close

Вход

Забыли?

вход по аккаунту

?

A broadband planar quasi -Yagi antenna: Characteristics and applications for waveguide transitions and microwave photonics

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 face, 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.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6" x 9" black and vtfiite
photographic prints are available for any photographs or illustrations appearing
in this copy for an additional charge. Contact UMI directly to order.
Bell & Howell Information and Learning
300 North Zeeb Road, Ann Arbor, Ml 48106-1346 USA
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.
UNIVERSITY OF CALIFORNIA
Los Angeles
A Broadband Planar Quasi-Yagi Antenna: Characteristics and Applications for
Waveguide Transitions and Microwave Photonics
A dissertation submitted in partial satisfaction o f the
requirements for the degree of Doctor of Philosophy
in Electrical Engineering
by
Noriaki Kaneda
2000
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number 9973165
___
__<g>
UMI
UMI Microform9973165
Copyright 2000 by Bell & Howell Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
Bell & Howell Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor. Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The dissertation o f Noriaki Kaneda is approved.
Reiner Stenzel
Ming. C. Wu
Tatsuo Itoh, Committee Chair
University o f California, Los Angeles,
2000
11
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CONTENTS
List o f Figures
Acknowledgments
Vita, Publications and Presentations
Abstract o f Dissertation
Chapter
1.
Introduction..................................................................................................................1
References 1 ............................................................................................................... 10
2.
A Broadband Quasi-Yagi antenna ........................................................................ 14
2.1
Quasi-Yagi Antenna Concept......................................................................... 14
2.2
Broadband Quasi-Yagi Antenna Design and Characteristics......................17
2.2.1
Antenna Configuration.......................................................................17
2.2.2
Antenna Simulation and Optimization............................................. 18
2.2.3
Antenna Return Loss..........................................................................22
2.2.4
Antenna Radiation Patterns...............................................................25
2.2.5
Antenna G ain..................................................................................... 28
2.2.6
Antenna Efficiency............................................................................32
2.3
Mutual Coupling Characteristics................................................................... 36
2.4
Gain-Enhanced Design................................................................................... 40
References 2 ................................................................................................................. 45
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.
A Broadband Microstrip-to-Waveguide Transition Using Quasi-Yagi
Antenna......................................................................................................................49
3.1
3.2
Transition Design........................................................................................... 49
3.1.1
Antenna (Radiator) Design in Waveguide.....................................49
3.1.2
Microstrip-to-CPS Baiun Design....................................................54
Simulation and Measurement Results..........................................................55
3.2.1
Simulation Results........................................................................... 55
3.2.2
Measurement Results.......................................................................58
3.3
Tolerance and Packaging Issues.................................................................. 61
3.4
W-band microstrip-to-waveguide transition using quasi-Yagi antenna... 65
References 3 .............................................................................................................. 69
4.
A Broadband CPW-to-Waveguide Transition Using Quasi-Yagi Antenna.. 71
4.1
4.2
4.3
Transition Design...........................................................................................71
4.1.1
CPW-to-CPS Baiun Consideration and Antenna D esign.............73
4.1.2
Single mode operation o f CPW lin e.............................................. 75
Simulation and Measurement Results..........................................................79
4.2.1
X-band prototype o f back-to-back transition.................................79
4.2.2
Single CPW-to-waveguide transition............................................ 82
4.2.3
Low insertion loss back-to-back transition design........................ 83
CPW-fed Quasi-Yagi A ntenna.................................................................... 84
References 4 ...............................................................................................................88
5.
Quasi-Yagi Antenna Applications in Microwave Photonics System..............92
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5.1
Millimeter-wave Generation by Photonic Devices......................................92
5.1.1 Background and M otivation..............................................................92
5.1.2 VMDP Measurement R esults............................................................94
5.2
Quasi-Yagi Waveguide Transition Combined with VMDP for Application
in Microwave Photonics................................................................................ 97
5.2.1 Transition Design and Scalemodel Demonstration..........................99
References 5 ............................................................................................................... 103
6.
Conclusions............................................................................................................... 105
Appendix A ............................................................................................................... 107
References A .............................................................................................................. 116
v
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List o f Figures
Fig. 1.1. A schematic o f the quasi-Yagi antenna.................................................................... 3
Fig. 1.2. Proposed Microstrip-to-waveguide transition using quasi-Yagi antenna............. 5
Fig. 1.3. CPW-to-Waveguide transition using quasi-Yagi antenna, (back-to-back).......... 6
Fig. 1.4. Velocity-matched distributed photodetectors (CPS type).......................................8
Fig. 1.5. Velocity-matched distributed photodetectors (CPW type).....................................8
Fig. 1.6. A schematic view of the VMDP-to-waveguide transition using quasi-Yagi
antenna......................................................................................................................................... 9
Fig. 2.1. Normalized power o f fundamental substrate modes for Duroid sr = 10.2 substrate
with 0.635 mm thickness.......................................................................................................... 16
Fig. 2.2. Picture o f a X-band quasi-Yagi antenna.................................................................. 17
Fig. 2.3. Quasi-Yagi antenna on the FDTD computation cells.............................................20
Fig. 2.4. Amplitude o f the electric field normal to the metal circuit patterns.....................21
Fig. 2.5. FDTD simulation and measured input return loss characteristic o f the prototype
quasi-Yagi antenna....................................................................................................................22
Fig. 2.6. FDTD simulation results o f the real and imaginary part o f the input impedance
o f the antenna.............................................................................................................................23
Fig. 2.7. FDTD simulation results o f the E and H plane’s co- and cross-polarization
radiation patterns o f the antenna at 9.5GHz........................................................................... 24
Fig. 2.8. (a) Measured E and H plane’s co- and cross-polarization radiation patterns o f the
antenna at 7.5 GHz....................................................................................................................25
vi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 2.8. (b) Measured E and H plane’s co- and cross-polarization radiation patterns o f
the antenna at 9.5 GHz............................................................................................................ 26
Fig. 2.8. (c) Measured E and H plane’s co- and cross-polarization radiation patterns o f the
antenna at 11.3 GHz................................................................................................................ 27
Fig. 2.9. Simulation and measured results o f the antenna’s absolute gain........................ 28
Fig. 2.10. Measured efficiency o f the antenna.....................................................................34
Fig. 2.11. Schematic views o f the two-element quasi-Yagi array, (a) coplanar and (b)
stacked array.............................................................................................................................36
Fig. 2.12. Measured results of the mutual coupling o f the two-element quasi-Yagi arrays.
................................................................................................................................................... 37
Fig. 2.13. Picture o f the gain-enhanced quasi-Yagi antenna............................................. 39
Fig. 2.14. Measured input return loss o f the gain-enhanced quasi-Yagi antenna............. 40
Fig. 2.15. Measured gain of the enhanced-gain quasi-Yagi antenna.................................. 41
Fig. 2.16. Measured radiation pattern o f the gain-enhanced quasi-Yagi antenna at (a) the
center frequency (9.55GHz), (b) the lower end frequency (9.09GHz) o f SI 1 < -lOdB
bandwidth................................................................................................................................. 42
Fig. 2.16. (c) Measured radiation pattern o f the gain-enhanced quasi-Yagi antenna at the
high end frequency (10.04GHz) o f SI 1 < -lO dB ................................................................. 43
Fig. 3.1. Individual components o f Quasi-Yagi antenna, (a) CPS Yagi-like antenna (b)
microstrip-to-CPS balun..........................................................................................................49
Fig. 3.2. Relative current amplitudes o f the three elements quasi-Yagi array. The current
amplitudes are observed at the edge closer to the reflector on driver and director element.
vii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(a) Observation points (b) Relative current amplitudes at observation points shown in (a).
...................................................................................................................................................50
Fig. 3.3. Insertion and return loss of the coupled microstrip line to waveguide transition
shown in Fig. 3.2 (a). Input o f the antenna is odd mode o f the coupled microstrip line and
the output is rectangular waveguide dominant mode...........................................................51
Fig. 3.4. Insertion and return loss of the microstrip-to-CPS balun shown in Fig. 3.2 (b).
The Input o f the balun is microstrip line and the output is coplanar strip line. Both even
and odd mode insertion loss o f CPS is simulated.................................................................53
Fig. 3.5. Simulated results o f insertion and return loss o f microstrip-to-waveguide
ransition on Duroid substrate with and without dielectric loss..........................................56
Fig. 3.6. Simulated results o f insertion and return loss o f microstrip-to-waveguide
transition on alumina substrate with and without dielectric loss....................................... 56
Fig. 3.7 Measurement setup o f the microstrip-to-waveguide transition......................... 57
Fig. 3.8. Measured insertion and return loss o f the quasi-Yagi microstrip-to-waveguide
transition using Duroid substrate..........................................................................................59
Fig. 3.9. Measured insertion and return loss o f the quasi-Yagi microstrip-to-waveguide
transition using alumina substrate........................................................................................ 59
Fig. 3.10. Quasi-Yagi microstrip-to-waveguide transition and three parameters
investigated for packaging tolerance issues......................................................................... 60
Fig. 3.11. Simulation results o f the transition with the inserted alumina substrate length o f
L=10mm, 15mm and 20mm..................................................................................................61
viii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 3.12. Simulation results o f the transition using the alumina substrate positioned at
height H=9.8mm and 11.8mm................................................................................................. 62
Fig. 3.13. Simulation results o f the transition with the width o f alumina substrate
W=10.13, 10.11 and 10.06mm................................................................................................ 63
Fig. 3.14. W-band microstrip-to-waveguide transition schematic view............................ 64
Fig. 3.15. Simulation results of W-band back-to-back transition. Both metal (gold) and
dielectric substrate (alumina) loss are not taken into account. The conductor is assumed to
have zero thickness................................................................................................................... 65
Fig. 3.16 Simulation results of W-band back-to-back transition. Both metal (gold) and
dielectric substrate (alumina) loss is taken into account. The conductor is assumed to
have zero thickness................................................................................................................... 66
Fig. 4.1. Schematic view o f the single CPW-to-waveguide transition using quasi-Yagi
antenna....................................................................................................................................... 70
Fig. 4.2. CPW-to-waveguide transition schematic (top view)............................................ 71
Fig. 4.3. CPW-to-waveguide transition schematic front view............................................ 72
Fig. 4.4. Schematic view o f CPW-to-CPS balun..................................................................73
Fig. 4.5. Sonnet simulation results o f CPW-to-CPS balun. CPW dimension [mm]: width
= 0.3, gap =0.15, open stub length 1.5. Air-bridges dimension [mm]: width = 0.15,
length =0.9, height=0.1. Substrate: thickness=0.635mm, er=10.2...................................... 74
Fig. 4.6. Proposed structure for the single mode operation o f CPW (a) schematic view (b)
propagation constant o f CPW mode. CPW dimension [mm]: width=1.44, gap=0.48,
thickness=0.635, sr=10.2. Trench dimension [mm]: width=2.76, depth=0.08..................75
ix
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 4.7. Propagation constant o f CPW. CPW dimension [mm]: w=1.44, g=0.48.
Substrate: t=0.635, er=10.2...................................................................................................... 76
Fig. 4.8. Propagation constants o f CBCPW in X-band waveguide. CBCPW dimension
[mm]: w=1.44, g=0.48. Substrate: t=0.635, sr=10.2............................................................77
Fig. 4.9. Simulation results o f the X-band transition excluding conductor and dielectric
loss..............................................................................................................................................78
Fig. 4.10. Measured results o f the X-band back-to-back transition. The waveguide and
adapter loss are calibrated out..................................................................................................79
Fig. 4.11. Measurement setup o f the single CPW-to-waveguide transition........................80
Fig. 4.12. Measured results o f the single CPW-to-waveguide transition............................ 81
Fig. 4.13. Photograph o f the CPW circuits with tapered ground planes for back-to-back
CPW-to-waveguide transition..................................................................................................82
Fig. 4.14. Measured results o f the back-to-back CPW-to-waveguide transition with
tapered CPW ground planes.....................................................................................................83
Fig. 4.15. Measured insertion loss o f the CPW-fed quasi-Yagi antenna.............................84
Fig. 4.16. E- and H-plane radiation patterns o f both co- and cross-polarization o f the
CPW-fed quasi-Yagi antenna.................................................................................................. 85
Fig. 5.1. A schematic o f the W-band VMDP measurement setup...................................... 94
Fig. 5.2. Output power o f W-band VMDP. The optical input power = 13.25dBm for both
solid and dashed line. The response includes both VMDP and W-band probe..................96
Fig. 5.3. Schematic o f a receiver for millimeter-wave radio astronomy and photonic local
oscillator.................................................................................................................................... 98
x
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 5.4. Schematic view of back-to-back CBCPS-to-waveguide transition......................99
Fig. 5.5. Measured results of the back-to-back CBCPS-to-waveguide transition............ 100
Fig. 5.6. Simulation results o f the W-band back-to-back CBCPS-to-waveguide transition
using InP substrate................................................................................................................. 102
Fig. A.l Sketch o f a slab waveguide.....................................................................................105
Fig. A.2. Effective guide thickness o f substrate modes. The phase shift at the interface is
caused by Goos-Hanchen shift.............................................................................................. 108
Fig. A.3. TE mode Reciprocity calculation for a dipole on an ungrounded substrate. .. 112
xi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Acknowledgments
I would like to express my sincere appreciation to my adviser, Professor Tatsuo
Itoh, for his guidance and support during my study at UCLA. I would also like to express
my appreciation to the members of my Doctoral Committee, Professor Ming C. Wu,
Professor Reiner Stenzel and Professor Harold R. Fetterman, for their kind help and
valuable discussions and comments. I would like to thank Professor Ming C. Wu and his
students Dr. Tai Chau, Mr. Saiful Islam and Mr. Sanjeeve S. Murthy for the valuable
discussions and excellent teamwork in the collaboration on microwave photonics
projects. I would also like to thank Dr. Anthony Kerr and Mr. Bill Shillue from National
Radio Astronomy Observatory for the strong support and valuable comments on the Wband photomixing project. Furthermore, I would also like to thank Dr. Yongxi Qian at the
UCLA Microwave lab for his guidance and valuable advice towards practical aspects of
my research and Dr. Bill Deal and Mr. James Sor for the discussions and collaborations
on several research topics and publications. Finally, I would like to acknowledge my
sincere pleasure and gratitude to all the members o f the UCLA Microwave lab and High
Frequency Center for the valuable discussions and kind assistance.
xii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
VITA
January 8, 1972
Bom, Ibaraki, Osaka, Japan.
1995
Electronics
Doshisha University
Kyotanabe, Kyoto, Japan
1995-1996
Exchange Abroad Program
Electrical Engineering Department
University o f California, Los Angeles
1997
Electrical Engineering
Doshisha University
Kyotanabe, Kyoto, Japan
1997-2000
Research Assistant
University o f California, Los Angeles
PUBLICATIONS AND PRESENTATIONS
1. N. Kaneda, W.R. Deal, Y. Qian and T. Itoh, “A broadband planar quasi-Yagi antenna”,
submitted for the publication in IEEE Transactions on Antennas and Propagation.
2. T. Chau, N. Kaneda, T. Jung, A. Rollinger, S. Mathai, Y. Qian, T. Itoh, M. C. W u, W.
P. Shillue and J. M. Payne, “Generation o f millimeter-waves by photomixing at 1. 55 |om
using InGaAs/InALAs/InP Velocity-Matched Distributed Photodetectors,” to be published
in IEEE Photonics Technology Letters, Aug. 2000.
3. W. R. Deal, N. Kaneda, J. Sor, Y. Qian and T. Itoh, “A New Quasi-Yagi Antenna for
Planar Active Antenna Arrays,” accepted for publication in IEEE Transactions on
Microwave Theory and Techniques, 1999
4. N. Kaneda, Y. Qian and T. Itoh, “A broadband microstrip-to-waveguide transition
using quasi-Yagi antenna”, IEEE Transactions on Microwave Theory and Techniques,
vol. 47, no. 9, pp. 2562-2567, Dec. 1999.
xiii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5. Y. Qian, W.R. Deal, N. Kaneda and T. Itoh, “ A microstrip-fed quasi-Yagi antenna
with broadband characteristics”, Electronics Letters, vol. 34, no. 23, pp. 2194-2196, Nov.
1998.
6. N. Kaneda, B. Houshmand and T. Itoh, “FDTD analysis o f dielectric resonators with
curved surfaces", IEEE Transactions on Microwave Theory and Techniques, vol. 45, no.
9, pp. 1645-1649, Sep. 1997.
7. N. Kaneda, Y. Qian and T. Itoh, “A broadband CPW-to-waveguide transition using
quasi-Yagi antenna”, to be presented in 2000 IEEE MTT-S International Microwave
Symposium, Boston, MA, June. 2000.
8. T. Chau, N. Kaneda, T. Jung, A. Rollinger, S. Mathai, Y. Qian, T. Itoh, M. C. Wu, B.
Shillue, J. Payne and D. Emerson, “High speed Velocity-Matched Distributed
Photodetectors”, 1999 IEEE LEOS Annual Meeting Conference Proceedings, San
Francisco, CA, pp. 834-835, Nov. 1999.
9. Y. Qian, W.R. Deal, N. Kaneda and T. Itoh, “ A uniplanar quasi-Yagi antenna with
wide bandwidth and low mutual coupling characteristics”, 1999 IEEE AP-S International
Symposium Digest, Orlando, FL, Vol. 2, pp. 924-927, July. 1999.
10. N. Kaneda, Y. Qian and T. Itoh, “A broadband microstrip-to-waveguide transition
using quasi-Yagi antenna”, 1999 IEEE MTT-S International Microwave Symposium
Digest, Anaheim, CA, Vol. 4, pp. 1431-1434, June. 1999.
