which are broader beamwidth for square
patch antenna, directional beam for 4 x 2 antenna array and omnidirectional pattern
for dipole antenna. With the existence of Butler Matrix, it has been proved that four
independent beams with four different directions have been generated.
In the case of square patch antenna, by integrating it with Butler Matrix, four
different beams have been generated, each of them directed at different angle which
are at +14.5°, -40.6°, +43.0° and -14.0° respectively. Each beam has approximately
about 30° HPBW, and total angular coverage is up to 120°. In the case of 4 x 2
antenna array, by integrating it with Butler Matrix, four different beams at four
different directions were also generated but the tilted angle of each beams only a few
degrees different between adjacent beams which are at +4.4°, -10.0°, +11.8° and 2.7° respectively. The beamwidth size of each beams also very narrow which
126
approximately about 7°, which then gives a small coverage area about 30°. For the
case of using dipole antennas, two types of beam can be generated which are narrow
beam and broader beam. The beamwidth size of the main beams is narrower while
for port 2L and 2R, the beamwidth size of the main lobe is broader. This feature
gives a result of 360° of coverage area.
Each of designed multibeam antennas has its own advantages and
disadvantages. In overall, all designed could reduce co-channel interference as each
of them has multiple beams that directed to different angle, which only focus to the
desired user location and null to undesired user. However, because the beamwidth
size is difference between each designed, the ability to suppress interference could be
varied as well. By using 4 x 2 antenna array, the beamwidth of the main lobe may be
narrow but it has more side lobes compare to others, which may cause interference
higher for this antenna. Same thing happen if use dipole antenna as an antenna array,
the beams that produced by this antenna is broader for port 2L and 2R and has 2
main beams for port 1R and 1L. Although the beamwidth is narrower compare to
omnidirectional antenna that used by traditional AP, this antenna still has less
capability to reduce co channel interference. As a conclusion, based on the measured
radiation pattern that obtained from this project, square patch antenna array produce
less side lobes and suitable to reduce co-channel interference issue on WLAN
environment.
As an overall conclusion, the objectives of this project have been successfully
implemented and achieved which are to design an antenna system that capable to
produce multiple beams of radiation pattern. Further more, the radiation
characteristics of multibeam antenna are compared between the theoretical and
measured result, and they correlates well.
127
5.2
Proposed Future Works
Further works should be carried out in order to improve the performance of
the multibeam antenna as suggested as below:
i.
Different beamforming network can be used such as Blass Matrix so that
its performance can be compared to the Butler Matrix
ii.
The development of Butler Matrix as presented in Chapter 2, Section
2.4.2 can be applied in future so that the phase errors contributed by
Butler Matrix can be reduced.
iii.
By using conventional Butler Matrix, it has been shown that side lobe
level is higher than -10 dB. Thus, techniques that proposed in [46]-[48]
should be studied also so that side lobes level of multibeam antenna can
be reduced.
128
REFERENCES
1. Yazd, M. H. and Faraji-Dana, R. Novel Smart Antenna Array for IEEE 802.11a
Wireless LAN Applications. IEEE Antenna and Propagation Society
International Symposium. July 3 – 8, 2005. IEEE. 2005. 295 – 298.
2. Yang-Seok Choi and Alamouti, S. M. Performance Analysis and Comparisons of
Antenna and Beam Selection Diversity. IEEE 60th Vehicular Technology
Conference. September 26 – 29, 2004. IEEE. 2004. 165 – 170.
3. Desmond, N. C. T. Smart Antenna for Wireless Applications and Switched
Beamforming. B. Eng. Project Report. University of Queensland; 2001
4. Salvatore, B., Balanis, C. A., Jeffrey, F. and Spanias, S. A. Smart-Antenna
Systems for Mobile Communication networks, Part 1: Overview and Antenna
Design. IEEE Antenna’s and Propagation Magazine. 2002. 44 (3): 145 – 154.
5. Alpesh U. B. and Patrick, L. P. An Overview of Smart Antenna Technology for
Wireless Communication. Aerospace Conference. March 10-17, 2001. IEEE.
2000. 2/875 – 2/883.
6. Bachman, H.L. Smart Antennas – The Practical Realities. IEEE Aerospace
Conference. February 1-8, 1997. IEEE. 1997. 63 – 70.
7. Alastalo, A. T. and Kahola, M. Smart Antenna Operation for Indoor Wireless
Local Area Networks using OFDM. IEEE Trans. On Wireless Communication.
2003. 2 (2). 392 – 399.
8. Seong-Sik Jeon, Yuanxun Wang, Yongxi Qian, Tatsuo Itoh. A Novel Smart
Antenna System Implementation for Broadband Wireless Communications. IEEE
Transaction on Antennas and Propagation. 2002. 50 (50): 600 – 606.
9. L. J. Pesik, M. A. Beach, D. P. McNamara and P. N. Fletcher. Performance
Analysis of Smart Antenna Systems for Indoor Wireless LANs. 3G Mobile
Communication Technologies. May 8-10, 2002. IEEE. 2002. 418 – 442.
129
10. Cladwell, F., Jr., Kenney, J. S. and Ingram, I. A. Design and Implementation of a
Switched-Beam Smart Antenna for an 802.11b Wireless Access Point. Radio and
Wireless Conference. August 11-14, 2002. IEEE. 2002. 55 – 58.
11. Lal, D., Jain, V. Zheng. Q.A, and Agrawal, D.P. Performance Evaluation of
Medium Access Control for Multiple-Beam Antenna Nodes in a Wireless LAN.
IEEE Transactions on Parallel and Distributed Systems. 2004. 15 (12): 11171129.
12. Shao-Hua Chu, Hsin-Piao Lin and Ding Bing Lin. Performance Enhancement by
using Switched Beam Smart Antenna in 802.11a WLAN System. Wireless
Communications and Applied Computational Electromagnetic International
Conference. April 3-7, 2005. IEEE. 2005. 1001 – 1005.
13. Jianfeng Wang, Yuguang Fang and Dapeng Wu. Uplink Medium Access Control
for WLAN with Multi-Beam Access Point. IEEE Global Telecommunications
Conference. November 28 – December 2, 2005. IEEE. 2005. 3012 – 3016.
14. Bevan, D. D. N. Beam Falsing in Switched Beam Smart Antenna Systems.
Vehicular Technology Conference. May 18-21, 1998. IEEE. 1998. 1034 – 1038.
15. Namboodiri, V., Lixin Gao and Janaswamy, R. Power Efficient Topology
Control for Wireless Networks with Switched Beam Directional Antennas. IEEE
International Conference on Mobile Adhoc and Sensor Systems. November 7 –
10, 2005. IEEE. 2005.
