close

Вход

Забыли?

вход по аккаунту

?

Novel passive microwave circuit components and simulation technique for wireless communications

код для вставкиСкачать
INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in typewriter free, while others may be
from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality
illustrations and photographs, print bleedthrough, substandard margins,
and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete
manuscript and there are missing pages, these will be noted. Also, if
unauthorized copyright material had to be removed, a note will indicate
the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and
continuing from left to right in equal sections with small overlaps. Each
original is also photographed in one exposure and is included in reduced
form at the back of the book.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6” x 9” black and white
photographic prints are available for any photographs or illustrations
appearing in this copy for an additional charge. Contact UMI directly to
order.
UMI
A Bell & Howell Information Company
300 North Zed) Road, Ann Arbor MI 48106-1346 USA
313/761-4700 800/521-0600
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
U n iv e r s it y
of
C alifo r n ia
Los Angeles
NOVEL PASSIVE MICROWAVE CIRCU IT
COMPONENTS AND SIMULATION TECHNIQUE FO R
W IRELESS COMMUNICATIONS
A dissertation submitted in partial satisfaction
o f the requirements for the degree
Doctor of Philosophy in Electrical Engineering
by
Kuang-Ping Ma
1999
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
UMI Number: 9926338
UMI Microform 9926338
Copyright 1999, by UMI Company. All rights reserved.
This microform edition is protected against unauthorized
copying under Title 17, United States Code.
UMI
300 North Zeeb Road
Ann Arbor, MI 48103
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The dissertation of Kuang-Ping Ma is approved.
C. -J. Kim
Itoh, Committee Chair
University of California, Los Angeles
1999
ii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
To myfam ily
in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table
of contents
vi
LIST OF FIG URES.................................................................................................
ACKNOWLEDGMENTS......................................................................................
VITA, PUBLICATIONS
x
.....................................................................................
xi
ABSTRACT OF THE DISSERTATION............................................................ xiv
1 INTRODUCTION ................................................................................................
2 NOVEL COPLANAR WAVEGUIDE CIRCUIT COMPONENTS
1
.......... 3
2.1 Introduction ...........................................................................................
3
2 2 FDTD Modeling of Mode Conversion...................................................
5
2.3 Design Concept of the New Transition..................................................
8
2.4 CPW-Slotline Transition Characteristics ..............................................
16
2.5 Applications Examples .........................................................................
22
2.5.1
A CPW-Fed Vivaldi A ntenna....................................................
24
2.5.2
A CPW Power Divider .............................................................
27
2.5.3
A CPW-Slotline Coupler ..........................................................
29
References 2 ................................................................................................. 32
3 NOVEL APPLICATIONS OF PHOTONIC BANDGAP STRUCTURE - 3 5
3.1 Introduction .......................................................................................
35
3.2 Novel Non-Leaky Conductor-Backed Coplanar Waveguide .................
36
3.3 Leakage Suppression in Stripline Circuits.............................................
43
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.4 Novel Magnetic Surface.......................................................................
50
References 3 ................................................................................................. 58
4 GLOBAL TIME-DOMAIN FULL-WAVE ANALYSIS OF
MICRO WAVE CIRCUITS.............................................................................
4.1 Introduction
60
.......................................................................................
60
4.2 The Equivalent Source Algorithm.........................................................
62
4.3 Crosstalk and Packageing Effects in Active Circuits ............................
63
4.4 EMC Characterizations for Active Circuits
........................................
70
4.5 A Mixed Full-Wave/Equivalent-Circuit Algorithm
for FDTD Method ...........................................................................
76
References 4 ................................................................................................. 82
5 CONCLUSION
...............................................................................................
v
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
84
L is t
o f f ig u r e s
2.1
The asymmetrical CPW circuit. .............................................................................. 8
2.2
The comparison of the S-parameters between the spectral domain and
FDTD results for the asymmetrical CPW circuit (a) Without air-bridge.
(b) With air-bridge....................................................................................... 9
2.3
Illustrations of CPW-slotline mode conversion, (a) The lengths difference is
180°. (b) The lengths difference is not 180°...........................................
11
2.4
Orthogonality of CPW mode and slotline mode.......................................
12
2.5
The proposed CPW-Slotline transition....................................................... 13
2.6
The mode conversion in the proposed transition without air-bridge
2.7
The effect of the air-bridge on the CPW mode and slotline mode. ............ 15
2.8
The dimensions and the calculated S-parameters for the CPW-slotline
transitions................................................................................................
2.9
14
19
Comparison between calculated and experimental result for the
back-to-back CPW-slotline transition......................................................
21
2.10 Calculated S-parameters for the CPW-slotline transition without
air-bridge.................................................................................................. 22
2.11 Field distributions in CPW section of the CPW-slotline transition with and
without air-bridges.................................................................................... 23
2.12 The CPW-fed Vivaldi antenna. .............................................................. 25
vi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.13 Measured retam loss oftheCPW-fed Vivaldi antenna. .......................... 25
2.14 The radiation patterns of the CPW-fed Vivaldi antenna at 13 GHz
for the cases with and without substrate cutting........................................
2.15 Measured S-parameters for power divider.
26
........................................... 28
2.16 Comparisons between simulated and measured S-parameters for
Power divider.
...........................................................................................28
2.17 A CPW-slotline coupler utilizing the CPW-slotline transition..................... 31
2.18 Calculated S-parameters of the coupler. ....................................................31
3.1
(a) The 2-D PBG lattice and (b) the insertion loss for a uniform
microstrip line with a PBG ground plane which shows the stopband
phenomenon.............................................................................................. 37
3.2
(a) The structural view of the CB-CPW. (b) FDTD simulated result for
showing the leakage in CB-CPW.............................................................
3.3
39
Conventional methods of suppressing the leakage o f CB-CPW.
(a) Shorting posts and (b) multi-layered structure...................................
40
3.4
The proposed nonleaky PBG-CB-CPW.................................................... 41
3.5
Measured insertion losses of conventional CPW, CB-CPW and
the newly proposed PBG-CB-CPW. A stopband from 9 GHz to 14 GHz
can be observed in this figure..................................................................
43
3.6
Propagation and leaky modes in stripline circuit structure.....................
44
3.7
The structure of a test circuit for characterizing the stopband of the
PBG lattice...............................................................................................
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
47
3.8
Simulated and measured S-parameters for the structure shown in
Fig.3.7....................................................................................................
3.9
47
Two test circuits for leakage suppression characterization, (a) No PBG
and(b) with PBG. ................................................................................. 48
3.10 Simulated and measured results for the two stripline test circuits...........
49
3.11 The characteristics of the PEC and PMC when both are illuminated
by a uniform plane wave, (a) PEC case and (b) PMC case.....................
51
3.12 The equivalent circuits of the PEC and PBC planes................................
52
3.13 The 2-D photonic bandgap structure used for realizing a magnetic
surface....................................................................................................
53
3.14 The experimental setup..........................................................................
55
3.15 The measured and simulated phase difference response. The phase
difference is defined as the difference between the phase of the PBG
surface and that o f the PEC surface.............................................................. 57
3.16 The measured and simulated magnitude response. The plot shows
the ratio of the two magnitudes of the reflection coefficients.................... 57
4.1
Configurations for (a) horizontal and (b) vertical equivalent
current-source implementation.................................................................. 64
4.2
Placement o f the equivalent source for CPW circuit configuration
4.3
Small signal model of NEC 76038 MESFET.............................................. 65
4.4
Libra code for finding the small signal equivalent circuit......................... 67
4.5
Configuration and dimensions of the CPW amplifier. ............................ 68
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
65
4.6 Comparisons between measured and simulated results for the CPW
amplifier. .................................................................................................................. 68
4.7 The calculated gain for different heights of the upper shielding metallic
plate. One FDTD cell equals 0.2116 mm..................................................
4.8 Crosstalk between two adjacent active circuits........................................
69
69
4.9 The spectrum at the output of the CPW amplifier for different lengths of
the slot. The Fslot denotes the resnance of the slot, and the width of the
slot is fixed at wS|0t = 0.2A.......................................................................
72
4.10 The spectrum at the output of the CPW amplifier for different widths
of the slot. The slot length is fixed at Fsiot = 6 GHz, and the one FDTD
cell equals 0.118 mm............................................................................
4.11 The measured power penetrated through the slotted shielding plate
73
75
4.12 The ratio o f the measured penetrated power to the measured power
picked up by the circuit without the shielding plate. This ratio
indicates the reinforcement of the slot on the penetrated power. ..........
76
4.13 The illustration of the original microstrip amplifier and the chopped
amplifier with equivalent loads..............................................................
4.14 A microstrip open end circuit and its equivalent LC distributed circuit
78
78
4.15 Comparison between the case using standard FDTD method and the case
using this algorithm................................................................................... 80
ix
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A ck n ow ledgm ents
I would like to express my sincere gratitude to my advisor Professor Tatsuo Itoh for
his guidance, encouragement, and patience during this research. I would also like to
express appreciation to the members o f my Doctoral Committee, Professors C. -J.
Kim, Behzad Razavi, and Frank M. Chang, for their kind help and valuable time.
I would like to thank Dr. Yongxi Qian, Dr. Dongsoo Koh, Dr. Chien-Nan Kuo,
Dr. Siou Teck Chew, Dr. Min Chen, Mr. Wei Fu, Ms. Dong-Lin Su for their
precious friendship and sincence help to make my personal and academic lives at
UCLA a warm and pleasant one. Thanks to all the members in Professor Itoh's
group, Ms. Jeannie Alt, Dr. Roberto Coccioli, Fei-Ran Yang, Mikko Sironen,
Yoichiro Miyamoto, Sylvia Lin, Juno Kim, Noriaki Kaneda, Cynthia Hang,
Jonathan Fredrick, Bill Deal, Sung-Hsien Chang, Chin-Chang Chang.
Finally, I will present my sincere gratitude to my advisor at University of
Missouri, Rolla, Dr. Jim Drewniak for his fairness and patient guidance during my
study in United State.
x
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
V it a
1967
Born, Miaoli, Taiwan
1985 - 1989
B.S., Aeronautical and Aerospace Engineering
National Chen Kung University, Tainan, Taiwan
1989 - 1991
M.S., Mechanical Engineering
National Taiwan University, Taipei, Taiwan
1991 - 1993
Army, Taiwan
1993 - 1995
M.S. Electrical Engineering
University of Missouri, Rolla
1995 - present
Graduate Student Researcher
Electrical Engineering Deptment
University of California, Los Angeles
P u b l ic a t io n s
M. Li, K. -P. Ma, D. M. Hockanson, J. L. Drewniak, T. H. Hubing, and T. P. van
Doren, "Numerical and experimental corroboration of an FDTD thin-slot model for
slots near comers of shielding enclosures,” IEEE Transactions on Electromagnetic
Compatibility, vol.39, pp.225-232, Aug. 1997.
xi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
K. -P. Ma, and J. L. Drewniak, "A comparison o f FDTD algorithms for sub-cellular
modeling of slots in shielding enclosures,” 1995 IEEE International Symposium on
Electromagnetic Compatibility, pp. 157-162, Atlanta, GA, Aug. 1995.
K. -P. Ma, J. L. Drewniak, T. H. Hubing, and T. P. Van Doren, "A comparison of
an FDTD thin-slot algorithm and method of moments for modeling slots near
comers," 1995 IEEE International Symposium on Electromagnetic Compatibility,
pp. 386-390, Santa Clara, CA, Aug. 1996.