11. N. Kaneda, Y. Qian and T. Itoh, “A novel Yagi-Uda dipole array fed by a microstripto-CPS transition”, 1998 Asia-Pacific Microwave Conference Proceedings, Yokohama,
Japan, pp. 1413-1416, Dec. 1998.
12. N. Kaneda, M. Tsuji and H. Shigesawa, “Near-field behavior and its related
characteristics o f coupled DR antennas: a new time-domain approach", XXVIII Moscow
International Conference on Antenna Theory and Technology, Sep. 1998.
13. N. Kaneda, M. Tsuji and H. Shigesawa, “FDTD analysis o f the complex frequency of
dielectric disk resonators”, 1997 IEICE general conference proceedings, March 1997.
xiv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ABSTRACT OF THE DISSERTATION
A Broadband Planar Quasi-Yagi Antenna: Characteristics and Applications for
Waveguide Transitions and Microwave Photonics
by
Noriald Kaneda
Doctor of Philosophy in Electrical Engineering
University o f California, Los Angeles, 2000
Professor Tatsuo Itoh, Chair
In this dissertation, a novel broadband planar antenna based on coplanar-strips
Yagi-Uda like dipole array is presented. This compact and single-layered “quasi-Yagi”
antenna fabricated on high dielectric constant substrate has an endfire radiation pattern.
The antenna achieves a measured 48 % frequency bandwidth for VSWR < 2, better than
12 dB front-to-back ratio, less than -15 dB cross-polarization, 3-5 dB absolute gain and a
nominal efficiency o f 93 % across the operating bandwidth. FDTD simulation is used for
optimization o f the antenna and the results agree very well with measurement. These
quasi-Yagi antennas are realized on a high dielectric constant substrate, and are
completely compatible with microstrip circuitry and solid-state devices. Furthermore,
utilizing the quasi-Yagi antenna’s broadband characteristics and good field alignment
xv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
with rectangular waveguide dominant mode, novel structures for microstrip line-towaveguide and coplanar waveguide (CPW) -to-waveguide transition are presented. The
X-band transition fabricated on an alumina substrate demonstrates broad bandwidth
(35%) with return loss better than -12dB and -0.3dB insertion loss at the center
frequency. This transition should find wide applications due to its high compatibility with
MIC/MMIC technology and very low fabrication cost. Finally, the consideration o f
microwave photonics system is presented along with the possible application o f the
quasi-Yagi antenna as a radiating element in a broadband microwave photonics system.
xvi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 1
Introduction
At microwave and millimeter wavelengths, planar antennas have a number of
advantages over waveguide antennas or wire antennas. These antennas are compact,
lightweight, low-cost and highly compatible with microwave/millimeter-wave circuits
and antenna array systems. In particular, antenna designers have for many years been
pursuing a broadband planar antenna on high dielectric constant substrate for a number of
reasons, including extremely compact size and high compatibility with monolithic
integration.
Existing planar antennas have a variety o f disadvantages when fabricated on high
permittivity substrates.
Microstrip patch antennas are typically narrowband and have
relatively low-efficiency due to the undesired substrate mode excited if high dielectric
constant substrates are used [1.1]. Although broadband performance can be achieved by
using more complicated schemes, such as the aperture coupled patch antenna [1.2], these
multi-layer designs add to the complexity in the antenna configuration and manufacturing
difficulties. The efficiency of the patch antenna can be also improved by chemical
etching o f the substrate at the expense o f the additional fabrication cost [1.3]. Slot
antennas with either microstrip or coplanar waveguide (CPW) feeding offer wider
bandwidth, but require additional design considerations and structural complexity such as
using cavities or reflectors to overcome the problem o f bi-directional radiation [1.4].
Alternatively, end-fire antennas such as the Vivaldi and linearly tapered slot antenna
1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(LTSA) are traveling-wave type structures that can achieve broad instantaneous
bandwidth [1.5]. However, they usually have larger electrical size than resonant-type
patches or slots and often suffer from the excitation o f the substrate mode which can
result in reduced efficiency, strong cross-talk between antennas in an array environment
and perturbed radiation patterns. Additionally, tapered slot antennas also require either
microstrip-to-slot or CPW-to-slot transitions as part o f the feeding network, which not
only increases the design complexity but also imposes a limit on the frequency response.
The Yagi-Uda antenna, first published in an English language journal in 1928
[1.6], has been used extensively as an end-fire antenna. However, only limited success
has been achieved at adapting this antenna to microwave/millimeter wave operation.
Several interesting approaches for this are a microstrip yagi array based on the microstrip
patch antenna [1.7], a monopole Yagi-Uda arranged into a multi-sector array covering the
entire horizontal plane at 19.5 GHz [1.8] and a coplanar-stripline fed printed Yagi-Uda
antenna with the reflector element printed on the back o f a thick, low permittivity slab at
60 GHz [1.9].
We have proposed and demonstrated a novel uniplanar quasi-Yagi antenna (Fig.
1) that has both the compactness o f resonant-type antennas and broadband characteristics
o f traveling-wave radiators [1.10]-[1.12]. The Yagi-Uda dipole array type o f antenna is
realized on a high dielectric constant substrate with a microstrip feed. Unlike the
traditional Yagi dipole design, we employ the truncated microstrip ground plane as the
reflecting element, thus eliminating the need for a reflector dipole [1.10]. This results in a
very compact design (<
X q/ 2
by Xo/2), which is totally compatible with any microstrip-
2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
based MMIC circuitry. Following the first experimental demonstration in [1.10], an Xband prototype with 17 % bandwidth, 6.5 dB gain, 18 dB front-to-back ratio and —15 dB
cross polarization level has been designed and tested successfully [1.11].
Microstrip Feed
Fig. 1.1. A schematic o f the quasi-Yagi antenna.
3
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The design and performances o f broadband quasi-Yagi antenna is presented in
detail in [1.13]. We have achieved extremely broad bandwidth
(measured 48% for
VSWR < 2), good radiation profile (front-to-back ratio >12 dB, cross-polarization < -12
dB), acceptable absolute gain (3-5 dB) and high efficiency (93%). Furthermore, the
mutual coupling between neighboring elements o f a quasi-Yagi antenna array is found to
be very low, with a measured level of < -1 8 dB for either stacked or side-by-side 2element structure with Xo/2 separation. Such a compact array with low mutual coupling
characteristics should find wide applications in modem communications and radar
systems as well as millimeter-wave imaging arrays. Since the antenna is very compact
and has a radiation field in good alignment with rectangular waveguide dominant mode,
the antenna can also be used in a part of the waveguide transitions.
Due to their low-loss characteristics, metallic waveguides are still essential
components in many microwave and millimeter-wave application systems. Since most
modem solid-state and photonics devices are based on planar fabrication technology,
waveguide transitions from microstrip, coplanar waveguide (CPW) and coplanar strips
(CPS) are critical for efficient integration o f waveguide with planar circuits. A number o f
techniques and schemes have been studied on microstrip-to-waveguide transition in the
past. One o f the broadest-band and Iowest-loss transitions is achieved by the ridgedwaveguide approach [1.14], at the expense o f extensive machining and special attention
to the DC-block problem. Using substrates with low dielectric constant, antipodal finline
[1.15],
[1.16]
and probe-type
[1.17],
[1.18]
transitions give quite
broadband
characteristics, but such a low dielectric constant substrate is difficult to integrate with
4
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
solid-state devices on high dielectric constant substrates. Meanwhile many recent works
have aimed at MMIC-compatible transitions where high dielectric constant substrates
along with low-cost fabrication techniques o f the transition are required [1.19]-[1.21].
While these slot-coupling type o f transitions demonstrated bandwidth of 10-20% for
Sn<-15dB, most o f these techniques demand a high degree o f mechanical complexity,
such as using multilevel substrates [1.19], [1.20] and backshort hermetic sealing o f the
waveguide [1.21].
Microstrip feed
Waveguide
Metal block
Fig. 1.2. Proposed Microstrip-to-waveguide transition using quasi-Yagi antenna.
5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
To address these issues, we recently developed a broadband MMIC compatible
microstrip-to-waveguide transition (Fig. 1.2) [1.22] [1.23] and CPW-to-waveguide
transition (Fig. 1.3) [1.24] using the quasi-Yagi antenna. A bandwidth o f 35% with return
loss better than -12dB has been achieved and the insertion loss is about -0.3dB at the
center frequency o f the X-band prototype of the microstrip-to-waveguide transition. The
transition is realized on a single layer o f high dielectric constant substrate (alumina with
er=9.9), which offers good compatibility with future GaAs or InP-based MMIC
implementation at millimeter-wave frequencies. The transitions significantly reduce the
amount o f mechanical complexity since it requires neither waveguide backshort nor
multilayer substrate configuration. This self-packaged transition structures should find
wide applications in microwave and millimeter-wave systems due to its low cost and
MMIC-compatible advantages.
Fig. 1.3. CPW-to-Waveguide transition using quasi-Yagi antenna, (back-to-back)
6
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Since most o f the semiconductor photonics devices are based on the high
permittivity substrate, the quasi-Yagi antenna and transitions are also applicable to the
microwave photonics system. Recent development o f Velocity-Matched Distributed
Photodetectors (VMDP) at UCLA [1.25] - [1.27] provides ideal solution for the high
frequency power generation and detection in a number o f applications in millimeter-wave
and submillimeter-wave region. VMDP consists o f array o f high-speed metalsemiconductor-metal (MSM) diodes fed by a passive optical waveguide (Fig. 1.4, 1.5).
The input optical signals are either two laser beating signals that differ in frequency by
the desired microwave output signal frequency or a single laser source that is optically
modulated at the output frequency. The MSM diodes are used to extract microwave
power by mixing or demodulating optical signals. Diodes are periodically loaded on the
passive microwave planar transmission line such as coplanar stripline (CPS) (Fig. 1.4)
and coplanar waveguide (CPW) (Fig. 1.5) to achieve 50 Q output impedance and at the
same time to slow down the microwave velocity to match with the optical wave velocity.
Due to these features, VMDP has achieved ultra-broadband and relatively high-power
characteristics.
Fig. 1.6 shows a schematic view o f the VMDP integrated waveguide transition
that is under current development for a W-band system. This new system can potentially
provide a stable millimeter-wave source for very broad bandwidth with lower cost
compared to existing systems based on microwave tubes or frequency multipliers.
7
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Passive Optical
Waveguide
Coplanar Strip
(CPS) Microwave
Transmission .
Line
/
NanoecaleMSM
Photodiodes for
Ultrafast Frequency
Response
Fig. 1.4. Velocity-matched distributed photodetectors (CPS type)
Optical Waveguide
Microwave
Output .
Optical
Input
Fig. 1.5. Velocity-matched distributed photodetectors (CPW type)
8
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 1.6. A schematic view o f the VMDP-to-waveguide transition using quasi-Yagi
antenna
9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
REFERENCES 1
[1.1] D. M. Pozar, “Considerations for millimeter wave printed antennas,” IEEE Trans.
Antennas Propagat., vol. 31, pp. 740-747, Sep. 1983.
[1.2] C. H. Tsao, Y. M. Hwang, F. Kilburg and F. Dietrich, “Aperture-coupled patch
antennas with wide-bandwidth and dual-polarization capabilities”, IEEE AP-S In t’l
Symposium Digest, New York, NY, pp. 936-939, June 1988.
[1.3] I. Papapolymerou, R. F. Drayton and L. P. B. Katehi, “Micromachined patch
antennas,” IEEE Trans. Antennas Propagat., vol. 46, pp. 275-283, Feb. 1998.
[1.4] Y. Yoshimura, “A microstripline slot antenna,” IEEE Trans. Microwave Theory
Tech., vol. 20, pp. 760-762, Nov. 1972.
[1.5] K. S. Yngvesson, T. L. Korzeniowski, Y. Kim, E. L. Kollberg, and J. F. Johansson,
“The tapered slot antenna - a new integrated element for millimeter-wave applications,”
IEEE Trans. Microwave Theory Tech., vol. 37, pp. 365-374, Feb. 1989.
[1.6] H. Yagi, “ Beam transmission o f the ultra short waves,” Proceedings o f IRE, vol.
16, pp. 715-741, June. 1928.
[1.7] J. Huang and A.C. Densmore, “Microstrip Yagi array antenna for mobile satellite
vehicle application”, IEEE Trans. Antennas Propagat., vol. 39, no.7, pp. 1024-1030, July
1991.
[1.8] T. Maruyama, K. Uehara and K. Kagoshima, “Analysis and design o f multi-sector
monopole Yagi-Uda array mounted on a ground plane using moment method”,
10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Proceedings o f 3rd Int 7 C onf on Computation in Electromagnetics, Bath, UK, pp.289294, April 1996.
[ 1.9] K. Uehara, K Miyashita, K.I. Matsume, K.H. Hatakeyama and K. Mizuno, “Lenscoupled imaging arrays for the millimeter and submillimeter-wave regions”, IEEE Trans.
Microwave Theory Tech., vol. 40, pp. 806-811, May 1992.
[1.10] N. Kaneda, Y. Qian and T. Itoh, “A novel Yagi-Uda dipole array fed by a
microstrip-to-CPS transition,” 1998 Asia Pacific Microwave Conf. Dig., Yokohama,
Japan, pp. 1413-1416, Dec. 1998.
[1.11] Y. Qian, W. R. Deal, N. Kaneda and T. Itoh, “A microstrip-fed quasi-Yagi antenna
with broadband characteristics,” Electronics Lett., vol. 34, no. 23, pp. 2194-2196, Nov.
1998.
[1.12] Y. Qian, W.R. Deal, N. Kaneda and T. Itoh, “ A uniplanar quasi-Yagi antenna with
wide bandwidth and low mutual coupling characteristics”, 1999 IEEE AP-S International
Symposium Digest, Orlando, FL, Vol. 2, pp. 924-927, July 1999.
[1.13] N. Kaneda, W. R. Deal, Y. Qian and T. Itoh, “A broadband planar quasi-Yagi
antenna”, submitted to IEEE Trans. Antennas Propagat..
[1.14] S. S. Moochalla and C. An, “Ridge waveguide used in microstrip transition,”
Microwaves and RF, pp. 149-153, March 1984.
[1.15] L. J. Lavedan, “Design o f waveguide-to-microstrip transition specially suited to
millimeter-wave applications”, Electronics Letters, vol. 13, no. 20, pp. 604-605, Sept.
1977.
11
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[1.16] G. E. Ponchack and A. N. Downey, “A new model for broadband waveguide-tomicrostrip transition design”, Microwave Journal, pp. 333-341, May 1988.
[1.17] Y. C. Shih, T. N. Ton and L.Q. Bui, “Waveguide-to-microstrip transitions for
millimeter-wave applications”, IEEE MTT-S Int 7 Symposium Digest, NewYork, NY, vol.
l,pp. 473-475, 1988.
[1.18] T. Q. Ho and Y. C. Shih, “Spectral-domain analysis o f E-plane waveguide to
microstrip transitions,” IEEE Trans. Microwave Theory and Techniques, vol. 37, pp. 388392, Feb. 1989.
[1.19] W. Grapher, B. Hudler and W. Menzel, “Microstrip to waveguide transition
compatible with MM-wave integrated circuits”, IEEE Trans. Microwave Theory and
Techniques, vol. 42, pp. 1842-1843, Sept. 1994.
[1.20] L. Hyvonen and A. Hujanen, “A compact MMIC-compatible microstrip to
waveguide transition”, IEEE MTT-S Int 7 Symposium Digest, San Francisco, CA, vol. 2,
pp. 875-878, 1996.
[1.21] F. J. Villegas, D. I. Stones and H. A. Hung, “A novel waveguide-to-microstrip
transition for millimeter-wave module applications”, IEEE Trans. Microwave Theory and
Techniques, vol. 47, pp. 48-55, Jan. 1999.
[ 1.22] N. Kaneda, Y. Qian and T. Itoh, “A broadband microstrip-to-waveguide transition
using quasi-Yagi antenna”, 1999 IEEE MTT-S In t’I Microwave Symposium Digest,
Anaheim, CA, vol. 4, pp. 1431-1434, June 1999.
12
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[1.23] N. Kaneda, Y. Qian and T. Itoh, “A broadband microstrip-to-waveguide transition
using quasi-Yagi antenna”, IEEE Trans. Microwave Theory and Techniques, vol. 47, pp.
2562-2567, Dec. 1999.
[1.24] N. Kaneda, Y. Qian and T. Itoh, “A broadband CPW-to-waveguide transition
using quasi-Yagi antenna”, submitted to 2000 IEEE M TT-S Int 7 Microwave Symposium.
[1.25] L. Y. Lin, M. C. Wu, T. Itoh, T. A. Vang, R. E. Muller, D. L. Sivco, A. Y. Cho,
“High-power
high-speed
photodetectors-design,
analysis,
and
experimental
demonstration ”, IEEE Transactions on Microwave Theory and Techniques, vol.45, pp.
1320-1331,.Aug. 1997.
[1.26] T. Chau, L. Fan, D. T. K. Tong, S. Mathai, M. C. Wu, D. L. Sivco, A. Y. Cho,
“Long wavelength velocity-matched distributed photodetectors for RF fibre optic links”,
Electronics Letters, vol.34, no.14, pp.1422-1424, July 1998.
[1.27] M. S. Islam, T. Chau, A. Nespola, S. Mathai, A. R. Rollinger, W. R. Deal, T. Itoh,
M. C. Wu, D. L. Sivco, A.Y. Cho, “Distributed balanced photodetectors for highperformance RF photonic links”, IEEE Photonics Technology Letters, vol.l 1, p.457-459,
April 1999.
13
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 2
A Broadband Planar Quasi-Yagi Antenna
2.1 Quasi-Yagi Antenna Concept
Planar antennas often suffer from the substrate surface-waves, such as TEo mode
in printed dipole antennas on ungrounded substrate and TM 0 mode in patch and tapered
slot antennas. Those substrate surface-wave modes result in low-efficiency and poor
mutual couplings o f the antennas and antenna arrays. These problems are more prominent
as the value o f the permittivity and the thickness o f the substrate increases [2.1]. The
quasi-Yagi antenna is a printed dipole array that utilizes both free space radiation and the
TEo surface-wave to achieve broadband operation with a good endfire radiation profile,
while many other dipole antennas have broadside radiation pattern and suffer from the
surface-waves.
The uniplanar quasi-Yagi antenna in Fig. 1.1 is constructed on a single piece o f
relatively high permittivity substrate (0.635 mm thick Duroid er = 10.2 for the X-band
prototype) with metalization on both sides. The top metalization consists o f a microstrip
feed, a broadband microstrip-to-CPS (Coplanar Stripline) balun and two dipole-elements,
one o f which is the driver element fed by CPS and the second dipole being the parasitic
director. The broadband microstrip-to-CPS transition was previously reported in [2.2].
The metalization on the bottom plane is a truncated microstrip ground, which serves as
14
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the reflector element for the antenna. The parasitic director element o n the top plane
simultaneously directs the antenna propagation towards the endfire direction, and acts as
an impedance matching parasitic element.
The broadband performance o f the quasi-Yagi antenna is partially due to the
effective excitation o f the TEo mode by the printed dipole array. To illustrate this, the
reciprocity theorem [2.3] has been applied to the computation o f the current elements and
surface-wave modes o f elementary printed dipoles in the dielectric substrate. Using the
technique described in [2.4] [2.5] and appendix A, the power density coupled into the
fundamental substrate-modes that are excited by an elementary dipole on an ungrounded
0.635mm thick Duroid with £r = 10.2 are plotted in Fig. 2.1 The plot is normalized to the
power density o f a dipole radiated into free-space and is in linear scale.
The figure
indicates that the quasi-Yagi antenna’s high permittivity substrate supports the TEo mode
excitation for quite a large frequency bandwidth, while the undesired TMo mode does not
couple efficiently to the substrate at the frequency range o f interest. In other words, for
the given substrate and operating frequency the elementary printed dipole generates TEo
surface-wave power at the same order o f the free space radiation power, w hile undesired
TMo surface-wave excitation is significantly suppressed.
One unique feature o f this antenna is the use o f the truncated ground plane as the
reflector element. The truncated ground plane acts as an ideal reflector for the TEo mode
that has cutoff frequency (39 GHz) much higher than the operating frequency o f the
antenna. The dipole elements o f the quasi-Yagi antenna are strongly coupled by the TEo
surface-wave that has same polarization and same direction with the dipole radiation
15
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
fields. The surface-wave coupling contributes to the capability o f the broadband input
impedance matching and good radiation pattern formation while keeping the antenna size
compact.
2.5
2
0.1.5
TEO
TMO
1
0.5
0
0
10
20
30
40
50
60
70
80
90
100
Frequency [GHz]
Fig. 2.1. Normalized power o f fundamental substrate modes for Duroid £r= 10.2
substrate with 0.635 mm thickness.
As with the classic Yagi-Uda antenna, proper design requires careful optimization
o f the driver, director and reflector parameters, which include element spacing, length
and width. By choosing the antenna parameters properly, the quasi-Yagi antenna
demonstrates broadband (40-50 % for VSWR < 2) characteristics with modest gains (~4
dB) or narrower bandwidth (10-20 % for VSWR < 2) with higher gains (~6.5 dB). The
current design features one director element. Incorporating additional elements has the
16
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
potential for increasing gain or bandwidth. However, this also increases the number o f
design parameters as well as the complexity o f design optimization, and has not been
investigated extensively.
2.2
Broadband Quasi-Yagi Antenna Design and Characteristics
2.2.1
Antenna Configuration
A photograph o f the X-band quasi-Yagi antenna is shown in Fig. 2.2 The
antenna’s dimensions are optimized by an in-house FDTD (Finite Difference Time
Fig. 2.2. Picture o f a X-band quasi-Yagi antenna.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Domain method) code to achieve broadband performance. The detailed issues in the
simulation of the antenna are presented in the following sub-section.
The antenna is realized on 0.635 mm thick Duroid with er = 10.2. The antenna’s
dimensions as defined in Fig. 1.1 are (unit: mm): Wj = W 3 = W 4 = W 5 = Wdn = Wdjr =
0.6, W2 = 1.2, W 6 = S 5 = S6 = 0.3, L, = 3.3, L2 = L5 = 1.5, L3 = 4.8, U = 1.8, Sref = 3.9,
Sdjr = 3.0, SSub = 1-5, Ldn = 8.7 and Ld,r = 3.3. The length o f the antenna’s director element
is shorter than the conventional Yagi-Uda antenna design where the director element’s
length typically remains 80 - 99.6% o f the driver element and the reflector element is
much longer than the on o f conventional Yagi-Uda arrays. This shorter director element
contributes to the broadband characteristics o f the antenna creating a resonance pole at
different frequency that driver element has. This bandwidth widening technique using
parasitic radiation element is also found in a broadband multi-layer patch antennas [2 .6 ]
[2.7]. The total area o f the substrate is approximately Xo/2 by ko/2 at the center frequency.
2.2.2
Antenna Simulation and Optimization
A number o f numerical simulation techniques are developed for the analysis of
electromagnetic
problems. The FEM
(Finite
Element Method)
is a powerful
electromagnetic numerical method and commercial software such as H P’s HFSS based
on this technique is capable o f analyzing antennas and waveguide structures. FEM is a
frequency domain method therefore an analysis over the wide range o f frequency band
often requires a long computation time. Also the modal field expansion with open-air
18
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
boundaries has its intrinsic problems and often generates wrong results. Mainly due to the
reasons stated above, FEM is not extensively used for quasi-Yagi antenna development.
The MoM (Method o f Moment) is another powerful numerical technique for
solving full-wave electromagnetic problems. In MoM, integral equations o f Green’s
function are constructed with unknown currents on conductor surfaces. Unknown
coefficients that define the unknown currents are solved with boundary conditions on
conductor surfaces. The method is also a frequency domain method and it often takes
long time to calculate results over a broad frequency range. Above all, the method has its
limitation since we have to find and derive the Green’s function for particular geometries.
To the author’s knowledge, the currently available commercial software based on MoM
only support two-dimensionally symmetric structures where the substrates are extended
until they are truncated at computational limit with PEC (perfect electric conductor).
Thus the MoM is not suitable for the particular simulation and design of quasi-Yagi
antenna where the truncated substrate is an essential part o f its concept.
The FDTD method [2.8] was extensively used for analysis and optimization o f the
quasi-Yagi antennas. The FDTD method discretizes the time-domain Maxwell’s
equations both in space domain and time domain. Thus the technique is a time-domain
electromagnetic analysis and suitable for the transient analysis o f microwave circuits. The
frequency domain characteristics are obtained by Fourier transforms o f the time-domain
computation results. Depending on the time-domain pulse width that is used in the
simulation, a single simulation generates the broadband frequency domain characteristics
19
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and it contributes to the much faster computation in comparison with frequency domain
methods like FEM and MoM.
When the antenna structures are simulated, the radiation boundary conditions o f
free space open boundary become a critical factor o f the simulation for successful
prediction of the antenna performances. The radiation boundary conditions are often
referred in FDTD as absorbing boundary conditions (ABC) and a number o f techniques
are demonstrated and available for programmers. Mur’s first order ABC [2.9] is the first
order approximation o f wave equations at radiation boundaries and it only absorbs the
wave normal to the boundary walls but the wave coming from other directions would be
reflected. This is improved in Mur’s second order ABC where the ABC suppresses the
reflection of the incoming waves with any incident angles. A perfectly matched layer
90
80
70
60
SO
40
ao
20
10
0
t
Fig. 2.3. Quasi-Yagi antenna on the FDTD computation cells.
20
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(PML) [2.10] significantly improves the level o f reflection by introducing the lossy
material layers at the radiation boundary. The drawback o f this technique is that it
requires a large computation memory and long computation time. It has been tested and
known by our experience that Mur’s second order ABC is sufficient for the antenna
simulations.
We have used the in-house FDTD code (available to public [2.11]) developed by
Dr. Yongxi Qian at UCLA. Fig. 2.3 shows the top plane o f the quasi-Yagi antenna in the
FDTD rectangular grids. The simulation typically uses a million cells or less and the
computation time for a simulation using million cells is approximately 45 minutes on PC
Fig. 2.4. Amplitude o f the electric field normal to the metal circuit patterns.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
with 500MHz CPU. Triangular cells are used for the bends o f the microstrip mitre. The
amplitude o f the electric field intensity normal to the conductors are shown in Fig. 2.4.
We can see two o f the antenna’s radiation elements, namely driver and director, are
excited with 180-degree phase difference.
2.2.3
Antenna Return Loss
The FDTD simulation and measured results of the return loss o f the antenna are
shown in Fig. 2.5. The simulated and measured bandwidths (VSWR < 2) are 43% and
48%, respectively. To the authors’ knowledge, no uniplanar antenna has ever achieved
0
•5
«-10
(0
fi -15
-20
CC. -25
Simulation
Measurement
-35
-40
6
7
8
9
10
11
12
13
14
Frequency (GHz)
Fig. 2.5. FDTD simulation and measured input return loss characteristic of
the prototype quasi-Yagi antenna
22
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
such a wide instantaneous bandwidth with such a compact design. The plot o f the
simulated input impedance, as shown in Fig. 2.6 further illustrates this. It can be seen that
both the real and imaginary components o f the input impedance are quite flat in the
operating band.
It is interesting to compare the bandwidth o f various planar antennas. A resonant
antenna such as patch antenna has an intrinsic narrow bandwidth, and the regular patch
antenna’s bandwidth is nominally 2-5%. This high-Q nature o f patch antenna can be used
for filtering undesired signals in an integrated system, but in general circumstances
200
150
Imaginary Part
Real Part
•50
-100
7
8
9
10
11
12
13
Frequency [GHz]
Fig. 2.6. FDTD simulation results o f the real and imaginary part of the input
impedance o f the antenna.
23
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
narrow bandwidth o f the antenna requires high precision in the design and fabrication and
may result in a costly system. The six-layered aperture-coupled patch antenna [2.6]
broadens the bandwidth o f the patch antenna to a noteworthy 70%. However, the
radiation pattern over the wide frequency range is not known and it sacrifices the
simplicity o f the original patch antenna structure.
The traveling wave antennas such as the tapered slot antenna and leaky wave
antenna can have significantly broad bandwidth. The bandwidth of the tapered slot
antenna such as the Vivaldi antenna can be extremely wide; it has been demonstrated that
the microstrip line matches with the Vivaldi antenna for astounding 8-40GHz [2.12]. This
90
135
45
225
315
180
270
Fig. 2.7. FDTD simulation results o f the E and H plane’s co- and crosspolarization radiation patterns o f the antenna at 9.5GHz.
24
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
90
135
45
225
315
E-CPOL
H-CPOL
E-XPOL
H-XPOL
270
(a)
Fig. 2.8. (a) Measured E and H plane’s co- and cross-polarization radiation
patterns o f the antenna at 7.5 GHz.
broad bandwidth and high gain o f the traveling wave antenna is very attractive in a
number o f applications but the electrically large size o f the antenna can be awkward in
some antenna array systems. Usage o f high dielectric constant substrate can contribute to
the reduction o f the tapered slot antenna size, however, in expense o f the surface-wave
excitation problem and detrimental effect on antenna efficiency.
2.2.4
Antenna Radiation Patterns
25
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 2.7 shows the simulated E- and H-plane radiation patterns o f the both co- and
cross- polarizations at 9.5 GHz. The figure indicates well-defined endfire radiation
patterns with front-to-back ratio o f -1 6 dB and the maximum cross-polarization level o f 18 dB. Further simulations reveal the radiation pattern is quite stable as the frequency
changes, with the front-to-back ratio >15 dB and cross polarization level < -12 dB across
the entire frequency band between
8
and 12 GHz. This is contrary to many other
90
H-CPOL
E-CPOL
H-XPOL
E-XPOL
180
0
270
Fig. 2.8. (b) Measured E and H plane’s co- and cross-polarization radiation patterns o f
the antenna at 9.5 GHz.
26
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
broadband planar antenna designs where increased bandwidth is usually realized at the
expense o f degradation in either backside radiation or cross polarization.
Measured radiation patterns of the broadband quasi-Yagi antenna are shown in
Figs 2.8 (a)-(c) The radiation patterns are measured at three different frequencies, 7.5
GHz, 9.5 GHz and 11.3 GHz. They approximately correspond to the lower end, center
and upper end frequencies o f the operating band o f the antenna. The front-back ratio is
better than 12dB and the cross polarization is better than -12dB for the three patterns.
H-CPOL
E-CPOL
H-XPOL
E-XPOL
90
-10
135
45
'-30
•35 ,
180
225
315
270
(C)
Fig. 2.8. (c) Measured E and H plane’s co- and cross-polarization radiation patterns
o f the antenna at 11.3 GHz.
27
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
These measured results are in close agreement with the simulated results shown in Fig.
2.7 at 9.5 GHz. As can be seen, the broadband quasi-Yagi antenna has a broad, single­
beam pattern in both the E and H-plane cuts.
We can also discuss the radiation patterns o f quasi-Yagi antenna in comparison
with other planar antennas and for use in an antenna array environment. In most o f the
applications, the unidirectional antenna radiation pattern is desired, for instance, to avoid
the interference between the radiation from antenna and feeding networks. Also, to insure
the symmetric beam-width between E- and H- plane o f the radiation patterns for twodimensional antenna arrays, the single antenna’s E- and H-plane’s radiation patterns
should reasonably symmetric, meaning the beam-width o f E- and H-plane’s radiation
patterns should be similar.
Due to its semi-infinite ground plane, the patch antenna can achieve front-to-back
ratio o f 20 dB or more. Also patch antenna’s E- and H-plane’s radiation patterns are
reasonably symmetric. The quasi-Yagi antenna has also good front-to-back ratio and its
E- and H-plane radiation patterns are reasonably symmetric, while the three-element
traditional Yagi antenna can only achieve 5.6dB front-to-back ratio [2.13]. The main
difference between traditional Yagi and the quasi-Yagi is its reflector element. It is
obvious from the radiation pattern that the truncated microstrip ground plane o f the quasiYagi antenna contributes largely to this good front-to-back ratio.
2.2.5
Antenna Gain
28
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The gain o f the antenna can be measured by several methods [2.14], The most
commonly used method is called one antenna gain measurement. This technique utilizes
a reference antenna calibrated by three antenna gain method and corresponding CAL
(calibration) table. The problem with this method is that the reference antenna must be
calibrated with the specific chamber that is used for the gain measurement. If the antenna
is calibrated in a different chamber, the CAL table does not calibrate out the error caused
by reflection from the chamber that is used for the measurement. Once the chamber’s
conditions are changed such as reconfiguration and replacement o f absorbers, the
reference antenna should also be recalibrated.
Measurement
Simulation
7.5
8
8.5
9
9.5
10
10.5
11
11.5
Frequency (GHz)
Fig. 2.9. Simulation and measured results o f the antenna’s absolute gain.
29
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The three-antenna gain method is very accurate method and there is no need for a
reference antenna with known gain. However, the problem with this method is that if the
antennas’ gains are very close each other’s, it is very sensitive to the error in the received
power measurement. To insure the successful measurement, the three antennas’ gain
should be approximately 2dB different from each other and within SdB from the
maximum to minimum gain.
The two-antenna gain method assumes that we have two identical antennas. This
assumption can be crude approximation for some type o f antennas but in most o f the
uniplanar antennas, duplication o f antennas is straightforward and the deviation among
fabrications is very small. This method can be less accurate compared to the threeantenna gain method, but the reliability and simplicity o f this measurement makes it very
attractive.
Since it is important to have the reliable measurement over the wide frequency
range in the broadband quasi-Yagi antenna measurement, the two-antenna gain method is
chosen. The gain o f the antenna has been determined both experimentally and through
simulation, as shown in Fig. 2.9. The solid line is the measured result and the dashed line
is the simulation result obtained by the HP’s HFSS. To measure the gain o f the antenna, a
pair o f identical quasi-Yagi antennas is used as both transmit and receive antennas in an
anechoic chamber. Once the input and output power o f the antenna are measured, the
gain is given by the Friis transmission formula [2.13]
30
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
where
PT : Time average transmitted power from the transmit antenna.
PK : Time average received power from the receiver antenna.
Gt : Gain o f the transmit antenna.
Gr : Gain o f the receiver antenna.
r : Distance between transmit antenna and receiver antenna.
A : Free-space wavelength.
The input return loss, and cable losses have been calibrated out o f the
measurement. However, no attempt was made to calibrate out the connector or microstrip
feed losses, which is estimated to be a fraction of dB. From the plot, the measured gain
varies from 3.4 dB to 5.1 dB while the simulated gain varies from 4.5 dB to 5.1 dB for
the entire pass-band. In general, fair agreement is observed. Additionally, the variation in
gain is relatively small across the wide operating bandwidth o f the antenna.
The gain o f the broadband quasi-Yagi antenna is not very high compared to the
other planar antennas. Patch antennas nominally have gain o f 6-7 dB, and the tapered slot
antenna with long tapering can achieve as much as 16dB [2.15]. The low gain
characteristic is observed as wide beam radiation pattern o f the antennas. One advantage
with wide beam-width is that the array’s radiation pattern would be readily predicted by
the antenna array factor. As in the traditional Yagi antennas, it is expected that additional
director elements would result in more gain of a single element o f quasi-Yagi antenna.
31
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Later in this chapter, the technique to achieve higher gain with the same number of
elements is described and the results are presented in detail.
2.2.6
Antenna Efficiency
Antenna efficiency, defined as the ratio o f radiated power to input power, is an
important antenna parameter but often one which is difficult to quantify in either
measurement or simulation. Printed antennas often have high efficiency, but surfacewave loss and feed network loss can reduce the efficiency. Three measurement methods
are commonly known, namely the Wheeler cap method [2.16] [2.17], the radiometric
method [2.18] and the directivity/gain method [2.19]. The excellent comparison o f these
three methods and the application to the patch antenna is found in [2 .2 0 ].
Wheeler originated the Wheeler cap method for measurement o f the efficiency of
electrically small antennas [2.16]. The experimental setup consists o f the test antenna
with ground plane that can be completely enclosed by the metallic hemisphere. The input
impedance of the test antenna is then measured at the resonant frequency o f the antenna
with and without the cap. In the original paper [2.16], the radius o f the cap hemisphere is
one radian wavelength (A. I I n ) but later it is reported that the radius o f hemisphere is not
critical and the shape o f the cap need not be spherical [2.17]. The simplicity made this
method one o f the most common methods for the antenna efficiency measurement.
However, this method is particularly good for resonant antennas with narrow bandwidth.
32
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The simple theoretical model o f this method is not applicable for quasi-Yagi antenna
where the multi-poles are expected in the antenna’s resonance mechanism.
The radiometric efficiency measurement is based on the measurement o f the noise
figure with a test antenna directed to cold and warm targets [2.18], The cold and warm
target practically is the clear sky and antenna chamber absorber and the relatively
complex experimental setup is required for both indoor and outdoor measurement. The
accuracy o f this method highly depends on the noise figures o f the R F amplifiers and
mixers o f the experimental setup. It is reported that the method obtains the accurate
results but the repeatability is not very high [2 .2 0 ].
The gain/directivity method utilizes the fact the antenna efficiency is obtained by
rj = G / D , where G is antenna gain and D is antenna directivity. In order to have
successful results by this method, the gain and directivity has to be measured very
accurately. However, the accurate measurement o f antennas gain is very difficult as
discussed in the previous section, and the directivity measurement often suffers from the
lack o f repeatability since the small errors in antenna radiation patterns can add up to a
large error in directivity calculation. Based on these reason, this method is neither
accurate nor repeatable.
The reflection method is recently proposed and it has demonstrated both accuracy
and repeatability in the measurement o f small antenna efficiency [2.21] [2.22]. This
technique is similar to Wheeler cap method except that the test antenna is inside the
sliding short waveguide. By adjusting the sliding short position, a complete circle can be
33
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
obtained in a Smith chart. The input return loss o f the antenna in free space should be
also plotted on the same Smith chart. The antenna efficiency is obtained as
<7 = 7
- A
-r v
,■<
(As,™) +(Av™) (j-S.J )
<2-2>
where Sufi is the free space input return loss and ASndn, A5max are the minimum and
maximum distance o f the circle from the input return loss [2.22]. This method maintains
the repeatability o f Wheeler cap method yet it is applicable to antennas with multi­
resonance.
The broadband quasi-Yagi antenna was placed in a waveguide-sliding short
structure and the input return loss was measured at 20 positions. This data and the input
return loss for the antenna radiating in free space were incorporated into a MATLAB
program. The results are shown in Fig. 2.10. From this, we see that the radiation
efficiency of the overall structure is quite high, nominally 93 % across the operating
band.
Two dips in efficiency are observed at 9.2 GHz and 11.0 GHz.
These dips
correspond to resonances in the waveguide-sliding short structure that were observed for
all sets o f sliding-short data. Additionally, since this phenomena is not observed for the
antenna radiating in free-space, we conclude that these dips in the radiation efficiency are
unphysical and represent a limitation o f the measurement technique at these resonant
frequencies. The error in measurement is estimated from the ripple to be roughly ± 3 %.
34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In comparison with the other planar antennas, the quasi-Yagi antenna achieves
high radiation efficiency. The patch antennas can achieve 80 - 96 % radiation efficiency
[2.20] [2.23] for with low permittivity substrate but nominally 65 % with high
permittivity substrate (£,=10.2) [2.23]. The tapered slot antenna is known to have a large
cross-polarization that affects adversely to the antenna’s radiation efficiency. Derived
from the directivity and gain reported in [2.24], the linear tapered slot antenna (LTSA) is
proved to have very low radiation-efficiency that is typically below 50%. The cross­
polarization radiation loss for the antenna on low permittivity substrate is as high as 30%
of the entire radiation power [2.24]. LTSA is expected to have even lower radiation
100
90
80
>*
7
8
9
10
11
Frequency [GHz]
Fig. 2.10. Measured efficiency o f the antenna.
35
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
12
efficiency when it is fabricated on high permittivity substrate.
2.3
Mutual Coupling Characteristics
As an end-fire radiator, the quasi-Yagi antenna can be easily configured into a 2D
array by simply stacking multiple cards o f sub-arrays, forming a sharp main beam useful
for adaptive arrays for communications, power combining or phased array radars, and
millimeter-wave imaging. Due to the linear design o f each card, minimal difficulties with
signal routing will occur. Additionally, because the antenna is completely compatible
with microstrip circuitry, components such as amplifiers and phase-shifters can be
incorporated easily into each card. Heat dissipation, often a problem in large planar 2-D
arrays, can also be addressed by adding additional heat sinking to the microstrip ground
plane. The stacked card array also readily allows air-cooling or routing for liquid coolant.
For these reasons, we believe that the quasi-Yagi antenna may prove to be a valuable
array antenna.
36
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.0 mm
(a)
(b)
Fig. 2.11. Schematic views o f the two-element quasi-Yagi array, (a) coplan ar and (b)
stacked array.
One important parameter in this context is the mutual coupling between elements
within an array environment, which may not only complicate the array design but also
become the cause o f the notorious scan-blindness problem in many existing planar
antenna arrays [2.25].
Despite its very compact design, the quasi-Yagi antenna
demonstrates very low mutual coupling when placed in an array environment. Fig. 2.11
(a) and (b) show the schematic figures o f the two-element stacked and coplan ar arrays.
The spacing between antennas is 15 mm, which corresponds to Xo/2 at 10 GHz. The
substrate for the coplanar arrays is truncated at both sides, with a total width of 30 mm.
The mutual coupling is determined by the measured direct transmission coefficient S21 o f
the arrays. As can be seen in Fig. 2.12, the coupling level is quite low across the entire
operating bandwidth with a maximum measured mutual coupling o f -18.5 dB at 12 GHz.
37
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Additionally, the coupling level is extremely flat over the operating bandwidth and only
creeps up at the edges o f the operating bandwidth where the antenna performance
degrades.
Furthermore, to emulate the situation o f two elements in a large array, two small
pieces o f absorbers are attached to the side edges o f the coplanar array so that the edge
effect o f the truncated substrate is eliminated [2.26]. The measured S21 in this case
reveals an even lower coupling level, with a maximum value o f -22.4 dB at the upper
end of the operating frequency range o f the antenna.
0
Iii
-
;
-10
ffl
to
%
S 20
11##1l
/I
I
#11
1
If
# w
Q.
§-30
\
Y
#•*
%
%
%
%
%
\'
%
-•
CO
-60
9
%
340
-50
1l
11#,
.
6
\ #
------Coplanar
-
8
9
10
\
t
\ t
{
i
i
i!
j
i
11
\t
m
12
13
Frequency [GHz]
Fig. 2.12. Measured results o f the mutual coupling o f the two-element quasi-Yagi
arrays.
38
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
It is interesting to compare the mutual coupling o f the quasi-Yagi antenna to that o f
other common planar antennas likely to be used in an array environment, such as the
microstrip patch antenna or linearly tapered slot antenna (LTSA). For several reasons, it
is difficult to make a systematic comparison o f the mutual coupling. One reason is that
references commonly compare the coupling as a function o f inner-edge to inner-edge
spacing, which can vary considerably depending on the type o f antenna or substrate
material. For this reason, we choose to compare mutual coupling as a function o f centerto-center antenna spacing. From Fig. 2.12, mutual coupling o f the quasi-Yagi antenna at
10 GHz where element spacing is
X q /2
has been measured to be -21 dB for horizontal and
vertical cases. For comparison purposes, a microstrip patch antenna on a 1.57 mm thick
substrate with e, = 2.55 with a center-to-center spacing of X q /2 at 5 GHz demonstrates an
E-plane coupling o f approximately —18 dB and H-plane coupling o f -13 dB [2.27]. Note
that these numbers are estimated by calculating the edge-to-edge spacing for the specific
case that the center-to-center spacing is
X q /2
in [2.27]. The same comparison can be
made with the published data on the LTSA antenna in [2.28]. In this case, the LTSA
antenna demonstrated between —10 and —14 dB coupling when the antennas were
fabricated side by side on the same substrate. This LTSA is fabricated on a 0.254 mm
thick substrate with sr = 2.2, with XJ2 spacing at 4.2 GHz calculated from the geometry.
Although these numbers will change as a function of frequency and substrate material,
both o f these examples show that the quasi-Yagi may be a superior array antenna in terms
of its mutual coupling characteristics.
39
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.4
Gain-Enhanced Design
The design o f quasi-Yagi antenna can be optimized so that the antenna
demonstrates higher gain at the expense o f reduced bandwidth. Fig. 2.13 shows the Xband gain-enhanced design o f quasi-Yagi antenna. As with the broadband version o f the
quasi-Yagi antenna, the antenna is realized on 0.635mm thick Duroid with er = 10.2. The
complete dimensions o f the antenna are (unit: mm): Wj = W3 = W4 = W5 = W6 = 0.58,
W2 = 1.12, W6 = 0.25, Wdri = Wdir = 1.17, S5 = S6 = 0.25, L, = 3.2, L2 = 2.5, L3 = 5.3, U =
2.0, Ls = 0.76, Sref = 7.7, Sdjr = 3.25, Ssub = 3.0, Ldn = 9.4 and L«jir = 5.5. In this design,
the distance between the driver element and the reflector element is still much larger and
the strip width is also considerably wider than that in the previous design.
Fig. 2.13. Picture o f the gain-enhanced quasi-Yagi antenna.
40
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 2.14 shows the measured input return loss o f the antenna. The bandwidth for
VSWR < 2 is 11 %, which is much wider than the regular patch antenna realized on the
similar substrate. The measured gain o f the antenna is shown in Fig. 2.15. The
measurement uses a broadband antenna as a reference and the corresponding calibration
table prepared by the manufacturer in a standard gain calibration measurement. The gain
o f the antenna varies from 5 to 7dB for the operating bandwidth. In comparison with the
broadband design, this design has 4 times narrower bandwidth but about 2 dB higher
gain. By adding more directors, we expect that the gain can be further increased. To
confirm the change in the absolute gain value, the radiation patterns are also plotted in
0
-2
-4
CD
-6
M
(0
o
-10
-12
&
-14
-16
-18
-20
8.5
9
9.5
10
10.5
11
Frequency[GHz]
Fig. 2.14. Measured input return loss o f the gain-enhanced quasi-Yagi antenna.
41
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 2.16 (a)-(c) for three different frequencies. The beam-width o f the radiation pattern is
narrower than that o f the broadband design while it keeping a front-to-back ratio o f 15 dB
and maximum cross-polarization level o f -15 dB. The change in the radiation pattern is
relatively small for the entire operating bandwidth.
7.5
6.5
4.5
3.5
8.4
8.9
9.4
9.9
FrequencyfGHz]
Fig. 2.15. Measured gain o f the enhanced-gain quasi-Yagi antenna
42
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10.4
E-plaM<Co-poi)
E-plaiw(Croee-poO
H-piaiM<Co-poi)
H-piana(Crou-pal)
1(0
270
(a)
E-pUn*(Co-po4)
E-pUn«<Cro«s-poJ)
H-plan*<Co-pol)
H-pUn«<Cro**-poJ)
1*0
Fig. 2.16. Measured radiation pattern o f the gain-enhanced quasi-Yagi antenna at (a)
center frequency (9.55GHz), (b) the lower end frequency (9.09GHz) o f SI 1 < -lOdB.
43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
90
-=10
13S
E-pteM(Co-pol)
E-plaiM(CroM-pol)
H-pUn«<Co-pol)
H-plafM(CroM-pol)
45
i1S
180
225
315
270
(C )
Fig. 2.16. (c) Measured radiation pattern o f the gain-enhanced quasiYagi antenna at the high end frequency (10.04GHz) o f SI 1 < -lO dB
bandwidth
44
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
REFERENCES 2
[2.1] G. M. Rebeiz, “ Millimeter-wave and terahertz integrated circuit antennas,”
Proceedings o f IEEE, vol. 80, pp. 1748-70, Nov. 1992.
[2.2] Y. Qian and T. Itoh, “A broadband uniplan ar microstrip-to-CPS transition,” 1997
Asia Pacific Microwave C onf Dig., pp. 609-612, Dec. 1997.
[2.3] C. A. Balanis, “ Advanced engineering electromagnetics,” John Wiley & Sons,
1989.
[2.4] N. G. Alexopoulos, P. B. Ketehi and D. B. Rutledge, “ Substrate optimization for
integrated circuit antennas,” IEEE Trans. Microwave Theory Tech., vol. 31, pp. 550-557,
July 1983.
[2.5] D. B. Rutledge, D.P. Niekirk, and D. P. Kashingam, “ Integrated-circuit antennas,”
in Infrared and Millimeter-waves Sereis, vol. 10, pp. 1-90, k.J. Button, Ed. New York:
Academic Press 1983.
[2.6] S. D. Targonski, R. B. Waterhouse and D. M. Pozar, “Design o f wide-band
aperture-stacked patch microstrip antennas,” IEEE Trans. Antennas Propagat., vol. 46,
pp. 1245-1251, Sept. 1998.
[2.7] D. M. Pozar and S. D. Duffy, “A dual-band circularly polarized aperture-coupled
stacked microstrip antenna for global positioning satellite,” IEEE Trans. Antennas
Propagat., vol. 45, pp. 1618-1625, Nov. 1997.
[2.8] A. Taflove, “ Computational electrodynamics: The finite difference time-domain
method,” Artech House, 1995.
45
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[2.9] G. Mur, “Absorbing boundary conditions for the finite-difference approximation o f
the time-domain electromagnetic field equations,” IEEE Trans. Electromagnetic
Compatibility, vol. 23, pp. 377-382, 1981.
[2.10] J. -P. Berenger, “A perfectly matched layer for the absorption o f electromagnetic
waves,” J. Computational Physics, vol. 114, pp. 185-200, 1994.
[2.11] Y. Qian and T. Itoh, FDTD analysis and design o f microwave circuits and
antennas -software and applications, Realize Inc., Tokyo, Japan, 1999.
[2.12] P. J. Gibson, “The Vivaldi aerial,” in Proc. &h European Microwave Conf,
Brighton, UK, pp. 101-105, 1979.
[2.13] W. L. Stutzman and G. A. Thiele, Antenna Theory and Design. 2nd ed. John Wiley
& Sons Inc., 1998.
[2.14] J. M. Schuchardt and W. A. Bohlman, “Automated methods provide antenna-gain
measurement,” Microwaves and RF, pp. 49-61, Sept. 1998.
[2.15] K. S. Yngvesson, T. L. Korzeniowski, Y. Kim, E. L. Kollberg, and J. F. Johansson,
‘The tapered slot antenna - a new integrated element for millimeter-wave applications,”
IEEE Trans. Microwave Theory Tech., vol. 37, pp. 365-374, Feb. 1989.
[2.16] H. A. Wheeler, “The radiansphere around a small antenna,” Proc. IRE, pp. 13251331, Aug. 1959.
[2.17] E. H. Newman, P. Bohley and C. H. Walter, ‘T w o methods for the measurement
of antenna efficiency,” IEEE Trans. Antennas Propagat., vol. 23, pp. 457-461, July 1975.
[2.18] J. Ashkenazy, E. Levine and D. Treves, “Radiometric measurement o f antenna
efficiency,” Electronics Lett., vol. 21, no. 3, Jan. 1985.
46
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[2.19] C. A. Balanis, Antenna Theory: Analysis and Design. John Wiley & Sons Inc.,
New York, 1982.
[2.20] D. M. Pozar and B. Kaufman, “Comparison o f three methods for the measurement
o f printed antenna efficiency,” IEEE Trans. Antennas Propagat., vol. 36, pp. 136-139,
Jan. 1988.
[2.21] R. H. Johnston and J. G. McRory, “An improved small antenna radiationefficiency measurement method,” IEEE Antennas and Propagation Magazine, vol. 40,
no.5, pp. 40-8, Oct. 1998.
[2.22] R. H. Johnston and J. G. McRory, “Small antenna efficiency by the reflection and
the Q measurement methods,” IEEE AP-S Int 7 Symposium Digest, Orland, FL, vol. 3, pp.
1810-1813, June 1999.
[2.23] I. Papapolymerou, R. F. Drayton and L. P. B. Katehi, “Micromachined patch
antennas,” IEEE Trans. Antennas Propagat., vol. 46, pp. 275-283, Feb. 1998.
[2.24] Y -S. Kim and K. S. Yngvesson, “Characterization o f tapered slot antenna feeds
and feed arrays,” IEEE Trans. Antennas Propagat., vol. 38, pp. 1559-1564, Oct. 1990.
[2.25] D. M. Pozar and D. H. Schaubert, “Scan blindness in infinite phased arrays o f
printed dipoles,” IEEE Trans. Antennas Propagat., vol. 32, pp. 602-610, June 1984.
[2.26] Y. Qian, W.R. Deal, N. Kaneda and T. Itoh, “ A uniplan ar quasi-Yagi antenna with
wide bandwidth and low mutual coupling characteristics”, 1999 IEEE A P S International
Symposium Digest, Orlando, FL, Vol. 2, pp. 924-927, July 1999.
47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[2.27] A. H. Mohammadian, N. M. Martin, D. W. Griffin, “A theoretical and
experimental study o f mutual coupling in microstrip anenna arrays”, IEEE Trans.
Antennas Propagat., vol. 37, pp. 1217-1223, Oct. 1989.
[2.28] R. Q. Lee and R. N. Simons, “Measured mutual coupling between linearly tapered
slot antennas”, IEEE Trans. Antennas Propagat., vol. 45, pp. 1320-1322, Aug. 1997.
48
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 3
A Broadband Microstrip-to-Waveguide Transition Using
Quasi-Yagi Antenna
3.1 Transition Design
The quasi-Yagi antenna consists o f two dipole antennas, a truncated ground plane
and a microstrip-to-CPS balun. The antenna is inserted in the E-plane o f the waveguide
as illustrated earlier in Fig. 1.2. The microstrip substrate and ground plane width is same
as the inside dimension o f the waveguide height. The waveguide wall is in contact with
the ground plane but not with the signal line o f the microstrip, ensuring that we do not
have DC block problem.
In this section of the chapter, we discuss the theoretical development and the
design procedure for individual parts o f the transition. The quasi-Yagi antenna consists o f
two components. One is the CPS antenna shown in Fig. 3.1 (a) and the other is the
microstrip-to-CPS transition shown in Fig. 3.1 (b). We explain the two parts separately in
order to understand the broadband characteristics of the transition. Optimizing those two
components separately makes the optimization easier.
3.1.1
Antenna (Radiator) Design in Waveguide
49
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The CPS antenna shown in Fig. 3.1 (a) has the coupled microstrip line input. This
coupled microstrip line is connected to the coplanar strip line with different strip width.
Since the coplanar strip line does not support the even mode, it acts as an open end for
the even mode o f the coupled microstrip line, and enables us to suppress the undesired
mode excited in the coupled microstrip line. The CPS line is connected to the printed
dipole antenna that has length o f approximately
where Xd = ^
(a)
+ lj and is
(b)
Fig. 3.1. Individual components o f Quasi-Yagi antenna, (a) CPS Yagi-like
antenna (b) microstrip-to-CPS balun
50
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
positioned approximately Xj/4 away from the reflector (truncated ground plane). Using a
high dielectric constant substrate makes it possible to excite the rectangular waveguide’s
dominant mode with a dipole antenna that would otherwise be too large to fit in the Eplane o f the waveguide. The truncated ground plane here works as a reflector o f the
quasi-Yagi radiating elements as well as a short circuit for the TEio-like dielectric loaded
waveguide mode. Unlike conventional transitions where substrate mode often causes
troubles, the transition utilizes the substrate mode as one o f radiating components.
0.9
\'
*0.7
A-8Ghz
•■-•10GHz
12GHz
ii
Driver.
0.3
Elem ent num ber
\:Reflector
(a)
Director. /
(b)
Fig. 3.2. Relative current amplitudes o f the three elements quasi-Yagi array. The current
amplitudes are observed at the edge closer to the reflector on driver and director element,
(a) Observation points (b) Relative current amplitudes at observation points shown in (a)
51
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Even though the dipole antenna’s radiation field aligns well with the dominant
mode o f the waveguide, the operation bandwidth o f the antenna tends to be narrow. Also
when it is operated in free space, the truncated ground plane reflector is not sufficient for
a good front-to-back ratio. In an attempt to improve the directivity of the antenna, a
shorter dipole element is added on the same side as the driver dipole element. This
director dipole not only increases the front-to-back ratio but it also plays an important
role in broadening the operation bandwidth o f the antenna structure as previously
demonstrated in other antenna structures with different types o f parasitic elements [3.1].
0
-2
i
j
i
;
a? -*
2,
««
S -8
!
i
' ;
!
O
65-12
•2-16
•18
-20
;
|
!
i
|
!
j
j
j
l
i
9
j
|
I
i
!
!
|
'
........ Return lo ss
I - - - - - - - - - - Insertion loss
•» •
T " '
"?
!
i
9.5
10
.
>
|
i
!
____i . . . . i . . ...
8.5
i
I
I
E
3-10
;
I
/
.............
j
i
I
i
i
i
10.5
11
I
I
11.5
i
12
12.5
Frequency [GHz]
Fig. 3.3. Insertion and return loss o f the coupled microstrip line to waveguide
transition shown in Fig. 3.2 (a). Input o f the antenna is odd mode of the coupled
microstrip line and the output is rectangular waveguide dominant mode.
52
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 3.2 shows the amplitude o f the x-component of induced currents on the three
elements o f the antenna. The amplitudes are normalized to that o f the driver element’s
peak current. From this figure, we know that the relative amplitudes o f the currents on
each dipole elements vary with the operation frequency. This mechanism can be
explained by referring to the log-periodic array antenna where the dipoles o f different
length operate at different frequencies according to their half-wavelength resonance. The
three-element quasi-Yagi antenna shows less dramatic change in induced current
amplitude compared to log-periodic antenna with many dipole elements [3.2], but the
power intensity o f the radiating elements apparently shifted from the longer element to
the shorter one as the frequency goes up. This partially explains the broadband
characteristic o f the antenna.
Although the initial guess for the length and the position o f the driver element is
given above, and director length can be found in conventional Yagi antenna design tables
[3.2], the optimal length and positions o f the dipole elements have to be found by full
wave simulations. We have found that the length o f the director and the distance between
director and driver has to be smaller than the conventional Yagi antenna design to
achieve broadband characteristic. Fig. 3.3 shows the optimized simulation result for the
antenna in Fig. 3.1 (a). It should be noted that even without the microstrip-to-CPS balun
the quasi-Yagi antenna shows a quite broadband characteristic. This indicates that the
antenna is a good candidate for waveguide spatial power combining array application
where the physically much larger tapered slot antenna has been used [3.1].
53
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.1.2
Microstrip-to-CPS Baiun Design
The microstrip-to-CPS balun shown in Fig. 3.1 (b) is also an essential part o f the
transition. The balun should be able to provide an odd mode in the coupled microstrip
line while suppressing the even mode over broad bandwidth. The balun’s phase shifter
creates 180-degree phase difference between the coupled microstrip lines around 10GHz,
-5
2,-10
i■**
CO
0-15
c
3-20
...........
&
>25
O
H-30
w
_c
-35
•40
•
Return loss
Even mode insertion loss
Odd mode insertion loss
» «
8.5
9.5
10
10.5
11
11.5
12
12.5
Frequency [GHz]
Fig. 3.4. Insertion and return loss of the microstrip-to-CPS balun shown in Fig. 3.2
(b). The Input o f the balun is microstrip line and the output is coplanar strip line. Both
even and odd mode insertion loss o f CPS is simulated.
54
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and offers a good impedance matching with the odd mode that has much smaller
impedance than the even mode when the coupling is tight. In addition to good impedance
matching, the CPS line provides an approximate open end for the even mode o f the
coupled microstrip line. By having the truncated ground plane close to the phase shifter,
the even mode impedance o f the couple line looking from the phase shifter becomes large
so that it suppresses the even mode over broad bandwidth.
The optimization o f the balun is also done with the full wave simulation. We have
used 50Q lines for the phase shifter and coupled line. Although there are many
parameters that can be optimized, the most critical parameters are the gap between the
CPS lines and the length o f the phase shifter [3.3]. Fig. 3.4 shows the simulation result o f
S-parameters o f the balun. We can see that the insertion loss o f the coupled line’s odd
mode is quite small (less than -ld B ) when the return loss is below -lOdB and that o f the
even mode is below -lOdB for almost entire X-band. The width o f the microstrip line at
the input port is left as a free parameter o f an impedance transformer to improve the
overall return loss when the balun is put together with the antenna.
3.2
Simulation and Measurement Results
3.2.1
Simulation Results
Two X-band prototypes have been designed using Duroid (er = 10.2) and alumina
(sr = 9.9) substrate respectively. The thickness o f the substrate is 0.635mm in both cases.
55
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The printed antenna pattern is located at the center o f the waveguide (H=l 0.8mm in Fig.
2) and the same dimensions o f the quasi-Yagi antenna are used for both transitions. The
simulation results using HP’s HFSS are shown in Fig. 3.5 and 3.6. Metal loss is not
included in simulations and it is assumed that the ground plane and the substrate o f the
microstrip are in contact with the waveguide’s inner sidewalls.
Fig. 3.5 shows the simulation result of the insertion and return loss o f the
transition using a Duroid substrate with and without the dielectric loss o f the substrate
(tan8~0.002) included. The return loss is below -lOdB over 35% bandwidth and the
maximum insertion loss is -0.4dB at the edge frequency for the simulation without
dielectric loss. However, when the dielectric loss o f the substrate is taken into account,
the insertion loss o f the transition degrades considerably at higher frequencies (-0.8dB at
12GHz).
Fig. 3.6 shows the simulation result on the transition using an alumina substrate
(tan§~0.0002). W ith the dielectric loss taken into account, 34% bandwidth is achieved for
return loss below -lOdB, and 30 % bandwidth is obtained for Su less than -15dB. The
insertion loss is improved with the small loss tangent substrate where S21 is -0.15dB at
the edge o f the Su smaller than -15dB frequency. From these simulation results, we
observe that the loss o f the dielectric substrate significantly affects the performance o f the
whole transition. This characteristic can be explained by the operation mechanism o f the
quasi-Yagi radiator, where the critical radiation takes place with the high dielectric
constant substrate including quarter-wavelength resonance o f the TEio-like mode o f the
dielectric-loaded waveguide.
56
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
t
M
- S21 without lorn
-S11 without loss
- S21 with loss
-S11 with loss
£-25
•30
_L
8.5
9.5
10
10.5
11
11.5
12
12.5
Frequency[GHz]
Fig. 3.5. Simulated results o f insertion and return loss o f microstrip-to-waveguide
transition on Duroid substrate with and without dielectric loss.
CD
*3f
g-10
2-15
s.
0-2°
r
M
!
;
|
:\ i
..:
\
I
!
i ..W
\
i
j
j
|
j
;
j
|
j
;
8.5
A
i
/
;
/»
!
'--r-O -.J / «
i
\ j " .*
?• '
----- S21 without loss
----- S11 without loss
II * « ;'
i *•
----- S11 with loss
i
--------1_____ i_____ I—. I
10 10.5
11
11.5
12
12.5
I
£-25
-30
i
/
~ -5
^
Y
' ^
9.5
Fraqusncy[GHz]
Fig. 3.6. Simulated results of insertion and return loss o f microstrip-to-waveguide
transition on alumina substrate with and without dielectric loss.
57
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.2.2
Measurement Results
Fig. 3.7 shows the measurement assembly o f the transition. Because of the nature
o f the transition, split block fabrication [3.4] [3.5] and backshort hermetic sealing [3.6]
[3.7] of the waveguide are not required. The standard waveguide as shown in Fig. 3.7 is
used in the X-band prototype demonstration. The microstrip substrate is inserted in the Eplane of the waveguide, and is extended outside the waveguide for simple mounting. In
order to support the antenna substrate a small copper block is attached to the waveguide
C o p p e r Bloblc
Fig. 3.7 Measurement setup o f the microstrip-to-waveguide transition
58
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
flange, and the microstrip ground plane and SMA connector are soldered to the copper
block. The assembly is very simple compared to all other microstrip-to-waveguide
transitions reported previously and is highly compatible with microstrip-based
MIC/MMIC circuitry.
The measured results o f both return and insertion loss o f the transition on Duroid
substrate are shown in Fig. 3.8. For accurate evaluation o f the insertion loss, we measure
the straight section o f the microstrip lines and the rectangular waveguide and we estimate
the SMA connector, microstrip feed line and the waveguide adapter losses and subtract
them from the measured insertion loss. The result shows 39% bandwidth with return loss
better than —12 dB and the insertion loss is -0.9 dB at the center frequency. This
broadband characteristic is comparable to that achieved by the antipodal finline approach
using low dielectric constant substrate. The slightly higher insertion loss is partly due to
the relatively large loss tangent o f Duroid substrate as we have seen in the simulation
results (Fig. 3.5). Fig. 3.9 shows the measured result o f the transition on alumina
substrate. The result shows 35% bandwidth with return loss better than -12dB, and the
insertion loss is -0.3dB at the center frequency. These results indicate that using low loss
dielectric substrate is critical in achieving optimal performance with this transition for
high frequency operation. The slightly higher insertion loss compared to the simulated
data can be due to the conductor loss o f the dipole antennas and the radiation loss o f the
semi-open input port. The insertion loss can be reduced by properly shielding the input
port.
59
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a-io
0-20
Insartion hM
Return loss
£-25
-30
9
8
10
10.5
11
11.5
12
12.5
Fraquency[GHz]
Fig. 3.8. Measured insertion and return loss o f the quasi-Yagi microstrip-to-waveguide
transition using Duroid substrate.
■o
Insertion loss
Rstum loss
3-15
5-20
-30
8
8.6
9
9.5
10
10.6
11
11.5
12
12.5
Froqusncy [GHz]
Fig. 3.9. Measured insertion and return loss o f the quasi-Yagi microstrip-to-waveguide
transition using alumina substrate.
60
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.3 Tolerance and Packaging Issues
We have demonstrated and proved that the transition’s unique topology enables
us to have a simple fabrication and assembly o f the microstrip-to-waveguide transition
with broad bandwidth and relatively low loss characteristic. However, the concern of the
sensitivity to mechanical assembly still remains. In this section, we extend our study to
the sensitivity o f the dimensions that determine the tolerance o f the mechanical assembly
and alignment o f the transition. The three parameters studied here are the inserted length
of the antenna substrate (L), the width o f the antenna substrate (W) and the distance
between the substrate and the sidewall o f the waveguide (H), as shown in Fig. 3.10.
M icrostrip feed
W aveguide
M etal block
Fig. 3.10. Quasi-Yagi microstrip-to-waveguide transition and three
parameters investigated for packaging tolerance issues.
61
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 3.11 shows the simulated results o f S-parameters on various lengths L o f the
inserted length o f the antenna substrate. The results show that as the inserted length
changes from 10mm to 20mm, the bandwidth is improved by 3%. Although the longer
inserted length generally improves the performance, the overall performance o f the
transition is not affected severely by this parameter.
Fig. 3.12 shows the simulation results o f the transition with the antenna substrate
±lm m offset from the center o f the waveguide (H=9.8mm and 11.8mm). Although the
offset we have assumed is relatively large (-10% ), the changes from the original results
■(£!
£9 *5
-25
8
8.5
9
9.5
10
10.5
11
11.5
12
12.5
Frequency [GHz]
Fig. 3.11. Simulation results o f the transition with the inserted alumina substrate
length o f L= 10mm, 15mm and 20mm.
62
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
are not significant. Slight improvement with the substrate 1mm lifted up from the center
indicates that the antenna’s radiation direction is tilted to substrate direction and more
optimization can be done on this parameter to improve the transition’s performance.
Although the substrate truncation error can be quite small using modem dicing
machine, the sensitivity on the substrate width is also studied. Fig. 3.13 shows the
simulation results o f the transition with substrate width 30pm, 50pm and 100pm shorter
than the original width (W =l 0.16mm). The inserted length o f the substrate L is kept
15mm. The results show that both return loss and insertion loss is not strongly affected by
the width o f the substrate for AW=30pm and 50pm. However, for AW= 100pm, a stop
«-! 0
•2 »
*3 O
•— S11; H=9.8mm
- S11; H=11.8mm
-40
8
8.5
9
9.5
10
S21; H=9.8mm
S21; H=11.8mm
10.5
11
11.5
12
12.5
Frequency [GHz]
Fig. 3.12. Simulation results of the transition using the alumina substrate positioned
at height H=9.8mm and 11.8mm.
63
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
band appears in band around 8.4GHz. This stop band is due to the half-wavelength
resonance of the ridged waveguide mode created by the small gap between the ground
plane and the sidewall o f the waveguide. If there are gaps on both sides, the quasi-TEM
mode creates half-wavelength resonance between open end o f waveguide and the open
end o f the ground plane at the similar frequency. When the dicing error is expected to be
high, one can push the resonant frequency out of the interesting frequency band by
inserting antenna substrate further into the waveguide.
- S21; W=10.06mm
■S21; W=10.11mm
•S21; W=10.13mm
a -s
S11; W=10.06mm
S11; W=10.11mm
S11; W=10.13mm
f}
20
-25
8
8.5
9
9.5
10.5
10
11
11.5
12
12.5
Frequency [GHz]
Fig. 3.13. Simulation results of the transition with the width o f alumina substrate
W=10.13, 10.11 and 10.06mm.
64
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.4
W-band microstrip-to-waveguide transition using quasi-Yagi
antenna.
As discussed in the introduction in chapter one, due to its low loss characteristics
metallic waveguides are still essential components in millimeter-waves frequencies. A
good example is that the input and output ports o f network analyzers for the frequencies
above 50 GHz are all realized by metallic waveguides. The 1 mm coaxial cable system
[3.8] can potentially replace the waveguide system up to the higher end o f W-band (110
GHz) but such a system is relatively new and currently too costly for most o f the tests
and applications.
Fig. 3.14. W-band microstrip-to-waveguide transition schematic view.
65
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
For the millimeter-wave frequencies, planar circuits frequently need to be
integrated with waveguide components. A high-speed WLAN (Wireless Local Area
Networks), for example, uses 40 GHz carrier frequency [3.9] and the transceiver at the
hub or base station must have high-gain waveguide hom antennas for sensitive reception.
Even though the antennas in those millimeter-wave systems are often based on the
waveguide antennas, the signal's down- and up-conversion mixing is most convenient
with planar circuits.
The microstrip-to-waveguide transition using quasi-Yagi antenna can be scaled to
any frequencies, but it has to be noted that the further optimization would be required
depending on the waveguide aspect ratio and the available substrate thickness. The
o
•5
-10
-15
e -25
S -30
-35
-40
75
80
85
90
95
100
105
110
Frequency [GHz]
Fig. 3.15 Simulation results o f W-band back-to-back transition. Both metal (gold)
and dielectric substrate (alumina) loss are not taken into account. The conductor is
assumed to have zero thickness.
66
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
waveguide aspect ratios are different for X-band (2.25:1) and W-band (2:1) waveguide
and the standard substrate thickness suitable for W-band transition is 3mil.
The simulation and optimization o f the W-band transitions are done by HFSS.
The back-to-back transitions are designed as shown in Fig. 3.14. The spilt block
waveguides are required for the back-to-back transition demonstration. The microstrip
circuits are designed on 3-mil thick alumina substrate and assumed to be on the metal
mount at the center of the split block waveguides. The radius o f the milling drill appears
as the round comer in the bottom-block o f the waveguide. The complete dimension o f the
transition are (unit: mm): Wi = W3 = W4 = W5 = Wdn = Wdjr = 0.067, W2 = 0.133, W6 =
-10
-15
c -25
S *30
•35
-40
75
80
85
95
90
100
105
110
Frequency [GHz]
Fig. 3.16 Simulation results o f W-band back-to-back transition. Both metal (gold)
and dielectric substrate (alumina) loss is taken into account. The conductor is
assumed to have zero thickness.
67
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
S5 = S6 = 0.033, U = 0.313, L2 = L5 = 0.166, L3 = 0.499, U = 0.211, Sref= 0.419, Sdir =
0.387, Ssub = 0.213, Ldn = 0.900 and Uir = 0.344.
Fig. 3.15 shows the simulated results o f the back-to-back W-band transition. The
metal and dielectric losses are not considered in this simulation. The results show that the
return loss is smaller than -lOdB for 26 % (81 - 105GHz) bandwidth and the insertion
loss is very small, nominally a fraction o f dB. Fig. 3.16 shows the simulation results of
back-to-back W-band transition with metal and dielectric loss. Gold is used for the
conductor but the thickness o f the conductor is assumed to be zero due to the simulation
difficulty. The simulation predicts the return less that -lOdB is achieved for the almost
entire W-band and the insertion loss is about -2dB.
The future fabrication of this circuit requires E-beam lithography for the gold
deposit and backside mask alignment for the alignment o f the truncated ground plane.
The thin substrate and backside metalization make this fabrication challenging.
68
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
REFERENCES 3
[3.1] C. H. Tsao, Y. M. Hwang, F. Kilburg and F. Dietrich, “Aperture-coupled patch
antennas with wide-bandwidth and dual-polarization capabilities”, IEEE AP-S I n t’l
Symposium Digest, New York, NY, pp. 936-939, June 1988.
[3.2] W. L. Stutzman and G. A. Thiele, Antenna Theory and Design. 2nd ed. John Wiley
& Sons Inc., 1998.
[3.3] A. Alexanian and R. A. York, “Broadband spatially combined amplifier array using
tapered slot transitions in waveguide”, IEEE Microwave and Guided Wave Letters., vol.
7, pp. 42-44, Feb. 1997.
[3.4] L. J. Lavedan, “Design o f waveguide-to-microstrip transition specially suited to
millimeter-wave applications”, Electronics Letters, vol. 13, no. 20, pp. 604-605, Sept.
1977.
[3.5] T. Q. Ho and Y. C. Shih, “Spectral-domain analysis o f E-plane waveguide to
microstrip transitions,” IEEE Trans. Microwave Theory and Techniques, vol. 37, pp. 388392, Feb. 1989.
[3.6] Y. C. Shih, T. N. Ton and L.Q. Bui, “Waveguide-to-microstrip transitions for
millimeter-wave applications”, IEEE M T T S Int 7 Symposium Digest, NewYork, NY, vol.
l,p p . 473-475, 1988.
[3.7] F. J. Villegas, D. I. Stones and H. A. Hung, “A novel waveguide-to-microstrip
transition for millimeter-wave module applications”, IEEE Trans. Microwave Theory and
Techniques, vol. 47, pp. 48-55, Jan. 1999.
69
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[3.8] K. Howell, K. Wong, “DC to 110 GHz measurements in coax using the 1 mm
connector,” Microwave Journal, Euro-Global Edition, vol.42, (no. 7), Horizon House
Publications, pp. 22 - 34, July 1999.
[3.9] D. J. Skellem, L. H. C. Lee, T. McDermott, N. H. E. Weste, J. Dalton, J. Graham,
T. F. Wong and A. F. Myles, “A high-speed wireless LAN,” IEEE Micro, pp. 40-46,
1997.
70
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 4
A Broadband CPW-to-Waveguide Transition Using QuasiYagi Antenna
4.1 Transition Design
CPW transmission line, first introduced in [4.1], is a uniplanar transmission line
where the circuit pattern including ground plane exist only on one side o f the substrate.
This uniplanar structure is frequently more convenient and compatible with solid-state
devices than the microstrip line structures. One disadvantage o f microstrip lines,
Fig. 4.1. Schematic view of the single CPW-to-waveguide transition
using quasi-Yagi antenna.
71
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
especially is that a very thin substrate is required. For example, at 100 GHz, 3mil thick
alumina is required to achieve 50Q line. In this section, the transition design o f the CPW
(coplanar waveguide) -to-waveguide transition using quasi-Yagi antenna shown in Fig.
4.1 is discussed in detail.
The common difficulty in designing the CPW circuits and its package is CPW’s
multi-modes characteristics. The solution o f the multi-modes problem o f CPW line is
presented later in this section. The quasi-Yagi antenna fed by CPW line also needs a
broadband balun to excite dipole antenna element. Selection on the broadband balun of
CPW-to-CPS transition is discussed in the following section.
L,
Director
Driver
Airbridge
R eflector
(metal block)
CPW
Fig. 4.2. CPW-to-waveguide transition schematic top view.
72
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.1.1 CPW-to-CPS Balun Consideration and Antenna Design
A schematic view o f the single CPW-to-waveguide transition is shown in Fig. 4.1
and the printed circuit o f the CPW quasi-Yagi antenna is shown in Fig. 4.2 (a). There are
a small number o f broadband CPW-to-CPS transitions reported [4.2], [4.3], though
narrowband transitions have been used in printed dipole antennas fed by CPW [4.4][4.6]. CPW-to-slotline (SL) transitions have been studied extensively [4.7]-[4.9] and one
o f CPW-fed quasi-Yagi antennas from our group is based on the technique using CPW,
SL and CPS [4.10]. This approach worked fine in antenna for free space radiation but the
suspected excitation o f surface wave mode from the CPW-to-SL transition may cause
severe effect when integrated in waveguide environment.
Substrate
CPW circuit
plane
Trenched metal
block
Fig. 4.3. CPW-to-waveguide transition schematic front view.
73
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In this work, the balanced signal on the coplanar strips is excited by a CPW-toCPS transition based on [4.2]. The simulation on the CPW-to-CPS balun is performed
prior to the construction o f the entire transition. The schematic view o f the balun and the
simulation results obtained by Sonnet are shown in Fig. 4.3 (a) and (b), respectively. The
simulation results indicate that the balun is very broadband, all the way from DC to
around 12GHz. The result also shows that better results are obtained for lower frequency,
indicating that the smaller dimensions o f CPW and balun (CPW width, gap and balun’s
open stub length) would improve the performance at higher frequency. The height o f air
bridges is also a critical parameter. Generally, the shorter the height o f air-bridges is, the
better they work [4.11], [4.12]. The balun is a uniplanar structure and the fabrication o f it
is relatively simple. The size o f the balun is also very small and it makes it attractive in
integrated circuits.
The field distribution o f the CPS fed dipole antenna is in good alignment with the
dominant mode o f the rectangular waveguide when the antenna is inserted along the E-
Air-bridges
CPW
CPS
Fig. 4.4. Schematic view o f CPW-to-CPS balun.
74
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
0
'
_ -5
m
2,
» >10
11
!
1
*
0
0
!
-15
E
3
£ -20
O
c• •25
m
c
0
0
0
0
#
*
0
........ Return loss
- - - - - - - - - - - - - - - Insertion loss
*
»
0
0
0
0
-30
!
!
1
i
i
t
1
-35 '
1
*
i
0
0
.
. .
.
.
4
.
.
.
. . . .
1
1
i
1
t
!
!
:
i
.
6
8
Frequency [GHz]
.
.
.
i
:
:
i
.
10
.
.
i
.
,
.
12
Fig. 4.5. Sonnet simulation results o f CPW-to-CPS balun. CPW dimension [mm]: width
= 0.3, gap =0.15, open stub length 1.5. Air-bridges dimension [mm]: width = 0.15,
length =0.9, height=0.1. Substrate: thickness=0.635mm, er=10.2.
plane o f the waveguide. The parasitic director element o f the antenna guides the direction
o f wave propagation into the rectangular waveguide. As is illustrated in the side-view of
the transition in Fig. 4.2 (b), the printed CPW circuit is placed face down on the trenched
metal block. The ground planes o f the CPW are placed in contact with the metal blocks to
provide support for the circuit. The supporting metal block also acts as a back short to
prevent the waveguide’s dominant mode from propagating backward, as well as serving
as the antenna’s reflector element.
4.1.2 Single mode operation o f CPW line
75
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Since the CPW can support two dominant modes, namely the CPW mode and the
CSL (coupled slotline) mode, any discontinuity on the circuit can potentially excite the
(a)
3.5
3
2.5
o
2
~ 1 .5
1
0.5
0
7
8
9
10
11
12
13
Frequency [GHz]
(b)
Fig. 4.6. Proposed structure for the single mode operation o f CPW (a) schematic view
(b) propagation constant o f CPW mode.
CPW dimension [mm]: width=1.44, gap=0.48, thickness=0.635, er=10.2.
Trench dimension [mm]: width=2.76, depth=0.08.
76
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
undesired CSL mode. Moreover, while the conductor-backed CPW (CBCPW) provides
mechanical strength and convenient packaging, the strongly excited parallel plate mode,
which is also frequently referred as Microstrip line like (MSL) mode, becomes an
inherent problem. This mode is regularly referred as leaky wave [4.13], [4.14] and it
hinders the various aspects o f circuit performances [4.15]—[4.19].
Thus, additional methods such as via holes [4.20], air-bridges [4.11], [4.12],
lateral walls [4.21], [4.22], multi-layer substrates [4.23], [4.24] and UC-PBG (uniplanar
compact photonic-bandgap) [4.25] are required to suppress those undesired modes of
CBCPW. The novel approach we propose in this work is the usage o f the trenched metal
block to suppress the CSL mode as well as the parallel plate mode. This ensures single
mode operation of the CPW structure for broad bandwidth, in addition to providing
3.5
■
3
2.5
»
Modes
------ CPW
1
0.5
0
7
8
9
10
11
12
13
Frequency [GHz]
Fig. 4.7. Propagation constant o f CPW.
CPW dimension [mm]: w=1.44, g=0.48. Substrate: t=0.635, er=10.2
77
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
support for the CPW-based circuit.
The schematic view o f the new approach is shown in Fig. 4.4 (a). The propagation
constant of this structure is plotted in Fig. 4.4 (b). Only one mode is propagated in this
structure for the entire X-band. It is interesting to investigate more on this trenched metal
approach in comparison with regular CPW and CBCPW. Fig. 4.5 and 4.6 are the
propagation constants o f the CPW and CBCPW in X-band waveguide. Fig. 4.5 shows
that CPW mode is the first dominant mode in the regular CPW transmission line. In the
CBCPW approach, however, the propagation constant of CPW mode differs considerably
from the original value due to the perturbation caused by backside metal (Fig. 4.6).
Moreover, the propagation constant o f CPW mode also suffers from the added
3.5
2.5
‘1.5
Modes
MSL
CSL
CPW
0.5
7
8
9
10
11
12
13
Frequency [GHz]
Fig. 4.8. Propagation constants o f CBCPW in X-band waveguide.
CBCPW dimension [mm]: w=l .44, g=0.48. Substrate: t=0.635, er=10.2.
78
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
dispersiveness with CBCPW approach. On the other hand the new approach maintains
the weakly dispersive nature o f CPW, as shown in Fig. 4.4 (b).
Additionally, the antenna’s dimensions have to be carefully optimized like in the
conventional Yagi-Uda antenna design. Since the metal block works as a reflector
element for the quasi-Yagi antenna, the trench depth and position have to be carefully
designed and optimized together with the other antenna parameters.
4.2 Simulation and Measurement Results
4.2.1 X-band prototype o f back-to-back transition
The transition for the X-band prototype is simulated and optimized using both
Return loss
Insertion loss
~ -3 0
1-35
-40
-45
8
8.5
9
9.5
10
10.5
11
11.5
12
12.5
Frequency [GHz]
Fig. 4.9. Simulation results o f the X-band transition excluding conductor and
dielectric loss.
79
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Ansoft and HP HFSS. Fig. 4.9 shows the simulation results o f the back-to-back X-band
transition. The simulation results do not include metal and dielectric loss. The result
shows that return loss is below -lOdB across all o f X-band with maximum insertion loss
o f -0.2dB.
The CPW circuit is fabricated on the 25mil thick RT Duroid substrate (^=10.2)
and thin copper foil is used to construct the air-bridges. As mentioned earlier, a trenched
copper block is utilized to support the CPW circuit in the E-plane o f the standard
waveguide. After the air-bridges were mounted, the CPW circuit was flipped over and
soldered onto the trenched copper block. Finally the block with the CPW antenna is slid
into a standard X-band waveguide, and waveguide adapters are used to measure the 2port S-parameters. The complete dimensions o f the CPW quasi-Yagi radiator in Fig. 4.2
3-20
i-25
Return loss
Insertion loss
•30
-35
8
8.5
9
9.5
10
10.5
11
11.5
12
12.5
Frequency [GHz]
Fig. 4.10. Measured results o f the X-band back-to-back transition. The waveguide
and adapter loss are calibrated out.
80
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
are (unit: mm): Wo = 1.44, Wi = W2 = 0.3, Wdrv = Wdrc = 0.6, So = 0.48, Si = S2 =0.15,
Ltap = Lfed = 2.0, Lstb = 1.5, Ldrc = 3.1, Ldrv = 7.95, Dref = 3.7, Ddrv = 3.35, and DSUb = 1.95.
The trenched metal block’s (Fig. 4.3) dimensions are (unit: mm): w = 2.76 and h = 0.08.
The measured results o f the back-to-back transition are shown in Fig. 4.10. The
results show 33% bandwidth with return loss better than -lOdB. Insertion loss ranges
from -l.OdB to -2.2dB in the pass-band. The waveguide adapter loss is measured to be
about -0.3dB for the entire pass-band and this loss is subtracted from the measured
insertion loss to evaluate the transition loss. The antenna parameters can alternatively be
optimized so that smaller insertion loss is achieved for narrower bandwidth as described
in a later section. The simulation and measurement results show that the transition is very
broadband and reasonably low-loss. These characteristics are comparable with the ridge
waveguide transition and much better than the aperture-coupled antenna type transition in
Fig. 4.11. Measurement setup o f the single CPW-to-waveguide transition
81
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[4.26] both in bandwidth and insertion loss.
4.2.2 Single CPW-to-waveguide transition
In practical applications, a single CPW-to-waveguide transition would be a
primary interest. To investigate the issues concerned with the single transition fabrication
and packaging, a single CPW-to-waveguide transition is constructed. The CPW circuit is
fabricated on Duroid substrate and is supported in the E-plane o f the standard waveguide
by the trenched copper mount shown in Fig. 4.11. The transition does not require split
blocks or waveguide backshort and the standard X-band waveguide straight section is
used for this test structure. The measured results o f the single transition are shown in Fig.
•20
8
8.5
9
9.5
10
10.5
11
11.5
12
Frequency [GHz]
Fig. 4.12. Measured results o f the single CPW-to-waveguide transition
82
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.12. The results show 20% bandwidth with return loss better than -lOdB. Insertion loss
ranges from -l.OdB to -1.8dB in the pass-hand. Relatively high insertion loss comes from
radiation loss of the CPW port and the CPW transmission line.
4.2.3 Low insertion loss back-to-back transition design
The insertion loss o f the transition can be improved by adjusting many parameters
such as the dipole elements o f quasi-Yagi antenna, air-bridges and CPW-to-CPS balun.
While it is impossible to discuss all possibilities for improvement, the tapered CPW
ground plane is introduced in this section to improve the insertion loss o f the transition.
As discussed earlier section, the tapered CPW ground plane gives a better transition from
Fig. 4.13. Photograph o f the CPW circuits with tapered ground planes for
back-to-back CPW-to-waveguide transition.
unbalanced CPW to balanced CPS transmission line. Fig. 4.13 shows the photograph o f
the CPW circuits with tapered ground plane for the back-to-back: transition. The air­
bridges are added to this circuit before it is mounted on a trenched copper block.
Fig. 4.14 shows the measured results o f the back-to-back transition with tapered
CPW ground planes. The results show 12% bandwidth with return loss less than -15dB.
83
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
•10
— -15
-20
e -25
-35
-40
8
8.5
9
9.5
10
10.5
11
11.5
12
12.5
Frequency [GHz]
Fig. 4.14. Measured results o f the back-to-back CPW-to-waveguide
transition with tapered CPW ground planes.
The insertion loss is improved to be less than —1.0 dB for the bandwidth with the return
loss less than -15dB. For some applications, the lower insertion loss is more critical than
the bandwidth and this system is capable o f providing low insertion loss transition.
4.3 CPW-fed Quasi-Yagi Antenna
Though CPW is commonly used in high frequency front end, the antenna based
on CPW has not been well established. The twin-slots antennas combined with diodes
[4.27] are frequently used in millimeter-wave imaging array, but it often requires either
84
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
multi-layered substrate [4.28] or silicon lens [4.29] to sufficiently suppress the surfacewave o f substrate. The quasi-Yagi antenna, on the other hand, partially utilizes the
substrate’s surface-wave and directs them into the endfire radiation direction. The
antenna is realized on high-permittivity substrate and it makes readily compatible to the
system like MMIC and millimeter-wave imaging arrays.
The CPW-fed quasi-Yagi antenna is built on the high-permittivity substrate (RT
Duroid, er=10.2). The antenna’s dimensions are the same as the ones in the microstrip-fed
quasi-Yagi antenna. The complete dimensions o f the antenna are (unit: mm) Wo = 1.44,
Wi = W2 = 0.3, Wdrv = Wdrc = 0.6, So = 0.48, S, = S2 -0.15, U p = Lfcd = 2.0, Ua> = 1.5,
Ldrc= 3.3, Ldrv = 9.35, Dref = 3.9, Ddrv = 3.0, and DSUb = 1-7.
®-20
-25
-30
8
8.5
9
9.5
10 10.5 11
Frequency [GHz]
11.5
12
12.5
Fig. 4.15. Measured insertion loss o f the CPW-fed quasi-Yagi antenna.
85
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 4.15 shows the measured insertion loss o f the antenna. The antenna achieves
20 % bandwidth (8-9.7 GHz) with return loss less than -1 0 dB. Though this bandwidth is
smaller than the 40-50 % bandwidth in microstrip-fed quasi-Yagi antenna but it is still
wider than dipole antenna fed by CPW. The radiation pattern is measured in Fig. 4.16.
The radiation patterns show front-to-back ratio better than 20 dB and cross polarization
better than -1 6 dB. This radiation pattern is quite good compared to 8 % o f a folded-slot
antenna fed by CPW [4.30].
In order to improve the antenna’s performance for broader bandwidth or better
so
135
H_CO
E .c ro ss
H .c ro ss
45
-15
180
225
315
270
Fig. 4.16. E- and H-plane radiation patterns o f both co- and cross-polarization o f
the CPW-fed quasi-Yagi antenna.
86
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
symmetry in E- and H-plane radiation patterns, a high-performing full-wave simulator is
compulsory. However, the accurate simulation o f the planar CPW circuits is still a
challenge for currently existing numerical methods. As discussed in chapter 2, FDTD was
the most effective numerical method both in accuracy and speed for the simulation o f
microstrip-fed quasi-Yagi antenna. However, CPW is often overmoded and time-domain
numerical method cannot clearly distinguish the modes in frequency-domain. This often
leads to the insufficient accuracy in FDTD simulation of CPW-fed planar antennas.
FEM-based numerical software is comprehensible with modes but the poor performance
o f the radiation boundaries makes the antenna simulation by the method significantly
unreliable.
87
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
REFERENCES 4
[4.1] C. P. Wen, “Coplanar waveguide: A surface strip transmission line suitable for
nonreciprocal gyromagnetic device applications,” IEEE Trans. Microwave Theory Tech.,
vol. 17, pp. 1087-1090, Dec. 1969.
[4.2] D. Prieto, J. C. Cayrou, J. L. Cazaux, T. Parra and J. Graffeuil, “CPS structure
potentialities for MMICs : a CPS/CPW transition and a bias network”, IEEE MTT-S Int.
Microwave Symp. Dig., Baltimore, MD, vol.l, pp. 257-260, June 1998.
[4.3] L. Zhu and K. Wu, “Hybrid FGCPW/CPS scheme in the building block design of
low-cost uniplanar and multilayer circuit and antenna,” IEEE MTT-S Int. Microwave
Symp. Dig., Anaheim, CA, vol. 3, pp. 867-870, June 1999.
[4.4] K. Tilley, X. -D . Wu and K. Chang, “Coplanar waveguide fed coplanar strip dipole
antenna”, Electronics Letters, vol. 30, no. 3, pp. 176-177, Feb. 1994.
[4.5] Y. -D . Lin and S. -N . Tsai, “Coplanar waveguide-fed uniplan ar bow-tie antenna,”
IEEE Trans. Antennas Propagat., vol. 45, pp. 305-306, Feb. 1997.
[4.6] A. T. Kolsrud, M. -Y. Li and K. Chang, “Dual-frequency electrically tunable CPWfed CPS dipole antenna”, Electronics Letters, vol. 34, no. 7, pp. 609-611, April 1998.
[4.7] H. Ogawa and A. Minagawa, “Uniplanar MIC balanced multiplier - a proposed new
structure for MIC’s,” IEEE Trans. Microwave Theory Tech., vol. 35, pp. 1363-1368, Dec.
1987.
[4.8] T. Q. Ho and S. M. Hart, “A broad-band coplanar waveguide to slotline transition,”
IEEE Microwave and Guided Wave Lett., vol. 2, pp. 415-416, Oct. 1992.
88
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[4.9] K. P. Ma, Y. Qian and T. Itoh, “Analysis and applications o f a new CPW-slotline
transition,” IEEE Trans. Microwave Theory Tech., vol. 47, pp. 426-432, April 1999.
[4.10] J. Sor, Y. Qian and T. Itoh, “Coplanar waveguide fed quasi-Yagi antenna”,
Electronics Letters, vol. 36, no. 1, pp. 1-2, Jan. 2000.
[4.11] C -Y . Lee, Y. Liu and T. Itoh, “The effects o f 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.
[4.12] J. Lee, H. Lee, W. Kim, J. Lee and J. Kim, “Suppression o f cpupled-slotline mode
on CPW using air-bridges measured by picosecond photoconductive sampling,” IEEE
Microwave and Guided Wave Lett., vol. 9, pp. 265-267, July 1999.
[4.13] H. Shigesawa, M. Tsuji and A. A. Oliner, “Conductor-backed slotline and coplanar
waveguide: dangers and full-wave analysis,” IEEE M TT-S Int. Microwave Symp. Dig.,
New York, NY, pp. 199-202,1988.
[4.14] M. Tsuji, H. Shigesawa and A. A. Oliner, “Simultaneous propagation o f both
bound and leaky dominant modes on conductor-backed coplanar strips,” IEEE MTT-S
Int. Microwave Symp. Dig., Atlanta, GA, vol. 3, pp. 1295-1298, June 1993.
[4.15] W. E. McKinzie and N. G. Alexopoulos, “Leakage losses for the dominant mode
of conductor-backed coplanar waveguide,” IEEE Microwave and Guided Wave Lett., vol.
2, pp. 65-66, Feb. 1992.
[4.16] Y. Liu and T. Itoh, “Leakage phenomena in multilayered conductor-backed
coplanar waveguides,” IEEE Microwave and Guided Wave Lett., vol. 3, pp. 426-427,
Nov. 1993.
89
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[4.17] C. -C. Tien, C. -K . C. Tzuang and S. T. Peng, “Effect o f finite-width backside
plane on overmoded conductor-backed coplanar waveguide,” IEEE Microwave and
Guided Wave Lett., vol. 3, pp. 259-261, Aug. 1993.
[4.18] 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, Sep. 1993.
[4.19] W. -T. Lo, C. -K . C. Tzuang, S. T. Peng, C. -C . Tien, C. -C . Chang and C- C.
Chang, “Resonant phenomena in conductor-backed coplanar waveguides (CBCPW’s),”
IEEE Trans. Microwave Theory Tech., vol. 41, pp. 2099-2108, Dec. 1993.
[4.20] M. Yu, R. Vahldieck and J. Huang, “Comparing coax launcher and wafer probe
excitation for lOmil conductor backed CPW with via holes and airbridges,” IEEE MTT-S
Int. Microwave Symp. Dig., Atlanta, GA, vol.2, pp. 705-708, 1993.
[4.21] C. -C. Tien, C. -K. C. Tzuang and J. Monroe, “Effect o f lateral walls on the
propagation characteristics of finite-width conductor-backed coplanar waveguides”,
Electronics Letters, vol. 29, no. 15, pp. 1357-1358, July 1993.
[4.22] N. K. Das, “Methods o f suppression or avoidance o f parallel-plate power leakage
from conductor-backed transmission lines,” IEEE Trans. Microwave Theory Tech., vol.
44, pp. 169-181, Feb. 1996.
[4.23] Y. Liu, K. Cha and T. Itoh, “Non-leaky coplanar (NLC) waveguides with
conductor backing,” IEEE Trans. Microwave Theory Tech., vol. 43, pp. 1067-1072, May
1995.
90
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[4.24] D. R. Jahagirdar and R. D. Stewart, “Nonleaky conductor-backed coplanar
waveguide-fed rectangular microstrip patch antenna,” IEEE Microwave and Guided
Wave Lett., vol. 8, pp. 115-117, Mar. 1998.
[4.25] K. -P . Ma, F. -R . Yang, Y. Qian and T. Itoh, “Nonleaky conductor-backed CPW
using a novel 2-D PBG lattice,” 1998 Asia Pacific Microwave Conf. Dig., Yokohama,
Japan, pp. 509-512, Dec. 1998.
[4.26] W. Simon, M. Werthen and I. Wolff, “A novel coplanar transmission line to
rectangular waveguide transition,” IEEE MTT-S Int. Microwave Symp. Dig., Baltimore,
MD, vol. 1, pp. 257-260, June 1998.
[4.27] A. R. Kerr, P. H. Siegel and R. J. Mattauch, “A simple quasi-optical mixer for
100-120GHz,” IEEE MTT-S Int. Microwave Symp. Dig., San Diego, CA, pp. 96-98, June
1996.
[4.28] Y. Qian and E. Yamashita, “A 60GHz imaging array using CPW-fed twin-slots on
multilayered substrates,” IEEE MTT-S Int. Microwave Symp. Dig., San Francisco, CA,
pp. 1007-1010, June 1996.
[4.29] G. P. Gauthier, W. Y. Ali-ahmad, T. P. Budka, D. F. Filipovic and G. Rebeiz, “A
uniplanar 90-GHz shottky-diode millimeter-wave receiver,” IEEE Trans. Microwave
Theory Tech., vol. 43, pp. 1669-1672, July 1995.
[4.30] H. -S . Tsai and R. York, “FDTD analysis o f CPW-fed folded-slot and multipleslot antennas on thin substrates,” IEEE Trans. Antennas Propagat., vol. 44, pp. 217-226,
Feb. 1996.
91
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter §
Quasi-Yagi Antenna Applications in Microwave Photonics
System
5.1 Millimeter-wave Generation by Photonic Devices
5.1.1 Background and Motivation
The klystron electron tubes were a significant invention in 1940s as it provided
the first practical microwave source that was suitable both electrically and mechanically
for applications such as aircraft navigation and airport landing systems [5.1]. Those
microwave tubes have achieved extremely high power (lOkW at 35GHz) [5.2], very low
phase noise [5.3] and small size as well. However, the microwave tubes require
extremely large DC voltages, typically a couple o f kV, and the size and weight o f the DC
power supply limits the application o f microwave tubes in most o f the commercial
microwave systems.
Due to the continuous improvement in the output power and phase noise, solidstates devices are the primary choices in millimeter-wave sources for high frequency
communication systems. Two-terminal solid-state devices such as Gunn devices and
IMPATT (impact avalanche transit time) diodes are capable of producing very large
power such as over 20dBm at 100 GHz [5.4]. Three-terminal devices such as InP or
92
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
GaAs-based HEMT (high electron mobility transistor) and HBT (heterojunction bipolar
transistor) have achieved low phase noise as millimeter-wave sources [5.5], Those solidstate devices are compact and capable o f producing high-power. However, those solidstate device oscillators are operated only at a single frequency and tunable oscillators are
very difficult to build at millimeter-wave frequency. Thus a broadband millimeter-wave
system using solid-state device oscillators can be significantly costly and complicated.
Though electro-optic theories and techniques have been developed for many years
for fiber-optic communication networks [5.6], the idea of using photonic devices as
microwave/millimeter-wave local oscillators is relatively new. In order to generate the
millimeter-wave signals with photonic devices, we need optical sources including lasers
(preferably phase-locked), amplifiers, filters, etc. in addition to the high-power/highspeed photodetectors. This system requirement would be too costly for many commercial
microwave systems.
However, an application such as millimeter-wave large antenna arrays for radio
astronomy observatory [5.7] requires millimeter-wave sources for each antenna receiver
that are spread out in a geographically large area. The oscillator frequency is desired to be
extremely high (close to Tera-hertz), and the frequency bandwidth is required to be
extremely wide (entire waveguide band, 40%). In this type of system, multiple solid-state
devices have to be prepared for a single waveguide band and the system can become
exceedingly costly. On the other hand, the photodetectors can be connected by optical
fibers to a centrally located optical source and such a system can be less expensive and
less complicated. In addition, oscillation frequency o f photodetectors is determined by
93
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the laser inputs that are mixed at the detectors. It is easier by a number of degrees to tune
the laser frequencies than implementing multiple oscillators into one circuit. The high
tunability o f the photonic local oscillators is tremendous advantage for a broadband
system.
5.1.2 VMDP Measurement Results
High-power and high-speed photodetector has been one o f the key elements in
fiber optics communication networks and a number o f device structures have been
:__ non
Optical
Band-pass
Filter
Polarization
Controller
3dB Coupler
1=5
Polarization
Controller
Power
Meter
0
A]
EDFA
Microwave
Spectrum
Analyzer
Power
Sensor
Attenuator
Picoprobe
ineim
(GGB)
HP4145B
DC Bias
-----
I
I
—
VMDP
Fig. 5.1. A schematic o f the W-band VMDP measurement setup.
94
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Pickup
Head
researched [5.8]. In a quest o f high-power and broad bandwidth structure, VMDP
(Velocity Matched Distributed Photodetector) has been proposed and demonstrated [5.9].
As described in chapter one and Fig. 1.4, VMDP is an array o f MSM photodiodes fed in
series by a single optical waveguide. The MSM photodiode when fabricated on LT-GaAs
is expected to have 3dB bandwidth o f 510GHz [5.10]. Using the high-speed MSM
photodiode fabricated on InP, the W-band VMDP is developed [5.11]. The complete
description of design, fabrication process and experimental results o f W-band VMDP are
found in [5.12]. In this section, the measurement results o f W-band VMDP are presented
and discussion on the future application o f the device in millimeter-wave local oscillator
follows.
Fig. 5.1 shows the schematic of the measurement setup o f W-band VMDP. Light
from two external cavity tunable lasers are combined by a 3 dB coupler. The combined
signals are amplified by a high power EDFA. The signal is coupled to the VMDP through
a fiber pickup head. A commercial probe with a built-in Bias-T was used to probe the
device, converting CPS to W-Band (75-100 GHz) waveguide. The millimeter-wave
output signal was delivered to a calibrated W-band power sensor and measured by a
power meter.
Fig. 5.2 shows the frequency response o f the W-band VMDP. This response
includes both VMDP and W-band probe. The solid lines show the broadband response of
the device at 3V and 5V DC bias, with the same input optical power of 13.25 dBm. The
wavelength of one o f the external cavity lasers is tuned to scan output millimeter-wave
frequency for entire W-band, from 75 to 110 GHz. The figure indicates that the signal
95
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
roll-off was small across the entire W-band. Maximum output power measurement was
conducted at two different frequencies and the maximum power o f -23 dBm at 75 GHz,
and -26 dBm at 95 GHz were measured. Taking into account the W-band probe, with a
nominal insertion loss o f 2dB from manufacturer’s data, the VMDP maximum output
signals are -21 dBm at 75 GHz and -24 dBm at 95 GHz.
The W-band VMDP demonstrates one o f the highest power obtained by optical
devices at 100GHz range. In comparison with up-to-date solid-state devices, the power
•10
Input optical power * 18.7 dBm
E
OQ -20
■O
IO
IL
0)
>
5VBias, I,** 11 mA
Input optical power >17 dBm
6V Bias, I*.* 15mA
-30
to
^ -40
5V Bias, U = 5.69 mA
3V Bias, L, = 5.33 mA
0)
E
•60
70
80
90
100
110
Frequency (GHz)
Fig. 5.2. Output power o f W-band VMDP. The optical input power = 13.25dBm for
both solid and dashed line. The resnonse includes both VMDP and W-hand nrobe.
96
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
obtained from VMDP is about 40dB lower. The low output power o f the device
originates primarily from very small AC efficiency. The AC efficiency at 90 GHz with
input optical power o f 13.5 dBm and DC bias o f 5 V is derived 0.03 % from Fig. 5.2, AC
output power and 4.2mA DC photocurrent. This efficiency is extremely low compared to
nominally several percent o f the solid-state devices at the same frequency.
However, a VMDP single device is capable o f producing same level o f power
over entire waveguide band. Microwave solid-state devices can never achieve this wide
tunable frequency with a single device and this is fundamental advantage o f the photonic
local oscillator approach. The phase noise o f the devices is expected to be low [5.12], but
measurement o f the phase noise remains as a future research. A number o f techniques
including reduction o f diffusion time are currently researched for further improvement of
the device efficiency.
5.2 Quasi-Yagi Waveguide Transition Combined with VMDP for
Application in Microwave Photonics
As previously discussed in chapter 3 and 4, due to its low loss characteristics the
rectangular waveguide is still a critical component in many millimeter-wave systems. The
receivers for millimeter-wave antenna array system are no exception at this point as the
receiver system is sketched in Fig. 5.4. High-frequency mixer utilized in this system is
SIS (superconductor insulator superconductor) tunnel junction mixer that has low noise
temperature and small conversion loss [5.13] [5.14]. The system is built with waveguides
97
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and components compatible with waveguide all the way up to IF output [5.14]. VMDP,
on the other hand, has planar microwave structure where CPS (coplanar stripline) is
utilized to collect the microwave signal from array o f photodiodes and the broadband
CPS-to-waveguide transition is indispensable to be integrated into a receiver system.
The W-band probe used in the measurement setup in Fig. 5.1 is a commercially
available CPS-to-waveguide transition. The manufacture’s spec sheet indicates that the
probe works fine with low insertion loss for the entire W-band. The problem with this
transition however, is that probing on a tiny millimeter-wave circuit is a standard
technique in laboratory but not in a practical application. When applied in practice, in
addition to the inconvenient setup, this approach requires a constant attention for
mechanical vibration and other environmental change. A more mechanically strong
RFin
[ran °
S S M xbt
\
LOin
KEMTAnp
-------- (
Photodetector
-------- (
D
D
3dBCoiffer
Optical
fiber
CWLaser
Fig. 5.3. Schematic o f a receiver for millimeter-wave radio astronomy and photonic
98
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
transition structure, ideally monolithic integration o f photonic devices and waveguide
transition is desired.
As we have witnessed in earlier chapters, quasi-Yagi waveguide transition has
compact design, broadband characteristics and high-permittivity substrate, and appears to
be a perfect candidate for a monolithic integration o f optical devices into waveguide
system in practical applications. Shown earlier in Fig. 1.6, the monolithic integration of
quasi-Yagi waveguide transition with VMDP is proposed. In this section, the quasi-Yagi
waveguide transition suitable for VMDP structure is presented and discussed.
5.2.1 Transition Design and Scalemodel Demonstration
Fig. 5.4. Schematic view o f back-to-back CBCPS-to-waveguide transition.
99
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The VMDP requires a heat sink to dissipate the heat generated from the
photodiodes. The conductor backed CPS can provide a good heat sink for device and a
well-defined reflector o f quasi-Yagi antenna. The conductor backed CPS (CBCPS) is
essentially a coupled microstrip line and only the latter term was used in chapter three.
However, we are only interested in the odd mode o f coupled microstrip fine and we refer
to this mode as CBCPS in this chapter. The CBCPS-to-waveguide transition using the
quasi-Yagi antenna is a part o f the microstrip-to-waveguide transition and its broadband
characteristics and optimization is discussed in chapter three. The X-band prototype of
the back-to-back CBCPS-to-waveguide transition is built and presented in this section.
Fig. 5.4 shows the schematic view o f the back-to-back CBCPS-to-waveguide
0
-5
-10
-15
c -25
Return loss
Insertion loss
S *30
-35
•40
8
8.5
9
9.5
10
10.5
11
11.5
12
12.5
Frequency [GHz]
Fig. 5.5. Measured results o f the back-to-back CBCPS-to-waveguide transition.
100
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
transition. A metal block supports the CBCPS circuit in the middle o f waveguide. The
ground plane o f the CBCPS is extended beyond the metal block end to make transition
insensitive to positioning error. The dimensions o f the quasi-Yagi antenna (Fig. 1.1) are
(unit: mm): Wdn = Wdir = W5 = 0.6, W6 = S5 = S6 = 0.3, Sref = 3.0, Sdjr = 2.45, Ssub = 2.04,
Ldn = 8.55 and L^r = 3.12. The characteristic impedance o f CBCPS line used here is
calculated by HFSS as 72Q. A transition with 50Q characteristic impedance can also be
achieved with the same antenna dimension and tapered CBCPS line at the input o f the
antenna.
Fig. 5.5 shows the measured results o f the back-to-back transition. A bandwidth
o f 26% is achieved for the return loss less than -lOdB. The insertion loss ranges from 0.5 to -1.2 dB for return loss less than —lOdB. The optimization in the transition design is
expected to further improve the bandwidth o f the transition. Two-terminal devices can be
easily integrated on the CBCPS quasi-Yagi antenna and the broadband and relatively low
loss characteristics o f the transition should find a wide range o f applications.
The W-band transition using 80pm thick InP substrate (Sr = 12.4) is also designed
and optimized. Fig. 5.6 shows the simulation results o f the W-band back-to-back
transition. The return loss less than -lO dB is achieved for the entire W-band, and the
insertion loss is quite small though the simulation does not include dielectric and metal
loss. With the dielectric and metal loss included, the insertion loss is expected to be
higher but return loss is generally not affected. The complete dimensions of the quasiYagi antenna (Fig. 1.1) are (unit: pm): Wdn = Wdir = W5 = 40, W6 = 33, S5 = S6 = 30, Sref
101
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
= 290, Sdir = 190, Ssub = 240, L*, = 975 and Ld,r = 385. The future fabrication o f this
transition requires backside metalization and thinning the frail InP substrate.
-
|
i
2 -15
\
1
1
•
•
4
•
•
••
f
•
%
S
:
•
.
4
•
_
*.
•
i
!
1
1
•5
5F
2.-10
| -i------
4
■
%--------------
0
%
I * 20
t
f*
*
|
•
*
i
I - 25
t
m
c
........Return loss
------Insertion loss
"30 i
-35
i
i
i
80
85
i
-40
75
i
!
|
j
,
90
95
100
105
110
Frequency [GHz]
Fig. 5.6. Simulation results o f the W-band back-to-back CBCPS-towaveguide transition using InP substrate.
102
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
REFERENCES 5
[5.1] J. H. Bryant, “Microwave technology and careers in transition: the interests and
activities o f visitors to speny gyroscope company’s klystron plant in 1939-1940,” IEEE
Trans. Microwave Theory Tech., vol. 38, pp. 1545-1558, Nov. 1990.
[5.2] M. L. Lopin, B. A. Belyavsky, K. G. Simonov, V. A. Cherepenin, A. D.
Zakurdayev, B. S. Grishin, A. A. Negiev and A. S. Pobedenostesev, “High-power
millimeter-wave tubes,” IEEE MTT-S Int. Microwave Symp. Dig., Atlanta, GA, vol.l, pp.
1119-1121, June 1993.
[5.3] C. Nilsen, M. Viant and M. Wong, “Low moise MM-wave transmitter tube
oscillators,” Microwave Journal, vol. 31, pp. 91-97, July 1988.
[5.4] H. Eisele and G. I. Haddad, “Two-terminal millimeter-wave sources,” IEEE Trans.
Microwave Theory Tech., vol. 46, pp. 739-746, June 1998.
[5.5] H. Wang, K. W. Chang and B. R. Allen, “Low phase noise millimeter-wave
frequency sources using InP-based HBT MMIC technology,” IEEE Jour. Solid-state
Circ., vol. 31, pp. 1419-1425, Oct. 1996.
[5.6] B. E. A. Saleh and M. C. Teich, Fundamentals o f photonics. John Wiley & Sons
Inc., New York, 1991.
[5.7] John Payne, Bill Shillue, and Andrea Vaccari, “ Photonic Techniques for Use on the
Atacama Large Millimeter Array,” IEEE MWP'99, Melbourne, Australia, November 1719, 1999.
103
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[5.8] K. Kato, “ Ultrawide-band, high-frequency photodetectors,” IEEE Transactions on
Microwave Theory and Techniques, vol. 47, no. 7, pp. 1265-1281, July 1999.
[5.9] L. Y. Lin, M. C. Wu, T. Itoh, T. A. Vang, R. E. Muller, D. L. Sivco, A. Y. Cho,
“High-power
high-speed
photodetectors-design,
analysis,
and
experimental
demonstration, ” IEEE Transactions on Microwave Theory and Techniques, vol. 45, pp.
1320-1331, Aug. 1997.
[5.10] S. Y. Chou and M. Y. Liu, “Nanoscale Tera-hertz metal-semicondcutor-metal
photodetectors,” IEEE Transactions on Microwave Theory and Techniques, vol. 28, pp.
2358-2368, Oct. 1992.
[5.11] T. Chau, 1.55 um Wavelength InGaAs/InAlAs/InP Velocity-Matched Distributed
Photodetectors, Ph.D dissertation, University o f California, Los Angeles, 2000.
[5.12] C. H. Cox, “Gain and noise figure in analogue fiber-optic links,” IEE ProceedingsJ., vol. 139, pp. 238-242, 1992.
[5.13] R. Blundell and C. -Y . E. Tong, “Submillimeter receivers for radio astronomy,”
Proceedings o f IEEE., vol. 80, pp. 1702-1720, Nov. 1992.
[5.14] J. M. Payne, J. W. Lamb, J. G. Cochran and N. Bailey, “A new generation o f SIS
receivers for millimeter-wave radio astronomy,” Proceedings o f IEEE., vol. 82, pp. 811823, May 1994.
104
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 6
Conclusions
A very compact and simple planar antenna based on the modification o f the
classic Yagi-Uda antenna has been presented. The antenna achieves extremely wide
frequency bandwidth and good radiation characteristics in terms o f beam pattern, frontto-back ratio, cross polarization and low mutual coupling. The antenna experimentally
demonstrated a bandwidth of 48 % for a VSWR < 2, better than 12 dB front-to-back
ratio, gain between 3-5 dBi, and nominal radiation efficiency o f 93%.
Additionally,
mutual coupling between antennas in a horizontal and vertical configuration was
measured to be better than -20 dB in the entire operating band. Finally, a higher gain
version o f the quasi-Yagi antenna has been presented. In this case, measured gain varies
between 5-7 dBi across the operating bandwidth, where the increased gain has been
achieved at the cost o f reduced bandwidth. Adding additional directors can potentially
further increase the gain.
The excellent radiation properties o f this antenna make it ideal as either a stand­
alone antenna with a broad pattern or as an array element. We believe that this antenna
should find wide applications in wireless communication systems, power combining and
phased arrays, as well as millimeter-wave imaging arrays. The broad pattern, low mutual
coupling and wide instantaneous bandwidth allow this antenna to be incorporated into
multi-frequency phased arrays with very large scanning capability.
105
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Utilizing this quasi-Yagi antenna, novel microstrip-to-waveguide and CPW-towaveguide transitions have been proposed and demonstrated. The transitions’
mechanisms are discussed in details for individual components to understand its
broadband characteristics. Design guidelines for the optimization of the transitions are
given and X-band prototype transitions are built on two different substrates.
Measurement results show both broadband and relatively low loss characteristics o f the
transition when low dielectric loss substrate is used. A tolerance study regarding
packaging issues is presented. The transitions provide compact design and require much
less fabrication effort than any existing techniques. These new transitions are readily
compatible with MMIC and RF photonics technology and should find a wide range o f
applications.
The application o f quasi-Yagi antenna in microwave photonics system was
discussed. The photonic local oscillator system using broadband optical devices such as
VMDP has an intrinsic advantage in high-tunability o f oscillation frequency. To meet the
specifications in current receiver systems for millimeter-wave radio astronomy, AC
efficiency o f the device needs to be improved in particular. A CBCPS-to-waveguide
transition can help the device compatibility in receiver systems but further
demonstrations including W-band fabrication o f the transition are needed in future
research.
106
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Appendix A
Calculation of substrate mode power excited by dipole and slot
antennas on dielectric substrates
Printed dipoles on a dielectric substrate are important elements in integrated
antennas especially in high frequency such as millimeter-wave imaging arrays. Antennas
on finite thickness dielectric substrate often suffer from the substrate modes that are
generated by radiating elements. The techniques and theories for the calculation o f the
substrate mode power [A.1] [A.2] [A.3] are summarized and presented in this appendix.
The fundamental discussion o f the slab waveguide modes are presented in the first
section and the discussions on the application o f the reciprocity theorem to elementary
nc
fc sp M *
Fig. A.l Sketch o f a slab waveguide
107
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
dipoles and slots are followed.
A.I. Slab waveguide modes
In accordance with the wave equation, various transverse decay (y) and
propagation constants (k) for the slab waveguide in Fig. A .l are expressed by
2
kc
2/2
=nck
- pr*Z
2
K'f = n2fk 2 - p 2,
(A .l)
K) =n]k2- p z = - y ) ,
where the subscript s,f and c refer to substrate, film, and cover respectively.
A.1.1 TE modes
For the TE-modes, the field components are expressed as
H y = Ex = E, = 0, H x = -(/? / (Ofu)Ey, H , = (y / oj/j)dEy / dx,
(A.2)
with the Ey component obeying the wave equation
d-Ey /dx2 = ( P 2 —n 2k 2) / E y,
(A.3)
108
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The boundary conditions demand that Ey and BEy / dx be continuous across the
media boundaries.
Ey = Ec exp [~yc(x - h)],
E = E f cos( k
'
,
x
1
- fa),
E v = E s e x p ^ x ),
for h<x
for 0<x<h
(A.4)
for x<0
Application o f the boundary conditions yields the formulas for the phase shifts
ta n £ = ys !tcf
tan& = yc/Kf
(A.5)
and the dispersion relation
K f h - < j > s -<f>c = V 7 l ,
(A.6)
where the mode label v is an integer.
The relation between the peak fields is also obtained as
E ) W f - N 2) = E 2( n } - n 2) = E]{n2f - n 2e ),
(A.7)
The effective thickness o f the waveguide is given by
109
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. A.2. Effective guide thickness o f substrate modes. The phase shift at the
interface is caused by Goos-Hanchen shift.
heff=h +— +—
Y*
(A.8)
Yc
where the effective thickness o f the waveguide is sketched in Fig. A.2.
From Fig. A.2, we can predict the propagation constant p
f3 = co/vp = knf sin#
(A.9)
Only a discrete set o f angles isallowedfor guided modes. The condition for 0
phase shift caused at each interfacesshould
is that the
add up to multiple o f 2n for oneperiod o f
reflection when the waves propagate. Thus the self-consistency condition is expressed as
2knf h cos 0 - 2<f>s - 2<f)c = 2vn
(A. 10)
where v is an integer (0,1,2...) which identifies the mode number.
110
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A. 1.2 TM modes
For the TM-modes, the field components are expressed as
Ey = H x = H s =0, Ex = (ft I o)/u)Hy, E, = - { j / (OfX)dHy / dx,
(A .ll)
with the Hy component obeying the wave equation
d2H y /dx2 = ( 0 2- n 2k 2) / H y,
(A. 12)
The boundary conditions demand that H y and dHy /dx be continuous across the
media boundaries.
Hy = H c exp [~yc( x -/»)], for h<x
Hy = H f cos( K f
x
-
j ,
),
H y = H s expf/jX),
^ {
for 0<x<h
for x<0
Application o f the boundary conditions yields the formulas for the phase shifts
tan <t>s
= { n f
/ n s ) 2y s / K f
ta n £ ={nf / n c)2yc / / c f
and the dispersion relation
111
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Kf h-<t>s -<f)c = vn ,
(A. 15)
where the mode label v is an integer.
The relation between the peak fields is also obtained as
(A. 16)
The effective thickness for the TM modes are defined as
(A. 17)
A.2 Application o f the reciprocity theorem to elementary dipoles and
slots.
The power carried in surface-wave modes, radiated power, and the gain of
elementary printed circuit antennas can be computed with the aid o f Lorentz reciprocity
theorem which states that for the two sets o f current density sources J, or J 2 , the
generated fields ( £ , , / / , ) and ( E2, H 2) satisfy over a region V enclosed by a surface S
the following relationship.
112
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The calculation for TE surface-wave modes, where the excitation is provided by
an electric current element as shown in Fig. A.3 is explained briefly. The source /, at
x
=
x 0
produces a TE surface-wave field with transverse components £,, H x at x = 0 .
We consider a second surface-wave field E2, H 2 propagating from right to left in the
figure. The Lorentz reciprocity theorem is applied to a large box S yields the total TE
surface-wave mode power as
C * =3
where
^
(A. 19)
he are found in (A.6), (A. 8), respectively.
While, the total power in a TM mode is found as
C .. = ^
Si" 2f i r COS ^
4h e
(A.20)
where <pm , he, 0d are found in equations (A. 15), (A. 17), (A.9) respectively. For a
grounded substrate at x = 0 , the effective thickness would be
113
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. A.3. TE mode Reciprocity calculation for a dipole on an ungrounded
substrate.
h ~ = h +— ,
Yc
hefr = h n— -—,
YcRc
for TE mode
(A.21)
for TM mode
Similar calculations give the power coupled into surface-modes by an infinitesimal slot in
a grounded substrate. The total power normalized to the power the slot would radiate in
the free space is given by
p% =
( A-22)
16«.
114
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and
pTM _ 3 s rA Q
~ 16A,
(A 2 3 )
The results for the dominant surface-wave modes in dipole antennas are plotted and
shown in Fig. 2.1.
115
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
REFERENCES A
[A.l] N. G. Alexopoulos, P. B. Ketehi and D. B. Rutledge, “ Substrate optimization for
integrated circuit antennas,” IEEE Trans. Microwave Theory Tech., vol. 31, pp. 550-557,
July 1983.
[A.2] D. B. Rutledge, D.P. Niekirk, and D. P. Kashingam, “ Integrated-circuit antennas,”
in Infrared and Millimeter-waves Sereis, vol. 10, pp. 1-90, k.J. Button, Ed. New York:
Academic Press 1983.
[A.3] H. Kogelnik, “ Theory o f dielectric waveguides,” in Integrated Optics, T. Tamir,
Ed. New York: Springer-Verlag, 1975.
116
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Документ
Категория
Без категории
Просмотров
0
Размер файла
4 464 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа