-.

1 卞

A Novel Branch-line Coupler Design

for Dual-band Applications

WONG Fai-leung

A Thesis Submitted in Partial Fulfilment

of the Requirements for the Degree of

Master of Philosophy

in

Electronic Engineering

• T h e Chinese University of Hong Kong

July 2004

The Chinese University of Hong Kong holds the copyright of this thesis. Any
person(s) intending to use a part or whole of the materials in the thesis in a proposed
publication must seek copyright release from the Dean of the Graduate School.
If11 m locT).
jL/统系‘書圆

々‘S>vL 丨BRArr/ SYSTEWt/^；^

Abstract

Branch line coupler is a popular passive device for microwave and millimeter-wave

applications. The standard design of branch line coupler provides specific power

dividing ratio and phase relationship (depending on types of branch line coupler)

between the output ports at the operating frequency. Due to its power dividing and

combining function, it's commonly used as building blocks for balanced amplifier,

mixer, etc.

However, conventional design of branch line coupler basically consists of quarter

wavelength transmission lines which lead to narrow band operation. A s a result, its

application to wide-band and multi-band system is thus greatly limited. In past years,

various reports concerning bandwidth enhancement and size reduction techniques

have been published in the literature. Recently, dual band branch line coupler

implemented by lumped elements was also reported.

In this thesis, two novel planar dual-band, 90° 3-dB branch line coupler designs are

presented. Both designs offer dual-band operation and low insertion loss. The first

design is slightly smaller than the conventional structure, while the second design uses

only quarter-wavelength long branch-lines, but with the insertion of shunt-stubs.

Rigorous analysis of both design with explicit design formulae are presented.

Moreover, practical issues such as the branch-line impedance realization, and effect of

characteristic impedance deviation are also addressed. Finally, for experimental

verification, both the simulated and measured results of a 900/1800 M H z branch-line

couplers constructed by microstrip technology are demonstrated.

ii
摘要

支綫稱合器是一種常用於微波和毫米波應用的無源器件。在其工作頻帶内，一

個標準的支錢鍋合器設計在其兩個輸出端口提供特定的能量和相位差（取決於

支綫锅合器的類型）。基於其能量結合的功能’支綫稱合器通常應用於不同的

電路結構中，例如平衡放大器，平衡混頻器等等。

但是，由於傳統的支綫鍋合器設計是由四分之一波長傳輸線構成，其工作頻帶

具有固有的窄帶特性，因此大大限制了支綫親合器在寬帶甚至多頻帶系統中的

應用。在過去的幾十年中，各種關於提升帶寬以及減小面積的技術在很多文獻

中發表。最近，一種用集總元件實現的雙頻帶支綫稱合器被報道。

本論文介紹了兩種新型的平面結構雙頻帶支綫稱合器的設計和實現，它們均爲

兩個特定工作頻帶提供了標準表現。第一個設計的電路面積比傳統的支綫親合

器稍大，而第二個設計則只由四分之一波長的傳輸線及分路短根組成°

本論文對支綫稱合器的精確分析提供了明確的設計公式°另外，一些實際問

題’例如支綫的阻抗等級以及偶合器最佳拓撲結構的選擇在本論文中也有闡

述。最後’基於驗證的需要’本論文給出了用微帶線技術搭建的900/1800MHZ

支帶稱合器的仿真和測量結果。

iii
Acknowledgment

I would like to express m y sincere thank to m y supervisor, Prof. Kwok-Keung

Michael Cheng for his guidance on m y research in past two years. In addition, I would

like to thank Prof. Ke-Li W u for his generous help and advice on m y postgraduate

study.

Mr. Kwok-kin Tse, our laboratory technician, is always helpful whenever I seek

assistances from him. M y colleagues, Tony Cheng, Kitty Leung, Alvis Au-yeung and

John Ng, always provide m e support and care by all means and m y life of

postgraduate study wouldn't be so fruitful without them. Thanks them so much and I

wish they have bright futures.

Last but not least, m y parents offer m e superior support and I really appreciate it very

much.

iv
Table of contents

ABSTRACT 11
mm "I
ACKNOWLEDGMENT IV
TABLE OF CONTENTS V
TABLE OF FIGURES VII

1 INTRODUCTION 1

2 BASIC THEORY OF BRANCH LINE COUPLER 4

2.1 FOUR-PORT NETWORK 4

2 . 2 E V E N - O D D MODE ANALYSIS 5

2.2.1 Even mode excitation 6

2.2.2 Odd mode excitation 7
2.2.3 Mathematical analysis 9
2.3 SIMULATION RESULTS 12

3 REVIEW OF ADVANCED BRANCH LINE COUPLER DESIGNS 15

3.1 BROAD-BAND UNIPLANAR BRANCH-LINE DESIGN 15

3.2 C O M P A C T BRANCH-LINE COUPLERS USING SLOW-WAVE S T R U C T U R E … … 1 7

3.3 MINIATURE BRANCH-LINE COUPLER USING EIGHT TWO-STEP STUBS 18

3.4 W I D E BAND LUMPED-ELEMENT 3 - D B QUADRATURE C O U P L E R 20

3.5 DUAL BAND BRANCH LINE COUPLER DESIGN USING LEFT-HANDED

TRANSMISSION LINES 22

4 DESIGN THEORY OF DUAL BAND BRANCH LINE COUPLERS 24

4• 1 D E S I G N 1 - DUAL BAND BRANCH LINE COUPLER WITH SUB-OPTIMUM

PERFORMANCE 25
4.1.1 Analysis and design 26
4.1.2 Size comparison 30
4 . 2 D E S I G N 2 - DUAL BAND BRANCH LINE COUPLER USING SHUNT S T U B S … … 3 4

4.2.1 Analysis and design 35

4.2.2 Size comparison 41

V
5 SIMULATION VERIFICATION 44

5.1 DESIGN 1 44

5.1.1 Schematic simulation 45

5.1.2 Schematic simulation with line width deviation 48
5.1.3 Schematic simulation with junction discontinuity 54
5.2 DESIGN 2 58

5.2.1 Schematic simulation 58

5.2.2 Schematic simulation with line width deviation 62
5.2.3 Schematic simulation with junction discontinuity 68
6 CIRCUIT IMPLEMENTATION AND CHARACTERIZATION 74

6.1 DESIGN 1 74

6.1.1 Circuit fabrication 74

6.1.2 Measurement results 75
6.2 DESIGN 2 78

6.2.1 Circuit fabrication 78

6.2.2 Measurement results 79

7 CONCLUSIONS 83

8 RECOMMENDATIONS FOR FUTURE WORK 85

9 REFERENCES 86

10 AUTHOR'S PUBLICATIONS 90

vi
Table of figures

FIGURE 2-1 ： GENERAL FOUR-PORT NETWORK 4

FIGURE 2-2： PROPOSED DESIGN OF QUARTER-WAVELENGTH TRANSFORMER 5
FIGURE 2-3: BRANCH-LINE COUPLER WITH LINE OF SYMMETRY WITH TWO IN-PHASE
SIGNAL GENERATORS 7
FIGURE 2-4： DECOMPOSITION OF THE BRANCH-LINE COUPLER IN EVEN MODE ANALYSIS?
FIGURE 2-5： BRANCH-LINE COUPLER WITH LINE OF SYMMETRY WITH TWO IN-PHASE
SIGNAL GENERATORS 8
FIGURE 2-6: DECOMPOSITION OF THE BRANCH-LINE COUPLER IN ODD MODE ANALYSIS . 8
FIGURE 2-7: SIMULATION RESULTS： RETURN LOSS 13
FIGURE 2-8： SIMULATION RESULTS: INSERTION LOSS 13
FIGURE 2-9: SIMULATION RESULTS： ISOLATION 14
FIGURE 2-10： SIMULATION RESULTS： PHASE DIFFERENCE 14
FIGURE 3-1： TRANSVERSE DIAGRAM OF (A) MICROSTRIPLINE AND (B) COPLANAR
WAVEGUIDE 15
FIGURE 3-2： PROPOSED DESIGN OF BRANCH LINE COUPLER USING COPLANAR
WAVEGUIDE AND SLOTLINE 16
FIGURE 3-3: (A) ARTIFICIAL TRANSMISSION LINE WITH SHUNT CAPACITANCES AND B)
ARTIFICIAL TRANSMISSION LINE WITH OPEN CIRCUIT SHUNT STUBS 17
FIGURE 3-4: GEOMETRY OF BRANCH LINE COUPLER REALIZED BY ARTIFICIAL
TRANSMISSION LINE 18
FIGURE 3-5： (A) PROPOSED HALF QUARTER WAVELENGTH TRANSFORMER AND (B)
EQUIVALENT STRUCTURE 19
FIGURE 3-6: RESTRUCTURE OF SINGLE OPEN STUB INTO TWO-STEP STRUCTURE 19
FIGURE 3-7: FINAL BRANCH LINE COUPLER WITH ALL PROPOSED SIZE REDUCTION
TECHNIQUES 20
FIGURE 3-8: (A) LUMPED ELEMENT QUADRATURE HYBRID AND (B) WIDE BAND
QUADRATURE HYBRID DESIGN 21
FIGURE 3-9: (A)PART OF LEFT-HANDED TRANSMISSION LINE AND (B) COMPOSITE
RIGHT/LEFT HANDED TRANSMISSION LINE FOR REPLACEMENT OF QUARTER
WAVELENGTH TRANSMISSION LINE 22
FIGURE 4-1 ： TOPOLOGY OF DESIGN 1 25
FIGURE 4-2： DESIGN 1 WITH DIFFERENT VALUES OF M AND N: (A) CONFIGURATION A
(M=L, N=2) AND (B) CONFIGURATION B FM=2, N=1} 28
FIGURE 4-3： RETURN LOSS, PORT ISOLATION AND INSERTION LOSS VERSUS FRACTIONAL
BANDWIDTH 29
FIGURE 4-4: CHARACTERISTIC IMPEDANCE OF SERIES ARM (Z；) AND SHUNT ARM (Z2)
AMONG FRACTIONAL BANDWIDTH 30
FIGURE 4-5 AND SAI VERSUS FRACTIONAL BANDWIDTH ( A ) 33
FIGURE 4-6: (A) PROPOSED QUARTER WAVELENGTH TRANSFORMER AND (B)
EQUIVALENT STRUCTURE 34
FIGURE 4-7： TOPOLOGY OF DESIGN 2 38
FIGURE 4-8: SIMPLIFICATION OF DOUBLE STUB INTO SINGLE STUB 39
FIGURE 4-9： PROPOSED STRUCTURE OF DUAL BAND BRANCH LINE COUPLER 40
FIGURE 4-10: COMPACT DUAL BAND BRANCH LINE COUPLER 40

vii
FIGURE 4-11 CHARACTERISTIC IMPEDANCE OF SERIES ARM (Z；), SHUNT ARM (Z2) AND
OPEN STUB (Z5) AMONG FRACTIONAL BANDWIDTH 41
FIGURE 4-12 SPI AND SA2 VERSUS FRACTIONAL BANDWIDTH ( A ) 43
FIGURE 5-1： A D S SCHEMATIC DIAGRAM FOR CONFIGURATION B , DESIGN 1 SIMULATION
45
FIGURE 5-2： SIMULATION - RETURN LOSS 46
FIGURE 5-3: SIMULATION - ISOLATION 46
FIGURE 5-4： SIMULATION - INSERTION LOSS 47
FIGURE 5-5: SIMULATION - PHASE DIFFERENCE 48
FIGURE 5-6: A D S SCHEMATIC DIAGRAM FOR LINE WIDTH DEVIATION SIMULATION …••49
FIGURE 5-7: SIMULATION - RETURN LOSS 50
FIGURE 5-8: SIMULATION - RETURN LOSS (LOWER BAND) 51
FIGURE 5-9: SIMULATION - RETURN LOSS (UPPER BAND) 51
FIGURE 5-10: SIMULATION - INSERTION LOSS (LOWER BAND) 52
FIGURE 5-11： SIMULATION - INSERTION LOSS 52
FIGURE 5-12： SIMULATION - INSERTION LOSS (UPPER BAND) 53
FIGURE 5-13： SIMULATION - PHASE DIFFERENCE 53
FIGURE 5-14： A D S SCHEMATIC DIAGRAM FOR SIMULATION WITH JUNCTION
DISCONTINUITY 54
FIGURE 5-15: SIMULATION - RETURN LOSS 55
FIGURE 5-16: SIMULATION - ISOLATION 55
FIGURE 5-17: SIMULATION - INSERTION LOSS 56
FIGURE 5-18： SIMULATION - PHASE DIFFERENCE 56
FIGURE 5-19: SCHEMATIC DIAGRAM FOR DESIGN 2 SIMULATION 58
FIGURE 5-20: SIMULATION - RETURN LOSS 59
FIGURE 5-21： SIMULATION - ISOLATION LOSS 59
FIGURE 5-22： SIMULATION - INSERTION LOSS 60
FIGURE 5-23: SIMULATION - INSERTION LOSS (LOWER BAND) 61
FIGURE 5-24： SIMULATION: INSERTION LOSS (UPPER BAND) 61
FIGURE 5-25: SIMULATION - PHASE DIFFERENCE 62
FIGURE 5-26: A D S SCHEMATIC DIAGRAM FOR LINE WIDTH DEVIATION SIMULATION ...63
FIGURE 5-27: SIMULATION - RETURN LOSS 64
FIGURE 5-28: SIMULATION - RETURN LOSS (LOWER BAND) 64
FIGURE 5-29: SIMULATION - RETURN LOSS (UPPER BAND) 65
FIGURE 5-30: SIMULATION - ISOLATION LOSS 65
FIGURE 5-31 ： SIMULATION - INSERTION LOSS 66
FIGURE 5-32: SIMULATION - INSERTION LOSS (LOWER BAND) 66
FIGURE 5-33： SIMULATION - INSERTION LOSS (UPPER BAND) 67
FIGURE 5-34： SIMULATION - INSERTION LOSS (MID-BAND) 67
FIGURE 5-35: SIMULATION - PHASE DIFFERENCE 68
FIGURE 5-36: A D S SCHEMATIC DIAGRAM FOR SIMULATION WITH JUNCTION
DISCONTINUITY 69
FIGURE 5-37： SIMULATION - RETURN LOSS 70
FIGURE 5-38: SIMULATION - ISOLATION LOSS 70
FIGURE 5-39： SIMULATION - INSERTION LOSS 71
FIGURE 5-40: SIMULATION - INSERTION LOSS (LOWER BAND) 71
FIGURE 5-41： SIMULATION: INSERTION LOSS (UPPER BAND) 12
FIGURE 5-42： SIMULATION - PHASE DIFFERENCE 72
FIGURE 6-1: CIRCUIT OF CONFIGURATION B (M=2, N二 V), DESIGN 1 74
FIGURE 6-2： MEASUREMENT： RETURN LOSS 75

viii
FIGURE 6-3: MEASUREMENT: ISOLATION FACTOR 75
FIGURE 6-4: MEASUREMENT: INSERTION LOSS 76
FIGURE 6-5: MEASUREMENT： PHASE DIFFERENCE 76
FIGURE 6-6: CIRCUIT OF DESIGN 2 78
FIGURE 6-7: MEASUREMENT: RETURN LOSS 79
FIGURE 6-8: MEASUREMENT: ISOLATION LOSS 79
FIGURE 6-9: MEASUREMENT: INSERTION LOSS 80
FIGURE 6-10: MEASUREMENT: INSERTION LOSS (LOWER BAND) 81
FIGURE 6-11： MEASUREMENT： INSERTION LOSS (UPPER BAND) 81
FIGURE 6-12: MEASUREMENT: PHASE DIFFERENCE 82

ix
1 Introduction

Branch line coupler is one of the passive components, which is commonly used in

microwave integrated circuits and R F measurement appliances. The fundamental

features of branch line coupler include coupling specific amount of power to coupling

port at certain phase difference between output ports, combining and dividing power

at specific frequency band. There are many R F applications involving branch line

coupler, such as balanced amplifier [1], balanced mixer [2] and modulator [3]. They

are all providing competitive functionalities when comparing to similar applications

in single-end structure.

However, due to the use of quarter wavelength transmission lines in the standard

design, the bandwidth is greatly limited. Specifically, if the bandwidth of operation is

defined as 1 dB output power mismatch between two output ports at the operating

frequency, the normalized bandwidth of a 90° 3-dB branch line coupler is about 10%

[4]. Such bandwidth can only satisfy the requirement of narrow band applications.

The size of branch line coupler is another critical drawback. As the use of quarter

wavelength transmission line, the dimension of a branch line coupler is significantly

large, and it would be even worse when it is designated to operate at lower frequency

band. For instance, the eventual size of the R F circuitry adopting branch line coupler

is probably greater than double to that of a single-end circuitry.

Various research had been conducted to study the method of improvement on the

branch line coupler performance, probably bandwidth enhancement and size reduction,

of branch line coupler. A number of reports proposed different broadband coupler

designs by different technologies [5-9]. However, either they are non-planar

(requiring external components) or they require further bonding process after

1
fabrication, which is not comparable to the conventional design to some extent.

Relatively, the reported improvement on size reduction of branch line coupler is more

encouraging. A m o n g those report, two of them propose a small planar design of

branch line coupler by artificial transmission line [10] and inducing extra open stubs

[11]. The two designs featured at smaller dimension, with ideally the same bandwidth

performance as that of conventional design. Apart from that, realization of branch line

coupler by integrated circuit technology is another bright solution in size reduction. In

[12-14], the designs of quadrature hybrid using lumped elements are proposed, when

the bandwidth of [14] is even up to 50%. However, its performance is severely limited

by the parasitic effect from lumped elements.

Recently, a dual band branch line coupler was proposed using lumped elements [16].

The coupler can operate at dual frequency bands by replacing quarter wavelength

transmission line with metamaterials formed by lumped elements and transmission

lines. Although it can be used in dual band application, again, the performance is

degraded because of the parasitic effect associated with the lumped element, and

fabrication cost is increased because of extra fabrication process for lumped elements.

In fact, the actual dimension of the circuit is even greater than the conventional design.

In this thesis, two designs of dual band 90° 3-dB branch line coupler are proposed.

Both of these two designs can be fabricated by microstrip technology and the

dimensions of the circuits are dramatically reduced. However, the performance of the

second design is significantly better than that of the first design in terms of size

reduction ratio and operating bandwidth. That design can be comparable to that of

standard design of 90° 3-dB branch line coupler too.

This thesis is divided into eight chapters. Chapter 2 is about the basic theory of 90° 3-

dB branch line coupler together with the corresponding mathematical analysis.

2
Chapter 3 introduces alternative designs reported in the literature with superior

performance than the standard structure. In chapter 4，two novel designs of dual band

90° 3-dB branch line coupler are discussed in terms of the theories and design

procedures. Simulation results are followed in chapter 5. Chapter 6 presents the

details of implementation, measurement results together with the discussion. Lastly,

chapter 7 concludes what this thesis contributes and chapter 8 proposes

recommendations for the possible future work.

3
2 Basic Theory of Branch Line Coupler

A branch line coupler, unlike other circuits, consists of four terminals and this can be

regarded as a four-port network. Introduction of four-port network and specifically,

analysis of branch line coupler will be shown in details in this chapter. Frequency

response of the design is simulated and presented in later part of this chapter for

illustration.

2.1 Four-port network

• •
Port 1 Port 2
Four-port network
• •

Port 4 Port 3

Figure 2-1: General four-port network

Fig. 1 shows a general four-port network with incident port, output port, coupling port

and isolated port. Let Pi be the input power to port 1 while P2, P3 and P4 are the power

delivered to port 2，3 and 4 respectively. Power delivered to specific port with respect

to input port, port 1 in this case, is defined as coupling coefficient, C. Which is

C 二-lOlogA n=2, 3, 4
Px (2-1)

For example, if port 3 is specific as coupling port, the coupling coefficient is

C = -101og^ dB
Pi

4
For the isolated port, say port 4，the power deliver at the port with respect to input

power is specified as isolation, I, which is

/ = -101og^ dB (2-2)

尸1

Moreover, directivity, D, is defined as the ratio between coupled power to isolation.

That is

D = I-C (2-3)

All four-port networks can be characterized by the above parameters. There are many

examples of four-port networks, such as directional coupler, rat-race hybrid and lange

coupler. Branch line coupler is also a kind of four-port device and thus it can be

characterized by the same approach.

2.2 Even-odd mode analysis

与，90°
V2
Port. 1 1 ^ ! - • Port 2

Z。，90。 Z。，90。

•，90°

Port 4 • - ^ ― P o r t 3

Figure 2-2: Proposed design of quarter-wavelength transformer

The topology of branch line coupler is shown in

Figure 2-2. W h e n all ports are connected to matched load, the power input at port 1

(as an input port) will deliver to port 2 (output port) and port 3 (coupled port) equally

with the phase difference of 90°. And there should be no power emerged at the port 4

5
(isolated port). Thus, this coupler has the coupling coefficient of 3 dB and zero

isolation.

It's not easy to analyze the behavior of branch line coupler by straightforward

approach as the topology of the 4-port network. Although there's transient analysis of

four-port network reported in [17], using A B C D manipulation under even-odd mode

analysis [18] is the typical method nowadays for symmetrical network.

Before applying even-odd mode analysis, there are some assumptions about the

network:

• The network is symmetric,

• The network is lossless, and

• The ratio of wavelength to line is sufficiently large so junction discontinuity

effect can be neglected.

To proceed even-odd mode analysis, it's first to apply even mode excitation and odd

mode excitation, respectively. And superposing both results derived and the behavior

of four-port network can be determined.

2.2.1 Even mode excitation

In Figure 2-3, if in-phase test signals with amplitude of ^ are applied at port 1

and port 4 respectively, there will be the same voltage change caused by two

sources at the point of line of symmetry. Therefore, the network can be simplified

as two two-port networks shown in Figure 2-4.

6
 , Port 1 Port 2

I Line of svmmetrv

— 乙

Z。,90° 90° Z。，90°

V2

Port 4 Port 3

Figure 2-3: Branch-line coupler with line of symmetry with two in-phase signal generators

^，90°
Port 1 芯 ’ Port 2

r II II �
• Z。，45° Z。，45°|

Ll，90°

Port 4 迟， Port .3
份 — H ^ H I
广 II II

卜，45° Z。’45°|

Figure 2-4: Decomposition of the branch-line coupler in even mode analysis

2.2.2 Odd mode excitation

A s shown in Figure 2-5，if out-of-phase test signals with half amplitude are

applied at port 1 and port 4 respectively, there will be the same amplitude of

voltage change but in different polarity caused by two sources at the point of line

of symmetry. Therefore the potential7 at the line of symmetry is always zero.

Therefore, the network can be simplified as two two-port networks shown in

Figure 2-6.

年 ’ 90°
Port 1 Port 2
欢 — H H H I
Line of symmetry

Z。’ 90° ，90° Z。，90°

V2

Port 4 Port 3

Figure 2-5: Branch-line coupler with line of symmetry with two in-phase signal generators

1，90°
Port 1 石， Port 2
1 I

广 II II 、

•，90°
Port 4 Port 3
— 欢 — I I

广 II II T

.z。，45° z。，45°-

Figure 2-6: Decomposition of the branch-line coupler in odd mode analysis

8
 2.2.3 Mathematical analysis

Even-odd mode analysis is adopted for derivation in order to explain the behavior of

branch-line coupler. Since the decomposed network can be treated as circuit cascaded

by two open shunt stub and a series quarter wavelength transmission line, the overall

A B C D matrix is just the multiplication of three A B C D matrixes.

The admittance of shunt open stub having electrical length of <9 is

Yos: J ^ ^ (2-4)
ZO

By A B C D matrix manipulation, A B C D matrix of Figure 2-4 is:

「A 「1 Ofcos: yZoSin^l …
= .sin 没 a
Lc D 丄 Ll^仍 ij 乂 cos 沒 ij
_ 1 oT cos6' ； " Z Q sin 6 [" 1 0 一
=.tan 没 , . s i n 没 . -tan^
1 1 / cos 6 J —丄
； Z o Z。 」L Z 。 」 （2-5)
1 oir 0 j z j ^ j 1 o_

二多。Aj^I^O 0 l"z。1_

=丄「
一 7 ? 1 / " 。 - 1 .

By A B C D to S conversion, reflection and transmission coefficients, Fe and 7；

respectively, in Figure 2-4 is,

J-i+y-y+i) (2-6)
- ( 一 l + y + y一1)
=0

9
 T= ？
‘Ae+BjZo+CeZo+De
2
=1 V~T= (2-7 )

A B C D matrix of Figure 2-6 can be evaluated by similar approach,

The admittance of shunt short stub having electrical length of ^is

= (2-8)
怒 ZO tan^ ^ )

By A B C D matrix manipulation, A B C D matrix of Figure 2-6 is:

「A 「1 o f cos没 ^^osin^lr ^ Q]
= .sin (9 ^
k Z)丄 k ij ] � cos沒 1」
_ 1 oT cos <9 J'z。sin(9「 1 0_
= j 1 .sin 没 ^ J 1
1 / cos 9 I
_ ZO TANO 」 , Z 。 」L ZOTAN没 」 （2-9)

1 oir 0 jzj^j 1 o_
假 ilyV^/z. 0 \[-j/z„ l_
=丄「1
一w L " z 。 1 _

By A B C D to S conversion, reflection and transmission coefficients,厂。and T。

respectively, in Figure 2-6 is,

厂 Ao+BjZ„-CoZ。-Do
。-A。+5„/Z。+C。Z。+D。
一 十 1) (2-10)

=0

10
T= ？
o A。+fi。/Z。+C;Z。+D。
2
=7 (2-11)

Superposing two sets of results and the following s-parameters can be obtained:

S - H j i - o

S -Te+T。- J
2'一 2 一
(2-12)
s -Z^一—丄
r 2 -

(2-12 ) shows that when port 1 is input port, port 4 is isolated port, port 2 is output

port and port 3 is coupled port, the branch line coupler demonstrates:

• N o reflected power at input port (^i, =0),

_ N o power is emerged at isolated port (54, = 0 )，

• Half power emerges to output port and coupled port respectively

(1^2. PH-^s. r = i ) , and lastly,

• There is phase difference of 90° between output port and coupled port.

Due to the symmetry of branch line coupler, any port can be assigned as input port

and the behaviour of branch line coupler is still the same based on the above

principle. For instance, if port 2 is used as input port, port 1 and port 4 are output

port and coupled port respectively, while port 3 will be the isolated port. S-

parameters of this case are: 5,2 = - 乂权 and S,^ = - By

11
 sweeping different port as input port, S matrix of branch line coupler can be

derived as:

"0 J 1 0'

R^ - 1 ； 0 0 1

[ 〜 乙 1 0 0 ;

0 1 j 0

2.3 Simulation results

In order to have a better illustration on the behavior of branch line coupler, frequency

responses of a branch line coupler are demonstrated in Figure 2-7 to Figure 2-10.

Simulation is performed by Advanced Design System, A D S , by Agilent Technologies

[15]. In principle, a branch line coupler has zero return loss and isolation loss. There

are -3dB power output at both output port (S21) and coupled port (S31) with the

phase difference of 90°. It should be noted that output port power, S21, is degraded

faster than coupled port power, S31, along frequency changes from centre frequency,

although they have the same power level at the operating frequency. Regarding the

phase difference, it shows that there's phase difference of 90° between signals of

output port and coupled port operating around the centre frequency.

12
 0

I:
-40 Y
！

_50 i - — - - - _ -- 1 - — "1 - — 1 . 1
0.8 0.9 1 1.1 1-2
Normalized frequency

Figure 2-7: Simulation results: return loss

I -3.5 / \ +|S21|

1 / 1 +|S31|

— —
1 T. — — 「-…p
^jI— -
0.8 0.9 1 1.1 1.2
Normalized frequency

Figure 2-8: Simulation results: insertion loss

13
 0

1 -20 ^^^^^^

i \f
一 -4° U
0.8 0.9 1 1.1 1.2
Normalized frequency

Figure 2-9: Simulation results: isolation

200 Ayyws/^

150
m 100
？ 50 -^phase(S21)

CO 0

《-1。。
-200 - -r —^ - i 丨

0.8 0.9 1 1-1 1.2

Normalized frequency
Figure 2-10: Simulation results: phase difference

14
3 Review of Advanced Branch Line Coupler

Designs

Since it is acknowledged that the conventional branch line coupler design suffers from

large dimension and narrow operating bandwidth, alternative solutions of branch line

coupler based on different circuit topologies and/ or implementation technologies

have been reported. In this chapter, several advanced designs with improved

performance (reduced size, wider bandwidth etc) are compared with the conventional

design.

3.1 Broad-band uniplanar branch-line design

Substrate Microstripline

Ground
(a)

Substrate Transmission line

Ground
(b)

Figure 3-1: Transverse diagram of (a) microstripline and (b) coplanar waveguide

15
For bandwidth enhancement, coplanar waveguide as well as slotline ring may be used

instead of microstripline [5]. There are better illustrations of the difference between

microstripline and coplanar waveguide in Figure 3-1.

In the conventional design of branch line coupler, the electrical length of transmission

line is not 90° when the signal frequency is different from operating frequency. Thus

amplitude and phase performance will be deviated significantly and it limits the

operating bandwidth. In this design, the frequency response of the junction formed by

coplanar waveguide and slotline is less sensitive to frequency variation than that of

microstripline and better phase and amplitude performance would be resulted.

臉〈_1
Figure 3-2: Proposed design of branch line coupler using coplanar waveguide and slotline

According to the measurement results, the bandwidth is improved to 4 0 % when the

amplitude mismatch tolerance is as large as IdB. This is much better than that of the

conventional branch line coupler and it is well suitable to be adopted in monolithic

integrated circuit. Nonetheless, the design is including slotline structure and additional

coplanar waveguide-to-slotline transition structures for connection to other circuit

components are necessary. This increases the difficulty as well as the cost of

operation.

16
 3.2 Compact branch-line couplers using slow-wave

structure

Miniaturization by capacitive loading is extensively used in microwave circuit designs

[19-21]. The main idea is to replace the quarter wavelength transmission line with

artificial transmission line of shorten length.

-J— — ^ ^ ^ ^ ：；;^^ ZoTL, VpTL

Cp T" T" 丁 丁
一 一 一 一
—一一
- - - - -

(a)

/ 、 Z o T L .

(b)

Figure 3-3: (a) Artificial transmission line with shunt capacitances and b) Artificial transmission line
with open circuit shunt stubs

Showing in Figure 3-3(a), the artificial transmission line is actually a transmission line

with additional periodic capacitive loads. The quarter wavelength transmission line

can be realized by a relatively high-characteristic-impedance transmission line with

shorter physical length. Since the open stub shows capacitive impedance in nature,

artificial transmission line can be modified by replacing capacitive loads with short

open stubs as shown in Figure 3-3(b). With this arrangement, the implementation of

branch line coupler can be truly planar.

17
The measurement results show that there is similar performance comparing to that of

the conventional design with around 50% size reduction. However, similar to the last

design, there are many junctions between stubs and further optimization should be

made before fabrication of the circuit, especially when it is operating at higher

frequency.

mmm

Figure 3-4: Geometry of branch line coupler realized by artificial transmission line

3.3 Miniature branch-line coupler using eight two-

step stubs

In [11], the author proposed a pi-transmission line network, shown in Figure 3-5(a),

which is equivalent to a single transmission line of 45°, as refer to Figure 3-5(b). By

cascading a couple of proposed structure, a quarter wavelength transformer would be

formed and is ready to replace the quarter wavelength transmission line in the

convention design.

To reduce the dimension of the design, the author tried to fabricate eight open stubs

inward. Instead of locating them in straightforward manner, a single open stub is

restructured as Figure 3-6. All open stubs could be fabricated evenly to reduce

coupling loss and the branch line impedance is easier to realize under this strategy.

18
 (a)

450’ Zo

(b)

Figure 3-5: (a) Proposed half quarter wavelength transformer and (b) Equivalent structure

^ib，没lb

z - ^ B H — > z <

Zib，没lb

Figure 3-6: Restructure of single open stub into two-step structure

Because the physical length of a couple of proposed structure is shorter than the

quarter wavelength transmission line by half, the dimension of the proposed design is

dramatically reduced by 75%.

Although it has the same performance as the conventional design theoretically, the

performance of this design is severely suffered from junction discontinuity effect.

Further optimization is necessary so that the adverse effect could be compensated.

19
 ，45。

I n — I _ / |

Z , ’ 45° _
十 • • 一 Z , , 45°

I
ZA，45°

Figure 3-7: Final branch line coupler with all proposed size reduction techniques.

3.4 Wide band lumped-element 3-dB quadrature

coupler

There are many proposed lumped element equivalent model [12-14] of branch line

coupler. A relatively compact design using lumped elements is reported in [12] and

the schematic is shown in Figure 3-8(a). This simple design provides exactly

equivalent performance of branch line coupler at the centre of operating frequency.

Nevertheless, the operating bandwidth is narrower than that of branch line coupler and

it is only useful in some applications with moderate bandwidth requirement.

20
 � �
c-L -Lc

o
Cm 7 ^ , Cm
J z ^ o
c•丄 丄c。

(a)

平 平 午
J
a cU
1/2]

。I KI I H °

。丄 l( I K 。

4c 丨 �
(b)

Figure 3-8: (a) Lumped element quadrature hybrid and (b) wide band quadrature hybrid design

In [13] the proposed design also consists of pure lumped elements as in Figure 3-8(b).

The topology includes altogether 13 lumped components. A s it's still a symmetric

network, even-odd mode analysis would be adopted for analysis. The derived Lp, Cm

and Cs are all easily realized (maximum values of inductance and capacitance are 3nH

and 2pF respectively). The reported measurement results show that the maximum

21
bandwidth of 5 0 % is achieved, with sufficient performance on return loss and

isolation loss.

This lumped quadrature hybrid design provides a pure lumped element solution with

easily realizable component values. It's suitable for monolithic integrated circuit

implementation. However, the massive use of inductor may be a problem as it

occupies extremely large space and the increase of production cost is caused.

3.5 Dual band branch line coupler design using left-

handed transmission lines

c/2 q/2

下
(a)

iS C/2
>-CIZHI ‘ )l L )f—CZ]-
Ir c/2 q I I,

f
(b)

Figure 3-9: (a)Part of left-handed transmission line and (b) Composite right/left handed transmission
line for replacement of quarter wavelength transmission line

22
In the conventional design, a branch line coupler can be operated at fundamental

frequency and odd harmonics. In other words, branch line coupler can function at

frequencies when the electrical length of the transmission line is 90n°, where n is any

odd number.

In [16], a left-handed transmission line is proposed. In Figure 3-9(a), it shows that

there are a series of capacitors and shunt inductors. It can be used together with

ordinary transmission line, or right-handed transmission line to form a composite

right/left handed transmission line as Figure 3-9(b) such that,

(ZK/I) = —90。

0if,) = -2W (3.1)

where/； and fi are arbitrary frequencies

With the use of composite right/left-handed transmission line, (3-1) can be satisfied at

arbitrary frequencies of/； and/2, with appropriate values of Cl, Ll and Ir.

This was the first reported dual band branch line coupler in the literature

demonstrating the use of metamaterials [22-24] implemented by lumped elements.

However, its performance is limited by parasitic and self-resonance of lumped

elements. Moreover, the dimension of this design is significantly larger than that of

conventional design of branch line coupler operating at lower frequency band. A s well,

the use of lumped elements and via hole in this design would cause a higher

production cost in comparison with that of planar design.

23
4 Design Theory of Dual Band Branch Line

Couplers

The use of branch line coupler in dual band applications is limited because of its

narrow operating bandwidth. Although there are several innovative wideband designs

reported and some of them successfully increase the bandwidth up to 50%, the

performance is degraded comparing to that of conventional design in terms of

insertion loss, return loss and isolation.

Despite the degraded performance, these wideband designs are not well suited for

dual band applications owing to the following reasons,

1. Since most of dual band applications feature large frequency separation

between two operating frequencies, the bandwidth requirement of the wide

band coupler is usually much higher than 50%. For instance, in order to use a

wideband branch line coupler for a dual band application operating at

fundamental frequency and second harmonic, the operating bandwidth of the

coupler, with respect to the centre frequency, is,

BW = ~!\ , =1 二 66.67%
3 (4-1)

The case is even worse for some dual band applications with farer frequency

separation.

2. These wideband branch line coupler designs provide aflatfrequency response

throughout the operating bandwidth. It may cause hazard to some dual band

applications when unwanted signal of mid frequency band induces

24
 interferences to the operation. Suppression of mid-frequency band signal is a

favourable feature for dual band operation.

Although there was proposed dual band branch line coupler which was implemented

by composite right/ left handed transmission lines [16], the dimension is significantly

larger than that of conventional design and the performance is degraded because of

parasitic effect from lumped elements.

In this chapter, two designs of dual band branch line coupler are proposed. Topology

of each designs, mathematical analysis and design procedures would be described in

details. Afterwards, the comparison of two designs will be presented too at the end of

this chapter.

4.1 Design 1 一 dual band branch line coupler with

sub-optimum performance

The conventional design of branch line coupler uses two pairs of quarter wavelength

transmission line with different characteristic impedances. Feasibility of branch line

‘ 2 2
Port 1 Port 2

Port 4 Ports

Figure 4-1: Topology of design 1

25
coupler realization using different electrical lengths of transmission lines is reported

in [25-26]. The main objective of such realization is for ease of M I C and M M I C

implementation.

In design 1 in Figure 4-1, the dual band branch line coupler consists of two pairs of

transmission lines with electrical length of——and——，respectively,at mid-

2 2

frequency where m and n are any integer. With different values of m and n, frequency

response of the configuration deviates from that of the conventional design.

In the text below, a completed analysis will be delivered in details so as to derive the

value of Zi and Z2 such that it can operate in dual frequency bands.

4.1.1 Analysis and design

The following conditions should be adopted in the design of dual band branch line

coupler:

(4-2)

1^2.1 = 1^3.1-

/ V2 (4-3)

ZS,, — ZS,. = ±90°

(4-4)

These are the features of a conventional branch line coupler at operating frequency,

except ( 4-2 ). The adoption of sub-optimum requirements is of ultimate importance to

attain dual-band operation of the coupler. Subsequently, from ( 4-3 ) and ( 4-4 )，the

formulas for evaluating Zi and Z2 can therefore be derived as:

,,仍 8K r 2c Y i 2c').
/ ⑷ = — — ； + 1+ — — 7 1+ — — 7 = 0

c +c V c +c c+c y (4-5)

_ c + c'

叫 丁 (4-6)

26
 P (4-7)

where

K = cot^-y^tan- cot0 + j8cot-

\ 2八 2J

c = l + y?tan苦 2 c o t < 9 t a n

c' = l-/?cot^ 2cot0 + j6cot^

2V 2>

没=竿(1 + 句

/2 + /1

n and m are positive integers; zi and Z2 are normalized values (with reference to Zo);/；

and/2 are the centre frequencies of the lower and upper bands.

It is interesting to note that the various s-parameters are also inter-related:

c 2_ e 2
^11 一 〜 1

e 2 c 2 1
+ '^21 = 2

Practically, the proposed structure should also exhibit small input reflection, good port

isolation and low insertion loss. For further investigation, the simulated responses of

the proposed coupler versus fractional bandwidth are plotted in Figure 4-4. There are

two different configurations for consideration:

1. A(n=2, m =1), and

2. B(n=l, m=2)，

27
which are the smallest possible designs.

Z,, 180o@,丨+,2
1 2
Port 1 Port 2

Z2’9。。@"^I，| I Z2’90°@~^I=,2

Port 4 Port 3
Z丨，

(a)

z,, 9 0 ° @ A 1 A
— Port 2

Z2，180° @ A l A Z2,180° @ ^ ^ ^ ^

Port 4
— Ports
Z丨，90。@力+72
1 2

(b)

Figure 4-2: Design 1 with different values of m and n: (a) Configuration A (m=7, n=2) and (b)
Configuration B ('m=2, n=J)

It can be seen that, by the proper choice of n and m integers, these configurations

offer sub-optimum return loss (better than 20 dB) and insertion loss (less than 3.2 dB)

over a fractional bandwidth ranging from 0.2 to 0.43 (1.5/} </2 < 2.5/；). Performance

28
of the design in terms of insertion loss, return loss and isolation loss among fractional

bandwidth is shown in Figure 4-3.

0 -1 -3

卜
” - / \ . \
；F-15- I V ^ > - -3.3二

\ V \ / - -3.4

込A
-30 - \ /
/ A
-一 B
…
- -3.6
-35 -I 1 1 -3.7

0.2 0.3 0.4 0.5

Fractional bandwidth (6)

Figure 4-3: Return loss, port isolation and insertion loss versus fractional bandwidth

In addition, the construction of the coupler is also constrained by the range of

impedance that can be realized. Figure 4-4 shows the variations of Zi and Z2 as a

function of the fractional bandwidth. These results indicate that the required

impedance level (35 - 80 ohm) is well within the fabrication capability of most

manufacturing technologies, such as microstrip.

29
 80
^ P x ^
§ 70 - -&-B

fee-
G 50
o
o
2 •
兰 40 ......• Z,
O •、.•.••
30 ‘ I I I I I
0.2 0.25 0.3 0.35 0.4 0.45 0.5
Fractional bandwidth (5)

Figure 4-4: Characteristic impedance of series arm (Z；) and shunt arm (Z2) among fractional bandwidth

To sum up, the design procedures for the proposed coupler are given as follows:

1. Select suitable values for n and m based on the fractional bandwidth (6)

calculated;

2. Determine the value of P by solving ( 4-5 ) numerically;

3. Compute the values of Z\ and Z2 using (4-6 ) and (4-7 ).

4.1.2 Size comparison

To have a better understanding on the size reduction ratio of the design, the size of

proposed design and conventional design operating at lower frequency band are

compared in terms of perimeter and area.

The perimeter of conventional design is

P _ = 4 X 冬 = 、 （4-8)

The perimeter of proposed design is

30
 Af Af
"cfgj =2x +
{ 2 4 j (4-9)

where

-fm is the mid frequency of/； and/2.

-A,丨 and are the wavelengths of signals with frequency of/； and/爪,respectively.

Since A, can be expressed in terms of 丨 as,

Jm r
J m

:2/1,1/i (4-10)
=/丨 +/2

Solving ( 4-8 )’ ( 4-9 ) and (4-10 )’ we get,

fA, A, ) 2/,
=2x i + i X ^
〜 [2 4 J /,+/,

川 2 ( 4 _ i "

r com f \ -f
J\ 十 Jl
-Pconv X Sp^

Where S^i is the perimeter scaling factor of this design.

Therefore, the proposed design features size reduction only if,

Pcom < Pc{g_l

/. +/2 (4-12)
2/丨 >/2

In other words, the perimeter of the design will be reduced from that of conventional

design when the upper operating frequency is smaller than second harmonic of lower

operating frequency.

The area of conventional design is

31
 Xr ^ Af
A = f\ =,丨

conv 一 4 J " 16 (4-13)

The area of proposed design is

Ar Af
A J m w Jm

^cru (4-14)

Solving ( 4-10 )’ ( 4-13 ) and ( 4-14 )’ we get,

A, Af ( 2f、V
A - Ji NX Ji sc

16 U1+/2J
2 (4-15)

-A J 2 彻 1
- U ^ f J
~ ^conv X 'S'^i

Where Sai is the area scaling factor of this design.

Therefore, the proposed design features size reduction only if,

^conv < ^cfg_l

、\ (4-16)
/2>/,(2V2-i)
«1.83/丨

Thus, the design features reduced size only when upper frequency is larger than 1.83

times of lower frequency.

In general, S^i and Sai versus fractional bandwidth are plotted in Figure 4-5. Area and

perimeter reduction can be achieved when fractional bandwidth is greater than 0.33.

In this connection, size reduction is possible over a wide fractional bandwidth,

although it's conditionally true.

32
1.5丨
i

0.2 0.25 0.3 0.35 0.4 0.45 0.5

fractional bandwidth (5)

Figure 4-5 S^i and S^； versus fractional bandwidth ( d )

33
4.2 Design 2 - dual band branch line coupler using

shunt stubs

The first design of branch line coupler features reduced size at specific fractional

bandwidth and provides —3dB output power at the exact operating frequencies.

In order to further improve the performance, the second design of branch line coupler

is proposed.

I • • • I

I ^m ‘ •^B _ 氏‘’ Zfi

(a)

90°, Z了

(b)

Figure 4-6: (a) Proposed quarter wavelength transformer and (b) Equivalent structure

Unlike changing the dimension and characteristic impedance of transmission line

from the original design, a quarter wavelength transformer is proposed in this design,

showing in Figure 4-6(a). It is a pi-transmission line network, with two open stubs

located at two ends of series branch. They all have the electrical length, 6’„，of 90° at

mid frequency. By forming the proposed transformer using quarter wavelength

transmission line with suitable characteristic impedances, Za and Zb, such that,

•^11=0 (4-17)

>^21=1 (4-18)

34
 ZS21 =90° or 270° (4_19)

The proposed transformer network can be used to replace the quarter wavelength

transmission line in the conventional design to provide dual band operation.

4.2.1 Analysis and design

Because the proposed transformer can be treated as cascaded structure by two shunt

open stub and a series transmission line, S-parameters of this transformer can be

easily found by manipulating its A B C D matrix.

The admittance of shunt open stub having electrical length of dm is

(4-20)

By A B C D matrix manipulation, A B C D matrix of Figure 4-6(a) is:

_A B'

� 1 o f cos 没J J Z . s m e l r J
= .sin (9,„
i^os 1 J-y- cos 没"I Yos 1
L ^A 」
- 1 Oir jZ, s i n ^ j r 1 Ol (4-21)
=:tan^ .sin 九 „ ..tan 没 ,
j—^ 1 j - ^ cos氏，
L ^B 」L ^A 」L 乙B 」
- 2 “
cos - tan sin 九 jZ^ sin e,„

jsmdi^ + ^-^tan'e] cosO„, - t a n 《 , 丨 sin《„

It's crucial to note that [A] = [D] because of symmetry.

The equivalent S-parameter matrix is therefore,

35
 -A + yZ。-CZo-D 2iAD-BC)

卜 ， 〜 ] 一 〃 ％ 严 。 〃 / z 。 + C Z 。 + D
522」"2 2 -A + ^^-CZo+D

A +^ +CZo+D A+^ + CZQ + D

From (4-21 ) and ( 4-22 )，（ 4-17 ) implies that

% =CZ。
/ ( 4-23 )

Furthermore with (4-18 ), it indicates that

S 2
2' ％ +CZo + D

2
= 2("%。）

= 5 (4-24 )
COS氏丨-|^tan《„ s i n t +

1
一 R + jS
二 R-jS
-R^+S^

z z
where R = cos ——-tan sin , 5 = —sin
Zfl ZQ

To fulfill the condition of (4-19),

R =0
5"=±1

Thus,

Z 、 1

Z。Isin 九 I (4-25)

Z。 cos《“

36
As refers quarter wavelength transmission line operating at mid of operating

frequencies, the electrical lengths of all transmission lines are

没 1 = 仏

”： (4-26)
e , 仏

where /,,, = A l A

To characterize the normalized operating frequency separation, fis defined as

fractional bandwidth such that

f - f
^ = (4-27)
Ji + J\

Rearranging ( 4-26 ) and ( 4-27 ), 61 and 62 can be expressed in terms of e.

6^2 - = en
(4-28)
+<9, =71

And finally,

jr n
e, = — - e - (4-29a)
1 2 2

^ n n
—+ (4-29b)
‘ 2 2

Substituting ( 4-29a) and (4-29a) into (4-25 ) subsequently, we obtain,

- 1
ZO “ (
COS e—
V 2j
(7t\ (4-30)
COS e—
Zg _ V
Zn . 2(双、
[ 2 )

37
There are the same values of ZA and Zb obtained when either 6i and 62 are substituted

into (4-25 ).

In general, the dual band branch line coupler with proposed structure can be

constructed according to ( 4-30 ) and the overall schematic and parameters are

illustrated in Figure 4-7 respectively.

Z2 ， 90°
Port 1 I I Port 2

二 I Z,’9。。1/
Z4 ’ 90°

Z3 , 90°

/ \ Z i ’ 9 。 。 | \

Porf 4 Port 3

Figure 4-7: Topology of design 2

And the design equations are,
(7l\
COS e—
Z, 二 1 Z, ^ I 2)
Z 。 V 2 cosf Z。
\ 2y V
( ( 4 - 3 1 )
COS S—
Z3 二 1 I 2)
Zf) ( Zq 21 ^ I
0 cos e- 0 sin" e-
l 2j I 2j

To further simplify the adjacent open stubs at each port, open stubs with characteristic

impedances of Z2 and Z4 are merged into a single stub, as illustrated in Figure 3-6.

38
 Z2，90。
r - i ^ H H I

z - z

Figure 4-8: Simplification of double stub into single stub

According to Figure 3-6, Z,„j can be expressed as

1 .tan <9 .tan 没

=J +J
Z Z 7
.(Zj+ZJtan^
=J
Z2Z4 (4-32)

Z - Z A 1
Z 2 + Z 4 ；tan^

O n the other hand, Zm can be expressed as

'"2 ； tan^ (4-33)

In comparison with (4-32 ) and ( 4-33 )，

^ Z2 + Z4 (4-34 )

It is necessary to note that the characteristic impedance of the single open stub (Z^) is

smaller than either Z2 or Z4. The simplified dual band branch line coupler is shown in

Figure 4-9.

The final topology of this design consists of two pairs of transmission lines (quarter

wavelength) and four shunt stubs (open end). With further optimization after rough

simulation, those four open stubs can be located inside the space at the centre of the

design and the size of the whole design is greatly reduced. However, attentions should

be paid to induce junction discontinuities and coupling between branches so that the

performance wouldn't be degraded.

39
 «
I Z3，90°

Z, ， 90°
I

_ Port•2

Z^ ’ 90。 @ 几 Z2，90。

Z. ’ 90°

• — — •
Port 4 • _ Port 3
I , 90° I

Figure 4-9: Proposed structure of dual band branch line coupler

I
Port 1 Port 2

Port 4 Port 3

Figure 4-10: Compact dual band branch line coupler

A n d (4-35 ) is the design equations for Z；, Z2 and Z5.
Z 丨 二 1 A =
Zn rr ( 7t\ Z^ (
u V 2 COS e— COS e—
I 2j I 2)
(n\
COS E— (4-35)
^S V

Z 。 +
V

Since the electrical length of proposed transformer structure is equal to 90° or 270° at

the frequencies of interest, insertion loss, return loss and isolation loss are excepted

the same as that of conventional design among fractional bandwidth.

40
 1
8 160 \

J 120 令 、

1 8 。 。 _ E ~ E ^ ^ ^ ^ ： ! ： ! !
2 40 # ^
CO
•C II
•
0 1 , ： 1

0.2 0.25 0.3 0.35 0.4 0.45 0.5

Fractional bandwidth (5)

Figure 4-11 Characteristic impedance of series arm (Z；), shunt arm (Z2) and open stub (Z5) among
fractional bandwidth

Similar to design 1，the implementation of design 2 is constrained by characteristic

impedance level of transmission line, which is changed among fractional bandwidth.

As shown in Figure 4-11，the impedance is realizable when its characteristic

impedance is between 30^2 to 90^2, and the corresponding fractional bandwidth is

valued between 0.3-0.5. Note that the characteristic impedances in both series

branches and shunt branches are well between 350 to 80^2 over a large fractional

bandwidth. The realization of induced open stub is the major limitation of the design

implementation.

All in all, the design procedures for branch line coupler design 2 are given as follows:

1. Determine the fractional bandwidth by solving (4-27 );

2. Compute the values of Zi, Z2 and Zs using (4-35 ).

4.2.2 Size comparison

In this subsection-comparison between the conventional and the proposed designs

would be made in terms of perimeter or area.

The perimeter of design 2 is

41
 A,
也 4 (4-36)

Solving ( 4-8 )，( 4-10 )and (4-36 )，we get,

fA, A, ) 2/,
〜 I 2 4j /1+/2

= 入 产 1

一 n y—
/1+/2
=/^conv X S

2f
Where Sp2 is the perimeter scaling factor of this design. As is always smaller
f\ + fi

than 1，the perimeter of design 2 is always smaller than that of the conventional

design.

The area of design 2 is

Xf ^ A> f
A — _

一 丁 J 一 (4-38)

Solving ( 4-10 )’ ( 4-13 ) and (4-38 )，we get,

‘ 2 - 16 "U.+/J

( 2 f V
\Ji 十 /2 y

( 2 f V
Where Sai is the area scaling factor of this design. A s — is always smaller

fi)

than 1，thus the area of design 2 is always smaller than that of the conventional

design.

In general, Sp2 and Sa2 versus fractional bandwidth are plotted in Figure 4-12.

Unconditional size reduction can be achieved in this design, in terms of perimeter and
42
area. It is crucial to know that the reduction ratio is better when the fractional

bandwidth is higher.

0.8 "•“ S"

o : 士 Sm

r
1 。 . 4 ！ ^ ^ ^ ^ ^ ^ ^ ^

0.2
！

0.2 0.25 0.3 0.35 0.4 0.45 0.5

fractional bandwidth (6)

Figure 4-12 Sp2 and Sa2 versus fractional bandwidth ( 6 )

43
5 Simulation Verification

Rigorous analysis of two dual band branch line coupler designs are presented in

details. In this chapter, computer simulations of two designs operating at 9 0 0 M H z and

1.8GHz would be carried out.

Simulator Advanced Design System, of Agilent Technologies [15], is used to verify

the frequency responses of proposed designs from the circuit parameters provided by

the derived equations. Moreover, it is impossible to fabricate transmission line with

exact dimension due to processing, such as over-etching, mask misalignment, etc. As

the accuracy of line width will affect the value of branch line impedance and thus the

performance of the design, the circuit performance due to slight deviation of physical

dimension would be studied by simulating the schematic with small line width

variation. Furthermore, parasitic effects due to substrate loss and junction

discontinuities, which are obvious in practice, should be taken into account too in the

simulation. To a first approximation, these effects would be included in simulations

with appropriate circuit models accordingly. The performance of the circuit thus can

be fine-tuned by computer optimization.

5.1 Design 1

For verification, one of the possible configurations, B (with m = 2 and n-1), operating

at 9 0 0 M H z and 1.8GHz is chosen for simulation.

44
5.1.1 Schematic simulation

/；, m and n
According to the procedures described in chapter 4 and the values of /�,

defined, Z； and Z2 can be calculated by equation ( 4-6 ) and ( 4-7 ). They are about 42

o h m and 61 ohm, respectively.

Figure 5-1 shows the schematic diagram of configuration B, design 1 simulation.

Ideal transmission line model is used and all ports of the circuit are terminated by

match load.

TUN
TL1
Z=Z1 Ohm
E=11
F=1.35GHz

‘ l erm + ,..
< Term1 TLIN TUN •[二rm,
• Num=1 nTL17 TL27 f] > 二
1 Z=50 0hm Z=Z2 Ohm Z=Z2 Ohm ？ Z=50 Ohm m VAR
- L^E=I2 E=I2 M !=•」 EF 0008
JL 丁 F=1.35GHz F=1.35 GHz | __ Z1;42.5
一 = Z2=61
11=90
,1 I 12=180

TLIN
TL25
llZ. Z=Z1 Ohm 产Term
S 丁6「m4 E=I1 < Terms
< = 二 m F=1.35GHz S Num=3
__ _ Z-50 Ohm < z=50 Ohm

Figure 5-1: ADS Schematic diagram for configuration B, design 1 simulation

T o characterize the performance of a branch line coupler, diagrams of return loss,

isolation loss, insertion loss and phase difference between two output ports versus

frequency are plotted after the simulation. Figure 5-2 to Figure 5-5 are the

corresponding simulation results.

45
 0
I

- 2 5 . 】 1 - I - - - -丁…"“ ”R— "1 I

0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2

Freq. (GHz)

Figure 5-2: Simulation - Return Loss

-hry i
I
—^^^^, , . — -- j - - - J - - - — --- - J - ——.-...-「 [ I

0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2

Freq. (GHz)

Figure 5-3: Simulation - Isolation

In principle, there is slight reflection at each port, when there is no reflected power in

the case of conventional design. By Figure 5-2, it's clear to see that the minimum

reflection exists at frequencies slightly deviated from designated operating

frequencies. Although only sub-optimal solutions are obtained, the return loss is still
46
as high as 20dB. For the case of isolation loss (Figure 5-3), it shows the same

phenomenon.

There are 3dB insertion losses at both frequencies of interest, showing in Figure 5-4.

This is in good agreement with one of initial conditions stated in previous chapter.

The bandwidth of each frequency band is narrower than that of conventional design

due to the imbalance in amplitude among frequency that slightly deviated from the

centre frequencies. At mid-frequency, it's crucial to note that 6dB insertion loss is

obtained and such attenuations may cause advantage to some applications.

Apart from that, the phase difference between two output signals is equal to 90° at

two operating frequency bands. There are deviations of ±3° over 250MHz around

centre frequencies, which is sufficient for most applications.

0
"•"S21

丨：着 •10
0.6 0.8
… T
1
'1
1.2
1
1.4 1.6
1“—
1.8 2
1 “ 1
2.2
Freq. (GHz)

Figure 5-4: Simulation - Insertion Loss

47
 200 ^ ,…、
fv --^phase(S21)
？ 150
I 100 -^Phase(S31)
1 50

N -50
召-100

i
-200 i- • T 厂 ” "！ - 厂 T --..r ... -1
0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Freq. (GHz)

Figure 5-5: Simulation - Phase Difference

5.1.2 Schematic simulation with line width deviation

The deviation of branch line characteristic impedance causes the frequency shift and

performance degradation of the design. Such deviation can be characterized by

simulation so that the effect can be forecasted. Instead of using ideal transmission line

mode, the schematic using real transmission model with maximum of ±0.1 m m change

from the designated line width would be simulated. The procedures of the simulation

is as follows:

1. Using LineCalc to calculate dimension of transmission line, with specific

substrate, characteristic impedance and electrical length. The chosen substrate

will be used for future implementation, while characteristic impedance and

electrical length should be calculated and chosen from the design equations;

2. Using the calculated dimension to fill in the real transmission line model and

performing simulation. The results will be a bit deviated from the ideal

schematic simulation due to substrate loss;

48
 3. Conducting the simulation again with the ±0.05mm and ±0.1mm change of

line width and recording the results. In the simulation, d is defined as times of

0.05mm width deviation. For instance, d=0 represents simulation with no line

width deviation while d=-2 represents simulation with decreased line width by

0.1mm.

ML IN
TL32
Subst="MSub1"
W=W1 mm
L=L1 mm

— 叫 I la^AR
EF_0008
Wl"=0.79+d*0.05
W2=0.36+d*0.05
•+4 Term MLIN ML IN J. jg^^i L1=21.08
Termi |lrL34 TL33 f l f ^ jerm2 L2=43.63
S Num=1 Bubst="MSub1" Subst="MS jb' “ > Num=2 d=-2
I Z=50 Ohm 卜=W2 mm W=W2 mm| ^ z=50 Ohm
—' n L = L 2 mm L=L2 mm 丁 L-J
=• 去 MSub

1 1 • MSUB

^ ‘ ！fnUb二
TL35 H=0.625 mm
Subst="MSub1" fr^^：'^
W=W1 mm Mur=J ,
一，Term L=L1 mm — Cond=5.8e7
jerm4 < Term3 Hu=1.0e+33mm
< Num=4 > Num=3 T=10um
5 z=50 Ohm I z=50 Ohm TanD=0.0023
——1 y 」 Rough=35 urn
——
I -i

Figure 5-6: ADS Schematic diagram for line width deviation simulation

Figure 5-6 shows the constructed schematic diagram. With the use of use of LincCalc,

dimensions of transmission lines are obtained. Since the dimension depends on

substrates, substrate information is embedded in the simulation. Rather than FR4,

which is a c o m m o n substrate used in microwave integrated circuit, RT/Duroid 6002 is

chosen for circuit fabrication of design 1.

49
 Parameters Duroid 6002

Dielectric constant 10.2

Thickness ( m m ) 0.625

Loss tangent 0.0012

Table 1: Major parameters of substrate RT/Duroid 6002

Simulation result of return loss is shown in Figure 5-7. The troughs of spectrum shift

slightly. A s there is sub-optimal solution even in the case that d=0, the slightly shift

of troughs would not significantly affect the performance in terms of return loss.

Close views can be found in Figure 5-8 and Figure 5-9. They show that there is the

highest return loss when d=2 at lower frequency band, while there is the highest

return loss w h e n d=-2 at upper frequency band. Same return losses are attained at

both frequency bands when d=0. Although asymmetry is demonstrated in this

simulation w h e n d is not equal to zero, the return loss is at least 18dB, which is still

sufficient for m a n y applications.

齐-20: 1 1 1 丨-—

t =丨 Y i f
—• - [ ^ 1 J ly
- I 11
: I ！
- I 1
- 4 0 — I 1 I I I I I I I I ~ I I ~ ~ I " " " "
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2

freq, GHz

Figure 5-7: Simulation - Return Loss

50
 -16

-18 一 — ： ”

I：
：- riTZ^^-- 丨
-26 j ^ ^ ^

- 2 8 — I ~ I ~ I I I ~ I “ ― I I ~ ~ I ~ ~ I I I I I I ~ I r ^ d=2
880 890 900 910 920

freq, MHz

Figure 5-8: Simulation - Return Loss (Lower band)

-16 1 ^ ^

-18 1 ~ d=0

I
r
-26 1~~I~I~I—~I~I~I""“IIIIIII“III~~
1.78 1.79 1.80 1.81 1.82

freq, GHz

Figure 5-9: Simulation - Return Loss (Upper band)

Figure 5-11 is the simulation result of insertion loss. It shows that operating

frequencies are shifted in the case of J=-2 to d=2.

Figure 5- and Figure 5-12 are the close views of insertion loss of both operating

frequency bands. With no significant degrade on insertion loss, the deviation of line

51
 0 ^ ^ ！ ^ 1 ！ ^

-d=2 d=l d40 d=-l \ d=-2 \ \

^― - ^^^ _ - . gi^-- - — -i — -- —
j — - •-—

•只__ . . . — — —Y
\
.\\
\ —————j /— -1 — —^^^

tt ^^ : … I ^ M 丨--巧
Q__ — j — t -j* —I

- . 丨 I 丨 I i I
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2

freq，GHz

Figure 5-11: Simulation — Insertion Loss

width would cause the shift of operating frequencies. In other words, there is greater

amplitude imbalance at designated operating frequencies when d is not equal to zero.

d=2 d=l d=0 d=-l d=-2

1
•4.5 Ir ^ I I I I I I I I I I I I 丨丨 I〉
850 860 870 880 890 900 910 920 930 940 950
freq, MHz

Figure 5-10: Simulation - Insertion Loss (Lower band)

52
 d=-2 d=-l d=0 d=l d=2

II
-4.0 I I I I I [ I I I I I I I I I [ I I I I I I I I I 7 ' ' ' '

1.77 1.78 1.79 1.80 1.81 1.82 1.83

freq, GHz

Figure 5-12: Simulation - Insertion Loss (Upper band)

Phase response is also simulated accordingly. Phase difference between both output

powers are plotted. It shows that there's no significant change of phase difference in

the case of d=-2 to d=2. Therefore, the impact of the line width deviation on phase

response is not too significant.

d=2 d=l d=0 d^l d=-2

？ ^QQ ^ ^ ^ ^ ^ ^ 1
^ ： 丨 I
CO - 丨
A . .•丨 . 1 - --• • J ……- -L - . - . - - - T - - -- — ： -•-

①
</) 丨 ！： I

I I I ！ ；
JC - , ： ‘
Q . - I ; J
_100 I -R— R • •— R
〜 - '丨/ 丨 ：
- ： 丨 丨 丨 I

^ ： ‘ ^ _ H — ^ 丨 丨
CO -200— I- 1 卜 、

1 ： ： W 丨 丨 丨 I
^ ^ — I I I I I I

Q- -300 1 j 1 1 1 1 1 1 j I I I I I
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2

freq, GHz

Figure 5-13: Simulation - Phase Difference

53
5.1.3 Schematic simulation with junction discontinuity

After verifying the idea of the design, detailed simulations have to be conducted by

including junction discontinuities and substrate information. It can be performed by

A D S as well as with the provision of transmission line model. Substrate information

is again embedded into simulation.

MLIN MLIN MLIN I S-PARAMETERS |

TL46 TL43 TL36 • •
Subst="MSub1" Subst="MSub1" Subst="MSub1" S.Parain
W=V\Cmm MTEt v\A=W1 mm ivirEE W=VVD mm SP1
L=LO mm Tee5 l=L1 mm Tee7 L=LO mm Start=
I ‘ ‘ 1—ij-1 H ‘ ^-"ir^ Step=50 MHz
, r r^ vAR
产 n 丨 e"” • , MLIN MLIN EF_0008 i — Term MSub
S Temi6 rHL41 TL42 fn WD=0.57 < TermS L____
> ； ub
s«="MSub1"Sub
st="M
J ubi" W1=0.79 > Num=2 MSUB
_」Z=50 Ohm I mm W=W2 mnM W2=0.36 J_ Z=50 Ohm MSub1
“ TL=L2 mm L=L2 mm L0=2 - H=0.625 mm
= L1=21.08 ± Er=10.2
L2=43.63 — Mur=1
rh _r*l_ _____ Cond=5.8e7
^ 1 [—' ^ J I ^ f /j 卜 Hu=1.0e+33 mm
fc•‘ '..TPP ‘ “ _ - • TFF T=10um
TL45 ！ ^
'T
EE TL40 ！^ TanD=0.0023
Subst="MSub1" Subst="MSubr' ®® s[ibi-'MSub1" Rough=35 urn
A . ,e..W=WOmm W=W1 mm VV=VVD mm 产];.erm
> Ten^jgLO mm L=U mm L=LO mm < Term?
< Num=4 > Num=3
T Z=50 Ohm
—
|_」Z=50 Ohm
—

Figure 5-14: ADS Schematic diagram for simulation with junction discontinuity

After including junction discontinuity effect, the operating frequencies shift to

9 0 4 M H z and 1.802GHz. According to Figure 5-15, there are at least 20dB return

losses at two operating frequencies. It demonstrates the sub-optimal solutions of

return loss. This is the same case as that of isolation loss shown in Figure 5-16.

54
 % YY
0 -T

-30

-35 i 1 1 1—I i 1——I 1 1

0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
Freq. (GHz)

Figure 5-15: Simulation - Return Loss

丨:hrv
•25 -I
0.6
1
0.8
r—
1
1
1.2
1
1.4
Freq. (GHz)
1
1.6
1
1.8
1
2 2.2

Figure 5-16: Simulation - Isolation

Simulated insertion loss result does agree with the conditions specified. Figure 5-17

shows that the design provides zero amplitude imbalance at the centre of operating

frequencies, and the imbalance becomes greater for nearby frequency signals. There is

about 6dB attenuation at the mid-frequency as well.

55
 -10

-1 2 1 1 1 1 1 1 r-
0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Freq. (GHz)

Figure 5-17: Simulation - Insertion Loss

Similar to ideal simulation results, Figure 5-18 shows a wide phase response at both

frequency bands. Table 2 provides a summary of the simulation results of schematic

diagram with junction discontinuity.

200 —

150 ^
^ 士 S21
1100 士 S31
虽 50

S -50 [^^Be^^^^

-200 I I 1 1 1~-—I 1
0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Freq. (GHz)

Figure 5-18: Simulation - Phase Difference

56
Frequency 904 M H z 1.8 G H z

Input Return Loss 21.4 dB 21.2 d B ~

Insertion Loss (1 to 2) 3.32 dB 3.49 dB

Insertion Loss (1 to 3) 3.30 dB 3.49 dB

Isolation factor (1 to 4) 21.4 dB 21.2 dB

Phase difference (2 & 3) ^ ^

Table 2: Simulation performance of microstrip coupler

57
5.2 Design 2

5.2.1 Schematic simulation

Using the equations derived in previous chapter, Zi, Z2 and Zs can be obtained. The

calculated values, which are about 40 ohm, 58ohm and 72 ohm respectively, are

entered into the schematic diagram and it is shown in Figure 5-19. Four ports are

terminated by match load and s-parameter response can be simulated accordingly.

• ^ I怨 I S-PARAMETERSI
丄 TLIN t TLIN I晰 I •
TL13 S Param
Z=Z4 Ohm Z=Z4 Ohm SP1
•J E=M y start=0.5 GHz
T F=1.35 GHz T F=1.35 GHz Stop=2.2 GHz
卜 Step=0,002 GHz
TUN
fhll'Z jUl：^ 二。hm i--
rp;i。hm -SO.
E=I1 L J
f —. J
- Y F = 1 . 3 5 GHZ T F = 1 - 3 5 GHZ J _

= 一
rzivAR
VAR1
J 1 j 21=57.7532
“ ‘ 11=90
丄 TUN 二 ± rilN i _ retm
i ‘ I emi plTLIO ；L】】。“ TL9 < Temi4 pBI VAR
RA Terms Z:Z4 0HM Z=Z4 Ohm | Num=4 ^
> Num=3 E=14 -u T Z=50 Ohm 73=40 8224
f Z=50 Ohm VF=135 GHZ F=1 .35 GHz f .35 GHz 」 gjJYfs二
- 13=90
"i 一 14=90

Figure 5-19: Schematic diagram for design 2 simulation

In Figure 5-20, the frequency response of return loss shows that there are the highest

return losses at two operating frequency bands with at least 50dB, which is

significantly higher than that of design 1. The lowest return loss is obtained at mid-

frequency band shows that there's greatest reflection for that frequency band signal.

For isolation loss performance, which is unlike that of design 1, losses as high as

50dB are obtained at two operating frequency bands within the boundary of mid-

frequency band. This phenomenon can be seen in details in Figure 5-21.

58
 0 产 “““<0^

- f v ^ v l
m -20 ii \i
1—

S -30

-40
-50 1 1 U I 1 1 i——M 1
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Freq. (GHz)
Figure 5-20: Simulation - Return Loss

」："^A A r
！-2� T \ /
1-30 \ /

-40 \

-50 1 —1""“UI 1 ^ i ^ I
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Freq. (GHz)
Figure 5-21: Simulation - Isolation Loss

Apart from return loss and isolation loss, this design shows excellent performance in

insertion loss. In Figure 5-22，both output ports shows 3dB insertion losses at both

operating frequencies with a strong mid-frequency band suppression of at least 50dB.

Even from frequency of 1.2GHz to 1.5GHz, there's at least 20dB signal suppression.

59
Such wide band suppression provides a good filtering characteristic and it would be

very suitable for many applications. It is consistent with the case of frequency

response of return loss and isolation loss that mid-frequency power will be totally

reflected at input port and no power will be delivered to other ports.

5: -40
\ \
—。
+S31

-50 1 1 1“ 1 1 1
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Freq. (GHz)

Figure 5-22: Simulation - Insertion Loss

Close views of insertion loss frequency response at lower and upper frequency bands,

in Figure 5-23 and Figure 5-24 respectively, are provided so that the frequency

responses at operating frequency bands can be shown in details. The frequency

response at each operating frequency band demonstrates similar spectrum as that of

conventional design at operating frequency band that 3dB insertion losses are

simulated at both ports at both operating frequencies. Since the amplitude imbalance

around two frequency bands is not significant, both bands show a relatively wider

bandwidth than that of design 1. To be more specific, bandwidth of operating band of

this design is about lOOMHz if 0.5dB amplitude imbalance is allowed.

60
 - 3 。 ^ ^ ^ ^ ^ _

..。/ \
^-4.5 / +S21 \

-5.0 ‘ r- 1 1
0.80 0.85 0.90 0.95 1.00
Freq. (GHz)
Figure 5-23: Simulation - Insertion Loss (Lower band)

!： /
-5.0 I" 1 1 ‘
\
-«-S31

1.70 1.75 1.80 1.85 1.90

Freq. (GHz)

Figure 5-24: Simulation: Insertion Loss (Upper band)

For the phase difference at both operating frequency bands, steady phase difference of

90° over 2 0 0 M H z is obtained and this is sufficient for the operation of this design.

61
 200 1

2 100 \ X ^ 、

I 50 \

h。-、、\J V
-200 -I 1 1 1 1 1 1 1 1
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Freq. (GHz)
Figure 5-25: Simulation - Phase Difference

In general, this design shows preliminary high return loss, high isolation loss and

similar performance as conventional design in terms of insertion loss and phase

difference, at both operating frequency bands.

5.2.2 Schematic simulation with line width deviation

Similar to design 1，the constructed schematic diagram shown in Figure 5-26 is for

above simulation, with computed dimensions of transmission lines. Again,

simulations are performed with width deviation of±0.05mm and ±0.1 m m .

62
 N = M S-PARAMETERS I
•丨 W=W3mm , W=W3 mm S Param
L=L3 mm L=L3 mm 『 1
Start=0.6 GHz
MLIN Stop=2.2 GHz
TL43 Step=0.05 GHz
Subst="MSub1"
W=W1 mm
L=L1 mm ,…
J U I El 糧
r' ^ ^ OptimizedJ
W1=2.546+d*0.05
W2=1.47+cr0.05
. , , W3=0.987+cr0.05
⑴ • ,MLIN MLIN 、 ferm L1=33 5356
< I 棚 6 [7141 TL42 n > Te触 L2=34.301
< , > u b s t = " M S u b 1 " Subst="MS ibl' > Num=2 L3=34.812
一」Z=50 Ohm M;v=W2 mm W=W2 m m M Z=50 Ohm ^=0
：一 |L=L2 mm L=L2 mm Lmax=-30
= 士 f1=0.9G
f2=1.8G

mIJI-J I “
TL40 MSub
Subst="MSub1" I
W=W1 mm i ‘ Term MSUB
f ^ L=L1 mm m Term7 MSub1

T Z=50 0h. 二38

- Concl=5.8e7
i i Minn ^MLOC = Hu=3.9e+34 mil
一 TL45 TL46 T=10 um
Subst="MSub1" Subst="MSub1" TanD=0.0021
W=W3 mm ~~ W=W3 mm Rough=35 um
L=L3 mm L=L3 mm

Figure 5-26: ADS Schematic diagram for line width deviation simulation

Different from design 1, RT/Duroid 4003C would be used for circuit fabrication. The

information is shown in Table 3.

Parameters Duroid 431OC

Dielectric constant 3.38

Thickness (mm) 0.813

Loss tangent 0.0021

Table 3: Major parameters of substrate RT/Duroid 4003C

Figure 5-27 shows that the troughs of spectrum shift slightly and therefore the change

of return loss is resulted. Close views can be seen in Figure 5-28 and Figure 5-29. For

lower frequency band, it is clear that the highest return loss can be obtained at both

63
operating frequencies when d=0. For the worse case when d=±2, degrade of return

loss can be as large as lOdB.

_1 0 一

I - . . 等 「 f 萨 -
-30—-———丨-TO——i 1 w
S ^ - I I , I I
^ ^ -- - i

- ： 丨 i I
一5Q . -. - - - -- — - - - I —
“ I i
- 6 0 — I I I I j I I I ~ " " " " I " " " " i I I

0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2

freq, GHz
Figure 5-27: Simulation - Return Loss

. 4 … - _ + - f l
^ -40 r- i - - y^j- y^ 1 " 1 _ -

£ 1 Z y ^ V ！

CO I Z
石-45—— d=0 .z i—- I
"J R
一• J —-— -"I —

1 ； ! I I

i ！ I I
-55 1 I I I I I I I I I I I I I I I I I I [ I I I I I I I I I

0.75 0.80 0.85 0.90 0.95 1.00 1.05

freq, GHz

Figure 5-28: Simulation - Return Loss (Lower band)

64
 -25 I I I J\J腸 I I ‘ I I
-30— - f 丨 - 斗 I f ’ ！ --! 「

P 35
^ — •
丨
-- -
\
.--「_
口誦
r VT —/•/——…1 "t
\ i i ‘
- I
忌 丨 i 韻 “ 厂 L
^jr^ -40— -- i - ! rI ！

CO j 1 丨 i 丨

•D ！ j I 1
； ！ I I i
_45— 1 r i I ！ 1 i
： 丨 1 ^ 丨 丨 丨 丨
-50 I I I I I I I I I I [ I [ I I I I I I I I I ‘
1 . 7 0 1 . 7 2 1 . 7 4 1 . 7 6 1 . 7 8 1.80 1.82 1 . 8 4 1 . 8 6 1 . 8 8 1.90 1 . 9 2

freq, GHz

Figure 5-29: Simulation - Return Loss (Upper band)

For upper frequency band, it seems that the case of d=2 provides the lowest return

loss where that of d=-2 provides the highest return loss among other cases. Although

return loss is degraded, there is still at least 25dB return loss obtained in general

Trend of isolation loss spectrum is similar to that of return loss except in mid-

0 ： ！ i 1 1

I -4。:. I t ： - -- ； ¥ … I — I j . 卜 -

-60 -• ……r -- -|j i • ： …丨--…..丨-…-

- ； ： 1 : 1 ： 丨 I
- 8 0 - ] " " “ I I ~ I I I I I I I I I " " " " I I I I
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2

freq, GHz

Figure 5-30: Simulation - Isolation Loss

65
frequency suppression and it is shown in Figure 5-30. A significant wide suppression

is achieved and therefore line width deviation wouldn't affect the performance much

in this aspect.

-20— - ^ ^ -//- i
- 1 丨
V — - .. - ^wuUk—l-m- 1 • •— - —t"

活 e

、_ 乂、uJ
舰
o -

•汽 IJ
-
. -
:

-.
I
W
-V
I f
awnHM-
丨
I丨
—--
i
——

DQ CQ O , ‘
"O "D -80 i nff -p- - "i
- I W I ！
_1〇〇 — n ：- —1 1— j—
F
- 1 2 0 - ] ~ ~ , I , j , ~ j ~ I ~ I I I j 1 !
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
freq, GHz
Figure 5-31: Simulation - Insertion Loss

Figure 5-31 to Figure 5-34 shows frequency spectra of insertion loss with different

value of d. A s an overview, Figure 5-31 demonstrates a frequency band shift with

different values of d when there's still significant wide mid-frequency band

suppression.

-2.8 i ‘ 1 1 ；

i I 1
- ； 丨 . 丨 ，
-3.0- ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^

0.78 0.82 0.86 0.90 0.94 0.98 1.02

freq, GHz

Figure 5-32: Simulation - Insertion Loss (Lower band)

66
Figure 5-32 shows the spectrum at lower frequency band with a close view. The case

of d=0 is the only case that zero amplitude imbalance is obtained at lower operating

frequency with the widest operating bandwidth. The same phenomenon happens at

upper frequency showing in Figure 5-33. For mid-frequency band, the impact is

insignificant again due to its wide suppression band in nature.

-3.0 ； ： ； ！ i
- : I i : ;

-3.2— df^2 A^^^'Sn^

-3 4一 d-0

藝：；：^^ 1.65 1.70 1.75 1.80

freq, GHz
1.85 1.90 1.95 2.00

Figure 5-33: Simulation - Insertion Loss (Upper band)

丨 l ^ l f y i 1

I I : : : _ : ^ y [ : 「
-110- ！ i I 」
- i : 丨 丨 丨 ：
-120
1.26
I I i
1.28
I i
1.30
I I
1.32
I i
1.34
I I
1.36
• I • I I
1.38 1.40 1.42

freq, GHz

Figure 5-34: Simulation - Insertion Loss (Mid-band)

67
Phase difference between both output powers is plotted accordingly. N o significant

change of phase difference for the case when d=-2 to d=2. It can conclude that the

impact of the line width deviation on phase response is insignificant.

二 200 ： 1 1 i 1 i
丨 I I

^^ ‘ ‘ N
I ： I J I

100 -- - -1- - - - •- — — • — — —. -ILL——；——-—J - — — — - — — — — • • ' — “ —

I^ Q _
: … I _ … M —
‘ I I
• - , . ！

^ ; 丨
"X - 1 0 0 - • -- -———.—— — — J — — —-- -J— — —

csT II iI
CO I

I -2。。：…！ - 一 「 T ^
Q- -300 1 1 1 J 1 1 1 1 1 1 1 1 J I
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2

freq, GHz
Figure 5-35: Simulation - Phase Difference

5.2.3 Schematic simulation with junction discontinuity

Similar to design 1, junction discontinuities and substrate information would be

included in simulations to foresee the measurement results which are affected by

junction discontinuity effect and substrate loss.

68
 " | 二 。 n = I 躲 丨 S-PARAMETERSj
• j i Subst="MSub1" L J Subst="MSubr'
MCROS , VV=W3mm , mm
rrr^cA L=L3 mm L=L3 mm 們
Cros4 Slart=0.6 GHz
Subst="MSub1" MLIN MLIN Stop=2 2 GHz
VV1 mm TL43 TL38 Step=0.002 GHz
W2=W3 mm Subst="MSub1" Subst="MSub1"
W3=W1 mm W=W1 mm \N=m mm
W4=W2 mm 1^-1 L=L1 mm r'n L=LO mm

— I V» H ^ ^ ^ 1 I ^ ^ Optimized_1
MLIN LJ L-J \A/D=1.9
TL36 MCROS W1=2.6
土， rg^.,, Subst="MSub1" Cros3 W2=1.5
T e r m ^ V W mm • , MLIN MLIN . Subst="MSub1" ± _ Term
< ^„^_L=LOmm |TL41 TL42 P] W1=W1 mm ^ Term8 [0=5
< J^nohm ；ubst="MSubr' Subst="MS Jb " W2=W3 mm S Num 二 2 L1=32 6874 opt{ 30 to 37}
yv=W2 mm W=W2 mm| | W3=W0 mm 1 Z=50 Ohm l2=31.1811 opt{31 to 38}
"{ L=L2 mm L=L2 mm W4=W2 mm - L3=37.4899 opt{31 to 38}
~ — Lmax=-30
f1=0.9G
r ^ f2=i.8G
" " " " 1 — — - 1
TL37 T V-tl ..M<= Kv. T TL39 _
Subst="MSubf-''C !OS MC^OS Subst="MSub1" —
i_|Temi 肿WD mm Cro 1 mm Cro 2 w=\M) mm 产 丨 em ^SUB
^ Terms L=LO mm Sut st="MSulif^1 mm Sul st="MSuli.t'Lo mm < Temi7 MSubl
^ Num=4 W1 WDmm W1 :W1 mm ^ Num=3 H=0.813 mm
] Z = 5 0 Ohm W2 W2 mm W2 :W2 mm Z=50 Ohm Er=3.38
W3 W1 mm W3 :W0 mm b 」 Mur=1
W4 W3 mm W4 :W3 mm Cond=5.8e7
^MLOC j i - MLOC ^ Hu=3.9e+34 mil
_ TL45 TL46 T=10 um
Subst="MSub1" Subst="MSub1" TanD=0.0021
W=W3 mm W=W3 mm Rough=35 um
L=L3 mm L=L3 mm

Figure 5-36: ADS Schematic diagram for simulation with junction discontinuity

The operating frequencies shift to 9 0 0 M H z and 1.79 G H z respectively. According to

Figure 5-37, the return loss is at least 20dB at both operating frequency bands. This is

the same case as that of isolation loss shown in Figure 5-38. At the mid-frequency

band, nearly all power is reflected and it shows that no power was delivered to

isolation port.

69
 iiKY -30 ！
0.6 0.8 1
‘ .
1.2
1-
1.4
Freq. (GHz)
1.6
-；-
1.8

Figure 5-37: Simulation - Return Loss

：---
2
- 丨--1
2.2

1 -30 \ j
:I \l —
0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Freq. (GHz)

Figure 5-38: Simulation - Isolation Loss

In Figure 5-39，a high mid-band suppression is shown of power at both output port

and coupled port for at least 50dB, with a frequency shift of lOOMHz. The

performance of insertion bandwidth and amplitude imbalance can be found in Figure

5-40 and Figure 5-41.

70
 1-30 I \
l II +S21

—-40 / -Hi-S31

•50 - 丨 ——「― - - -T —II—I— - — I — --[-— 1 ]

0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Freq. (GHz)
Figure 5-39: Simulation - Insertion Loss

In Figure 5-40, the amplitude imbalance of lower frequency band is less than 0.3dB,

while insertion losses of both ports are as low as 3.4dB. The bandwidth is near

-3
+S21

I：/ ... J 1 — _ — ^
\\\ I

0.8 0.85 0.9 0.95 1

Freq. (GHz)

Figure 5-40: Simulation - Insertion Loss (Lower band)

71
lOOMHz assuming 0.5dB amplitude imbalance can be tolerated. Similar case is

happened for upper band, shown in Figure 5-41.

- 3 ] - ^ S 2 1

J • + S 3 1

！ 35

i. -4 [ / N y

^ I \

1.7 1.75 1.8 1.85 1.9

Freq. (GHz)
Figure 5-41: Simulation: Insertion Loss (Upper band)

Similar to design 1，the deviation of phase difference is nearly the same as ideal

results. It shows phase difference of nearly 90° over a wide frequency band.

200 丨 +phase(S21) ^ ^

I 二 1 \ 1 ^ 一 ) p v f

叫NKir
、-150
•200 -
> i N J
1 - - - … - 丨 … … I — — . … r
\J
厂 — I
\
「…..”‘1

0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2

Freq. (GHz)
Figure 5-42: Simulation - Phase Difference

72
Table 4 provides a summary of simulation results.

Frequency 900 M H z 1.79 G H z

Input Return Loss 20.58 dB 23.62 dB

Insertion Loss (1 to 2) 3.42 dB 3.39 dB

Insertion Loss (1 to 3) 3.28 dB 3.28 dB

Isolation factor (1 to 4) 21.32 dB 23.64 dB

Phase difference (2 & 3) 90.00° 90.02。^

Table 4: Simulation performance of microstrip coupler

73
6 Circuit Implementation and Characterization

After schematic simulations, circuits of design 1 and design 2 can be fabricated

accordingly. The implementation details would be explored in this chapter, and

measurement results would be presented as well.

6.1 Design 1

6.1.1 Circuit fabrication

For verification, the configuration B (m=2 and n=1) of design 1 operating at 900MHz

and 1.8GHz is constructed by substrate RT/Duroid 6002. Four S M A connectors are

connected for measurement.

• (a)
Figure 6-1: Circuit of configuration B (m=2, n=l), design 1

74
6.1.2 Measurement results

�::Kn
-25
0.6
1
0.8 1
1 -1
1.2
1
1.4
Freq. (GHz)
‘
1.6
Figure 6-2: Measurement: Return Loss
1
1.8
‘
2 2.2

tfj]
0 —

-30 ‘ 1 —1 1 ‘ r 1 ‘
0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Freq. (GHz)
Figure 6-3: Measurement: Isolation Factor

75
The measurement results show that the operating frequencies of the circuit shift to

9 0 6 M H z and 1.82GHz respectively. Showing in Figure 6-2 and Figure 6-3，there's at

least 15dB return loss while there's over 17dB isolation obtained at both operating

frequency bands. It offers satisfactory performance in many applications.

0厂

\
o3 -8

-10
V \ f \ f

-12 I 1 1 1 ！ 1 1 1

0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2

Freq. (GHz)
Figure 6-4: Measurement: Insertion Loss

200 I ^
-^150 -E-S21 ‘

I 100 + S 3 1
呈 50
§ 0

- 2 U U H 1 1 1 1 1 I I 1

0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2

Freq. (GHz)
Figure 6-5: Measurement: Phase Difference

76
Insertion loss performance shown in Figure 6-4 is well agreed with simulation results.

Nearly no amplitude imbalance is measured between two out power with the loss at

the maximum of 3.5dB. There is mid-band attenuation of 6dB and it is also similar to

simulation results.

The phase differences between two output power are 90.2° and 92.4° at lower and

upper frequency bands respectively. Similar to simulation results, it shows phase

difference equal to 90° over a wide frequency range.

A summary of measurement is presented in Table 4. The measurement results show a

low insertion loss and phase error, with return loss higher than 15dB and isolation loss

higher than 17dB.

Frequency 906 M H z 1.82 G H z

Input Return Loss -15.6 dB -20.4 dB

Insertion Loss (1 to 2) -3.40 dB -3.45 dB

Insertion Loss (1 to 3) -3.43 dB -3.42 dB

Isolation factor (1 to 4) -17.0 dB -20.2 dB

Phase difference (2 & 3) 90^

Table 5: Measurement performance of microstrip coupler, design 1

77
6.2 Design 2

6.2.1 Circuit fabrication

Design 2 circuitry is constructed accordingly using substrate Duroid RT/4003C. Four

S M A connectors are connected for measurement. Four open stubs are fabricated

straight for illustration purpose and a compact version of the design is fabricated by

bending all stubs into the centre area of the circuit, after further optimizations.

(b)

Figure 6-6: Circuit of design 2

78
6.2.2 Measurement results

I： Y Y
n 1

-30 1 1 I -N 1 1 • 1

0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
Freq. (GHz)
Figure 6-7: Measurement: Return Loss

-50 - y

- 6 0 -I— 1 1 1 1 1 1 1— 1
0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2

Freq. (GHz)

Figure 6-8: Measurement: Isolation Loss

The measurement results show that operating frequencies shift to 900MHz and

1.77GHz respectively. From Figure 6-7 and Figure 6-8’ both return loss and isolation

loss are as high as 25dB. For the measured isolation loss, although the centre

frequency of mid-band suppression is lowered to about 1.25GHz, a wide band

79
isolation loss of at least 20dB is shown over 3 0 0 M H z from l.lGHz to 1.4GHz is still

obtained.

! 3 。 I I

1-40 +S21

-50 J +S31

- 6 0 ！ 1 1 亡 1 1 ‘ 1

0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
Freq. (GHz)

Figure 6-9: Measurement: Insertion Loss

Wide band mid-band suppression is also shown too in insertion loss measurement.

Figure 6-9 shows that the insertion losses at both frequency bands are at a maximum

of 3.5dB. Close views can be found in Figure 6-10 and Figure 6-11，where the

amplitude imbalance is less than 0.1 dB. The bandwidth of each operating frequency

band is about lOOMHz, for 0.5 dB tolerance of amplitude imbalance.

80
 -3.0 I

1： f飞 -5.0 U a
0.80
,
0.85
1———
0.90
1
0.95
^
1.00
Freq. (GHz)
Figure 6-10: Measurement: Insertion Loss (Lower band)

-3.00 I —

t
w -4.00 /

/
CO /

—-4.50 “ -H-S21
+ S31
-5.00 -J T 1 ‘
1.70 1.75 1.80 1.85 1.90
Freq. (GHz)
Figure 6-11: Measurement: Insertion Loss (Upper band)

Similar to design 1, a nearly 90° phase difference is measured between two output

power, which is similar to simulation results.

81
 200

|100^phase(S(2.1))l X . 、
B 50 +PHASE(S(3，1)) \ X

运 ° V \
、-50 \ Y V

-200 i 1 1 1 1 1 1 — 1 i

0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
Freq. (GHz)
Figure 6-12: Measurement: Phase Difference

A summary of design 2 measurement results is listed in Table 6. Input return loss and

isolation loss are at least 25dB while the insertion losses both output power are at

most 3.4dB. A small amplitude balance is measured at centre operating frequencies.

Nearly 90° phase difference is measured too as well.

Frequency 900 M H z 1.77 G H z

Input Return Loss -25.63 d B - 2 6 . 4 2 dB

Insertion Loss (1 to 2) -3.24 dB -3.33 dB

Insertion Loss (1 to 3) -3.27 d B - 3 . 3 0 dB

Isolation factor (1 to 4) -25.76 dB -26.36 dB

Phase difference (2 & 3) 88.31。 87.94°

Table 6: Measurement performance of microstrip coupler, design 2

82
7 Conclusions

In this thesis, two designs of dual band branch line coupler are proposed, constructed

and characterized. In the first design, dual band operation is achieved by employing

two pairs of transmission line which the electrical lengths are integral multiple of 90°.

Design equations are derived and sub-optimal solutions can be obtained by relaxing

requirements of return loss and isolation loss. Simulation results show that -3dB

output power can be delivered to both output port and coupled port at the operating

frequencies, with the same phase shift between two signals as that of the conventional

design. Under relaxed conditions, there are still at least return loss and isolation loss

of20dB.

In the second design, a proposed transformer is used to replace quarter wavelength

transmission line so that the topology is similar to that of conventional design with

four open stubs introduced. It provides high return loss and isolation loss, 90° phase

shift and 3dB insertion loss at operating frequencies.

Both of the designs can operate over a wide range of fractional bandwidth with easy

realization of transmission line by major manufacturing technologies.

The major differences between two designs are the insertion loss performance and

size of the design. Regarding to insertion loss performance, both simulation and

measurement results show that the second design provides optimum solution with a

relatively larger bandwidth than that of the first design. It can be illustrated when

comparing the frequency response of insertion loss of both designs. Moreover, mid-

frequency attenuation shows in the second design provides a good filtering purpose in

some applications whereas that of first design shows only 6dB signal attenuation. Size

83
of the second design is more compact that it shows an unconditional size reduction in

comparison of the conventional design. While in the case of the first design, only

conditional size reduction can be achieved. When comparing with the dual band

coupler design using lumped-distributed structure [16], design 2 is featured compact

size, planar structure and lower losses because of the absences of parasitic effects

from lumped elements.

In general, both designs are suitable for dual band applications with moderate

bandwidth. Design 2 shows lower losses and smaller size when design 1 shows a

simpler structure.

84
8 Recommendations for Future Work

Regarding the presented results, research work can be extended in different ways.

Firstly, it's about the bandwidth enhancement. In design 2’ the bandwidth of each

frequency band is similar to that of conventional design, and the frequency response

of insertion loss shows that that pass band may not be flat and wide enough for some

applications. Further investigation should be focused on the feasibility of further

bandwidth improvement so that the usefulness of the design can be increased.

Moreover, both designs are implemented in planar structure in this thesis and they are

proved to be able to realize in most planar manufacturing technologies. However, the

feasibility of lumped element implementation is still a question. Further investigations

can be devoted on lumped element implementation so that the design can be applied

by popular integrated circuit technologies such as C M O S and LTCC, with smaller

dimension than that of the proposed designs and integrated with active devices.

Apart from the above recommendations on coupler design, the performance of dual

band building blocks such as balanced amplifier using dual band coupler can be

analyzed, such as bandwidth, circuit size, etc. Although additional circuit techniques

may be required as well for dual band operation, it is believed that the building block

using dual band coupler has superior performance than that of conventional design

when using smaller (or fewer) active devices.

85
9 References

[1] Badr, K.H.; El-Motaafy, H.A.; El-Badawy, E,S.A.; Bayoumy, M.E.-S.A.,

“Design and implementation of M W balanced amplifier using branch-line

couplers，，，7996 Radio Science Conference, pp. 385 - 394’ March 1996.

[2] Man-Long Her; Phun-Fei Teo; Fan-Hsi Kung; Chi-Yuan Xie; Yu-Zhen Wang;

Yi-Chyun Chiou; Yu-Lin Wang; Chun-Wei Kuo, "Design and analysis for three

applications of 90-degree single balanced mixer", 2002 The S"丨 International

Conference on Communication Systems, vol. 1’ pp. 25 - 28’ November 2002.

[3] Jokanovic, B , Stojanovic, S.; Peric, M.，"Direct Q P S K modulator for point-to-

point radio link at 23 GHz，，，2001 The International Conference on

Telecommunications in Modem Satellite, vol. 1，pp. 217 - 220, September 2001.

[4] P382, Microwave Engineering

[5] Ho, C,H.; Fan, L.; Chang, K.，“A broad-band uniplanar branch-line coupler

using a coupled rectangular slotline ring，，, IEEE Microwave and Guided Wave

Letters, vol. 3，no. 6’ pp. 175 - 176’ June 1993.

[6] Chieng-Tang Lin; Chun-Lin Liao; Chun Hsiung Chen, "Finite-ground coplanar-

waveguide branch-line couplers，，，IEEE Microwave and Wireless Components

Letters, vol. 11’ no. 3’ pp. 127 - 129’ March 2001.

[7] Fisher, R.E, "Broad-Band Twisted-Wire Quadrature Hybrids (Short Papers)，，，

IEEE Trans. Microwave Theory and Techniques, vol. 21, no. 5, pp. 355 - 357，May

1973.

[8] Andrews, D.P.; Aitchison, C.S., "Wide-band lumped-element quadrature 3-dB

couplers in microstrip", IEEE Trans. Microwave Theory and Techniques, vol. 48’

86
 no. 12, pp. 2424 -2431, December 2000.

[9] Chien-Hsun Ho; Lu Fan; Kai Chang, “New uniplanar waveguide hybrid-ring

couplers and m a g i c - T ' s " , T r a n s , Microwave Theory and Techniques, vol. 42，

no. 12, pp. 2440 - 2448, December 1994.

[10] Eccleston, K.W.; Ong, S.H.M., “Compact planar microstripline branch-line and

rat-race c o u p l e r s " , T r a n s . Microwave Theory and Techniques, vol. 51，no. 10，

pp. 2119-2125, October2003.

[11] Sakagami, I.; Haga, M.; Munehiro, T.，“Reduced branch-line coupler using eight

two-step stubs", lEE Proc. Microwaves, Antennas and Propagation, vol. 146，no. 6’

pp. 455 - 460, December 1999.

[12] Frye, R.C.; Kapur, S.; Melville, R.C., “A 2-GHz quadrature hybrid implemented

in C M O S technology", IEEE Journal of Solid-State Circuits, vol. 38’ no. 3，pp. 550

-555，March 2003.

[13] Yi-Chyun Chiang; Chong-Yi Chen, “Design of a wide-band lumped-element 3-

dB quadrature coupler", IEEE Trans. Microwave Theory and Techniques, vol.

49，no. 3，pp. 476 - 479，March 2001.

[14] Wei-Shin Tung; Hsu-Hsiang W u ; Yi-Chyun Chiang, “Design of microwave

wide-band quadrature hybrid using planar transformer coupling method", IEEE

Trans. Microwave Theory and Techniques, vol. 51, no. 7’ pp. 1852 - 1856，July

2003.

[15] www.agilent.com

[16] I-Hsiang Lin; Caloz, C.; Itoh, T.，“A branch-line coupler with two arbitrary

operating frequencies using left-handed transmission lines", 2003 IEEE MTT-S

International Microwave Symposium Digest, vol. 1，pp. 325 - 328, June 2003.

[17] Ross, G.F.，"The Transient Analysis of Certain T E M M o d e Four-Port

87
 Networks", IEEE Trans. Microwave Theory and Techniques, vol. 14’ no. 11’ pp.

528 - 542, November 1966.

[18] Reed, J.; Wheeler, GJ.，“A Method of Analysis of Symmetrical Four-Port

Networks", IEEE Trans. Microwave Theory and Techniques, vol. 4’ no. 4，pp.

246 - 252, October 1956.

[19] Gorur, A , Karpuz, C ; Akpinar，M.’ “A reduced-size dual-mode bandpass filter

with capacitively loaded open-loop arms，，，IEEE Microwave and Wireless

Components Letters, vol. 13，no. 9’ pp. 385 - 387，September 2003.

[20] Jia-Sheng Hong; Lancaster, M J.，"Theory and experiment of novel microstrip

slow-wave open-loop resonator filters", IEEE Trans. Microwave Theory and

Techniques, vol. 45，no. 12，pp. 2358 - 2365, December 1997.

[21] Jia-liang Chen; Sheng-fuh Chang; Shi-wei Kuo, “A low phase-noise oscillator

with a planar capacitively loaded microstrip resonator", 2001 Asia Pacific

Microwave Conference, vol. 3，pp. 1092 - 1095’ December 2001.

[22] Caloz, C ; Itoh, T.，"Application of the transmission line theory of left-handed

(LH) materials to the realization of a microstrip “LH”’’，2002 IEEE Antennas

and Propagation Society International Symposium, vol. 2’ pp. 16 - 21’ June

2002.

[23] Caloz, C ; Itoh, T.，"A novel mixed conventional microstrip and composite

right/left-handed backward-wave directional coupler with broadband and tight

coupling characteristics", IEEE Microwave and Wireless Components Letters,

vol. 14，no. 1, pp. 31 - 33，January 2004.

[24] Okabe, H.; Caloz, C.; Itoh, T.’ “A Compact Enhanced-B and width Hybrid Ring

Using an Artificial Lumped-Element Left-Handed Transmission-Line Section",

IEEE Trans. Microwave Theory and Techniques, vol. 52, no. 3，pp. 798 - 804,

88
 March 2004.

[25] Vogel, R.W., "Analysis and design of lumped- and lumped-distributed-element

directional couplers for M I C and M M I C applications", IEEE Trans. Mc薩ave

Theory and Techniques, vol. 40’ no. 2，pp. 253 - 262, February 1992.

[26] Toker，C; Saglam, M.; Ozme, M.; Gunalp, N.’ “Branch-line couplers using

unequal line lengths，，, IEEE Trans. Microwave Theory and Techniques, vol. 49，

no. 4, pp. 718-721, April 2001.

89
 10 Author's Publications

1. F. L. W O N G , and K. K. M . Cheng, “A Novel Planar Branch-Line Coupler For

Dual-Band Applications", IEEE MTT-S International Microwave Symposium,

Texas, June 2004.

2. K. K. M . Cheng, and F. L. W O N G , “A Novel Approach to the Design and

Implementation of Dual-Band, Compact, Planar, 90。Branch-Line Coupler",

accepted in IEEE Trans. Microwave Theory and Techniques, 2004

90
 .‘
.r-；-

•.

• • • .
• . - ‘ . •

. , , ‘ . . .••

•.. ‘
, ‘ •

‘ • . • •

. “ - .

.•. . , - •
/':,.....‘，•.
‘ . .
. . • • . . . ,
. .
、• ...：....• .. ,,. - 、 ，•，•
. •

• • . • . ^ • • • , “

• - .
. .‘ - .
, • • ..
、 ... ，• ,’ ，
. . . •‘‘ . - .
 •- 一 ：
，
-

j ^^^ D
；Q ii
h

i『
. Is屋
n M L.
ff d
s
y- >； I
々“ vi- • V... .
-rI •- • r
？ . .. \ .
-• s
. •. . - .•
f .、
.:’.』 :...’
. ’ . ” - . . • . . . • .
• • • • . .
, f -
• .. ..
,
'- r; •
. . . . - ' - • •

./-•-.• . . . . • . .
‘ • • ‘ - •.. . . .
- 、 、
••
：八.
-’•
.

... . .
.、•
，•：
.-•
:•. .

•、. .. . .
v..
-、4
.:v • .
-..• • , V: .......‘ 一.. .
. .
.;.
...
. ？ V...
；
. . - . . . , - • , .
• r's ； • •- • • /
.T - 乂 .
..
，： . .- .\ ..
.-, ‘
<-. ： I. • ••••••• - . •‘ •• " •
••.
.
‘， ,• - . . . - J, ,•/••. *•-,『-.——.-,/,•- -M ：
： -