16. Tayeb A. Denidni and Taro Eric Libar. Wide Band Four-Port Butler Matrix for
Switched Multibeam Antenna Arrays. The 14th IEEE 2003 International
Symposium on Personal, Indoor and Mobile Radio Communication Proceedings.
September 7 – 10, 2003. IEEE. 2003. 2461 - 2464
17. Hiranandani, M. A. and Khishk, A. A. Widening Butler Matrix bandwidth within
X-band. Antennas and Propagation Society International Symposium. July 3-8,
2005. IEEE. 2005. 321 – 324.
18. Zheng, S., Chan, W. S., Leung, S. H. and Xue, Q. Broadband Butler Matrix with
flat coupling. Electronic Letters. 2007. 43 (10): 576 – 577.
19. Hayashi, H., Hitko, D. A., and Sodini, C. G. Four-Element Planar Butler Matrix
Using Half-Wavelength Open Stubs. IEEE Microwave and Wireless Components
Letters. 2002. 12 (3): 73-75.
130
20. Nedil, M., Denidni, T. A., and Talbi, L. Novel Butler Matrix using CPW multilayer Technology. Antennas and Propagation Society International Symposium.
July 3 – 8, 2005. IEEE. 2005. 299 – 302.
21. Nedil, M., Denidni, T. A., and Talbi, L. Novel Butler Matrix using CPW multilayer Technology. IEEE Transaction on Microwave Theory and Techniques.
2006. 54 (1): 499 – 507
22. Kwang, T. K. Gary and Gardner, Peter. 4 x 4 Butler Matrix Beamforming
Network using Novel Reduced Size Branchline Coupler. European Microwave
Conference. October 31, 2001. IEEE. 2001. 1 – 4.
23. Collado, C., Grau, A. and De Flaviis, F. Dual Band Butler Matrix for WLAN
systems. IEEE MTT-S International Microwave Symposium Digest. June 12 – 17,
2005. IEEE. 2005. 2247 – 2250.
24. Ying Jung Chang and Ruey Bing Hwang. Switched Beam System for low-tier
Wireless Communication Systems. Asia Pacific Microwave Conference.
December 3 – 6, 2001. IEEE. 2001. 946 – 949.
25. Dau-Chyrh Chang and Shin-Huei Jou. The study of Butler Matrix BFN for four
beams antenna system. Antennas and Propagation Society International
Symposium. June 22 – 27, 2003. IEEE. 2003. 176 – 179.
26. Pham, N. T., Gye-An Lee and De Flaviis, F. Microstrip Antenna Array with
Beamforming Network for WLAN. Antennas and Propagation Society
International Symposium. July 3 – 8, 2005. IEEE. 2005. 299 – 302.
27. Ahmad, S. R and Seman, F. C. 4-port Butler Matrix for Switched Multibeam
Antenna Array. Asia Pacific Conference on Applied Electromagnetics. December
20 – 21, 2005. IEEE. 2005. 69 -73
28. Lal C. Godara. Applications of Antenna Array to Mobile Communications, Part
I: Performance Improvement, Feasibility, and System Considerations.
Proceedings of IEEE. July, 1997. IEEE. 1997. 1031 – 1060.
29. Tsoulus, G. V. Smart Antenna for mobile communication systems: benefit and
challenges. Electronics & Communication Engineering Journal. 1999. 11 (2): 84
- 94.
30. Institute of Electrical and Electronics Engineers. IEEE standard definition of
terms for antenna. IEEE STD 145-1983. 1983.
131
31. Siachalou, E., Vafiadis, E., Goudos, S. S., Samaras, T., Koukourlis, C. S. and
Panas, S. On the design of Switched-beam wideband base stations. IEEE
Antennas and Propagation Magazine. 2004. 46 (1): 158 – 167.
32. Kraus, J. D. and Marthefka, R. J. Antenna for All Applications.3rd ed. Singapore:
McGraw Hill. 2003.
33. Balanis, C. A. Antenna Theory. 3rd ed. Hoboken, N. J.: John Wiley & Sons. 2005.
34. Chang, K. RF and Microwave Wireless Systems. New York: John Wiley & Sons.
2000.
35. Golio, C. M. The RF and Microwave Handbook. Boca Raton: CRC Press LLC.
2001.
36. Hall, P. S. and Vetterlein, S. J. Review of Radio Frequency Beamforming
Techniques for Scanned and Multiple Beam Antennas. IEEE Proceedings.
October, 1990. IEEE. 1990. 293 – 303.
37. Blass, J. Multidirectional antenna – New Approach to Stacked Beams. IRE
International Convention Record. March, 1960. IEEE. 1960. 48 – 50
38. Koubeissi, M., Decroze, C., Monediere, T. and Jecko, B. Switched-beam
Antenna based on Novel Design of Butler Matrices with Broadside Beam.
Electronics Letter. 2005. 41 (20): 1097 – 1098.
39. Denidni, T. A. and Libar, T. E. Experimental Investigation of a Microstrip Planar
feeding network for a Switched-Beam Antenna Array. IEEE Antennas and
Propagation Society International Symposium. June 16 – 21, 2002. IEEE. 2002.
130 – 133.
40. Pozar, D. M. Microwave Engineering. 3rd ed. Hoboken, N. J.: John Wiley &
Sons. 2005.
41. Hasan, E. Microstrip Antenna Array for Wireless Local Area Networks
Applications. M. Eng. Project Report. Universiti Teknologi Malaysia; 2007
42. Mayer, B. and Knochel, R. Branchline Coupler with improved design flexibility
and broad bandwidth. IEEE International Microwave Symposium Digest. May 8 10, 1990. IEEE. 1990. 391 - 194
43. Mariadoss, P. Q. Design of a Beamforming Network using Butler Matrix. B. Eng.
Project Report. Universiti Teknologi Malaysia; 2005.
44. Asrokin, A. Design and Development of Dual Band Microstrip Antenna using
Scaling Factor and Inset Feed for Wireless Local Area Network Application. M.
Eng. Thesis. Universiti Teknologi Malaysia; 2007.
132
45. Splitt, G. and Davidovitz, M. Guidelines for design of electromagnetically
coupled microstrip patch antennas on two-layer substrates. IEEE Transactions on
Antennas and Propagation. 1990. 38 (7): 1136 -1140.
46. Li, W.-R., Chu, C. –Y., Lin, K. –H. and Chang, S. – F. Switched beam antenna
based on modified Butler Matrix with low side lobe level. Electronics Letters.
2004. 40 (5): 290 – 292.
47. Gruszezynski, S., Wineza, K. and Sachse, K. Reduced side lobe four-beam Nelement antenna array fed by 4 x 4 Butler Matrix. Antennas and Wireless
Propagation Letters. 2006. 5 (1): 430 – 434.
48. Wineza, K., Gruszezynski, S. and Sachse, K. Reduced side lobe four-beam
antenna array fed by modified Butler Matrix. Electronic Letters. 2006. 42 (9):
508 -509.
49. Mohamad A. Haidar. Dynamic Load Balancing and Channel Allocation in
Indoor Wireless Local Area Networks. Ph.D Tesis. University of Arkansas; 2005
50. Institute of Electrical and Electronics Engineers. IEEE 802.11b. IEEE 802.11b1999. 1989.
51. Malaysian Communications and Multimedia Commission (MCMC). Spectrum
Plan. Kuala Lumpur (Malaysia). 1st Edition. 2002.
133
APPENDIX A
134
APPENDIX B
Schematic Simulation Result of Butler Matrix
1) Return Loss and Isolation from each input port.
a) Input Port 1
b) Input Port 2
c) Input Port 3
d)Input Port 4
2) Coupling value from each input port
a) Input Port 1
b) Input Port 2
c) Input Port 3
d)Input Port
135
Momentum Simulation Result of Butler Matrix
1) Return Loss and Isolation from each input port.
a) Input Port 1
b) Input Port 2
c) Input Port 3
d)Input Port 4
2) Coupling value from each input port
a) Input Port 1
b) Input Port 2
c) Input Port 3
d)Input Port 4
136
APPENDIX C
H-Co Measured radiation pattern for square patch antenna case
a) Port 1R
b) Port 1L
c) Port 2R
d) Port 2L
137
APPENDIX D
E-Co Measured radiation pattern for square patch antenna case when multiple
inputs activated simultaneously
a) Port 1R & 2L
b) Port 1R and 2R
c) Port 2R & 1L
d) Port 1L & 2L
138
e) Port 2R & 2L
f) Port 1R & 1L
g) Port 1R & 1L & 2R & 2L
139
APPENDIX E
MULTIBEAM ANTENNA ARRAY WITH BUTLER MATRIX
FOR WLAN APPLICATIONS
1
S .Z. Ibrahim 1M.K.A.Rahim, 1T Masri, 1M.N.A.Karim, 2M.Z.A.Abdul Aziz
1
Wireless Communication Centre, Faculty of Electrical Engineering,
Universiti Teknologi Malaysia, 81310 Skudai Johore Baharu.
[email protected], [email protected], [email protected], [email protected],
Fax: 607-5535252
2
Faculty of Electronic and Computer Engineering
Universiti Teknikal Malaysia Melaka
75450 Ayer Keroh, Melaka
[email protected]
Keywords: Microstrip, Multibeam Antenna, Butler Matrix.
Abstract
In this paper, four beam patterns generated by four identical
planar antennas connected to a 4x4 Butler Matrix are
presented. Each of the planar antennas consists of a 4x2
radiating array patch antenna which has the capability to
produce a directional beam pattern. Butler Matrix 4 x 4 is
chosen as a beamforming network and was designed to
produce four independent beams in four different directions.
The measurement patterns result achieved by this project is
then compared with the theoretical calculation.
inset feeding. It was shown in [6] that only 1L and 1R
appeared as directional patterns and the others showing
unidentified direction. The directions of beam patterns are 19º and 19º, which represent about 13% errors from the
theoretical calculation.
In this project, four units of 4x2 rectangular patch microstrip
antenna arrays are connected to a beamforming network and
their radiation patterns are observed. The designed was aimed
for WLAN application at 2.4GHz. Butler Matrix 4 x 4 is
chosen as a beamforming network and was designed to
produce four independent beams in four different directions.
The configuration of the Butler Matrix together with the
antenna array is shown in Figure 1.
1 Introduction
The application of smart antenna system has become more
popular nowadays, as it is capable to increase the signal to
noise (SNR) ratio, reduce interference and increase the
channel capacity of the system [3]. One method of
implementing these smart antennas is by using switched beam
antenna arrays provided with fixed number of beams, and one
or more beams can be selected from the array for transmission
or reception. By selecting only one or more beams at an
instant, power received in desired direction can be
maximized.
In recent years, the topic of multibeam antenna constructed
using Butler Matrix as a beamforming network has received
much attention due to its simplicity and low cost of
implementation. Four ports Butler Matrix incorporated with 4
linear antenna array has been presented in several papers such
as in [2], [4] and [6]. In paper [2], 4x2 bowtie antennas have
been used as a radiating antenna. It was verified that four
beams have been created with beam angles at 12º, -40º, 40º
and -12º. The design configuration constructed in [4] can
generate four beams; 1R, 2L, 2R and 1L with the beam angles
at 16º, -39º, 38º and -15º. A similar design configuration also
presented in [6], but they brought some changes on the
feeding techniques where coaxial probe is used instead of
(a)
(b)
Figure 1: The configuration of the project
(a) The Block Diagram. (b) The construction of the project.
2 Design Configuration
2.1 4x2 Rectangular Array Patch Antenna
As shown in Figure 1, Array1, Array2, Array3 and Array4 are
constructed by a 4x2 radiating array patch antenna. They are
designed based on the rectangular patch shape and each of the
140
4x2 rectangular patch antennas has the capability to produce a
directional beam pattern. The effective length and width of
the patch antennas are calculated using the theory of
microstrip patch antenna, and these dimensions are then
optimized by using the simulation software Microwave
Office. Coaxial probe is used as a feeding technique and
placed at the centre of the array as this technique produces
less radiation effect that caused by the transmission line. Each
of the 4x2 element antenna arrays is then fabricated onto four
individual identical FR4 board with relative permittivity of
4.5. Figure 2 shows a layout and implementation of a single
4x2 rectangular array patch antennas.
Table 1: Design Specification of the Butler Matrix
Ports
A1
A2
A3
A4
β
1R
0º
-45º
-90º
-135º
45º
2L
-90º
45º
-180º
-45º
-135º
2R
-45º
-180º
45º
-90º
135º
1L
-135º
-90º
-45º
0º
-45º
2.3 A 6dB Power Divider
A 6dB power divider was also designed in this project using
T-junction theory. It is used to combine the four of 4x2
element array antennas so that the radiation pattern of four
4x2 antenna array without beamforming network can be
observed. Figure 4 illustrates the layout and implementation
of 6dB power divider that has been implemented using
microstrip transmission line and the block diagram of the
connection is shown in Figure 5.
Figure 2: Layout and implementation of a 4x2 rectangular
array patch antennas.
2.2 The Butler Matrix
Butler Matrix is a passive feeding NxN network with beam
steering capabilities for phased array antennas with N outputs
connected to an antenna elements and N inputs or beam ports
[3]. This network is able to control the phases and amplitudes
of the excitation current. It was designed in such a way so that
four different phases with same amplitudes of the excitation
current can be coupled to the antenna arrays. It consists of
four 90° hybrid coupler, two 0 dB crossover and two -45°
phase shifter that has four input and four output ports. Four
outputs A1, A2, A3 and A4 are used as inputs to the antenna
to produce four beams. This network is also fabricated on the
same material as the antenna array which is then combined
with four units of 4x2 planar antenna arrays connected using
four equal length coaxial cables. Figure 3 illustrates the block
structure and layout of the Butler Matrix and Table 1 shows
the design specification of the Butler Matrix.
Figure 4: The layout and implementation of the 6dB power
divider.
Figure 5: Antenna array together with 6dB power divider
block diagram
2.3 Array Factor Calculation
The radiation characteristic of the antenna array is
subsequently studied using phased scanning array principle.
The normalised array factor (AF) of a beam angle array is
given by equation [5]:
AF =
sin( N π
Nπ
d
λ
d
λ
(sin θ − sin θ 0 )
(1)
(sin θ − sin θ 0 )
where N is the number of antenna element, d is the distance
between antenna elements, θ0 is the beam angle and λ is the
wavelength at the free space. Beam angle can be calculated
by using this equation [5]:
⎛ βλ ⎞
(2)
θ = sin −1 ⎜
⎟
0
a) Block Structure
b) Layout
Figure 3: Block structure and layout of the Butler Matrix.
⎝ 2π d ⎠
where β is the phase difference of excitation current between
antenna elements that provided by the Butler Matrix. Table 2
shows the computed beam angle, θ0 correspond to the phase
difference that provided by the Butler Matrix.
141
Table 2: Computed beam angle.
Phase
Computed
Difference , ß Beam Angle, θ0
1R
45º
14.5º
2L
-135º
-48.6º
2R
135º
48.6º
1L
-45º
-14.5º
(HPBW) of approximately 27°. Figure 7 shows the
comparison of the simulated four measured radiation pattern
of the 4x2 antenna. Return Loss value at the operating
frequency is shown in Table 4. It is shown that good
impedance matching is obtained at the operating frequency as
S11 values are lower than -10 dB.
Ports
As shown in Figure 1, four of the 4x2 antenna arrays are
connected to the beamforming network with half wavelength
spacing between antennas. This half wavelength is chosen so
that the array will have the largest gain and directivity that
does not have grating lobes [1]. By substituting θ0 = 14.5º, 48.6º, 48.6º,-14.5º, N=4 and d=λ/2 to the equation (1), AF for
this project can be computed. The computed radiation pattern
of AF corresponds to each ports can be shown in Figure 6.
Figure 7: Radiation Pattern of each antenna array
Table 4: Return Loss of antenna array
S11 (dB)
Return Loss
at 2.4GHz
Array1 Array2 Array3 Array4
Simulation
-9.37201
Measurement -11.80 -11.90 -11.70 -13.35
Figure 6: Computed AF radiation pattern corresponds to each
port.
3 Simulation and Experimental Result
Table 3 shows the simulated and measured results of the 6dB
power divider. The simulation of the 6dB power divider was
done using Microwave Office software. It was shown that the
outputs of each arm of the power divider are approximately
about -6.5 to -6.8 dB. The coupling value of the power divider
was measured using a vector network analyzer. The result
shows that the measurement value is about -1 dB lower than
simulation value. This may be due to the physical loss
contributed by the microstrip transmission line and the
substrate used.
Table 3: The Result of 6dB power divider.
Return Loss and Coupling at 2.4GHz
S11
S21
S31
S41
(dB)
(dB)
(dB)
(dB)
Simulation
-28.8
-6.5
-6.8
-6.8
Measurement
-16.6
-7.5
-7.7
-7.9
S51
(dB)
-6.5
-6.4
The simulation of the 4x2 antenna array has been done using
Microwave Office software. It is verified through
measurement that this 8 element (4x2) antenna array can
produce a directional beam with the Half Power Beamwidth
The simulation of the Butler matrix has also been done using
Microwave Office software and the obtained result has less
phase error, approximately about 2°. Table 5 shows the
simulated results of the corresponding magnitudes and phase
shifts between the inputs and the outputs of the Butler Matrix.
Table 5: The simulated result of Butler Matrix.
Ports
A1
A2
A3
A4
Mag (dB)
-7.2
-8.1
-8.4
-8.3
1R
Phase
0.7º
-43.4º
-91.3º
-134.5º
Mag (dB)
-8.0
-7.5
-7.8
-7.9
2L
Phase
-90.4º
43.5º
-178º
-44.8º
Mag (dB)
-7.9
-7.8
-7.5
-8.0
2R
Phase
-44.8º
-178º
43.5º
-90.4º
Mag (dB)
-8.3
-8.4
-8.1
-7.2
1L
Phase
-134.5º
-91.3º
-43.4º
0.7º
It can be observed from the results that the coupling values
for each output are almost equal which is about 7.2 to -8.4dB.
As mentioned previously, the variation of the values may be
due to the physical loss factor that contributed by the
microstrip transmission line and the substrate used. It can be
observed that the Butler Matrix also provide constant phase
increment between its output ports which are 45° for port 1R,
-135° for port 2L, 135° for port 2R and -45° for port 1L.
These results show that the design specification that shown in
Table 1 is achieved.
142
The radiation characteristics of the beams are measured using
far-field method in the anechoic chamber. All input ports are
fed with the same signal but only one port is activated at one
time while the other ports are terminated with 50Ω. Figure 8
illustrates the measured radiation patterns of 1L, 2R, 2L and
1R beams and the directions of the main beam patterns for
Figure 8 are shown in Table 8. Figure 9 shows the measured
radiation patterns of antenna array with 6 dB power divider.
4 Result Analysis
From Figure 7, it can be observed that there is presently a
slight difference between simulated and measured radiation
patterns. The beamwidth of the main lobes are similar for
both simulated and measured result. The side lobes levels and
beamwidth size for measured results are slightly higher than
those simulated results. This is may be due to the fabrication
process that cause the occurrence of a mutual coupling and
mismatch between feeding network and antenna.
The measured results of each antenna also provide similar
radiation characteristics. Since the radiation pattern of each
antenna is identical, the antenna array can be found according
to the pattern multiplication theorem that is given as follows
[1]:
Array pattern = Single unit pattern x Array Factor (AF)
(3)
The data of a single array unit pattern is obtained from the
measured result of Array1 (Figure 7) while the data of the
array factors are obtained from the computed results of
equation 1 (Figure 6). The radiation patterns of the built
project can be predicted using equation (3). Figure 10
illustrates the comparison between the calculated and
measured radiation patterns of the array.
Figure 8: Measured beam patterns when input ports of Butler
Matrix are fed individually.
Table 8: The directions of beam patterns correspond to input
ports.
Port
Direction of beam pattern
1R
5º
2L
-9 º
2R
13 º
1L
-2 º
a) Input Port 1R
Figure 9: Measured radiation patterns of antenna array with 6
dB power divider.
b) Input Port 1L
143
5 Conclusion
The implementation of the 4x4 Butler Matrix together with
four units of 4x2 antenna array is presented. The Butler
Matrix with four inputs and four outputs has been designed to
excite the four units of antenna array to steer the beams in
different directions. The experimental results obtained show
that the constructed Butler Matrix is able to produce four
different beams in four different directions. The radiation
characteristics of the antenna array are compared between the
theoretical and measured result, and they correlates well.
References
c) Input Port 2R
d) Input Port 2L
Figure 10: a)-d) Comparison between calculated patterns and
measured patterns of the array.
From Figure 10 above, it can be observed that the calculated
patterns have a bigger beamwidth compared to measured
result. The direction of beams of 1R and 1L are similar to the
calculated patterns which are directed to 5º and -2º
respectively. The characteristic of the beams of 1R and 1L
also looks similar to the ones in Figure 9. It shows that the
beamforming network only cause the main lobe to be directed
to a certain direction.
The main beam of port 2R and 2L are directed to 13º and -9º
respectively, which are a little bit different from the
calculated patterns. Calculated patterns show that the main
beam appears at the centre of the polar plot with lower
magnitude as the beam position of the array factor, 2R and 2L
are directed to 48.6º and -48.6º (Figure 6). The measured
result of 2L and 2R has a few grating lobes, higher than 10dB. This may be due to the mutual coupling effect that
happens when the entire antennas are placed on the
measurement field, as the surface of the antenna holder is not
perfectly uniform.
[1] C. A. Balanis. “Antenna Theory Analysis and Design”,
John Wiley, Inc., New Jersey, pp. 299, (2005).
[2] D. C. Chang, S. H. Jou. “The study of Butler Matrix
BFN for four beams antenna system”, IEEE Antenna
and Propagation Society International Symposium,
volume 4, pp. 176-179, (2003).
[3] H. L. Bachman. “Smart antennas-the practical realities”,
Proc. IEEE Aerospace Conference, volume 1, pp. 6370, (1997).
[4] N. T. Pham, G. A. Lee, F. D. Flavis. “Microstrip antenna
array with beamforming network for WLAN
applications”, IEEE Antenna and Propagation Society
International Symposium, volume 3A, pp. 267-270,
(2005).
[5] R. Tang, R. W. Burns. “Phased Array” in R. C. Johnson
(eds.), Antenna Engineering Handbook, McGraw Hill,
Inc., New York, pp. 20-5, (1993).
[6] S. R .Ahmad, F. C. Seman. “4-port Butler Matrix for
Switched Multibeam Antenna Array”, IEEE Applied
Electromagnetic Proceedings, pp. 69-73, (2005).
144
Multibeam Antenna for WLAN Application
Siti Zuraidah Ibrahim and Mohamad Kamal A.Rahim
Wireless Communication Centre,
Faculty of Electrical Engineering,
Universiti Teknologi Malaysia,
81310 Skudai Johor Bahru.
Tel No: 607-5536088 FaxNo: 607-5535252
[email protected], [email protected]
Abstract
In this paper, four beam patterns generated by
integrating a linear antenna array with a 4x4 Butler
Matrix are presented. Butler Matrix 4 x 4 is
functioning as a beamforming network which was
designed to produce four independent beams that
directed to four different directions. The research
aims to design and implement a fixed multi beam
antenna array for the application of wireless local
area networks (WLAN). The measurement patterns
result achieved by this project is then compared
with the theoretical calculation.
Key words: Multibeam Antenna, 4x4 Butler Matrix,
Microstrip
Antenna
Array,
Wireless
Communications
1. Introduction
The current trend in most access point in
conventional wireless local area network (WLAN)
is to use omnidirectional antennas, which radiates
and receives power equally in all directions. This
attributes however give a result of lower power
efficiency as most of transmitted power is directed
towards location where no network node is located.
They also decrease network performance due to cochannel interference from nodes or clients that are
transmitting from undesired directions [2]. One of
the proposed solutions to overcome these
constraints is to use multibeam antenna on WLAN
access points [2]. Multibeam antennas are antenna
array that make use of beamforming network to
produce multiple independent beams that directed
to different directions. By offering independent
beams or channels, access point will switch
between these channels to select the channel that
has the highest received power. This feature assist
the antenna system to maximize the power received
in the desired directions. Figure 1 illustrates the
idea of having four beams coverage pattern
produced by multibeam antenna.
Figure 1: Four beams coverage pattern produced
by multibeam antenna.
2. The Theory of Multibeam Antenna
The concept of the multibeam antenna can be
studied using phased array scanning theory. This
type of array has the capability of steering a beam
from center polar plot to any desired angle, θ0. The
normalized array factor (AF) of a beam steered
array is given as [3]:
AF =
sin( N π
Nπ
d
λ
d
λ
(sin θ − sin θ 0 )
(1)
(sin θ − sin θ 0 )
where N is the number of antenna element, d is the
distance between adjacent antenna elements, θ0 is
the beam angle and λ is the wavelength at the free
space. Beam angle, θ0 can be calculated by using
this equation [3]:
⎛ βλ ⎞
⎟
⎝ 2π d ⎠
θ 0 = sin −1 ⎜
(2)
where β is the phase difference of excitation current
between adjacent antenna elements. By knowing the
value of β, the direction of beam, θ0 can be
predicted by using equation (2). Then, AF can be
145
calculated by substituting the value of θ0 to the
equation (1).
In the case of multibeam antenna, the number of
required beam is more than one, thus multiple value
of β should be provided to the antenna array as well
in order to generate this multiple beams. Several
studies had used Butler Matrix to provide multiple β
to the antenna array since its implementation using
microstrip element is easier and required less cost
as compare to other techniques [4].
Butler Matrix also known as beamforming
network as it has a capability to steer a beam
electronically to a certain direction by providing
multiple β through different transmission paths. The
conventional Butler Matrix can provide four
different value of β which are 45º, -135º, 135º, -45º.
Table 1 shows the computed beam angle when
substituting β = 45º, -135º, 135º, -45º and d= λ/2 to
the equation (2).
prototype is implemented by using microstrip
transmission line technique and fabricated on FR4
board with relative permittivity 4.5.
Table 1: Computed beam angle
Figure 3. The fabricated single antenna.
Phase
Difference , ß
45º
-135º
135º
-45º
Computed
Beam Angle, θ0
14.5º
-48.6º
48.6º
-14.5º
Beam
Position
1R
2L
2R
1L
By substituting θ0 = 14.5º, -48.6º, 48.6º,-14.5º,
N=4 and d=λ/2 to the equation (1), AF for this
project can be computed. The computed radiation
pattern of AF corresponds to each ports can be
shown in Figure 2.
3.1. Antenna Array Configuration
For antenna array part, at first a single square
patch antenna was designed at 2.4GHz using
proximity coupling as its feeding technique. This is
to ensure that this type of antenna is functioning
well at the designed frequency. The fabricated
single antenna is shown in Figure 3.
Then, four identical square patch antennas were
designed at 2.4GHz, arranged in a linear form and
spaced at half-wavelength apart. This half wave
spacing ensures that the array will have the largest
gain and directivity that does not have grating lobes
[1]. The antenna is fed using proximity coupling
technique as this technique can offer a bigger
bandwidth as compare to standard inset feed
technique [1]. This antenna array is constructed on
two layer substrates. Square patch antenna are
printed on the upper substrate while the feeding line
are printed on the lower substrate but both of the
substrate have the same specifications. The
implementation of the antenna array is shown in
Figure 4.
Figure 4. The fabricated antenna array
3.2. Butler Matrix Configuration
Figure 2. Computed radiation pattern of AF.
3. Design Configuration
In this project, two major components are
needed to be designed which are antenna array and
Butler Matrix part. Each component is designed and
simulated using Microwave Office Software. The
The Butler Matrix is a 2n x 2n network consisting
in 2n inputs and 2n outputs, 2n-1 log2 2n hybrids and
some phase shifters [5]. In this project, the designed
Butler Matrix consists of four 90° hybrid coupler,
two 0 dB crossover and two -45° phase shifter. It
has four inputs 1R, 2L, 2R and 1L, and four outputs
A1, A2, A3 and A4. The four outputs are used as
inputs to antenna elements to produce four beams.
146
Figure 5 shows the block structure
implementation of the Butler Matrix.
and
3.3. Integration of the project
The ports of the antenna array are connected to
the output ports of the Butler Matrix by using four
equal length coaxial cables. The picture of the
integrated project is shown in Figure 6.
Figure 6.The implementation of the project
The linear antenna array also has been
connected to 6 dB power divider so that the
radiation pattern of linear antenna array can be
observed. The implementation of connection is
shown in Figure 7.
Figure 5. The block structure and implementation
of the Butler Matrix.
The input ports of the Butler Matrix are named
according to the beam position that corresponds to
the given β, as calculated and shown in Table 1. As
shown in Figure 2, first beam that appear on the
right side of the polar plot is named as beam 1R.
This beam is created when β = 45º, in the sense that
when current excited to port 1R of the Butler
Matrix, the phase excitation of the current that
coupled to the antenna will be 45º different between
adjacent elements.
Figure 7. Linear array with 6dB Power Divider
4. Simulation and Experimental Result
Table 2. Design target of the output phase of the
Butler Matrix.
Por
t
1R
2L
2R
1L
A1
A2
A3
A4
β
0º
-90º
-45º
-135º
-45º
45º
-180º
-90º
-90º
-180º
45º
-45º
-135º
-45º
-90º
0º
45º
-135º
135º
-45º
This can be clearly seen in Table 2 which has
been referred as a design target of the output phase
of the Butler Matrix. It was designed in such a way
so that when current excited to any input ports, the
phase different between adjacent output ports will
only has one constant β as shown in Table 2.
Figure 8. The simulated return loss values
correspond to the each patch antenna
Figure 8 shows the simulated return loss values
correspond to the each patch antenna that is shown
in Figure 4. It is shown that good impedance
matching is obtained at the operating frequency as
S11 values are lower than -10 dB.
147
Table 3. The simulated result of Butler Matrix.
A1
0.7º
0.7º
-90.4º
0.4º
-44.8º
0.2º
-134.5º
0.5º
Phase
| Error|
Phase
| Error|
Phase
| Error|
Phase
| Error|
1R
2L
2R
1L
A2
-43.4º
1.6º
43.5º
1.5º
-178º
2º
-91.3º
1.3º
A3
-91.3º
1.3º
-178º
2º
43.5º
1.5º
-43.4º
1.6º
A4
-134.5º
0.5º
-44.8º
0.2º
-90.4º
0.4º
0.7º
0.7º
Table 3 shows the simulated results of the
corresponding phase shifts between the inputs and
the outputs of the Butler Matrix and the obtained
result shows that it has less phase error,
approximately about 2°. It can be observed that the
Butler Matrix also provide constant phase
increment between its output ports which are 45°
for port 1R, -135° for port 2L, 135° for port 2R and
-45° for port 1L. These results show that the design
specification that shown in Table 2 is achieved.
Table 4. The measured result of Butler Matrix
Port
1R
2L
2R
1L
A1
-82º
-166º
-128º
141º
A2
-130º
-53º
82º
-172º
A3
-169º
84º
-54º
-131º
A4
145º
-125º
-166º
-82º
Table 4 shows the measured result of the Butler
Matrix. Even though it gives a big different
between measured and simulated result, in fact,
these output phases are not important. The main
concern about the result of the Butler matrix is the
phase differences between output ports. As long as
the phase differences are following the values in
Table 1, this Butler Matrix can be trusted. Thus, the
phase differences are calculated and shown in Table
5. It can be observed that the error occurs up to 23°.
The difference may happen due to the imperfect
fabrication process as it only done by human being.
Table 5. The computed phase error of the measured
result of the Butler Matrix
Port
1R
| Error|
2L
| Error|
2R
| Error|
1L
| Error|
β1
β2
β3
(A2-A1)
(A3-A2)
(A4-A3)
-48º
3º
113º
22º
-150º
15º
48º
3º
-39º
6º
137º
2º
-136º
1º
41º
4º
-46º
1º
151º
16º
-112º
23º
48º
3º
The radiation characteristics of the beams are
measured using far-field method in the anechoic
chamber. At first, the radiation pattern of the single
antenna is measured. The obtained radiation pattern
of the single antenna can be used later to calculate
the predicted radiation pattern of the integrated
project. The measured radiation pattern of the single
antenna is shown in Figure 9.
Figure 9. The measured radiation pattern of the
single antenna.
For the measurement of the integrated project,
all input ports are fed with the same signal but only
one port is activated at one time while the other
ports are terminated with 50Ω. Figure 10 illustrates
the measured radiation patterns of 1L, 2R, 2L and
1R beams and Figure 11 shows the measured
radiation patterns of antenna array with 6 dB power
divider.
Figure 10. The measured radiation pattern of the
integrated project.
Target
-45º
+135º
-135º
+45º
Figure 11. The measured radiation pattern of the
linear array with 6dB power divider.
148
5. Discussion
Without Butler Matrix, the radiation patterns of
the array only appear at the center of the polar plot
as shown in Figure 11. By integrating linear array
with the Butler Matrix, it has been demonstrated
that the Butler Matrix manage to steer the main lobe
to be directed to the desired direction. This
multibeam can cover 120° of coverage area (-60° to
60°). The radiation pattern of the integrated system
can be predicted using pattern multiplication
theorem that is given as follows [1]:
Array pattern = Single unit pattern x (AF)
[4] S.R. Ahmad and F.C. Seman, “4-port Butler Matrix
for Switched Multibeam Antenna Array”, IEEE Applied
Electromagnetic Proceedings 2005, Dec. 20-21, pp. 6973.
[5] T.A. Denidni and T. E. Liber, “Wide band Four Port
Butler Matrix for Switched Multibeam Antenna Arrays”,
14th IEEE 2003 Personal, Indoor and Mobile Radio
Communication Proceedings, Sept. 7-10, pp.2461-2464.
(3)
The data of a single array unit pattern is
obtained from the measured result of single antenna
(Figure 9) while the data of the array factors are
obtained from the computed results of equation 1
(Figure 2). The radiation patterns of the built
project can be predicted using equation (3). Figure
12 illustrates the comparison between the calculated
and measured radiation patterns of the array. It can
be observed that the measured radiation has a
similar pattern to the computed pattern. This is
verified that the experiment is reliable as it has a
good agreement with the theoretical calculation.
a) Input Port 1R
6. Conclusion
The implementation of linear array with Butler
Matrix is presented. The Butler Matrix with four
inputs and four outputs has been designed to excite
the four units of antenna array to steer the beams in
different directions. The experimental results
obtained show that the constructed Butler Matrix is
able to produce four different beams in four
different directions. The radiation characteristics of
the antenna array are compared between the
theoretical and measured result, and they correlates
well.
7. References
b) Input Port 2L
c) Input Port 2R
[1] C.A. Balanis, Antenna Theory Analysis and Design,
John Wiley, Inc., New Jersey, 2005.
[2] D. Lal, V. Jain, Q. A. Zeng, and D.P. Agrawal,
“Performance Evaluation of Medium Access Control for
Multiple-Beam Antenna Nodes in a Wireless LAN”,
IEEE Transactions on Parallel and Distributed Systems,
IEEE Computer Society, Dec. 2004, pp. 1117-1129.
[3] R. Tang and R.W. Burns. “Phased Array” in R. C.
Johnson (eds.), Antenna Engineering Handbook, McGraw
Hill, Inc., New York, 1993.
d) Input Port 1L
Figure 12. a)-d) Comparison between calculated
and measured patterns of the integrated project.
149
Switched Beam Antenna using Omnidirectional Antenna Array
Siti Zuraidah Ibrahim and Mohamad Kamal A.Rahim
Wireless Communication Centre,
Faculty of Electrical Engineering,
Universiti Teknologi Malaysia,
81300 Skudai, Johor, Malaysia
[email protected], [email protected].
Abstract – This paper presents the result of using
omnidirectional antenna array on the Switched Beam
Antenna System. The 4x4 conventional Butler Matrix
has been used in this project as the beamforming
network to provide four different values of the phase
difference that coupled to antenna elements. By
integrating the omnidirectional antenna array with 4x4
Butler Matrix, this system has the capability to
generate four different beams at four different
directions with the coverage area of 360º. Two types
of beam can be generated by the system which are
narrow beam and broader beam. The comparison
between the measured and computed radiation pattern
of the switched beam antenna also presented.
Keywords: Switched Beam Antenna; 4x4 Butler Matrix;
Microstrip; Antenna Array
1. Introduction
The topic of switched beam antenna as a smart
antenna has been discussed vigorously as the
implementation of it is simple and requires less cost as
compared to adaptive antenna array. Unlike the
adaptive antenna, switched beam antenna only
constructed by a number of radiating elements, a
beamforming network and RF switch [7] while
adaptive array systems provide more intelligent
operation and needs more advanced signals processing
to function. For simplicity, this paper presents the
switched beam antenna that comprises of antenna array
and beamforming network only since the combination
of these blocks are good enough in order to observe
the radiation pattern characteristics of the system.
Recently, most of the papers were discussing
about switched beam antenna that uses microstrip
antenna array which only accomplished to serve 120º
of angular coverage area [2, 3, and 5]. In this project,
omnidirectional antenna has been used as the radiating
elements and the radiation pattern characteristics of the
system have been observed. The conventional Butler
matrix is used as the beamforming network of the
system. The measured radiation patterns attained by
this project are then compared to the theoretical
calculation.
2. Project Development
The development of the project involves three
stages which are implementation of the antenna array,
design and implementation of Butler Matrix and the
integration between antenna array and Butler Matrix.
2.1 Implementation of Antenna Array
The antenna array is consists of four radiating
antenna spaced at half-wavelength apart at the carrier
frequency and arranged in a linear form. The antenna
elements spaced at d= λ/2 = 6.25 cm apart. This half
wave spacing ensures that the array will have the
largest gain and directivity that does not have gating
lobes [1].
The radiating elements that have been used in this
project are commercial dipole antennas manufactured
by D-Link Company. The antenna is designed to
operate at 2.4 GHz frequency band, and is aimed for
WLAN applications.
2.2 Design and Implementation of Butler
Matrix
The Butler Matrix is a 2n x 2n network consisting
in 2 inputs and 2n outputs, 2n-1 log2 2n hybrids and
some phase shifters [6]. In this project, the designed
Butler Matrix consists of four 90° hybrid coupler, two
0 dB crossover and two -45° phase shifter. Each
component is designed at the operating frequency of
2.4GHz and simulated using Agilent ADS schematic.
The prototype is implemented using microstrip
transmission line technique and fabricated on FR4
board with relative permittivity 4.5, dissipation factor
tan δ = 0.019, and thickness of 1.6 mm.
The Butler Matrix has four inputs 1R, 2L, 2R and
1L, and four outputs A1, A2, A3 and A4. The four
outputs are used as inputs to antenna elements to
produce four beams. The input ports of the Butler
Matrix are named according to the beam position
which will be generated by activating one of the input
ports of the Butler Matrix. Figure 1 shows the block
structure and layout of the Butler Matrix.
Butler Matrix also known as beamforming
network as it has a capability to steer a beam
electronically to a certain direction by providing
multiple phase differences, β through different
n
150
transmission paths. The conventional Butler Matrix
can provide four different value of β which are -45º,
+135º, -135º, +45º. It was designed in such a way so
that when current excited to any input ports, the phase
different between adjacent output ports will only has
one constant β as shown in Table 1.
Table 2: Simulated results of the corresponding phase
shifts between the inputs and the outputs of the Butler
Matrix.
Port
1R
2L
2R
1L
A1
1º
-90º
-44.4º
-134.1º
A2
-44.8º
42.2º
-179.7º
-93º
A3
-93º
-179.7º
42.2º
-44.8º
A4
-134.1º
-44.4º
-90º
1º
Table 3: Computed phase error.
Port
1R
| Error|
2L
| Error|
2R
| Error|
1L
| Error|
Figure 1: The block structure and layout of the Butler
Matrix.
β1
β2
β3
(A2-A1)
(A3-A2)
(A4-A3)
-45.8º
0.8º
132.2º
2.8º
-135.3º
0.3º
41.1º
3.9º
-48.2º
3.2º
138.1º
3.1º
138.1º
3.1º
48.2º
3.2º
-41.1º
3.9º
135.3º
0.3º
-132.3º
2.8º
45.8º
0.8º
Target
-45º
+135º
-135º
+45º
The radiation characteristics of the beams are
measured using far-field method in the anechoic
chamber. At first, the radiation pattern of the single
antenna is measured. The obtained radiation pattern of
the single antenna can be used later to calculate the
predicted radiation pattern of the integrated project.
The measured radiation pattern of the single antenna is
shown in Figure 2.
Table 1: Design Target of the Butler Matrix
Port
1R
2L
2R
1L
β
-45º
+135º
-135º
+45º
2.3 Integration of the project
The ports of the antenna array are connected to the
output ports of the Butler Matrix by using four equal
length coaxial cables.
3. Result
Figure 2: Measured radiation pattern of individual
antenna
For the measurement of the integrated project, all
input ports are fed with the same signal but only one
port is activated at one time while the other ports are
terminated with 50Ω. Figure 3 illustrates the measured
radiation patterns of 1R, 2L, 2R and 1L beams.
Table 2 shows the simulated results of the
corresponding phase shifts between the inputs and the
outputs of the Butler Matrix. The main concern about
the result of the Butler Matrix is the phase differences
between the output ports, not the value of phase at the
output ports. Thus, the phase differences between
output ports are calculated and the difference between
the obtained value and the target also being compared
as shown in Table 3.
(a)
(b)
151
antenna elements and λ is the wavelength at the free
space. By substituting N=4, d=λ/2 and β = -45º, +135º,
-135º, +45º, to the equation (2), AF for this project can
be computed. The computed radiation pattern of AF
can be plotted by substituting θ from 0º to 360º to
equation (2). The computed radiation pattern of AF
corresponds to each ports can be shown in Figure 5.
(c)
(d)
Figure 3: Measured radiation patterns, (a) Port 1R, (b)
Port 2L, (c) Port 2R and (d) Port 1L
It can be observed that when port 1R and 1L is
activated, two main beams appeared and directed to
the upper (+14º, -10º) and bottom (+168º, -158º) part
of the polar plot. The beamwidth size of the main
beams is narrower while for port 2L and 2R, the
beamwidth size of the main lobe is broader. Figure 4
shows the overlapping radiation pattern of the Port 1R,
2L, 2R and 1L on the same plot. It can be shown that
the system capable to cover up to 360º of coverage
area.
(a)
(b)
(c)
(d)
Figure 5: Computed radiation pattern of AF, (a) β = -45º,
(b) β = +135º, (c) β = -135º, (d) β = +45º
Figure 4: Measured radiation pattern of the integrated
project
4. Result Analysis
The radiation pattern of the integrated system can
be predicted using pattern multiplication theorem that
is given as follows [1]:
(a)
(b)
(c)
(d)
Array pattern = Single Unit pattern x Array Factor (1)
The data of a single array unit pattern is obtained
from the measured result of the single antenna (Figure
2) while the data of the array factors can be calculated
using following equation [4]:
⎛ − βλ ⎞
(sin θ − sin(sin −1 ⎜
⎟))
⎝ 2πd ⎠
d
⎛ − βλ ⎞
Nπ (sin θ − sin(sin −1 ⎜
⎟))
λ
⎝ 2πd ⎠
sin( Nπ
AF =
d
λ
(2)
where N is the number of antenna element, d is the
distance between adjacent antenna elements, β is the
phase difference of excitation current between adjacent
Figure 6: Comparison between the computed and
measured radiation patterns (a) Port 1R, (b) Port 2L, (c)
Port 2R and (d) Port 1L
The radiation patterns of the built project can be
predicted using equation (1). Figure 6 illustrates the
152
comparison between the calculated and measured
radiation patterns of the array. It can be observed that
the measured radiation has a similar pattern to the
computed pattern. This is verified that the experiment
is reliable as it has a good agreement with the
theoretical calculation.
5. Conclusion
The implementation of linear array with Butler
Matrix is presented. The Butler Matrix with four inputs
and four outputs has been designed to excite the four
units of antenna array to steer the beams in different
directions. The experimental results obtained show that
the constructed Butler Matrix is able to produce four
different beams in four different directions and
accomplish to serve 360º of coverage area. It has two
types of beam which are narrow beam (port 1R and
1L) and broader beam (port 2L and 2R). The radiation
characteristics of the antenna array are compared
between the theoretical and measured result, and they
correlates well.
References
[1] C.A. Balanis, Antenna Theory Analysis and
Design, John Wiley, Inc., New Jersey, 2005.
[2] D. C. Chang, S. H. Jou. “The study of Butler
Matrix BFN for four beams antenna system”, in
IEEE Antenna and Propagation Society
International Symposium, June 2003, pp. 176-179.
[3] N. T. Pham, G. A. Lee, F. D. Flavis. “Microstrip
antenna array with beamforming network for
WLAN applications”, in IEEE Antenna and
Propagation Society International Symposium,
July 2005, pp. 267-270.
[4] R. Tang and R.W. Burns. “Phased Array” in R. C.
Johnson (eds.), Antenna Engineering Handbook,
McGraw Hill, Inc., New York, 1993.
[5] S. R. Ahmad and F. C. Seman, “4-port Butler
Matrix for Switched Multibeam Antenna Array”,
in Proceedings of the 2005 Asia-Pacific
Conference on Applied Electromagnetics, Johor,
Malaysia, November 2005, pp. 69-73.
[6] T. A. Denidni and T. E. Liber, “Wide band Four
Port Butler Matrix for Switched Multibeam
Antenna Arrays”, in 14th IEEE 2003 Personal,
Indoor and Mobile Radio Communication
Proceedings, September 2003, pp.2461-2464.
[7] Y. J. Chang and R. B. Hwang “Switched beam
System for low-tier wireless communication
systems,” in Proceedings of the 2001 Asia Pacific
Microwave
Conference,
Taipei,
Taiwan,
December 2001, pp. 946-949.