K. -P. Ma, M. Li, J. L. Drewniak, T. H. Hubing, and T. P. van Doren, "Comparison
of FDTD algorithms for subcellular modeling of slots in shielding enclosures," IEEE
Transactions on Electromagnetic Compatibility, vol.39, pp.147-155, May 1997.
K. -P. Ma, and Tatsuo Itoh, "A new broadband coplanar waveguide to slotline
transition," in 1997 IEEE International Microwave Symposium Digest, pp. 16271630, Denver, CO, June. 1997.
K. -P. Ma, K. Hirose, F. -R. Yang, Yongxi Qian, and Tatsuo Itoh, "Realisation of
magnetic conducting surface using novel photonic bandgap structure," Electronics
Letters, pp. 2041-2042, vol. 34, IEE, 15 Oct. 1998.
K. -P. Ma, F. -R. Yang, Yongxi Qian, and Tatsuo Itoh, "Nonleaky conductorbacked CPW using a novel 2-D PBG lattice," 1998 Asian-Pacific Microwave
Conference, pp. 1627-1630, Yokohama, Japan, Dec. 1998.
K. -P. Ma, J. Kim, F. -R. Yang, Yongxi Qian, and Tatsuo Itoh, "Leaky suppression
in stripline circuits using a 2-D Photonic Bandgap lattice," to be presented in 1999
IEEE International Microwave Symposium, Anaheim, CA, June. 1999.
xii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
K. -P. Ma, Yongxi Qian, and Tatsuo Itoh, "Analysis and applications of a novel
CPW-slotline transition," accept for publication on IEEE Transactions on
Microwave Theory and Technique.
K. -P. Ma, M. Chen, Bijan Houshmand, Yongxi Qian, and Tatsuo Itoh, "Global
Time-Domain Full-Wave Simulation Involving Highly Nonlinear Phenomena and
EMC Effects,” accept for publication on IEEE Transactions on Microwave Theory
and Technique
xiii
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
A b s t r a c t o f t h e D is s e r t a t io n
NOVEL PASSIVE M ICROW AVE CIRCUIT
COMPONENTS AND SIMULATION TECHNIQUE FOR
W IRELESS COMMUNICATIONS
by
Kuang-Ping Ma
Doctor of Philosophy in Electrical Engineering
University of California, Los Angeles, 1999
Professor Tatsuo Itoh, Chair
In this dissertation, several novel microwave circuits are proposed and demonstrated
to conquer some of the problems occurred in conventional microwave circuits. First,
a new CPW-slotline transition utilizing CPW-slotline mode-conversion phenomenon
is demonstrated and the concept of the new transition is verified both numerically
and experimentally. Both results reveal broadband performance of the new transition
as a result of the CPW-slotline mode conversion and the elimination of slotline
mode by air-bridge, as well as the optimization of the transition geometry. Potential
applications of this novel transition have been demonstrated by designing three types
xiv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
of passive structures, i.e., a CPW-fed Vivaldi antenna, a CPW power divider and a
CPW-slotline coupler.
Several novel microwave circuits utilizing a novel 2D PBG lattice are
demonstrated in this study including leakage suppression in CB-CPW and stripeline,
as well as the realization o f a PMC ground plane. The proposed PBG lattice can be
easily realized by standard planar process, readily to lower fabrication cost than
conventional shorting posts. The PBG pattern is also employed to realize a magnetic
surface which has significant influence in the constructions o f some wireless
communication antennas such as loop antennas.
The fmite-difference trme-domain (FDTD) method is applied to predict the
parasitic effects such as housing and crosstalk effects when the circuit is densely
integrated. A new technique for alleviating the computational burden of FDTD
method is also proposed and demonstrated in this study.
xv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 1
INTRODUCTION
With the high demand of infonnation exchange, the performance requirements placed
on the communication equipment become more stringent. The requirements of
providing both the environment for multi-users and larger bandwidths for the high
speed transmission of multi-media data streams make the system design of the
communication devices nontrivial. Traditionally, a huge amount of communications
have been done using wireline types of systems with the transmission media such as
twist pair, optical fiber or coaxial cable. Due to the high installation expense and nonmobile characteristics of the wired-type communications, information exchanging
wirelessly becomes an alternative method to establish the communication link
effectively.
Even though wireless communication has been becoming more promising and
gaining more attentions, there are some difficulties which must be solved or at least
alleviated. The difficulties stem from the fact that wireless channel resource is very
precious in terms of accommodation of more users and bandwidth occupation. In
order to accommodate more users while meeting the bandwidth requirement,
1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
alternative approaches must be thoroughly considered. Several example o f such
alternatives is by using an effective multiple access scheme, such as Code-Division
Multiple Access (CDMA), to accommodate more users. One of the proposal for the
third-generation cellular standard which adopts a wide-band CDMA and operates at
900/1800 MHz is one such example. The other alternative is to move to higher
operating frequencies in order to have larger bandwidths at higher frequencies, such
as the multi-media standard proposed in Japan. Moving toward higher operating
frequencies (combined with an effective multiple access scheme) is definitely the
trend in order to provide high quality services for the customers.
Traditionally, most of the microwave circuits were built in microstrip line
configurations. As the operating frequencies become higher, the configuration o f high
frequency circuits should be considered carefully. Moreover, some of the difficulties
encountered in traditional designs can be relieved using other approaches which could
be either a different circuit configuration or a novel concept of realizing the circuit. It
is the purpose of the first part of this study to seek alternative ways to realize high
frequency circuit components and explore the potential applications for these novel
circuit components. And in the second part of this study, the extended FiniteDifference Time-Domain (FDTD) method is applied to predict the parasitic effects
such as crosstalk, housing effects which are very critical in a hostile wireless
environment.
2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 2
NOVEL COPLANAR WAVEGUIDE C IR C U IT
COM PONENTS
2.1 Introduction
Coplanar waveguide (CPW) has been an important building block for MMICs due to
several attractive features in comparison with microstrip lines. No via hole is needed
for mounting the circuit components, and a low wave propagation dispersion can be
obtained up to very high frequency operations are two such features.
There have been intensive studies on CPW for various configurations including
symmetric and asymmetric CPW, CPW with air-bridges, multi-layered CPW,
conductor-backed CPW (CB-CPW) and finite ground plane CPW (FGCPW) [2.12.4]. One phenomenon which always occurs in asymmetric CPWs is the CPWslotline mode conversion due to different distances the waves propagate through the
two slots. In practical CPW circuits the two slots cannot always be kept to the same
lengths, and a mode conversion from CPW mode to slotline mode often occurs.
Previously, efforts have been placed on the methods of eliminating this parasitic
slotline mode [2.5, 2.6]. The most commonly employed method is using air-bridges
3
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
to keep two ground planes at the same potential, thereby eliminating slotline mode
because the air-bridge exhibits a short circuit for slotline mode [2.2]. The
effectiveness of the suppression of slotline mode by air-bridges has been shown by
various hill-wave analyses in both frequency and time domains [2.1-2.2, 2.7-2.8],
and design data are also given by Lee [2.2].
Despite the fact that slotline mode is undesirable for most of the waveguiding
purposes, the mode conversion effect can be utilized to design a CPW to slotline
transition. Several CPW-slotline transitions have been proposed and analyzed
previously [2.9,2.10]. As proposed in [2.9], the transition is designed by terminating
one slot with a hollow circular patch for matching the impedance of the CPW and
the slotline. Several other types of transitions using different combinations of CPW
short and slotline open circuits are provided by Ho [2.10] and an experimental
approach is employed to verify the concept. Three types of transitions are proposed
in [2.11] and several improved transitions based on similar ideas have been reported
to have a maximum relative bandwidth of 7.7:1 for a 10 dB return loss [2.12]. Most
of the realized transitions are of the short circuit type in which the signal line of the
CPW is shorted to one ground plane. This might cause problem in DC biasing an
amplifier when connecting the amplifier to the transition.
In this study, a new transition based on CPW-slotline mode conversion
characteristic is proposed, and the transition bandwidth is broadened by the use of
air-bridges. This is an open circuit type transition, and possesses intrinsic DC
isolation characteristic. The Finite-Difference Time-Domain (FDTD) method is
4
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
employed to analyze the transition, and experimental results are also given to verify
the design concept Furthermore, the applications of this new transition are
demonstrated by building several CPW passive circuits based on this transition.
Numerical and measured results for a CPW-fed Vivaldi antenna, a CPW power
divider and a CPW-slotline coupler are provided to demonstrate the effectiveness of
these new CPW passive circuits.
2.2 FDTD Modeling of Mode Conversion
There have been several studies in the literature to examine mode conversion and the
ability of the air-bridge to eliminate slotline mode in frequency domain [2.1-2.2,
2.13—2.14]. The incorporation of the air-bridge effect is proceeded by either a
capacitive model [2.1] or a simplified Green’s function [2.2]. Due to the complexity
in deriving the dyadic Green's function for a 3D structure, frequency domain method
is relatively difficult to be used to analyze a complex circuit. On the contrary, the
time domain method FDTD is a good candidate to analyze a relatively complex
circuit. However FDTD method hasn't been employed to analyze the mode
conversion intensively so far [2.7, 2.8]. In this section, FDTD is used to explore the
mode conversion of an asymmetric CPW with and without air-bridges and the
results are compared to those obtained by spectral domain method.
FDTD method has been intensively applied in a wide variety of electromagnetic
studies and the formulations can be found from many references [2.15,2.16]. In this
5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
study, a 3D FDTD code with the Perfectly Matched Layer (PML) ABC [2.17] has
been written and verified. The excitation used in this study is of the form
e
(2.1)
where t 0 is the time shift, T controls the width o f the pulse, x q is the center of the slot
and / is half width o f the slot. The pulse has a Guassian variation in time and a
singular function in space to account for the edge condition. This field variation is
used to excite the fields of the two slots with a 180° phase difference for launching a
CPW-mode field. Since the PML ABC requires a large amount of memory and a
long run time, the choice of the parameters of the PML ABC is very important. In
this study, the number of PML layers and the artificial conductor loss are chosen
according to the criterion suggested by Wu and Fang [2.18]. An 8-layer PML ABC
and a reflection coefficient of 1O'6 are used throughout this study for a compromise
between memory requirement and performance of the PML ABC.
Due to the multi-modal field nature of the CPW, a post-processing procedure as
proposed in [2.8] is used to separate the CPW and slotline modes. Since the CPW is
excited in a complementary fashion, the incident wave is a mixture of CPW modes
(fundamental and higher order). Due to the different electric lengths of the two
slots, a mixed CPW and slotline modal fields are generated as reflected wave and
transmitted wave. The CPW and slotline modes can then be separated by using the
simple manipulations
6
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
y r tf* ^ _
* 2 Unit)
K jn (0
(2 2 )
2
Vnrait)
r ^ iO + V ^ i t )
(2.3)
(2.4)
y t r j
^
_
K
^ u < ( 0 + * 2 .o « tf ( 0
(2.5)
2
where the subscript I, 2 are the indices of the two slots, in, out are for input and
output ports, c and s indicate the CPW mode and slotline mode, and inc, ref and tr
correspond to incident, reflected and transmitted waves, respectively. Since this
transition is a multi-port and multi-mode structure, the S-parameters used in this
paper are represented as S^, where m, n denote the indices of the ports and x, y are
the modes at port m and port n, in which port n is the excitation port, x, y can be
either c for CPW mode or s for slotline mode. For example, S £ represents the Sparameter for slotline mode at port 3 with a CPW mode excited at port 2.
An asymmetric CPW circuit, as shown in Fig.2.1, is analyzed using FDTD. The
circuit has been analyzed previously by Lee [2.2], with which a large amount of
slotline mode is expected to be excited. The comparisons of the S-parameters are
shown in Fig.2.2 for the cases with and without air-bridge. Both the magnitude and
the phase of the S-parameters agree very well for the circuit without air-bridges,
however a relatively large deviation is observed for S“ of the case with air-bridges.
7
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3+/%+%$%$
Figure 2.1: The asymmetrical CPW circuit
It is found that in deriving the Green's function in spectral domain for the air­
bridge, the current density on the air-bridge is assumed to be ID only in order to
simplify the derivation [2.2]. However with the same structure, it has been shown by
FDTD that the current density is indeed 2D. Therefore the spectral domain results
could differ from the FDTD results since a different current distribution is assumed
which is believed to be the reason for the deviation.
2.3
Design Concept of the New Transition
One phenomenon which always occurs in CPW circuits is the CPW-slotline
mode conversion when the electric lengths of the two slots are different [2.19]. At
8
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
0
-1
-2
S1 1(CC)
-3
dB
-4
-5
-6
-7
-8
-9
►
♦
S11(sc L
♦
...........
♦— mA.
4
♦
i
'
1
li
«<*___
6
7
Freq (GHz)
(a)
5
i
0
.
.
it
t
1
I
-5
•10
dB
-15
♦
♦
i
♦
♦
S11(sc]
-20
♦
-25
-30
........
♦
'! ♦
a
i
i
\
6
7
Freq (GHz)
(b)
Figure 2.2: The comparison of the S-parameters between the spectral domain and
FDTD results for the asymmetrical CPW circuit, (a) Without air-bridge, (b) With
air-bridge.
♦ SDM [2.2]
—
...... FDTD.
9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
frequency where the electric length difference between the two slots is exactly 180°,
the CPW-mode field converts to slotline-mode field completely, as shown in Fig.2.3
(a). However, for other frequencies whose corresponding electric length differences
between the two slots o f CPW are away from 180°, the resultant modal field at one
location is a mixture o f CPW-mode and slotline-mode, as shown in Fig.2.3 (b). The
slotline-mode field becomes weaker when the electric length difference between the
two slots is farther away from 180°, and this decrement of slotline-mode field is
gradual in comparison with the response found in other resonant type structures.
Also note that the orthogonality of the CPW and slotline mode in CPW, as shown in
Fig.2.4, is a very distinct feature of this multi-mode CPW transmission line.
The proposed new transition utilizes the mode-conversion feature to fulfill the
transition function. As shown in Fig.2.5, the proposed transition consists of three
sections: a CPW, a 180° phase shifter and a slotline. The 180° phase shifter
connected between the CPW section and the slotline section converts CPW mode to
slotline mode. Since a slotline cannot support CPW mode, all the CPW-mode fields
are bounced back from the front edge o f the slotline section. The transmission
efficiency is centered at one frequency which corresponds to a 180° electric length
difference between the two slots, and the roll-off of the transmission efficiency is
slow. This scenario can be well explained in Fig.2.6 in which the slotline mode
propagates to the slotline while the CPW mode is reflected back by the edge of the
mode conversion section.
10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
76
(a)
I
9999999
(b)
Figure 2.3: Illustrations of CPW-slotline mode conversion, (a) The lengths
difference is 180°. (b) The lengths difference is not 180°.
11
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
©
©
s
©
a
Orthogonality between CPW mode and Slotline mode
+
III
©
■©
o
E
©
a
©
-a
B
at
£
Bn
u
o
©
b
s
Figure 2.4: Orthogonality of CPW mode and slotline mode.
12
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
62ZZZZ
Air-bridge
CPW feed line
Figure 2.5: The proposed CPW-SIotline transition.
Moreover, when the reflected CPW-mode field travels back to the front edge of
the mode conversion section, a mixture of CPW-mode and slotline-mode field is
again generated. An air-bridge can be placed at the front edge of the mode
conversion section to reflect the slotline-mode field back toward the slotline section,
as shown in Fig.2.7. Using this 'bouncing-back-and-forth' scheme, the transition
efficiency is increased, thereby broadens the transition bandwidth. It should be
pointed out that the success of bandwidth broadening is due to the presence of the
13
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
* Slotline mode ( " • ) propagates through
* CPW mode ( ■►) reflected back
V r'fe = °"
Figure 2.6: The mode conversion in the proposed transition without air-bridge.
14
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Slotline mode reflected back
CPW mode propagates
Figure 2.7: The effect of the air-bridge on the CPW mode and slotline mode.
15
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
air-bridge, which is also a very unique feature of this transition.
The above explanation is based on the assumption that an ideal 180° phase
shifter is used to realize the mode conversion. In reality, the phase delay section not
only provides a phase shift but also creates discontinuities to the fields. The actual
dimension of the phase shifter is also a very important factor for maximizing the
transition bandwidth. The design of the phase shifter, however, is not
straightforward because this transition itself represents a very complex boundary
value problem. A full-wave method is essential for analyzing this structure in order
to obtain a maximized transition bandwidth. In this study, the Finite-Difference
Time-Domain (FDTD) method is employed for full-wave analysis of this transition.
2.4 CPW-Slotline Transition Characteristics
The new CPW-slotline transition is designed to operate at 13 GHz by having a
length difference of X/2 for the two slots on a substrate with dielectric constant of
10.2 and thickness of 25 mil. The characteristic impedance of the CPW is chosen to
be 50 Q. In FDTD simulation, the CPW is excited by launching a CPW mode, and
time domain voltage waveforms are recorded at the input and output reference
planes. The different modes (CPW and slotline modes) are extracted using Equation
(2.2)-(2.5), and S-parameters are obtained by performing a fast Fourier transform
(FFT).
16
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5
S11(sc)
S11(cc)
S21(sc)
Freq (GHz)
Stub:
# Number of stub :
2
136 mm
# Center conductor width :
0.42 mm
# Gap width:
CPW:
# Slot width :
03429 mm
# Center conductor w idth: 1.0287 mm
Air-Bridge:
# Height:
# W idth:
03116 mm
0.4900 mm
(«)
17
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
S11(sc)
S11(cc)
S21(sc)
10
15
20
30
Freq (GHz)
Stub:
# Number of stu b :
1
# Center conductor width :
0.9144 mm
# Gap width :
0-3429 mm
CPW*
# Slot width :
0.3429 mm
# Center conductor width : 1.0287 mm
Air-Bridge:
# Height:
# Width :
0.2116 mm
0.4000 mm
(b)
18
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
5
0
5
10
15
20
25
30
Freq (GHz)
(C)
Figure 2.8: The dimensions and the calculated S-parameters for the CPW-slotline
transitions.
In order to obtain a wider transition bandwidth, intensive numerical simulations
using FDTD have been made to determine the dimensions of the phase shifter. The
number, width and gap of the stubs have been adjusted to optimize the performance
of the transition. It has been found that one-stub design yields the widest bandwidth
among all cases. This could be due to the stronger discontinuity effect produced if
more stubs are introduced. Two transitions and the simulated results of S-
19
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
parameters, as shown in Fig.2.8, are used throughout this study for demonstrating
the applications of this transition.
The new CPW-slotline transition is analyzed for a back-to-back configuration
using FDTD and numerical results are compared to experimental results. The circuit
is fabricated on RT/Duroid 6010 substrate with a relative permittivity o f 10.2 and
thickness of 25 mil. The experiment is performed using an HP 8720A Network
Analyzer which is calibrated from 300 MHz to 20 GHz. Since SMA connectors
which represent short circuits for slotline-mode fields are used in measurements, it is
not easy to distinguish the S-parameters for different modes [2.19]. The measured
results will be somewhat deviated because of the multiple reflection caused by the
SMA connectors and the inability of the experimental setup to separate different
modes. The comparison of
between numerical and experimental results is shown
in Fig.2.9 where a good agreement in magnitude and trend over a wide frequency
range is observed. Some deviations are noticeable in high frequency region and
some dips are also shown in the figure. This is mainly because a lossless structure is
assumed in FDTD simulations. Furthermore, since the SMA connectors used in this
study cause reflections when frequency becomes higher, some dips are observable
for S2 icc. Moreover, part of the deviation is caused by the residual slotline-mode
field which is not completely eliminated by the air-bridges used in the experiment. It
has been pointed out in [2.19] that the slotline-mode field is bounced back and forth
which may cause a noticeable error in measurements.
20
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
-10
FDTD
-25
Measured
Uniform CPW
-30
0
5
10
15
20
Freq (GHz)
Figure 2.9: Comparison between calculated and experimental result for the back-toback CPW-slotline transition.
The single transition is also analyzed for the case without air-bridge to show the
bandwidth broadening effect by the air-bridge. The result, as shown in Fig.2.10,
indicates that without the air-bridge, the transmission is maximum at the design
frequency and falls off gradually. This well demonstrates the operating principle of
this transition which has been stated earlier. A comparison between the case with
air-bridge (Fig.2.8(b)) and without air-bridge (Fig.2.10) clearly shows the
improvement in transition performance by the addition of the air-bridge. The field
distribution plots in CPW section with and without air-bridge for a 17 GHz
sinusoidal excitation are also obtained from FDTD simulation, as shown in Fig. 2.11
These plots indicate that the slotline mode is suppressed greatly by air-bridge, and
the transmission bandwidth is increased as a result.
21
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
-10
-15
dB
-20
-25
-30
S11(sc)
S11(cc)
S21(sc)
-35
-40
0
5
10
15
Freq (GHz)
20
25
30
Figure 2.10: Calculated S-parameters for the CPW-slotline transition without air
bridge.
2.5 Applications Examples
Three new passive CPW circuit components based on the proposed novel CPWslotline transition are demonstrated and analyzed in this section. All the numerical
procedures and FDTD parameters as well as the measurement setup are the same as
those in the previous section. The working principle, simulated and measured results
are presented in this section as well. The comparison between simulated and
measured results is also made in this section.
22
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
II
Without air-bridges
40
20
.///nm nm iiillW v
"/////»
With air-bridges
5
10
15
20
25
30
35
40
45
50
55
Figure 2.11: Field distributions in CPW section of the CPW-slotline transition with
and without air-bridges.
23
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.5.1
A CPW-Fed Vivaldi Antenna
Belonging to the family of tapered slotline antennas (TSA), Vivaldi antenna has
been known for having a broad bandwidth since its introduction by Gibson [2.20].
For feeding the antenna by a coplanar waveguide while retaining the wide band
characteristic, a broadband CPW-slotline transition is required. Previously, there
have been intensive studies on the tapered slotline antennas, including LTSA and
CWSA [2.21, 2.22]. While most of the studies have been focusing on the radiation
pattern, there is not much information about the impedance for TSAs [2.23]. In this
study, the new broadband transition utilizing the CPW-slotline mode conversion is
employed to feed the Vivaldi antenna. Although it has a wide band characteristic,
the beamwidth of the Vivaldi antenna is the largest among the threetapered slot
antennas [2.24]. In this study, the effect of substrate cutting is examined for the input
impedance matching and radiation pattern. The principle and design of the Vivaldi
antenna can be easily referenced [2.20, 2.21]. In this study, the exponential function
described by y ( x ) = ± A e 61 where A = .05 mm and b = 0.118 is used. The
antenna is designed to operate at 13 GHz, for which the length and the opening
width of the antenna are two free space wavelengths and one free space wavelength
respectively, for having an effective radiation while keeping a compact size. Another
exponentially tapered transition is employed to transform the impedance of the
slotline to the input impedance of the Vivaldi antenna. The length of the tapered
transition is chosen as one free space wavelength. The entire configuration for the
24
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 2.12: The CPW-fed Vivaldi antenna.
5
with cutting
no cutting
0
-5
v: ?r*
v . *: :
-10
**
d
i fI
i iy
|S 1 1 |’15
(dB) .20
-25
-30
-35
i—1-i f
» t I .t » ■ . I . t r «-
-40
0
2
4
6
8
10
12
14
16
18
Freq (GHz)
Figure 2.13: Measured return loss of the CPW-fed Vivaldi antenna.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
90
45
135
■25
i-25 --20 '1 5 V*° r5
180
no cutting
with cutting
E-Plane
225
315
270
180
cutting
cutting
H-PIane
270
Figure 2.14: The radiation patterns of the CPW-fed Vivaldi antenna at 13 GHz for
the cases with and without substrate cutting.
26
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
transition, tapered transition and the Vivaldi antenna is shown in Fig.2.12. The
measured return loss, as shown in Fig. 2.13, indicates a wideband characteristic of
the Vivaldi antenna which is now fed with a standard 50 Q CPW. The measured
return loss o f the transition-fed Vivaldi antenna with substrate cutting is also shown
in Fig.2.13. It is indicated that the return loss is slightly degraded than that of
without substrate cutting. Fig.2.14 shows the measured E-plane and H-plane
radiation pattern for both cases at 13 GHz. It is shown that the beamwidth is
narrower for the case of cutting the substrate.
2.5.2 A CPW Power Divider
A CPW can be viewed as two coupled slotlines excited out o f phase. A CPW power
divider can be built by placing two CPW-slotline transitions side by side, as shown
in Fig.2.15. The two CPW-slotline transitions split the input RF signal and convert
the slotline-mode fields into CPW-mode fields. Because the transitions are
broadband, the resultant response for the power dividing is also very wideband. The
power divider is analyzed using FDTD and the simulated results are compared to
measured results. The power divider is designed to have a 50 Q characteristic
impedance at each port. The distance between the two transitions must be chosen far
enough to prevent the induction of the slotline mode in the two output CPWs and
close enough to have a minimized tapered transition. In this study, the spacing is
chosen as 9.144 mm and other parameters are the same as those in Fig.2.8(b). A
27
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
RF o u t ( -
3 dB f o r each )
//* - s v /
AH the other (Hmcasioiu
e as those In
n*.2A (h).
RF in
Figure 2.15: The CPW power divider.
o
I" 1
-5
-10
-15
-20
dB
-25
-30
-35
1
/
¥
S21
S21
S11
S11
/
-40
•45
8
10
12
Freq (GHz)
14
(Measured)
(FD-TD)
(Measured)
(FD-TD)
18
20
Figure 2.16: Comparisons between simulated and measured S-parameters for power
divider.
28
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
linear tapered transition is used to connect the input CPW and the two transitions,
and the length is chosen as 9.144 mm. hi FDTD simulation, all ports are terminated
by perfectly matched layer (PML), and one of the output ports is terminated by a 50
Q load during the measurement. The measured results, compared with the simulated
results, shown in Fig.2.16, indicate good agreements and also show the wideband
characteristic of this novel power divider.
2.53
A CPW-SIotline Coupler
A CPW possesses better propagation characteristics than a slotline because the
radiated fields produced by the two slots of the CPW tend to cancel each other as a
result of opposite field polarization [2.25]. Therefore the coupling between a slotline
and a CPW should be weaker than the coupling between two slotlines. In many
uniplanar circuit designs which employ a mixed structure o f CPWs and slotlines, an
efficient CPW-slotline coupler should be of practical significance. By coupling a
slotline to the slotline section of our proposed transition, a very effective coupling
structure can be easily realized. As shown in Fig.2.17, using a back-to-back
configuration, the energy in the main CPW is first converted to slotline-mode field
and coupled to another slotline from the slotline section. The field in the main guide
is then converted back to CPW-mode field by another slotline-CPW modeconversion section. In this fashion, CPWs can be used throughout the entire circuit
while the energy can be coupled to slotlines from the slotline section of the back-to-
29
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
back CPW-slotline transition. Fig.2.18 shows the calculated S-parameters for a
slotline side coupled to a back-to-back CPW-slotline transition. A reasonably good
coupling efficiency is clearly demonstrated for this novel coupler.
30
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
SfatKna
field
Z1
CPW
mode
/
AH the other dimensions
ere the a n as those in
Fig.2J(a).
Figure 2.17: A CPW-slotline coupler utilizing the CPW-slotline transition.
Freq (GHz)
Figure 2.18: Calculated S-parameters of the coupler.
31
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
References 2
[2.1] N. I. Dib, M. Gupta, G. E. Ponchak, and L. P. B. Katehi, "Characterization of
Asymmetric Coplanar Waveguide Discontinuities,” IEEE Trans. Microwave
Theory Tech., vol.41, pp.1549-1558, Sept. 1993.
[2.2] C.-Y. Lee, Y. Liu, and T. Itoh, "The Effects of the Coupled Slotline Mode and
Air-Bridges on CPW and NLC Waveguide Discontinuities/’ IEEE Trans.
Microwave Theory Tech., vol.43, pp.2759-2765, Dec. 1995.
[2.3] C.-C. Tien, C.-K. C. Tzuang, S. T. Peng, and C.-C. Chang, "Transmission
Characteristics o f Finite-Width Conductor-Backed Coplanar Waveguide,”
IEEE Trans. Microwave Theory Tech., vol.41, pp.1616-1624, Sept. 1993.
[2.4] G. E. Ponchak and L. P. B. Katehi, "Open- and short-circuit terminated series
stubs in finite-width coplanar waveguide on silicon,” IEEE Trans. Microwave
Theory Tech., vol.45, pp.970-976, June. 1997.
[2.5] M. Riaziat, I. Zubeck, S. Bandy, and G. Zdasiuk, "Coplanar waveguides used
in 2 - 18 GHz distributed amplifier,” in 1986 IEEE MTT-S Dig., Baltimore,
MD, pp. 337-338.
[2.6] A. A. Omar and Y. L. Chow, "Coplanar waveguide with top and bottom
shields in place o f air bridges,” IEEE Trans. Microwave Theory Tech., vol.41,
pp. 1559-1563, Sept. 1993.
[2.7] P. Mezzanotte, G. Pompei, L. Roselli, R. Sorrentino, "FD-TD Analysis of
Coplanar Waveguide to Slotline Transitions Accounting for Air-Bridge,
Shielding Effects and Coaxial Connectors,” in Proc. 24th European Microwave
Conf., vol.2, pp.1929-1932,1994.
[2.8] M. Rittweger, N. H. L. Koster, S. Kosslowski, R. Bertenburg, S. Heinen, and I.
Wolff, "Full-Wave Analysis of a Modified Coplanar Air Bridge T-Junction,”
in Proc. 21th European Microwave Conf., pp.993-998, Stuttgart, 1991.
[2.9] T. Q. Ho and S. M. Hart, "A Broad-Band Coplanar Waveguide To Slotline
Transition,” IEEE Microwave and Guided Wave Letters, vol. 2, pp.415-416,
Oct. 1992.
[2.10] T. H. Ho, L. Fan, and K. Chang, "Experimental Investigations of CPWSlotline Transitions for uniplanar MICs,” in 1993 IEEE MTT-S Dig., Atlanta,
GA, pp.877-880.
32
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
[2.11] W. Grammer and K. S. Yngvesson, "Coplanar Waveguide Transitions to
Slotline: Design and Microprobe Characterization,” IEEE Trans. Microwave
Theory Tech., vol. 41, pp.1653-1658, Sept 1993.
[2.12] K. Hettak, J. Ph. Coupez, T. LE Gouguec, S. Toutain, P. Legaud, and E.
Penard, "Improved CPW to Slotline Transitions,” in 1996 IEEE MTT-S Dig.,
San Francisco, CA, pp.1831-1834.
[2.13] Thomas Becks and I. Wolff, "Full-Wave Analysis o f Various Coplanar
Bends and T-Junctions with Respect to Different Types o f Air-Bridges”, in
1993 IEEE MTT-S Dig., June 1993, pp.697-700.
[2.14] E. Vourch, M. Drissi, J. Citeme and V. Fouad Hanna,” A Full-Wave
Analysis of Coplanar Waveguide-Slotline Transition", in 1994 IEEE MTT-S
Dig., June 1994, pp.1309-1312.
[2.15] K. S. Kunz and R. J. Luebbers* The Finite Difference Time Domain Method
for Electromagnetics. CRC Press, Inc., 1993.
[2.16] A. Taflove, Computational Electrodynamics: The Finite-Difference
Time-Domain Method. Artech House Publishers, 1995.
[2.17] J. P. Berenger, ” A Perfectly Matched Layer for the Absorption of
Electromagetic Waves,” J. Comput. Physics, 114, pp.185-200, Oct. 1994.
[2.18] Z. Wu and J. Fang, "Numerical Implementation and Performance of Perfectly
Matched Layer Boundary Condition for Waveguide Structures,” IEEE Trans.
Microwave Theory Tech., vol. 43, pp.2676-2683, Dec. 1995.
[2.19] M.-D. Wu, S.-M. Deng, R.-B. Wu, and P.-W. Hsu, "Full-Wave
Characterization of the Mode Conversion in a Coplanar Waveguide RightAngled Bend,” IEEE Trans. Microwave Theory Tech., vol. 43, pp.2532-2538,
Nov. 1995.
[2.20] P. J. Gibson, "The Vivaldi aerial,” in Proc. 9th European Microwave Conf.,
(Brighton, U.K.), 1979, pp.101-105.
[2.21] K. S. Yngvesson, D. H. Schaubert, T. L. Korzeniowski, E. L. Kollberg, T.
Thungren, and J. F. Johansson, "Endfire tapered slot antennas on dielectric
substrates,” IEEE Trans. Antennas Propagat., vol. 33, pp.1392-1400, Dec.
1985.
[2.22] K. Sigfrid Yngvesson, T. L. Korzeniowski, Young-Sik Kim, Erik L. Kollberg
33
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and Joakim F. Johansson, "The Tapered Slot antenna - A New Integrated
Element for Millimeter-Wave Application,” IEEE Trans. Microwave Theory
Tech., v. 37, pp.365-374, Feb. 1989.
[2.23] R. N. Simons and R. Q. Lee, "Linearly tapered slot antenna impedance
characteristics,” 1995 IEEE Antennas and Propagation Society International
symposium, pp.170-173,1995.
[2.24] S. Sugawara, Y. Maita, K. Adachi, K. Mori, K. Mizuno, " A mm-wave
tapered slot antenna with improved radiation pattern,” 1997 IEEE MTT-S
International Microwave Symposium Digest, v.2, pp.959-962,1997.
[2.25] T. Itoh, "Overview of Quasi-Planar Transmission Lines,” IEEE Trans.
Microwave Theory Tech., vol. 37, pp.275-280, Feb. 1989.
34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 3
NOVEL APPLICATIONS O F PH O TO N IC BANDGAP
STRUCTURE
3.1 Introduction
A novel two-dimensional Photonic Bandgap (PBG) lattice is employed to construct
several novel passive circuits to conquer several difficulties occurred in some
microwave structures. A PBG lattice is a periodic structure which possesses a very
distinctive stopband phenomenon. There are many different ways to construct a PBG
lattice for microwave applications. Previously, PBG pattern was realized by either
drilling holes in the substrate or etching holes in the ground planes. The holes can be
of any shape, for which the property of the realized PBG structure will be different.
Due to the difficulty of drilling holes in the substrate, etching PBG pattern in the
ground planes becomes a better choice. The PBG lattice used in this study is twodimensional and etched in the ground planes. Due to this uni-planar characteristic,
the fabrication process is really simple and standard. Fig.3.1 shows the PBG pattern
used in this study and the stopband characteristic for a microstrip line with PBG
ground plane configuration. It should be noted that the stopband will be different for
35
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
other types of circuit configurations, in which the full-wave simulations must be
employed to examine the stopband.
3.2 Novel Non-Leaky Conductor-Backed Coplanar Waveguide
Belonging to one member of the CPW family, conductor-back CPW (CB-CPW) has
been used for some applications due to its stronger mechanical strength and the
increased gain for antenna applications. One undesired feature of CB-CPW is the
parallel-plate mode leakage, which may deteriorate the circuit performance and
cause crosstalk with the neighboring circuits [3.1]. Previously, two methods have
been employed to suppress the leakage. One method uses shorting posts to block the
leakage and the other method uses multi-layered structure to shift the dispersion
curve of the parallel-plate mode [3.1]. From the view point of a planar circuit
process, both methods exhibit a disadvantage in terms of the production cost due to
the more complex circuit configurations.
In this study, a novel uni-planar structure is proposed for which is realized by
etching a 2-D photonic band gap (PBG) lattice in the upper ground planes to
suppress the leakage in CB-CPW. The effectiveness o f this newly proposed structure
is demonstrated experimentally by comparing the measured results for a
conventional CPW, a conventional CB-CPW and the proposed PBG-CB-CPW. Our
experiments show that the parallel-plate mode leakage is almost completely
suppressed in the stopband of the PBG lattice.
36
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
d :0.127 mm
w : 2d
g :5 d
s i : 10 d
s2 :6 d
I : 2d
(a)
o
-10
20
fi-30
-40
-50
S21(FDTD)
.-a - - S21 (M easured)
S21(Ref_MS)
-60
0
2
4
6
10
8
12
14
16
18
20
Frequency (GHz)
(b)
Figure 3.1: (a) The 2-D PBG lattice and (b) the insertion loss for a uniform
microstrip line with a PBG ground plane which shows the stopband phenomenon.
37
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig.3.2 shows the structural view o f a CB-CPW, where the two top ground
planes and the center conductor form a CPW while a parallel-plate waveguide is
formed between top and the bottom ground planes. The energy will leak along a
particular angle once the wave is launched. This energy leakage is significant even at
low frequencies, which may cause severe effect on the circuit performance.
It is apparent that the leakage can be stopped if a guard ring type method is used
along the signal propagation direction, as shown in Fig.3.3(a). A commonly
employed guard ring is via, which not only increases the production cost but also
makes the design procedure more complex. Another method uses multi-layered
substrate to accelerate the phase velocity of the parallel-plate mode to stop the
leakage over a specific frequency range, as shown in Fig.3.3(b), which is dictated by
the dispersion curve of the multi-layered structure [3.1]. Although its effectiveness
for suppressing the leakage, multi-layered structures possess a fabrication
complexity which precludes its usage for low cost products.
It is well known that a periodic structure exhibits a stopband phenomenon which
can be used to stop the wave propagation [3.2]. Several planar PBG circuits have
been demonstrated to utilize the stopband characteristic to either stop the surface
wave or eliminate the harmonics generated by a power amplifier [3.3, 3.4]. Fig.3.4
shows our newly proposed PBG lattice for leakage suppression in CB-CPW. Our
recent study by using this 2-D PBG lattice for the ground plane of a microstrip line,
reveals a wide and distinctive stopband for an X-band prototype [3.5].
38
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ParmUei-plate waveguide
Regular CPW
(a)
•20
12
14
18
20
Freq (GHz)
(b)
Figure 3.2: (a) The structural view o f the CB-CPW. (b) FDTD simulated result for
showing the leakage in CB-CPW.
39
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(a)
(b)
Figure 3.3: Conventional methods of suppressing the leakage of CB-CPW. (a)
Shorting posts and (b) multi-layered structure.
40
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Signal propagation
Figure 3.4: The proposed nonleaky PBG-CB-CPW.
41
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
For the present study, the PBG lattice is employed to stop the propagation o f the
parallel-plate mode. As can be seen from Fig.3.4, this PBG lattice should have little
or no effect on CPW and slotline modes because most of the energy is concentrated
between the two slots of the CPW. On the other hand, this PBG lattice exhibits an
effective stopband over which the leakage can be stopped for the parallel-plate
mode. The PBG lattice can be fabricated using standard etching process and can be
adopted in MMIC designs without any modification, which is much easier and
cheaper than all other methods proposed previously.
Three CPW through lines, including a conventional CPW, CB-CPW, and the
proposed nonleaky PBG-CB-CPW, are fabricated on RT-Roger Duroid 6010
substrates with an effective dielectric constant of 10.2 and a thickness of 25 mils.
The measurements are done using HP 8720A which is calibrated from 130 MHz to
20 GHz. The measured results, shown in Fig.3.5, indicate that for a CB-CPW, the
leakage is observed over all frequencies and is significant even at lower frequencies.
The conventional CPW, meanwhile, shows moderately low insertion loss as
expected, with some ripples at higher frequencies caused by the reflections in
between the two SMA connectors. As shown by the solid curve in Fig.3.5, the
leakage loss due to the back conductor has been suppressed almost completely by
using the PBG ground planes. The insertion loss o f the new PBG-CB-CPW structure
is comparable to that of the conventional CPW between 9 GHz and 14 GHz,
which corresponds to the stopband of the 2-D PBG lattice for the parallel-plate
mode.
42
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
stop b an d o
0E
-5 j
-10 i
-15 :
-20 1
|S21|
-25 :
-30 |
CB-CPW
CPW
Nonleaky PBG-CB-CPW
-35 ;
-40 |
45 t
0
2
4
6
8
10
12
14
16
18
20
GHz
Figure 3.5: Measured insertion losses of conventional CPW, CB-CPW and the
newly proposed PBG-CB-CPW. A stopband from 9 GHz to 14 GHz can be observed
in this figure.
3.3 Leakage Suppression in Stripline Circuits
Stripline circuits have been used in a variety o f microwave systems, including
stripline-fed slotline-coupled patch antennas [3.6]. In some MMIC designs, due to
43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the high integration density offered by a multi-layer structure, the stripline may also
be formed [3.7]. One undesired feature of stripline circuits is the parallel-plate mode
leakage existing when the symmetry in the cross-section is destroyed [3.8], which
may deteriorate the circuit performance and cause crosstalk with neighboring
circuits. The commonly used method to suppress the parallel-plate mode leakage is
to use shorting posts. The disadvantage o f this method is twofold. Firstly, the design
of the shorting posts must be performed carefully since the shorting posts will either
change the amount of the crosstalk between the two lines [3.9] or change the input
StripDne Mode
Parallel-plate waveguide
Regular Stripline
Figure 3.6: Propagation and leaky modes in stripline circuit structure.
44
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
impedance o f the stripline-fed antennas which makes the design a difficult and time
consuming task [3.6]. Secondly, from the view point of planar process, shorting
posts increase significantly fabrication complexity.
In this study, a novel uniplanar structure which is realized by etching a 2-D
photonic bandgap (PBG) lattice in the ground planes to suppress the leakage is
proposed and demonstrated. The effectiveness of this newly proposed structure is
demonstrated both numerically and experimentally by comparing the coupling
between two short-circuited uniform striplines. Our results show that the parallelplate mode leakage is suppressed by over 30 dB in the stopband of the PBG lattice.
Fig.3.6 shows the structural view of a stripline circuit, where the two ground
planes and the center conductor form a stripline. Meanwhile a parallel-plate
waveguide is formed between the top and bottom ground planes. This parallel-plate
mode has a zero cut-off frequency, which will cause energy leakage once it is
excited inside the structure. This leakage will cause interference between adjacent
circuits, and also lower the efficiency of stripline-fed antennas.
It is well known that a periodic structure exhibits a stopband phenomenon which
can be used to prohibit the wave propagation over a certain frequency range. Several
planar PBG circuits have been demonstrated to utilize the stopband characteristic to
either stop the surface wave, eliminate harmonics generated by power amplifiers, or
suppress the leakage in conductor-backed coplanar waveguides (CB-CPW). For the
present study, the PBG lattice is employed to stop the propagation of the parallelplate mode as what has been demonstrated in the CB-CPW case. Fig.3.7 shows a
45
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
uniform stripline with the PBG lattice etched in the ground planes. The simulated
and measured S-parameters are plotted in Fig.3.8, showing a distinctive stopband
around 8 GHz to 12 GHz. It is also interesting to note that the test circuit used in this
case is built on a Duroid substrate with a dielectric constant o f 2.33 and a thickness
of 31 mil. The PBG lattice is scaled as twice large as the one used in Figure 3.1. The
stopbands for both cases (Figure 3.1(b) and 3.8) are almost the same even though
one is microstrip and the other one is stripline. This indicates that the PBG lattice
can be scaled quite nicely and a simple linear model could be employed to have a
first order estimation of the stopband frequency.
In order to demonstrate the leakage suppression capability of the proposed PBG
lattice for stripline circuits, two test circuits, shown in Fig.3.9, are constructed to
verify the proposed concept of leakage suppression. First task in this study is to
excite the parallel-plate mode. The method for exciting parallel-plate mode in this
study is by connecting a via at the end of an open-ended stripline to destroy the
symmetry of the structure. Two open-ended striplines are connected in a back-toback configuration with a gap in between. All the circuits are fabricated on RTRoger Duroid 5870 substrates with an effective dielectric constant of 2.33 and a
thickness of 31 mils. In order to confirm that the coupling is mainly caused by
parallel-plate mode, a test circuit without vias and PBG lattice is first simulated and
measured. In this fashion, the amount of coupling caused by the stripline mode can
be examined since the symmetry in cross-section o f this circuit is not destroyed and
only stripline mode exists. The results for the circuit without via and PBG lattice
46
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 3.7: The structure o f a test circuit for characterizing the stopband of the PBG
lattice.
ffl
.1 0
VZ
•20
-25
-30
-35
-40
0
2
4
6
8
10
S11
S11
S21
S21
(Meas.)
(FD-TD)
(Meas.)
(FD-TD)
12
14
Freq (GHz)
Figure 3.8: Simulated and measured S-parameters for the structure shown in Fig.3.7.
47
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
(b)
Figure 3.9: Two test circuits for leakage suppression characterization, (a) No PBG
and(b) with PBG.
48
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Stopband
CD
■o
£ -20
o>
©
1 -3 0
S21
S21
S21
S21
(0
Q.
C O -40
(w/o PBG) (Meas.)
(w/o PBG) (FD-T )]
(PBG) (Meas.)
(PBG) (FD-TD) l
-50
-60
6
8
10
12
14
16
Freq (GHz)
Figure 3.10: Simulated and measured results for the two stripline test circuits.
show that the amount of coupling caused by the stripline mode is non-observable
since the gap in between the two striplines is large enough to prevent the coupling of
the stripline mode.
Fig.3.10 shows the simulated and measured results for the test circuits with and
without PBG lattice in between the gap of the two striplines. FDTD simulation
results indicate that the PBG lattice can suppress the coupling caused by parallel-
49
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
plate mode leakage by about 25 dB in the stopband region. The measured Sparameters are in a reasonable agreement with this prediction where an over 30 dB
suppression of leakage coupling has been achieved in the stopband from 8 GHz to
12 GHz which correspond to the stopband region observed in Fig.3.8.
3.4 Novel Magnetic Surface
An electric dipole placed above a perfectly electric conducting (PEC) ground plane
is a canonical electromagnetic problem and has been analyzed completely in the
literature [3.10]. The results can be used for constructing antennas such as a
monopole antenna with a PEC ground plane. The dual of this canonical problem is a
magnetic dipole radiating above a perfectly magnetic conductor (PMC). Contrast to
the realization of a PEC which is not difficult in practical situations, the realization
of a PMC remains a difficult task. The difficulty stems from the fact that there has
not been a suitable material found which can be used as a PMC.
In the this part of the study, an effort has been devoted to realize a magnetic
surface for the use of a magnetic dipole type antenna. The novel 2-D uni-planar PBG
structure as the one presented in Section 3.2 and 3.3 is employed to realize the
magnetic conducting surface. This novel PBG structure has the advantage of easing
the fabrication over other types of PBG structures which always call for vias [3.11].
The characteristics of the constructed PBG magnetic surface is measured and the
FDTD method is employed to analyze this novel realization of magnetic surface.
50
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
E lectric F ield
PEC
E lectric F ield
= 0
(a)
Electric Field
PMC
Electric Field
1= 0
(b)
Figure 3.11: The characteristics o f the PEC and PMC when both are illuminated by a
uniform plane wave, (a) PEC case and (b) PMC case.
51
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Perfectly Electric Conductor
(PEC)
LC Resonator
Perfectly Magnetic Conductor
(PMC)
Figure 3.12: The equivalent circuits of the PEC and PBC planes.
52
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
89
39
Z2ZZ
7777
777 7
7777
7x77.
Figure 3.13: The 2-D photonic bandgap structure used for realizing a magnetic
surface.
Good agreement found in the comparisons between measured and numerical results
verifies this novel concept of realizing a magnetic surface.
The basic difference in electrical property between an infinitely large PEC
and PMC can be characterized by the reflection coefficient for a uniform plane wave
incidence. The magnitudes of the reflection coefficient for both cases are the same
and equal to one, while the phase differs by 180°, as shown in Fig.3.11. Another
view of the difference between PEC and PMC can be made by looking at the surface
impedance of PEC and PMC. PMC calls for an open circuit seen from the incoming
53
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
plane wave, and one way to provide the open circuit condition is to create a periodic
pattern. Each element o f this periodic pattern provides an equivalent L and C parallel
connection which changes the surface impedance, as illustrated in Fig.3.12. At the
frequencies where the periodic loading becomes open, a magnetic surface is created
at those frequencies. The newly proposed 2-D uni-planar PBG pattern, as shown in
Fig.3.13, is one of the candidates for fulfilling this purpose. This novel PBG pattern
has the advantage of being simple and can be easily fabricated using any standard
planar process without the need of using vias which are necessary for other type of
PBG structures proposed previously [3.3,3.11]. The experimental setup, as shown in
Fig.3.14, consists of two X band hom antennas and an HP 8720A network analyzer.
The two hom antennas, one for transmitting and the other for receiving, forms an
(almost) monostatic scattering measurement. A plastic foam with relative dielectric
constant of 1.0 inserted in between o f the two hom antennas is used to hold the
scatterer. Two types of scatterers used in this study are an intact copper (PEC) and a
PBG surface, and both surfaces are fabricated on a conductor-backed Rogers Duroid
6010 substrate with 8r= 10.2 and thickness = 25 mil.
The FDTD method is employed for the analysis of this newly proposed PBG
magnetic surface. Due to the 2-D periodic nature of this PBG structure, the
computational burden can be greatly reduced by applying the Periodic Boundary
Condition and only one period of the PBG elements has to be analyzed. The detailed
description of this implementation can be found in [3.12].
Two scatterers as described in the previous section are fabricated and measured,
54
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Scatterer
Foam
Horn Antennas
Figure 3.14: The experimental setup.
the same structures are simulated using FDTD method. The phase difference is
obtained from the ratio of the two curves displayed in HP8270A network analyzer,
and by taking the phase difference of the fourier transformed values in FDTD for
PEC and PBG cases.
The results, as shown in Fig.3.15, indicate the magnetic surface is successfully
realized at around 14 GHz since a 180° phase difference is observed at that
frequency. Below 14 GHz, the surface is inductive while it is capacitive above 14
GHz. The agreement of the comparison between measured and simulated phase is
quite well with an error of around 6%. The comparison between measured and
55
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
simulated magnitude responses, as shown in Fig.3.16, also shows a good agreement,
and the magnitude is unity over all frequencies.
Several PMC ground planes built on substrates with different dielectric
constants and with different PBG dimensions have also been measured and
simulated. The results show a good scalability of this PMC ground plane. The
comparisons between measured and simulated responses for PMC ground planes
with different dimensions also show a good agreement. Several measurements with
different spacing between the two hom antennas, and distances between antennas
and scatterer are also performed. The results show insignificant changes which
indicates the robustness of the realized magnetic surface.
In the current wireless communications, monopole or loop antennas are used
intensively in the handsets. The performance of these antennas is expected to be
improved with magnetic ground plane which can be realized using the proposed
PMC ground plane. The impedance and radiation characteristics of these antennas
with this novel PMC ground plane is currently under investigation, with which the
results can be utilized for the design of antennas for wireless communications
devices.
56
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
-60
>120
<PBG - <PEC
(Deg)
180
•240
-300
-360
GHz
Figure 3.15: The measured and simulated phase difference response. The phase
difference is defined as the difference between the phase o f the PBG surface and that
of the PEC surface.
10
5
|T(PBG)|
Q
|T(PEC)|
(dB)
-5
-10
-15
10
11
12
13
14
15
16
17
18
GHZ
Figure 3.16: The measured and simulated magnitude response. The plot shows the
ratio of the two magnitudes of the reflection coefficients.
57
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
References 3
[3.1] Yaozhong Liu, Kimin Cha, T. Itoh, "Non-leaky coplanar (NLQ waveguides
with conductor backing,” IEEE Trans. Microwave Theory Tech., v. 43, n. 5,
pp.1067-1072, May 1995.
[3.2] J. D. Joannopoulos, R. D. Meade and J. N. Winn, Photonic Crystals, Princeton
University Press, 1995.
[3.3]- D. Sievenpiper and E. Yablonovitch, "Eliminating surface currents with
metallodielectric photonic crystals,” to be presented at 1998 IEEE MTT-S Int.
Microwave Symp., Baltimore, MD, June 1998.
[3.4] V. Radisic, Y. Qian, and T. Itoh, "Broadband power amplifier using dielectric
photonic band-gap structure,” IEEE Microwave and Guided Letters, v.8, pp. 1314, Jan. 1998.
[3.5] Fei-Ran Yang, Yongxi Qian, Roberto Coccioli, and T. Itoh, " A novel low
loss slow-wave microstrip structure,” submit to IEEE Microwave and Guided
Wave Letters.
[3.6] A. Bhattacharyya, OJordham, and Y. Liu, "Analysis of stripline-fed slotcoupled patch antennas with vias for parallel-plate mode suppression,” IEEE
Transactions on Antennas and Propagation, vol.46, pp.538-545, April 1998.
[3.7] J. Kim, H.-Y. Lee, and T. Itoh, "Novel Microstrip-to-StripIine Transitions for
Leakage Suppression in Multilayer Microwave Circuits,” IEEE 7th Topical
Meeting on Electrical Performance in Electronic Packages, West Point, New
York, pp. 252-255, O ct 26-28,1998.
[3.8] D. Nghiem, J. T. Williams, D. R. Jackson and , A. A. Oliner, "Suppression of
Leakage on Stripline and Microstrip Structures," in 1994 IEEE MTT-S Dig., San
Diego, CA, pp. 145-148, May 1994.
[3.9] G. E. Ponchak, D. Chen, J.-G. Yook, and L.PJB. Katehi, "Characterization of
Plated Via Hole Fences for Isolation Between Stripline Circuits in LTCC
Packages,” in 1998 IEEE MTT-S Symp., Baltimore, MD, pp. June 1998.
[3.10] Constantine Balanis, Advanced Engineering Electromagnetics, pp.314-323.
John Wiley & Sons Publisher. 1989.
[3.11] Yongxi Qian, D. Sievenpiper, V. Radisic, E. Yablonovitch, and T. Itoh, "A
58
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Novel Approach for Gain and Bandwidth Enhancement o f Patch Antenna”,
1998 IEEE Radio and Wireless Conference, Colorado Spring, CO, Aug. 9-12,
1998. *
[3.12] A. Alexanian, N. J. Kolias, R. C. Compton, and R. A. York, "Threedimensional FDTD analysis of quasi-optical arrays using Floquet boundary
conditions and Berenger's PML", IEEE Microwave and Guided Wave Letters,
vol.6, pp. 138-140. March 1996.
59
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 4
GLOBAL TIM E-DOM AIN FULL-WAVE ANALYSIS OF
M ICROW AVE CIRCUITS
4.1 Introduction
With the high demand of information exchange, the performance requirement placed
on the high frequency circuits is becoming more stringent. Traditionally, circuit-type
simulators have been extensively employed for the design of high frequency circuits.
When microwave circuits become more highly integrated and operate at higher
frequencies, designers have to tackle the problems caused by various
electromagnetic effects, including electromagnetic interference, surface wave
leakage, coupling between different sub-circuits and packaging effects. Previously,
the way to characterize these harmful effects and find a solution to eliminate these
effects strongly depended on the experience of the designers since there were very
few effective tools existing which could be used to solve these problems. The
difficulty stems from the fact that all numerical full-wave methods for characterizing
the electromagnetic effects require extensive computer resources which were quite
expensive. However, with the progress o f the computer technology, solving
60
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
these problems has been becoming more practical. Circuit designers may
characterize unwanted EM emission, or design nonstandard circuit components by
utilizing full-wave methods on a desktop personal computer.
Among available full-wave approaches, the Finite-Difference Time-Domain
(FDTD) method has received much attention because of its versatile simulation
capability and high accuracy. FDTD has been applied to many electromagnetic
problems in the last two decades due to the significant advancement of the computer
technology. One important application of FDTD method is the analysis of
microwave circuits by incorporating passive and active devices into the simulation
[4.1-4.4]. An extended FDTD algorithm based on the equivalent sources concept for
incorporating active devices into FDTD method was presented in [4.5]. Applications
o f this algorithm to large-signal analysis of an FET amplifier [4.6] and a diode mixer
[4.7] have been reported. The packaging and crosstalk effects which are significant
in densely integrated circuits have not been considered extensively in the literature.
Moreover, an important issue in wireless communication environment is the
electromagnetic compatibility (EMC), since the wireless environment is full of
electromagnetic waves. The compliance with the regulations of various standards is
extremely important for RF circuit designs. A circuit design without taking these
EMC effects into consideration could cause deterioration of the circuit performance
or even malfunction. It is the purpose of this paper to investigate parasitic effects,
and to extend the range of applicability of this equivalent source algorithm for
analyzing different microwave circuits from different aspects.
61
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.2 The Equivalent Source Algorithm
Consider a multi-port network composed of lumped elements, such as a microwave
MESFET, when the size o f the network is small compared to the shortest signal
wavelength of interest, it can be treated as zero dimension and represented by an
equivalent circuit model. Even though the dimension of the network is considered to
be zero, the effects caused by spatial distribution and packaging structure can still be
taken into account by including parasitic elements in the equivalent circuit model.
The important issue is the establishment o f the communication between
electromagnetic quantities, i.e., electric and magnetic fields, and circuit quantities,
i.e., voltage and current, at each port terminal. As will be shown below, the
employment of equivalent sources concept provides a convenient way of
communicating between these two types of quantities.
There are two types of equivalent sources, namely equivalent current source and
equivalent voltage source [4.5, 4.8], which are equivalent in nature. The placements
of these two equivalent sources are shown in Fig.4.1. The equivalent sources are
related to the equivalent circuit of the device by the state equations derived from
KirchofFs current and voltage laws. The state equations are first-order differential
equations expressed in a matrix form as
A(X)— = B(X)X + F(X ),
dt
where X is the state variable and A, B, F are matrices.
62
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(4.1)
This matrix equation can be expressed as a time-stepping matrix equation by
employing a forward or backward differencing scheme as
— - BCX**1) X"+I - A(X - X" - F(X*+I) = 0.
At
J
At
(4-2)
For linear devices, the elements of the matrices are constant and can be calculated
before the time stepping begins. For nonlinear devices, the elements are varied with
each time step and a Newton-Raphson method can be applied to solve this nonlinear
matrix equation [4.6]. Once the system equations are solved, the terminal voltages at
each port are known and can be used as the feedback to update electric field
components.
The extension of this equivalent source algorithm to CPW circuits is
straightforward. Since the active device is mounted on the top surface of the
substrate, the incorporation of the equivalent current sources can be realized as
shown in Fig.4.2. The reason for using equivalent current source in this
implementation instead of equivalent voltage source is because it is the current
which flows through the center conductor, although the implementation using
equivalent voltage source might be also feasible.
4.3 Crosstalk and Packaging Effects in Active Circuits
A coplanar waveguide (CPW) linear amplifier designed to operate at 4.5 GHz is
employed in the study o f the parasitic effects and electromagnetic interference.
63
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
H-field
Integration
Microstrip
Via
Equivalent Current Source
Ground
(a)
H-field
Integration
Microstrip
Equivalent Current Source
Ground Plane
(b)
Figure 4.1: Configurations for (a) horizontal and (b) vertical equivalent currentsource implementation.
64
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
H-field
Integration
Vias
CPW
Equivalent Current Source
Figure 4.2: Placement of the equivalent source for CPW circuit configuration.
Ltotal,g |
Lg
G
Rg
Cgd
— i\
„Cgs__vg
JL +
Rd
Ltotal,d
Ld
Di
Cds
VtouI.dQ
Qvtoud.g:
Figure 4.3: Small signal model of NEC 76038 MESFET.
65
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
An NE76038 FET transistor in small-signal operation is used in this study. The small
signal linear equivalent circuit of this FET, as shown in Fig.4.3, is optimized from 2
GHz to 12 GHz using Libra. The Libra code for finding the equivalent is shown in
Fig.4.4. The CPW amplifier, shown in Fig.4.5, is measured and simulated using the
modified equivalent source algorithm for CPW circuit. Fig.4.6 shows the
comparison between measured and simulated results, where good agreements can be
seen.
The CPW amplifier is used to study the effects of housing and crosstalk on the
circuit performance when the circuit layout becomes dense. A metallic plate placed
above the amplifier exhibits a shielding property to protect the circuit against the
electromagnetic interference. However, the plate also changes the circuit
performance electromagnetically. The simulated results for varying the distance
between the plate the amplifier is shown in Fig.4.7, in which the performance of the
circuit can be seen to change significantly when the plate is placed closer to the
amplifier. Meanwhile, two simplified CPW amplifiers placed side by side are
simulated to study the crosstalk effect o f a very dense circuit layout. Fig.4.8 shows
the simulated results, in which a non-negligible crosstalk is observed.
It can be concluded from these two cases that the circuit designer should
consider the packaging and layout issue along with many other design issues of the
circuit itself. FDTD method provides an effective tool for evaluating the
performance of the circuit without adopting a more expensive and time consuming
trial-and-error procedure.
66
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
DIM
FREQ GHZ
RES OH
COND /OH
IND NH
CAP pF
LNG MM
TIMEnS
ANG DEG
VOL V
CUR MA
PWR DBM
VAR
11
cl
12
rl
13
r2
14
c2
15
r3
c3
r4
c4
r5
c5
# .01 05385 4 2
.8
# .0 1 0.121514 a
# 5 0.424034 .8
# 0 . 0.850704 2.0
# .1 0.21791 .4
# .0 0.0334494 .8
# .1 0.108244 .7
# .0 4.92255e-006 .3
# .2 0.402695 .7
# .0 0.184805 2.8
# .1 0.237268 .8
# .0 0.0920541 1.2
# .0 0.0494546 .1
# 100.327.865 800.
# 0 . 0.17467 5
CR T
s2p 1 2 0 e:\group\kpm a\academ ic\libra\n76038b.s2p
def2p 1 2 m easured
ind 1 2 M l 10.01034
cap 2 0 c \ l ! 0.16683
ind 2 3 1*12 1054858
res 3 4 r* rl 11.70843
ind 9 0 1*13 10.20135
res 6 9 r*r2 10.58269
ind 12 II 1*14 10.47327
cap 11 0 c*c2 1 0.22610
ind 8 II 1*15 1058952
res 7 8 r*r3 11.26977
cap 4 5 c*c3 1 0.55262
res 5 6 r*r4 10.85596
cap 4 7 c*c4 1 0.06049
vccs 4 7 5 6 m =0.04886 a=0 rl= 1000000000 r2*r5 M O O 1=0
cap 6 7 c*c5 10.07294
def2p 1 12 fetm odel
OUT
fetm odel s i I
fetm odel s22
m easured s i I
m easured s22
FREQ
SW EEP 4 10 0.1
opt
fetmodel model measured
Figure 4.4: Libra code for finding the small signal equivalent circuit.
67
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
L s l : 7.14 mm
L s 2 :4 J 2 m m
W sl: 1.54 mm
W s2 :1.22 nun
D s l: 15.0 mm
D s 2 :1.62 mm
{ : 0-35 nun
w : 0 .94 mm
b : 0 .64 mm
(tr : 10.2
|e|w|
|b
Figure 4.5: Configuration and dimensions of the CPW amplifier.
20
FDTD (S21)
Measured (S21)
FDTD (S11)
Measured (S11)
-30
F,
0
I ■ ■ I . j - U L l ■ ■ I ■ I - - .L U
1
2
3
■ . . . I . . . .
4
5
6
I ... . L U
7
. U .1.U
8
9
, I
i
, I , , I
10
Frequency (GHz)
Figure 4.6: Comparisons between measured and simulated results for the CPW
amplifier.
68
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30
_30
0
1 • ■ ■ • ‘-1_■ ■ ■ • 1 ■ • ■ ■ 1 ■ ■ ■ ■
5
10
15
20
25
Frequency (GHz)
Figure 4.7: The calculated gain for different heights of the upper shielding metallic
plate. One FDTD cell equals 0.2116 mm.
Frequency (GHz)
Figure 4.8: Crosstalk between two adjacent active circuits.
69
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.4 EMC Characterizations for Active Circuits
A metallic enclosure is always needed to prevent RF circuits from unwanted EM
emission and interference caused by other nearby equipment The integrity of typical
shielding enclosure designs is compromised by numerous slots and apertures for heat
vents, and I/O cables among others. The electromagnetic interference through these
slots and seams in conducting enclosures is of greater concern in meeting regulations
specified by various wireless standards. These slots might be thinner than one FDTD
cell, thus in order to keep the computational load to an acceptable level, special
treatment for the thin slot must be taken. For the study of the effects of the EM
interference in this paper, an effective and well-proven thin slot algorithm is used for
the cases that the slot width is smaller than one FDTD cell [4.9-4.11]. The thin slot
can be viewed as a capacitance from a quasi-static point of view, and equivalent
capacitors can be used for modeling the thin slot. The equivalent sr of the thin slot
can be found as [4.11]
s
r,
(4.3)
where K() is the complete elliptic integral of the first kind, w and A are width of the
slot and FDTD cell size, respectively. The equivalent capacitors with the material
properties
70
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(4.4)
ri*-c*a
(4-5)
can be placed in the thin slot region. In this fashion, the thin slot is replaced by
equivalent sources which represent the equivalent capacitance, and the FDTD cell
can be made as large as possible without the need to refine the cell.
A metallic plate with a thin slot placed on top of the amplifier and interfered
with by EM emissions from nearby equipment is simulated using FDTD method.
Three cases with slot lengths of 15.218 mm, 25.252 mm and 37.052 mm are
simulated, and the corresponding slot resonances are roughly 10 GHz, 6 GHz and 4
GHz. The width of the slot fixed at wsi0t = 0.2A and A = 0.118 mm is used in this
case. The thin slot algorithm is employed to model the slot and s _jjr is found to be
2.018 in this case. An electric dipole radiating above the shielding plate simulates
the interferer. The spectrum at the output o f the amplifier, as shown in Fig.4.9,
indicates the interference is picked up by the slot and the effect is more significant
when the slot resonance is closer to the operating frequency of the amplifier. It
should be noted that the interference is non-negligible even when the slot resonance
is not at the operating frequency of the amplifier. Three cases with different widths
of 1A, 0.2A and 0.004A are also simulated in this study. The corresponding values
for sr
are found to be 1,2.018 and 4.6141, respectively. The length of the slot is
kept at 25.252 mm which corresponds to a resonance of 6 GHz. The results for
71
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
250
200
F slo t = 10 GHz
F slo t = 6 GHz
F slo t = 4 GHz
7
Mrtattic Sftcdtoa
Plat*
T3 150
a>
Q.
CO
1 . 100
TY ' t Y
3
r.a.fra.’ih.tB
4
5
6
7
Frequency (GHz)
Figure 4.9: The spectrum at the output of the CPW amplifier for different lengths of
the slot. The Fslot denotes the resnance of the slot, and the width of the slot is fixed
at wsiot = 0.2A.
72
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Resonance due to matching circuit
W slot = 1 cell
W slot = .2 cell
W slot = .004 cell
Frequency (GHz)
Figure 4.10: The spectrum at the output o f the CPW amplifier for different widths of
the slot. The slot length is fixed at FS|0t = 6 GHz, and the one FDTD cell equals 0.118
mm.
73
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
showing the effect of the slot width on the received interference, as shown in
Fig.4.10, indicates an increased level of interference as the slot becomes wider. It is
interesting to observe that the slot resonance also produces an undesired interference
at the slot resonance which could deteriorate the desired signal and reduce the signal
to noise ratio (SNR) as well as increase the bit error rate (BER) in digital
communication.
Experiments are also conducted to study the EM interference effect. To facilitate
the measurement, a hom antenna is used as the interference source. Fig.4.11 shows
the measured power penetrated through the slotted shielding plate, in which a
maximum at around 4.5 GHz and another peak at 6 GHz which corresponds to the
slot resonance can be seen.
The ratio of the penetrated power (the case with slotted shielding plate) to the
power picked up by the circuit (without shielding plate) for a 6 GHz slot resonance
and wsiot= 0.2A is shown in Fig.4.12. In this figure, the reinforcement of the slot on
the penetrated power is presented with an interesting observation that this ratio is
maximum at 6 GHz which corresponds to the slot resonance in this case. This result
should not be surprising because the circuit has more gain at the operating
frequency, the power picked up by the matching circuit at the operating frequency of
the circuit should be the largest for the cases with and without the slotted shielding
plate. However the power picked up by the circuit should be reduced with the slotted
shielding plate for all frequencies except the slot resonant frequency since the slot
74
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3
3.5
4
4.5
5
5.5
6
6.5
F requency (GHz)
Figure 4.11: The measured power penetrated through the slotted shielding plate.
reinforces the penetrated power due to the slot resonance.
A short summary can be made at this point. Although current commercial circuit
simulators provide a fast and accurate tool for circuit designs, to meet the more
stringent performance requirements of high frequency circuits, greater details of
electromagnetic effects must be taken into consideration. The above examples
demonstrate the capability of the extended FDTD method for providing a useful and
effective tool for this purpose. It enables potential circuit designers to identify
the cause of problems, and to design circuits which operate in a harsh wireless
environment.
75
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.5
4
4.5
5
5.5
Frequency (GHz)
6
6.5
Figure 4.12: The ratio o f the measured penetrated power to the measured power
picked up by the circuit without the shielding plate. This ratio indicates the
reinforcement of the slot on the penetrated power.
4.5 A Mixed Full-Wave/Equivalent-Circuit Algorithm for FDTD
Method
The FDTD method has been extensively used to analyze various electromagnetic
problems during the last decade. One important application for FDTD method is the
analysis of microwave circuits, including passive circuit, linear active circuit and
nonlinear active circuit. Despite its flexibility in implementation and high accuracy,
one well known disadvantage of FDTD is the huge memory requirement and
substantially long run time. Several methods have been proposed to reduce either the
computational domain, such as nonuniform mesh method [4.12], or the run time,
76
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
such as Prony method [4.13]. Those methods have some limitations which preclude
their common acceptance.
In this study, a new mixed full-wave/equivalent-circuit algorithm for the FiniteDifference Time-Domain (FDTD) method is proposed. This algorithm partitions a
microwave circuit into several parts and replaces some parts by their equivalent
circuits while the remaining parts are discretized by the standard Yee's mesh. The
equivalent circuits which are obtained by extracting from the corresponding Sparameters are incorporated into FDTD method by the equivalent source method
[4.5]. This method has the advantage of reducing the size of the computational
domain while keeping a high accuracy. A microwave amplifier is analyzed using this
algorithm and the results agree well with those obtained by using standard FDTD
method. A significant reduction for both memory consumption and run time is
apparent and this well demonstrate the effectiveness o f this algorithm.
A typical microwave circuit, as shown in Fig.4.13, consists of several matching
circuits and an active component and can be analyzed using FDTD method as
demonstrated in [4.5]. Since the matching stubs extend over a long distance laterally,
the number of the discretizations along x-direction is very large. The overall
computational domain becomes very large which results in a huge memory
requirement and a long simulation time. Due to this high computational cost,
analyzing the microwave circuits using FDTD method can just be a final verification
step so far.
77
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
«■
Chopped
igstub,
Open ended ndcrostrip
Equivalent load
Figure 4.13: The illustration of the original microstrip amplifier and the chopped
amplifier with equivalent loads.
Figure 4.14: A microstrip open end circuit and its equivalent LC distributed circuit.
78
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
It is apparent that if the stubs can be chopped somewhere and appropriate loads
which represent the open ended circuits are placed at the end o f the chopped stubs,
the computational domain can be reduced laterally, results in a smaller
computational domain, as shown in Fig.4.13. The S-parameters of the open-ended
circuits can be found by any full wave method and an LC distributed circuit can be
used as the equivalent circuit for the open ended circuits, as shown in Fig.4.14. By
using the optimizing function provided by Libra, the values for the equivalent
circuits can be found as a best fitting to the corresponding S-parameters. These
equivalent circuits then are incorporated into the standard FDTD method using the
equivalent source method as described in [4.5]. In this fashion, the memory
requirement and the simulation time are reduced significantly. It should be noted that
error will be introduced with this algorithm since the incorporation of the equivalent
circuit using equivalent source method can not fully represent the actually field
distribution. Fortunately, the error is small and the results are acceptable considering
the significant saving of computational resources. Another advantage of this
algorithm is that the equivalent circuits are reusable in the sense that other designs
using similar matching circuits can use the equivalent circuits found from a wellestablished component library. There is no need to analyze the whole circuit which
results in a long design cycle. This algorithm can be also extended to any circuit
components provided that the equivalent circuits of those components can be found.
A microstrip linear amplifier has been designed using Libra software package to
79
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
S21 -10
(dB)
Original
C hopped
•20
-30
-40
0
2
4
6
8
10
12
14
Freq (GHz)
Figure 4.15: Comparison between the case using standard FDTD method and the
case using this algorithm.
operate at 4 GHz. The complete circuit and the chopped one, as shown in Fig.4.13,
are analyzed using both standard FDTD method and this algorithm. The computation
domain size for the complete circuit is 87 * 100 * 30 and 27 * 100 * 30 for the
chopped one. The memory requirements for both computation domains are 120
Mbytes and 40 Mbytes respectively with an 8-layer PML ABC. A reduction of twothird for the memory consumption and run time is apparent. The simulated Sparameters for both cases, as shown in Fig.4.15, indicate the capability and accuracy
of the proposed algorithm. This algorithm can be also extended for any type of
80
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
passive components as long as the equivalent circuits of those passive circuit are
found and being built into a library.
81
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
References 4
[4.1] W. Sui, D. A. Christensen, and C. H. Dumey, "Extending the two-dimensional
FD-TD method to hybrid electromagnetic systems with active and passive
lumped elements,” IEEE Trans. Microwave Theory Tech., vol. 40, pp. 724-730,
Apr. 1992.
[4.2] M. Piket-May, A. Taflove, and J. Baron, "FD-TD modeling of digital signal
propagation in 3-D circuits with passive and active loads,” IEEE Trans.
Microwave Theory Tech., vol. 42, no. 8, pp. 1514-1523, Aug. 1994.
[4.3] P. Ciampolini, P. Mezzanotte, L. Roselli, and R. Sorrentino, "Accurate and
efficient circuit simulation with lumped-element FDTD technique," IEEE Trans.
Microwave Theory Tech., vol. 44, no. 12, pp. 2207-2215, Dec. 1996.
[4.4] Q. Chen and V. F. Fusco, "Hybrid FDTD large-signal modeling of threeterminal active devices," IEEE Trans, on Microwave Theory and Tech., vol.45,
no.8, pp. 1267-1270, August 1997.
[4.5] C. -N. Kuo, V. A. Thomas, S. T. Chew, B. Houshmand, and T. Itoh, "Smallsignal analysis of active circuits using FDTD algorithm,” IEEE Microwave
Guided Wave Lett., vol. 5, no. 7, pp. 216-218, July 1995.
[4.6] C. -N. Kuo, B. Houshmand, and T. Itoh, "Full-wave analysis of packaged
microwave circuits with active and nonlinear devices: an FDTD approach,”
IEEE Transactions on Microwave Theory and Tech., vol.45, no. 5, pp. 819-826,
May 1997.
[4.7] M. Chen, S. T. Chew, and T. Itoh, "Nonlinear analysis of a microwave diode
mixer using the extended FDTD,” in 1997 IEEE MTT-S International
Microwave Symposium Digest, Denver, CO, USA, pp. 67-70, June 1997.
[4.8] C. -N. Kuo, R. -B. Wu, B. Houshmand, and T. Itoh, "Modeling of microwave
active devices using the FDTD analysis based on the voltage-source approach,”
IEEE Microwave Guided Wave Lett., vol. 6, no. 5, pp. 199-201, May 1996.
[4.9] J. Gilbert and R. Holland, "Implementation of the thin-slot formalism in the
finite-difference EMP code THREDII," IEEE Trans. Nucl. Sci., vol. NS-28, no.
6, pp. 4269-4274, Dec. 1981.
[4.10] K. -P. Ma, M. Li, J. L. Drewniak, T. H. Hubing, and T. P. vanDoren,
82
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
"Comparison o f FDTD algorithms for subcellular modeling of slots in shielding
enclosures," IEEE Transactions on Electromagnetic Compatibility, vol.39,
pp.147-155, May 1997.
[4.11] C. D. Turner, and L. D. Bacon, "Evaluation of a thin-slot formalism for
finite-difference time-domain electromagnetics codes," IEEE Trans.
Electromagn. Compat., vol. EMC-30, pp. 523-528, Nov. 1988.
[4.12] V. J. Brankovic, D. V. Krupezevic, and F. Arndt, " An efficient twodimensional graded mesh finite-difference time-domain algorithm for shielded
or open waveguide structures," IEEE Transactions on Microwave Theory and
Techniques, vol.40, no. 12, pp.2272-2277, Dec. 1992.
[4.13] J. A. Pereda, L. A. Vielva, A. Vegas, and A. Prieto, " Computation of
resonant frequencies and quality factors o f open dielectric resonators by a
combination o f the finite-difference time-domain (FDTD) and Prony's
methods,’’ IEEE Microwave and Guided Wave Letters, vol.2, no.l 1, p.431-433,
Nov. 1992.
83
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTERS
CONCLUSION
Several novel microwave circuits have been proposed and demonstrated in this
study. First, a new CPW-slotline transition utilizing CPW-slotline mode-conversion
phenomenon is demonstrated successfully and the concept of the new transition is
verified both numerically and experimentally. Both results reveal broadband
performance of the new transition as a result of the CPW-slotline mode conversion
and the elimination of slotline mode by air-bridge, as well as the optimization of the
transition geometry. Potential applications of this novel transition have been
demonstrated by designing three type of passive structures, i.e., a CPW-fed Vivaldi
antenna, a CPW power divider and a CPW-slotline coupler.
Several novel microwave circuits utilizing a novel 2D PBG lattice are
demonstrated in this study including leakage suppression in CB-CPW and
stripeline, as well as the realization of a PMC ground plane. For leakage
suppression in CB-CPW and stripline, the leakage is blocked by the stopband of a
2-D PBG structure etched in the ground planes. The proposed PBG lattice can be
easily realized by standard planar process, readily to lower fabrication cost than
84
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
conventional shorting posts. This novel structure should find wide applications such
as stripline-fed antennas and other various stripline based circuits. The PBG pattern
is employed to realize a magnetic surface which has significant influence in the
constructions of some wireless communication antennas such as loop antennas
which is magnetic dipole type antennas in nature.
Effective simulation tools are needed for accurate design and verification of
microwave and millimeter-wave circuits to minimize the conventional trial-anderror process. Existing commercial circuit simulator might fail to provide important
information for designers, especially when electromagnetic interference issues
within active circuit environments are concerned. The FDTD method has shown its
flexibility and versatility in analyzing various types o f complex microwave and
millimeter-wave structures, and has evolved to be an excellent candidate for filling
in the insufficiency of commercial CAD tools, especially for evaluation of circuit
performance in hostile wireless communication environments.
The studies reported here mainly demonstrate the applications of the FDTD
method in overcoming the above-mentioned problems of current commercial
software. The parasitic effects such as housing and crosstalk effects when the circuit
is densely integrated are characterized successfully with the extended FDTD
approach. The electromagnetic interference study also provides circuit designer a
useful tool to identify the source of problems if the circuit performance disagrees
with expectation. This technique should prove useful in the analysis and design of
microwave circuits when various electromagnetic effects become crucial, such as in
85
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
active antenna circuits where any part of the circuit is exposed to radiation and can
affect the performance of the entire RF front end.
86
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
IMAGE EVALUATION
TEST TARGET (Q A -3 )
KM
4ia
&
150mm
IM/IC3E. In c
1653 E ast Main S treet
R ochester. NY 14609 USA
Phone: 716/482-0300
Fax: 716/288-5989
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A. '
< > 4
V.
Документ
Категория
Без категории
Просмотров
0
Размер файла
3 577 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа