Académique Documents
Professionnel Documents
Culture Documents
A Designer's Guide
Microstrip Patch Antennas:
A Designer's Guide
Dr R. B. Waterhouse
RMIT University
AlI rights reserved. No part of this publication may be reproduced, stored in a retrieval
system or transmitted in any form or by any means, electronic, mechanical, photo-copying,
microfilming, recording, or otherwise, without the prior written permission of the
publisher, with the exception of any material supplied specifically for the purpose of being
entered and executed on a computer system, for exclusive use by the purchaser of the
work.
Permissions for books published in the USA: perrn; ss; ODs@wkap corn
Permissions for books published in Europe: permissions@wkap.nl
Printed on acid-free paper.
Dedication
Chapter 1 Introduction 1
1.1 History 1
1.2 Advantages and Issues 7
1.3 Applications 10
1.4 Summary of Book 16
1.5 Bibliography 19
3.1 Introduction 69
3.2 Intuitive Procedures 70
3.3 Horizontally Coupled Parasitic Patches 75
3.4 Stacked Patches 83
3.5 Large Slot Excited Patches 111
3.6 Aperture Stacked Patches 123
3.7 Ultra-wideband ASPs 150
3.8 Summary 162
3.9 Bibliography 164
Chapter 1: Dr R. Waterhouse
Chapter 8: Dr R. Waterhouse
Acknowledgements
Firstly, I would like to thank all my colleagues and former students who
directly helped contribute to the material presented in this book. These people
are listed on page x and I very much enjoyed all the interactions we had over
the years. I would also like to thank all my technical colleagues who
contributed indirectly to this book by providing feedback on work presented
at conferences or reviewed in journals, or just discussed at technical meetings.
I would also like to thank members of the antenna community whose work
over the years inspired me to undertake some of the work presented herein. A
special thank you goes to the professorial staff from the University of
Massachusetts and their former graduate students for their kindness
throughout the years.
I would like to thank RMIT University for supporting the activities that led to
the body of work presented within this book. I would also like to thank the
institutions that helped fund much of this work, including the Australian
Research Council, CSIRO Australian National Telescope Facility, British
Aerospace and the Australian Army Agency.
I would also like to thank my family and friends whose support over the years
is much appreciated. Special mention must be made to two wonderful
woman: Ilona and Margaret for all their love and devotion to Dalma and
myself over the years. And finally a special thank-you goes to Renee for
suggestions on the title of the book - maybe next time.
Chapter 1 Introduction
1.1 History
1970s
In the mid-1970s probably saw the first real contributions to the area
of microstrip patch technology. Although there were some patents on this
subject filed many years beforehand, it was really in this timeframe when the
first significant advancements were made. In particular, two styles of
exciting/feeding a microstrip patch antenna were developed: the edge-fed
patch and the probe fed patch [1]. These two types of excitation methods are
really the 'god fathers' of all patches, highlighting many of the inherent
advantages of microstrip patch antennas and unfortunately some of their
disadvantages too. Despite the many cynical researcher and developer (there
are still some people today who do not see the all the potential glory of the
microstrip patch!), the early pioneers continued in their endeavors developing
patches of all shapes, linear and planar arrays of these elements [2], both large
and small and even microstrip patches that could be shaped to reside on the
curved surface of a missile: a 'wrap-around' antenna. Albeit, all these
developments involved printed antennas of extremely narrow operation
bandwidth, typically 1 - 2%, not at all suitable for most communication
systems and hence the early skepticism. It is interesting to note that one of
the first attempts to enhance the impedance bandwidth, stacking patches was
proposed in the late 1970s [3], however, the method was unsuccessful until
2
the nineties, when it was fmally mastered. One key disadvantage of the
microstrip patch antenna was the relative difficulty in accurately predicting its
input impedance response. This was because of its resonant nature and the
inhomogeneous medium in which the patch resided. Despite these difficulties
two modeling techniques were introduced with respect to the microstrip patch
in the 1970s: the transmission line model and the cavity model. These
analytical tools, although limited in what they can 'accurately' model,
provided a means of understanding the behavior of the microstrip patch and
how its performance can be improved upon, to a limited extent. To this day,
no other software tools can provide the same intuition to the designer as the
transmission line and cavity models, although some of the more advanced
software packages, with their impressive visualization interfaces can provide
many clues to still unsolved issues related to the microstrip patch.
1980s
4
1990s
2000s
By the tum of the century microstrip patches were pretty well being
used in most 'free-space ' communication systems or at least one of the first
considered for them! (Probably with the exemption of multi-octave radar and
earth stations) Further enhancements in bandwidth have been achieved (some
using variations of the excitation method). More and more examples of
integration are surfacing, including for spatial power combiner applications
[26], as well as other high gain solutions that had minimal feed network loss.
Utilizing Photonic Bandgap (PBG) structures to enhance the efficiency and
improve the overall radiation performance of the patch antenna have been
explored [27]. Attempts have been made to use optimization procedures (for
example, genetic algorithms) to ensure the 'optimum' performance could be
achieved [28], [29].
Now that a critique of the R&D that has impact the area of microstrip
patches has been given, its appropriate to give the advantages and demerits of
this form of radiator in some detail. Of course this list is dynamic and so
some issues have resolved, at the expense of some other feature. Having said
this, these disadvantages will still be listed here, with their possible
resolutions.
Advantages
--
Stba.!,a:.
rr ~
_v..
e,
GI ~P tarlO
8
Efficient: Microstrip patch antennas are basically efficient radiators.
This can be attributed to one simple fact: they are resonant-style radiators.
Resonant antennas are more efficient than their broadband traveling-wave
style counterparts. Compared to other resonant antennas microstrip patches
are generally reasonably efficient, although this will depend on whether good
microstrip procedures are held to in the design procedure. There are three
loss mechanisms that need to be addressed in printed antenna designs:
conductor loss; dielectric loss and surface wave loss. All are material
dependent and so careful selection of the material can lead to good overall
efficiency. The efficiency of the microstrip patch is dependent also upon
what feeding procedure is adopted: direct contact techniques are more
efficient than the non-contact methods.
Disadvantages
9
Size (again!): The size of a microstrip patch antenna appears in
both its advantages and disadvantages, simply because there are applications
where even this form of printed antenna is too large. These applications are
associated with the wireless industry, in particular for the development of
hand-held communication terminals operating at frequencies below 2 GHz.
As a microstrip patch is a resonant antenna, it must have a length comparable
to half a guided wavelength at the relevant operation frequency. Of course at
low microwave frequencies and for small volumes, as required for handset
terminals, there is not a simple solution to this problem using microstrip patch
technology. Over the years several techniques have been developed and are
currently in use for commercial mobile communication handsets and these
will be presented in a later chapter.
1.3 Applications
10
in more prevalent systems. Fixed frequency/low bandwidth Radar
applications were such examples where microstrip patch technology could
readily be incorporated. Several spacebome systems were launched using
microstrip patches as the fundamental radiator to monitor environmental
issues on Earth [32]. The light-weight feature of printed antenna technology
was very advantageous reducing the payload of the launched satellites.
Figure 1.3.1 shows an example of a microwave patch array used for this
application.
More recently low cost motion detection systems and tagging systems
have been developed using microstrip patch arrays. These systems do not
require large bandwidths and probably the most important issue is to keep the
cost of development low. As microstrip technology allows for the antenna to
be easily integrated with the feed network and any active devices (surface
mount oscillators, mixers and detectors), patch antennas are an obvious
choice. Also, these systems tend to be in the moderate microwave frequency
bands (5 GHz - 10 GHz) and so the overall sizes of these antennas are quite
small, although there has also been many millimeter-wave printed antennas
developed for collision avoidance systems. Figure 1.3.2 shows an example of
a microstrip patch array developed for this application.
11
Figure 1.3.2 MicrostripPatch Array for Collision Avoidance System
Missiles: One of the first applications of the microstrip patch was for
telemetry systems communicating with missiles or other forms of projectiles.
The requirements for such antennas on the missiles were/are: low proflle and
conformal nature (they could not impact the aerodynamics of the projectile)
and a broad radiation pattern (to maximize communication spatial coverage).
Thus from the previously given characteristics of the microstrip patch it is
apparent that this antenna can easily conform to these specifications.
Microwave dielectric layers used in the construction of microstrip patches are
usually soft laminates and can easily be molded to comply the skin of a
missile. An added feature of the microstrip patch specifically applicable to
military systems is that it has a low impedance bandwidth, or large
electromagnetic signature. This means that it only operates effectively at a
very narrow range of frequencies and therefore is less susceptible to possible
jamming frequencies or interference. The initial telemetry applications were
in the mid to late 1970s and variations of the 'wrap-around' patch are still
being used today. A photograph of a wraparound microstrip patch antenna
array is shown in Figure 1.3.3.
12
Figure 1.3.3 TelemetryMicrostripPatch Antenna
14
Figure 1.3.6 Mobile Communications Base Station Antenna
15
Figure 1.3.7 Internal Antenna for Cellular Systems
16
packages available today that accurately analyze these printed antennas and a
microstrip patch antenna designer will require one of these to successfully
design their antenna . Having said this, the author encourages the reader to
develop their own analysis tool based on these well-established techniques to
get a better appreciation of what is involved in developing these somewhat
complicated software tools. Also fundamental theories, such as the cavity
model and the transmission line model should be consulted but will not be
presented here in their entirety. As stated before, one of the objectives here is
to give a designer's prospective of microstrip patch antennas.
1.5 Bibliography
[1] R. E. Munson, "Conformal Microstrip Antennas and Microstrip Phased Arrays,"
IEEE Trans. Antennas Propagat., vol. AP-22, pp. 74 - 78, January 1974.
[2] R.1. Mailloux, 1. F. Mcllvenna and N. P. Kemweis, "Microstrip array technology,"
IEEE Trans. Antennas & Propagation , vol. 29, pp. 25 - 37, Jan. 1981.
[3] S. A. Long and D. M. Walton, "A dual-frequency stacked circular-disc antenna,"
IEEE Trans. Antennas Propagat., Vol. AP-27, pp. 270-273, March 1979.
[4] D. M. Pozar, "A Microstrip Antenna Aperture Coupled to a Microstrip Line,"
Electronics Letters, Vol. 21, pp. 49 - 50, January 1985.
[5] D. M. Pozar and B. Kaufman, "Increas ing the Bandwidth of a Microstrip Antenna by
Proximity Coupling," Electronics Letters, Vol. 23, pp. 368 - 369, April 1987.
[6] 1. R. Mosig and F. E. Gardiol, "General Integral Equation Formulation for Microstrip
Antennas and Scatterers,' Proc. lnst. Elect. Eng., pt. H, Vol. 132, pp. 424 - 432,
1985.
[7] D. M. Pozar, "Input Impedance and Mutual Coupling of Rectangular Microstrip
Antennas," IEEE Trans. Antennas Propagat., Vol. AP-30, pp. 1191 - 1196,
November 1982.
[8] D. M. Pozar and D. H. Schaubert, "Scan blindness in infinite arrays of printed
dipoles ," IEEE Trans. Antennas & Propagation, vol. 32, pp. 602 - 610, June 1984.
[9] J. T. Aberle and D. M. Pozar, "Analysis of infinite arrays of probe fed rectangular
microstrip patches using a rigorous feed model", Proc. Inst. Elec. Eng., Pt. H, vol.
136, pp. 110-119, April 1989.
[10] D. L. Rascoe , et al, "Ka-band MMIC beam-steered transmitter array", IEEE Trans.
Microwave Theory & Techniques, vol. 37, pp. 2165 - 2168, Dec. 1989.
[11] E. Levine , G. Malamud, S. Shtrikman and D. Treves, "A study of microstrip array
antennas with the feed network," IEEE Trans. Antennas & Propagation, vol. 37, pp.
426-434, April 1989.
[12] 1. Huang, "A Technique for an Array to Generate Circular Polarization with Linearly
Polarized Elements", IEEE Trans. Antennas Propagat., AP-34, pp. 1113 - 1124,
September 1986.
[13] P. S. Hall, 1. S. Dahele and 1. R. James, "Design Principles of Sequentially Fed Wide
Bandwidth, Circularly Polarized Microstrip Antennas," IEE Proc. H, vol. 136, pp.
381- 389, October 1989.
19
[14] S. Sabban, "A New Broadband Stacked Two-Layer Microstrip Antenna," 1983 IEEE
Antennas Propagat. Symposium, pp. 63 - 66, June 1983.
[15] Kumar and K. C. Gupta, "Non radiating edges and four edges gap-coupled multiple
resonator broad-band microstrip antennas", IEEE Trans. Antennas Propagat., Vol.
AP-33, pp. 173-178, February 1985.
[16] J.F. Zurcher, 'The SSFIP: A global concept for high performance broadband
planar antennas", Electronics Letters, vol. 24, pp. 1433-1435, Nov. 1988.
[17] XFDTD, Remcom: www.fdtd.com
[18] J.-M. Jin and J. L. Volakis, "A Hybrid Finite Element Method for Scattering and
Radiation by Microstrip Patch Antennas and Arrays Residing in a Cavity," IEEE
Trans. Antennas Propagat., vol. AP-39, pp, 1598 -1604, November 1991.
[19] S. D. Targonski, R. B. Waterhouse and D. M. Pozar, "Design of wideband aperture-
stacked patch microstrip antennas", IEEE Transactions Antennas & Propagation,
vol. 46, pp. 1246 - 1251, Sept. 1998.
[20] D. R. Jackson, J. T. Williams, A. K. Bhattacharyya, R. L. Smith, S. J. Buchheitt and
S. A. Long, "Microstrip Patch Designs That Do Not Excite Surface Waves," IEEE
Trans. Antennas Propagat., vol. AP-41, pp. 1026 -1037, August 1993.
[21] Y. Qian and T. Itoh, "Progress in Active Integrated Antennas and Their
Applications", IEEE Trans. Microwave Theory and Techniques, 1998, vol. 46, no.
11, pp. 1891-1900.
[22] E. Lee, P. S. Hall, P. Gardner and D. Kitchener, "Multi-band Antennas," 1999 IEEE
Antennas Propagat. Symposium, Orlando Florida, July 1999.
[23J D. M. Pozar, S. D. Targonski and H. D. Syrigos, "Design of millimeter-wave
microstrip reflectarrays," IEEE Trans. Antennas & Propagation, vol. 45, pp. 287 -
296, Feb. 1997.
[24J R. B. Waterhouse, "Small microstrip patch antenna," Electronics Letters, vol. 31, pp.
604-605, Apr. 1995.
[25] I. Park and R. Mittra, "Aperture-coupled small microstrip antenna", Electron. Lett.,
Vol. 32, pp. 1741-1742, Sept. 1996.
[26] M. P. DeLisio and R. A. York, "Quasi-optical and Spatial Power Combining," IEEE
Trans. Microwave Theory Techn., Vol. MIT-50, pp. 929 - 936, March 2002.
[27] R. F. Jimenez Broas, D. F. Sievenpiper, E. Yablonovitch, "A High-Impedance
Ground Plane Applied to a Cellphone Handset Geometry," IEEE Trans. Microwave
Theory and Tech., vol. 49, pp, 1262-1265, Jul. 2001.
[28] H. Choo and H. Ling, "Design of Multiband Microstrip Antennas Using a Genetic
Algorithm," IEEE Microwave and Wireless Components Letters, Vol. 12, pp, 345 -
347, September 2002.
[29] A. Mitchell, M. Lech and R. B. Waterhouse, "Optimization of broadband microstrip
patch antennas," 2000 Asia Pacific Microwave Conference, APMC'oo, Sydney
Australia, pp, 711 - 714, Dec. 2000.
[30] C. R. Rowell and R. D. Murch, "A capacitively Loaded PIFA for Compact Mobile
Telephone Handset," IEEE Trans. Antennas Propagat., vol. AP-45, pp, 837 - 843,
August 1997.
[31] R. B. Waterhouse, D. Novak, A. Nirmalathas and C. Lim, "Broadband printed
sectoral coverage antennas for millimetre-wave wireless applications," IEEE
Transactions on Antennas & Propagation, vol. 50, pp. 12 -16, Jan. 2002.
[32] K. R. Carver, "Antenna Technology Requirements for Next-Generation Spacebome
SAR Systems," 1983 IEEE Antennas Propagat. Symposium, pp. 365 - 368, July
1983.
[33] J. J. Schuss, J. Upton, B. Myers, T. Sikina, A. Rohwer, P. Makridakas, R. Francois,
L. Wardle and R. Smith, 'The IRIDIUM main mission antenna concept", IEEE
Trans. Antennas & Propagation, vol. AP-47, pp. 416 -425, Mar. 1999.
20
Chapter 2 Fundamental Properties of Single
Layer Microstrip Patch Antennas
2.1 Introduction
From this chapter it is hoped that the reader understands the general
properties of the microstrip patch and has some insight into the features and
issues associated with changing the original structure, whether it be changing
the feeding mechanism or the conductor shape or the material used. An
understanding of the material presented here is imperative to the techniques
discussed in subsequent chapters on how to enhance the performance of the
patch antenna . For example, there will be circularly polarized patch arrays
presented in Chapter 6 using the techniques presented here.
2.2-1 Introduction
Patch conductor
Edge feeding
Grounded substrate
22
Simplistically when a voltage is applied to the feeding point the
radiator, a current (or mode) will be excited on the patch and vertical electric
fields will be generated between the patch and the ground plane. Because the
now formed slots (between the edges of the patch and the ground-plane) are
A/2 apart, the radiated fields will add constructively creating an efficient,
resonant radiator. The radiation efficiency will depend upon the material
used, as the confinement of the fields will be determined by the dielectric
constant of the substrate and its height. The efficiency will be discussed
further in Section 2.4.
23
analysis techniques can be sub-divided into two camps, depending on how
Maxwell's equations are solved: differentially or by integrating. Of the full-
wave analysis procedures, the Integral Equation approach is the most mature
with the Spectral Domain formulation and the Spatial Domain formulation
both being applied to predict the performance of a microstrip patch antenna
since the early 1980s. Over the years very good agreement between theory
and experiment has been achieved using both procedures and the degree of
accuracy really depends on the degree in which the programmer wants to
represent the structure to be analyzed. The initial problem of using the
Integral Equation procedures was the large computation time required to solve
the antenna under question. However, the speed of present day computers has
resolved this issue. As mentioned in Chapter 1 there are several commercially
available software packages based on these procedures.
Of all the analytical tools applied to the microstrip patch antenna and
its variants, the Spectral Domain Integral Equation technique is the one I am
most familiar with having formulated and developed several versions of it.
Thus my opinion may be somewhat biased, however I have found it to be
extremely useful in the design of patch antennas and accurate, when the
surrounding environmental conditions are not too adverse (such as when the
substrate needs to be truncated directly below the patch radiator). For such
conditions, "patch modifiers" will need to be applied. Close to half of the
designs presented in this book were based upon using the spectral domain
technique. One distinct advantage it has over other full-wave analysis
methods is its speed when conically shaped patches are considered. For such
cases entire domain basis functions can be used significantly reducing the
computational time. A fast analysis tool is imperative when attempting a
design, especially when it is quite complicated.
2.3-1 Introduction
Over the years there have been many conductor shapes proposed and
investigated for a microstrip patch antenna. Schematic diagrams of these are
shown in Figure 2.3.1 and a brief summary of the advantages and
disadvantages are given below. Here we are assuming operation only in the
lowest order mode, a condition a patch is normally operated under simply
because its size is smallest.
The first and probably the most utilized patch conductor geometry
were the rectangular and square shapes. Figure 2.3.1a and b show these
geometries. For the rectangular patch the antenna is excited at some point
26
along the resonant dimension, L, to generate a mode in this direction. Figure
2.3.2 shows the currents excited on a rectangular patch, in particular the
direction they are excited. In general the length of the patch controls the
resonant frequency and the width of the patch affects the impedance level at
resonance as well as the bandwidth (a second order effect): the larger width of
the patch, the smaller the input impedance of the antenna. Keep in mind this
statement is only valid under certain conditions (relatively thin substrate
material). As the thickness of the material increases to greater than 0.03 An,
these relationships are not mutually exclusive and the feeding procedure and
location can dramatically change all the measures of performance. The case
shown in Figure 2.3.2 is for a linearly polarized radiator, so the currents
should predominantly be in one direction. By moving the feed location off
the symmetry point with respect to the W dimension, currents can also be
generated in this direction. Once again, under conservative conditions, if a
square patch is considered and the feed is located along the diagonal plane,
dual polarized radiation at the frequency of operation can be generated.
A circular patch, like a square patch only has one degree of freedom
in its conductor shape and that's its radius. Thus changing the radius will
control the resonant frequency of the circular patch. Once again under
conservative conditions, the feed position will control the input impedance of
the antenna at the chosen resonant frequency.
Triangular and disc sector patch geometries are smaller than their
rectangular and circular counterparts, although at the expense of further
reduction in bandwidth and gain. Figure 2.3.le and f show schematic
diagrams of these conductor shapes. Triangular patches also tend to generate
higher cross-polarization levels, due to their lack of symmetry in the
configuration. Figure 2.3.4 shows the current distribution on a triangular
patch conductor. Contributions to cross-polarization fields come from
currents directed in the orthogonal direction to the main polarization. The
triangular and disc sector patches have a similar number of design freedoms
as the rectangular patch. For a disc sector antenna, if the TMot mode is
excited, then the radius of the disc controls the resonant frequency and the
sector angle controls the bandwidth and impedance. Depending on the aspect
of the triangle and the disc sector, dual frequency and dual polarized patches
can be developed using either of these conductor shapes, however the
bandwidth is typically very narrow. Also for the dual frequency patch, the
polarizations for each frequency band are orthogonal.
29
Figure 2.3.4 Typical Fields Excited on a Triangular PatchAntenna
Annular ring geometries are the smallest conductor shape, once again
at the expense of bandwidth and gain. Figure 2.3.1g shows this conductor
shape. One problem associated with an annular ring is that it is not simple
process to excite the lowest order mode and obtain an input impedance close
to 50 Q at resonance. In fact, impedance values ranges from 150 - 250 Q are
quite normal. Non-contact forms of excitation (see Section 2.5) are typically
required to feed this element at the expense of antenna efficiency. The
symmetry issues mentioned for the circular patch cases also apply here. The
current distribution of a probe fed annular ring is shown in Figure 2.3.5.
The annular ring has one more design variable than the circular patch
and therefore its response should be easier to control. Both the inner and
outer ring dimensions can be used to control the resonant frequency of the
printed antenna, which is very advantageous. However, as the inner radius
approaches the outer radius dimension, the impedance bandwidth becomes
narrower.
30
Figure 2.3.5 Typical CurrentslFields Excited on an Annular RingPatch
10D .....---------------"""'T"-----.
~. 2.2 •••••
. .....
...... "
"......
.>
. .....
.
. ......
.......
...... '
..- ~=
-"
102_-
(a)
10.0 r - - - - - - - - - - - - - - - - - - - - ,
tr-ID _ t -22
-
: ! .
~ -- -- ---------
;: :; fo--------- --- -
t.-
~
t 5D f- 102
l!!
0.050
(b)
32
ID """""..-=------
~ ~ ~
------------,
"" " .
" , • • • •••• •••• +
t.=22
~ .
-,
"" -,
"" ... ...
,
M'
... ...
•~ 0.5 I- ....
~
"-
... ...
" ... E.= 10.2
"
I
0.05 0.10
sm.irate tJddmess, dI10
Figure 2.4.1 Performance trends of single layered patch antennas: (a) impedance
bandwidth; (b) directivity; (c) surface wave efficiency
180
From Figure 2.4.1 the initial design procedure of any microstrip patch
solution can be established. If only a low impedance bandwidth solution is
required, say less than 5 % of the operating frequency, then the material to be
used for the patch can be selected by looking at Figure 2A.la. Once this has
been determined the next phase: the conductor shape can be selected. This
34
will be based upon the space available for the antenna and the polarization
required. Typically the gain specified for the antenna is considered much
later in the process as the gain a single microstrip patch element is relatively
constant and can be easily increased by using arraying techniques. Once the
above has been chosen, the next stage is to look into the feeding technique, or
how to couple power to and from the patch antenna.
2.5-1 Introduction
35
Microstrip feedline, Wf
Grounded substrate
36
discontinuities associated with the contact point between the microstrip line
and the patch antenna.
L=Ae!2 (2.1)
where Ae is the effective wavelength given by:
37
(2.2)
Introduction
Two design cases are presented here for single layered edge fed
patches: an edge fed rectangular patch etched on a high dielectric constant,
electrically thick material [13]; and a dual frequency sectored disc patch for a
self-diplexing transceiver [14]. The example is a case of essentially what not
to do, as it highlights some of the shortcomings of this feeding procedure.
The second design gives an example of where edge feeding is very useful and
easy to do. More fundamental designs of edge fed patches can be readily
found in the literature and so they will not be repeated here.
38
Millimeter-wave Edge Fed Rectangular Patch Antenna
11 = 1 - P sw I Plot (2.3)
where Plot is the total power delivered by the source to the antenna
and Psw is the power delivered by the source to the surface waves.
Introduction
40
Antenna Design
When the sector angle ¢o is 70°, the first and second modes of the
circular sector patch antenna are TM vII' VI =2.571 and TM01 mode,
41
respectively, and the electric field intensity of the two modes, calculated by
using the FDTD method [18], are shown in Figure 2.5.5. From the electric
field distribution of TM vII and TM 01 mode, we found that the two modes
have orthogonal polarization and broadside radiation patterns. The feeding
position is critical to get good isolation between two modes as can be seen
from field plots.
". ""
..,
D'
o
In Figure 2.5.4, port 1 excites the TM 01 mode and port 2 excites the
TM vII mode. The measured S parameters of the two-port patch antenna are
shown in Figure 2.5.6. The isolation between the two ports (S21) is greater
than 30dB. The radiation patterns of the proposed antenna at the two
operating frequencies are measured and shown in Figure 2.5.7. It can be seen
that the two operating modes have broadside radiation patterns and orthogonal
planes. The cross-polarization radiation levels are more than 20 dB below the
co-polar radiation levels. The difference in the antenna gain in the broadside
direction at the transmit and receive frequencies is approximately 0.2dB.
o .......\ x~·7 ··
·10
~ V
!
522
-20 L - 5 11
! -30
_.- .- .... ... ......-. .
-40
""
\
_. ........ . -.....
,, 521
I~
\
·50
-60
5 5.2 5.4 5.6 5.8 6
Frequency [GHz]
42
10 10
110 0
270 270
(a) (b)
Figure 2.5.7 Measured Radiation Patterns at: (a) 5.4 GHz; (b) 5.5. GHz
2.5-3.1 1ntroduction
There are several key advantages of the probe-fed patch. Firstly, the
feed network, where phase shifters and filters may be located, is isolated from
the radiating elements via a ground-plane. This feature allows independent
optimization of each functional layer. Of all the excitation methods, probe
feeding is probably the most efficient as the feed mechanism is in direct
contact with the antenna as well as most of the feed network is isolated from
the patch, minimizing spurious radiation. The high efficiency of this printed
antenna has seen a somewhat renaissance of the probe-fed styled patch,
despite the added complexity of developing a via connection. For a probe-fed
43
patch mounted on the same material considered for the edge-fed case
presented in Section 2.5-2.1 is approximately 96 %.
y'
The design procedure for probe fed patches is similar to that for edge
fed patches and so it will not be repeated here, however some additional
defmitions and comments will be made. Figure 2.5.10 shows the effect on the
input impedance response of varying the position of the probe feed for a
44
microstrip patch antenna. As mentioned before, varying this parameter
(equivalent to varying the position of the insert for an edge fed patch) changes
the input impedance of the antenna from a high value when the probe is
located close to the edge, to a low value when the probe is positioned near the
center. When the input impedance response is will matched at resonance (50
Q) we will define this as a case where the patch is critically coupled. This
means that at this feed location the fields are at the appropriate levels to give
this impedance. For the case where the input impedance is less than 50 Q at
resonance we will define this as being under coupled and for when the
impedance is greater than 50 Q we will say the patch is over-coupled. Figure
2.5.10 displays this representation. These definitions will be used throughout
this book and are very important when it comes to investigating the effects of
parameter variation on the response of the antenna.
There are several points that should be made here with respect to
designing probe fed patches and these are related to how the probe feed
should be connected to patch antenna. SMA connectors are the usual choice
for this application, except when very high frequencies, into the mm-wave
range are considered. Here K-type connectors are typically used. Firstly, the
feed pin radius should be small to minimize conductor loss associated with
the pin. Also, thinner pins yield low cross-polarization levels. Before
soldering the pin to the patch conductor, ensure it does not extend much
beyond the top of the copper track, relative to the frequency of operation. If
the pin extends beyond the patch conductor, it is likely to contribute to
generating more cross-polarized radiation. Care should be taken when
drilling the hole through the patch conductor to the ground-plane to avoid
potential short-circuits occurring from copper ribbons coming in contact with
the ground-plane. It is also important to make sure the hole in the ground-
45
plane, where the center conductor of the SMA comes through the ground-
plane physically matches the outer radius of the connector.
Under-coupled Crilically-coupled
Over-coupled
2.5-3.4 Design Example - Low Cost Printed Antenna for Indoor Wireless
Base Stations
Introduction
46
communication topology and a microcellular or indoor wireless system is
required to ensure no loss of continuity for the user.
47
x
•
r,
+ ~
(X", Y,,) (x". Y,, )
(a) (b)
Ftgure 2.5.11 Schematic uf Arueuua Configuration (a) Top View; (b) Side View
Figure 2.5.12 shows the predicted return loss of the two circular
patches as a function of frequency. Also shown in this diagram are the
measured ISIII and ISzzI results obtained using a Wiltron 360 B network
analyser connected to the two ports of the module. As the results show, the
amount of coupling between the patches is relatively small, since experiment
and theory are in good agreement. The coupling between the ports was
measured and the maximum was -33 dB at 1.56 GHz. Figure 2.5.12 also
shows that the impedance bandwidth requirement for both transmission and
reception bands has been satisfied using this substrate.
-5
-10
-15
~
·20
~
a.
E -25
a:: · 30
••••• 'Trans. Theory
- - Trans. Exp.
- •• - Reo. Theory
- • - • Rec. Exp.
-35
-40
-45
1.5 1.53 1.56 1.59 1.608 1.638 1.668 1.698
Frequency (Gllz)
48
The E- and If-plane far-field radiations patterns of the antenna
module were measured at the center frequencies of both transmission and
reception bands. Figure 2.5.13 presents the measured co-polar E- and H-
plane patterns of the transmit patch at 1.54 GHz (with the receive patch
terminated with a 50 Q load). The E-plane results are typical for a microstrip
patch, whereas the H-plane pattern displays an interesting phenomenon. A
'reflector-like' behavior is experienced due to the presence of the other
element. This is a similar pattern distortion to that experienced in a printed
Yagi antenna [22]. The patterns for the receive patch at 1.632 GHz are very
similar to those presented in Figure 2.5.13, with the 'broadening' of the H-
plane pattern on the other side of the maximum, as expected. The measured
cross-polarization levels were less than -25 dB at all frequencies in both
planes. From the results presented in Figure 2.5.13 it is evident that the usual
requirement for wireless base stations has been satisfied, namely a relatively
broad 3 dB beamwidth, in this case 90 0 in the E-plane and 850 in the H-plane.
o
225
180
49
carriers from the one port, typically using a diplexer which increases the cost
of the RF circuitry.
2.5-4.1 Introduction
pal.chCL, W)
Full-wave analyses are not too difficult to develop for the proximity-
coupled patch due to the lack of a current discontinuity between the feed
network and the radiating element [15]. Thus unlike for edge-fed and probe-
fed patches, special attention does not need to be undertaken near the feed
point, and no 'attachment mode' or heavy segmentation of the conductor area
in this vicinity is necessary. It is interesting to note that the formulation of the
Spectral Domain Integral Equation in [15] that can be applied to both edge-
fed and proximity-coupled patches is more accurate for the latter case, simply
because of the lack of a current discontinuity.
The proximity coupled patch is slightly less efficient than the edge
fed patch, given the same material parameters for the design. In the case
51
considered previously for the edge-fed patch, the proximity coupled solution
has an efficiency of approximately 88 %. The efficiency for a proximity
coupled patch is an issue, especially when thicker material is to be used to
enhance the bandwidth (considering the trends presented in Section 2.4). This
will be looked into further in Chapter 3.
One of the salient features of proximity coupled patches are the extra
degrees of freedom available in this form of microstrip patch, compared to the
direct contact feeding procedures. Thus to achieve a well-matched impedance
response, the designer can change several parameters, relatively
independently. Here we will consider 6 parameters and their effects on the
input impedance response.
Antenna substrate
Feed substrate
Feed-linewidth
This parameter, along with the thickness of the materials used, is the
most critical in determining whether the patch is critically coupled or not.
Moving the location of the open circuit termination closer to beneath one of
the radiating edges of the patch, increases the coupling between the feed-line
53
and radiator. Simplistically this action can be interpolated as attempting to
overlap the location of the maximum fields from the resonant patch and the
fields associated with the open circuit termination. Once again if a proximity
coupled patch has been designed to be critically coupled and then the location
of the open circuit termination is moved closer to the patch edge, the antenna
will become over-coupled.
Introduction
54
was approximately 72 %. Thus comparing this result to that presented in
Section 2.5-2, larger bandwidths can be achieved with this configuration
without serious degradation to the efficiency. The gain of the proximity
coupled patch was 6 dBi. More designs incorporating proximity coupled
patches are presented later in this book.
2.5-5.1 Introduction
55
Oroun~plan.
Antenna substrate
The stub termination allows the designer to readily match the input
impedance response of the aperture coupled patch. The microwave circuitry
analogy to this parameter is the single stub matching procedure. For the patch
configuration, the terminated stub allows the frequency response to be rotated
on a constant conductance circle on the Smith chart.
Slot length
The slot length controls the amount of coupling to the patch radiator.
As it is increased in size, the amount of coupling to the patch increases. The
slot length is directly related to the antenna substrate used: the thicker the
substrate; the longer the slot.
58
The design procedure for the aperture coupled patch, is similar to that
for all microstrip antennas, in that it is conducted by examining the impedance
performance of the antenna on a Smith chart. The starting points mentioned
for the other feeding techniques, namely the patch length and substrate
selection is the same. The key to an aperture coupled design is to get a
starting impedance response that not confined to the edge of the chart. Once
there is a hint of a finite real part to the input impedance, it is relatively
straightforward to design a critically coupled patch using the parameter study
given above. I typically start with a slot length about half of the length of the
patch, although this will depend on the thickness of the material being used.
The start point for the open circuit stub, is usually near directly under the
center of the patch. As can be seen from the above 'parameter dependence of
the aperture coupled patch, different parameters can have countering effects
on the impedance and radiation performance. This is very useful as it gives
the designer the capability to tailor their design for the given application. For
example if minimal backward directed radiation is important, then rather than
increasing the slot length to couple to the patch mounted on a relatively thick
layer, the width of the patch can be decreased, to a degree.
Introduction
The final configuration investigated for the base station antenna was
the aperture coupled microstrip patch (refer to Figure 2.5.17 for details of the
layout). Here the microstrip line and hence photonic and mm-wave devices is
coupled to the radiating element via a slot in the ground plane. Importantly
once again, the antenna substrate can be optimized for efficient radiation and
bandwidth, relatively independently of the photonic/rf circuitry material.
59
Results and Discussion
2.6-1 Introduction
60
approach. Each method has its merits and disadvantages of which will be
summarized herein.
By far the best CP bandwidths and lowest axial ratios can be achieved
using a synchronous subarray [27]. A four-element synchronous subarray is
shown in Figure 2.6.3. Here each patch is spatially rotated 90° with respect to
the previous patch. In addition to the spatial rotation, the feeds are also phase
rotated sequentially by 90°. Thus port 2 leads port 1 by 90°, port 3 leads port
1 by 180° and port 4 leads port 1 by 270°. Having such a feed arrangement
cancels the generated cross-polarization fields of the individual patches. It
has been shown in the past that very good axial ratio bandwidths can be
achieved using a synchronous subarray of elements that individually do not
radiate CP [28]. Also the impedance bandwidth of this CP antenna is
62
typically larger than that of the single element due to the feed network.
Synchronous subarrays do not necessarily consist of four elements; cases of 3
(0°, 120° and 240°) and higher order cases have been investigated. As long as
there is a spatial symmetry with respect to the center of the configuration as
well as the appropriate phase rotation, this approach should provide good
axial ratio performance. A typical axial ratio plot of a 4 element synchronous
subarray is shown in Figure 2.6.4. Note the very good axial ratio, in
particular between ± 45°.
Palch4 Patch 3
Patch I
Patch2
10 , I.
,,
,, ,
,:"
--3dBAR
,
,,
, ,:" --SFednf.
,, t ........... OF eonf.
,, I
,
"
J --- SS eonf.
, "
___ f'f
o
'" <, ",'
, '"
"I
0.8 0.9 I.I 1.2
Normalized Frequency
2.7 Summary
As mentioned before, this chapter lays the foundation for the rest of
the book. To get maximum benefit from this book, it is necessary to have a
thorough understanding of the material presented here. Many concepts based
on the material and trends of chapter 2 will be assumed in subsequent
chapters.
2.8 Bibliography
[1] K. R. Carver and J. W. Mink, "Microstrip Antenna Technology," IEEE Trans.
AntennasPropagat., Vol. AP-29, pp. 2 - 24, January 1981.
[2] E. H. Van Lil and A. R. Van de Capelle, ''Transmission Line Model for Mutual
Coupling Between Microstrip Antennas," IEEE Trans. AntennasPropagat., Vol. AP-
32, pp. 816 - 821, August 1984.
[3] Y. T. Lo, D. Solomon and W. F. Richards, "Theory and Experiment on Microstrip
Antennas," IEEE Trans. Antennas Propagat. , Vol. AP-27, pp. 137 - 145, March
1979.
[4] W. F. Richards, Y. T. Lo and D. D. Harrison, "An Improved Theory of Microstrip
Antennas with Applications," IEEETrans. AntennasPropagat. , Vol. AP-29, pp. 25 -
27, January 1981.
[5] M. D. Deshpande and M. C. Bailey, "Input Impedance of Microstrip Antennas,"
IEEE Trans. Antennas Propagat., Vol. AP-30, pp. 651- 656, July 1982.
65
[6] D. M. Pozar, "Input Impedance and Mutual Coupling of Rectangular Microstrip
Antennas," IEEE Trans. Antennas Propagat., Vol. AP-30, pp. 1191 - 1196,
November 1982.
[7] J. R. Mosig and F. E. Gardiol, "General Integral Equation Formulation for Microstrip
Antennas and Scatterers," Proc. lnst. Elect. Eng., pt. H, Vol. 132, pp. 424 - 432,
1985.
[8] A. Reineix and B. Jecko, "Analysis of Microstrip Patch Antennas Using Finite
Difference Time Domain Method," IEEETrans. AntennasPropagat., vol. AP-37, pp.
1361 - 1369, November 1989.
[9] J.-M. Jin and J. L. Volakis, "A Hybrid Finite Element Method for Scattering and
Radiation by Microstrip Patch Antennas and Arrays Residing in a Cavity," IEEE
Trans. AntennasPropagat., vol. AP-39, pp. 1598 -1604, November 1991.
[10] High Frequency Structure Simulator - Version 5.5, Agilent Technologies, 2000.
[11] R. F. Munson, "Conformal Microstrip Antennas and Microstrip Phased Arrays,"
IEEE Trans. AntennasPropagat., Vol. AP-22, pp. 74 -78, January 1974.
[12] D. M. Pozar, "Input Impedance of Microstrip Antennas," Proc. IEEE, Vol. 80, pp. 74
- 81, January 1992.
[13] R. B. Waterhouse and D. Novak, "Design of Patch Antennas for Integration in OEICs
for Optical Fiber Picocellular Systems," International Topical Meeting Microwave
Photonics (MWP96), Kyoto Japan, pp. 89 - 92, Dec. 1996.
[14] D.-K. Park, R. Waterhouse, Y. Qian and T. Itoh, "Self-diplexed integrated antenna
transceiver for wireless applications", 2001 IEEE Antennas & Propagation
Symposium, Boston, USA, July 2001.
(15] D. M. Pozar and S. M. Voda, "A rigorous analysis of a microstripline-fed patch
antenna", IEEE Trans. Antennas & Propagation, vol. AP - 35, pp. 1343-1349, Dec.
1987.
[16] W. F. Richards, J.D. Ou, and S. A. Long, "A theoretical and experimental
investigation of annular, annular sector, and circular sector microstrip antenna," IEEE
Trans. AntennasPropagat., vol. 32, pp. 864-867, Aug. 1984.
[17] L. C. Shen and S. A. Long, "The resonant frequency of a circular disc, printed-circuit
antenna," IEEETrans. AntennasPropagat., vol. 25, pp. 595-596, July 1977.
[18] K. S. Yee, "Numerical solution of initial boundary value problems involving
Maxwell's equations in isotropic media," IEEE Trans. Antennas Propagat., vol. 14,
pp.302-307, 1966.
(19] Special Issue on Fiber-Optic Microcellular Radio Communication Systems and Their
Technologies, IECETrans. Commun., vol. E76B, September 1993.
[20] F. Ali and J.B. Horton, "Introduction to special issue on emerging commercial and
consumer circuits, systems, and their applications", IEEE Trans. Micro. Theory &
Tech., MTT-43, pp. 1633-1638, July 1995.
[21] J.T. Aberle, D.M. Pozar and C.R. Birtcher, "Evaluation of input impedance and radar
cross section of probe fed microstrip patch elements using an accurate feed model",
IEEE Trans. AntennasPropagat., AP-39, pp. 1691·1697, December 1991.
[22] J. Huang, A. Densmore and D.M. Pozar, "Microstrip Yagi array for MSAT vehicle
antenna application", Int. Mobile Sat. Conj; pp. 554-559, June 1990.
[23] D. M. Pozar and B. Kaufman, "Increasing the Bandwidth of a Microstrip Antenna by
Proximity Coupling," Electronic Letters, Vol. 23, pp. 1070 - 1072, September 1987.
[24] D. M. Pozar, "A Microstrip Antenna Aperture Coupled to a Microstrip Line,"
Electronic Letters,Vol. 21, pp. 49 - 50, January 1985.
[25] J. R. James and P. S. Hall, Handbook of Microstrip Antennas, Vols. 1 and 2, Peter
Peregrinus, London, UK, 1989.
[26] C. A. Balanis, Antenna Theory: Analysisand Design, 2nd Edition, Wiley, New York,
1997.
66
(27) 1. Huang, "A Technique for an Array to Generate Circular Polarization with Linearly
Polarized Elements", IEEE Trans. Antennas Propagat., AP-34, pp, 1113 - 1124,
September 1986.
(28) D. M. Pozar, "Scanning characteristics of infinite arrays of printed antenna
subarrays," IEEE Trans. Antennas Propagat., AP- 40, pp. 666 - 674, June 1992.
67
Chapter 3 Enhancing the Bandwidth of
Microstrip Patch Antennas
3.1 Introduction
70
questions arise, how far can this envelope be pushed and what are the limiting
factors?
The bandwidth presented in this case is pretty well the inherent limit
of a direct contact patch antenna, without doing any other modifications to the
patch (stacking, match structures, etc..). This can be attributed to the
inductive nature of the feeding mechanism for the antenna. As the thickness
of the substrate is increased, the inductance of the feedline becomes larger and
it shifts the impedance locus of the antenna to a more inductive state. This
can be graphically represented on a Smith chart, as shown in Figure 3.2.2. As
the thickness is increased, it becomes more and more difficult to make the
antenna resonate, thereby reducing the efficiency of the radiator. For a probe-
fed solution, as the thickness of the substrate is increased, the location of the
probe for a 50 .Q match at resonance must be moved closer to the edge of the
patch. In fact a similar trend must be applied to an edge-fed patch, namely to
71
achieve a 50 Q matched resonance on thicker material, the insert length, must
be reduced, if not removed. Another consequence is that the patch conductor
size becomes smaller. This is because the antenna is more inductive in nature
and so for resonance, the operating frequency is where a thinner antenna
would normally be in a capacitive state, namely at a frequency above its
natural first resonance.
73
feeding a microstrip patch radiator are relatively straightforward: there is
more than one radiator/resonator in the broadband solutions. More will be
discussed and presented on this issue in subsequent sections, however it is for
this reason that has seen aperture coupled patches (and variants of this printed
antenna) become the most widely used microstrip antenna.
···................... ....
i~~~~~~~----l-
: :
- i .\ i
Input
50 n microstrip line
Patch (on coupled layer)
-.
-10
~
L/- r--.
~15
:::'20
\ / \
o
..J
...c-25
1- Experiment / \
~
I
'G>-30
II:
-35
-40
-45
32 33 34 35 36 37 38 39 40
Frequency (GHz)
3.3-1 Introduction
1:d
f
Figure 3.3.1 Horizontal Parasitically Coupled Patches
Figure 3.3.2 shows a case where this technique can give comparable
bandwidth enhancement with that of a stacked patch configuration (which we
will give more detail on in the next section). From Figure 3.3.2, for moderate
bandwidths the horizontally coupled procedure can function quite well,
although the overall size of this antenna is close to 3 times as large as the
stacked version, making it difficult to use as an element of an array (as
mentioned before). Thus a very critical limitation of this procedure is the
large horizontal dimension.
f',
0 .6 '"
r" \
~
.... O.S
" "
~, \ VSWR. H
I '
/
/
t::: 0 .4
\... .. - - Con"YCDlionai /
~ '\, ' \ - - Gap c<>upled ' /
U t .. ... Stecked I' /
=
0
0.3
'':::
~
'ii
c:.: 0.2
0 .1
76
3.3-2 Annular Ring Loaded Shorted Patch
3.3-2.1 Introduction
Figure 3.3.3 Probe-fed circular microstrip patch antenna with a concentric annular ring
and a single shorting pin
78
The predicted and measured input impedance behavior of a shorted
patch with an annular ring incorporating 5 rom ROHACELL 71 HF is shown
in Figure 3.3.4 (refer to the caption for dimensions). The experimental results
were measured using a Wiltron 360B network analyzer. As seen from these
results very good agreement between experiment and theory was achieved.
The predicted and measured 10 dB return loss bandwidths were 6.6% and
6.8% respectively. Itshould be noted that to fabricate the antenna, the shorted
patch and ring conductors were etched on 0.254 rom thick Rogers 5880
Duroid which was adhered to one side of the foam and a brass plate was
affixed to the other to provide a ground plane. The electrically thin Duroid
layer was not considered when designing the antenna and may be the
contributing factor to the minor frequency shift between the measured and
analytic results (approximately 1%) as well as the slight closure of the
resonant loop. The co- and cross-polar patterns in the cardinal planes of the
shorted patch with a concentric ring were measured and the results were very
similar to a conventional microstrip patch antenna mounted on foam The
measured 3 dB beamwidth in the E- and H-planes were 84° and 75°,
respectively compared to the predicted values of 86° and 70°. The radiation
patterns are shown in Figure 3.3.5. The measured H-plane cross-polarization
level was always more than 20 dB below the co-polar level.
Figure 3.3.4 Predicted and Measured Impedance of the Annular Loaded Shorted
Microstrip Patch (Er = 1.13, tan 0 =0.001, d =5.0 mm, R = 11.8 mm, xp =8.0 mm, Yp =0, xps
=9.3 mm, Yps =0, TO =0.325 mm, ros =0.325 rnm, a =31.5 mm, b =17.0 mrn)
, --
-2 .."j,."
i -4
:,. /
.,. ~'
~
If -6
.:,,',,
" - - E-planeExp .
: I - •• - ,. E-plane Theory
~ .. I
" I
'; -8 I
.. . . . . .. "-plane Exp.
~ I
I
- - - - " -plane Theory
-10
-12 L- ----.J
3.3-3.i introduction
Figure 3.3.7 Predicted and measured impedance of the annular loaded shorted
microstrip patch (Er 1 =Er2 =2.2, tan oJ =tan ~ =0.001, d J =6.096 mm, d2 =0.254 mm, R =
8.2 mm, xp =4.1 mm, Yp =0, xps =5.3 mm, Yps =0, r o =0.325 mm, ros =0.325 mm, a =25.8
mm, b = 9.6 mrn)
The co- and cross-polar patterns in the cardinal planes of the CARL
shorted patch were measured and the co-polar E- and H-plane results are
shown in Figure 3.3.8 with the antenna mounted on a 30 x 30 cm2 ground
plane. The measured 3 dB beamwidth in the E- and H-planes were 85° and
80°, respectively compared to the predicted values of 90° and 78°. The slight
asymmetry in the E-plane pattern can be attributed to the presence of the
shorting pin. The measured H-plane cross-polarization level was always more
than 20 dB below the co-polar level. The measured and calculated gains of
the printed antenna were 6.2 dBi and 6.4 dBi, respectively. It is pertinent to
compare these results with a conventional circular probe-fed rnicrostrip patch
designed using the same material. Here the 10 dB return loss bandwidth is
only 4.7 %, the diameter is approximately 6 em and the gain is 7.1 dBi. Thus
the CARL shorted patch offers a bandwidth enhanced variation of the single
layer microstrip patch.
82
o
180
Figure 3.3.8 Measured E-plane and H-plane Radiation Patterns of CARL Patch
3.4-1 Introduction
A fundamental limit of the direct contact patch antenna has been the
achievable impedance bandwidth. Here, 10 dB return loss bandwidths
approaching 8 % have been obtained for a single element configuration if
foam is used as the substrate, however this is by no means adequate for most
wireless systems. Vertically parasitically coupling another printed antenna to
the driven patch, or stacking, is one procedure that can alleviate this problem
to a degree. It is interesting to note that for early versions of the direct feed
stacked patch, only moderate bandwidths, less than 15 % were achieved [10].
It can be postulated that this barrier was due to the inductive nature of the feed
and so impedance control of the radiator when thick substrates were used was
very difficult. Also due to complex nature of the currents associated with the
discontinuity between the patch conductor and the probe feed, to accurately
model this antenna configuration required sophisticated, accurate numerical
tools. As was pointed out in the Chapter 1, not only do designers need
accurate analysis tools, the software needs to be fast, such that many iterations
can be processed. It wasn't until the late 1990s that such analysis tools were
available and that the true potential of the direct fed stacked patch antenna
could be realized.
83
In this section design trends for broadband direct contact feed stacked
patches are established. Here broadband is defined as a 10 dB return loss
bandwidth greater than 25 %. The objective of this section is to develop a
robust design philosophy for stacked patch designs including how to start the
design procedure and which dimensions of the printed antenna should be
varied to achieve the desired impedance response. The characteristics and
impedance trends of probe-fed stacked circular, rectangular and annular rings
will be presented. It will be shown that all three configurations behave
effectively in the same manner and varying one parameter on say the circular
patch geometry has the same effect when changing a similar parameter on the
other geometries . It will be shown that annular rings and rectangular stacked
patches do offer an extra degree of freedom compared to circular patches .
This gives easier impedance control at the expense of slightly reduced
bandwidth.
Figure 3.4.1 Side-views of Probe-fed Stacked Patch Geometries: (a) Rectangular Patches; (b)
Circular Patches; (c) Annular Ring Patches
84
Stacked Rectangular Patches
A stacked circular patch antenna was designed for 1.9 GHz using the
full-wave analysis and the parameters are as follows: Erl =2.2, d, =3.048 mm,
R1 =31 mm, x, =27.5 mm, YP =0, ro =0.325 mm, Erz =1.15, dz =9 mm, R z =34
mm. The simulation and experimental impedance behavior are shown in
Figure 3.4.3. The theoretical 10 dB return loss bandwidth was 20% compared
to a measured bandwidth of 19%.
85
Figure 3.4.3 Input Impedance of Stacked Probe-fed Circular Patches
It can be seen from the Smith chart plot that the resonant impedance
loop for experimental result was slightly shifted from the predicted curve.
This can be attributed to the thin layer of material required for mounting the
upper patch on the foam that was not included in the simulation. Overall,
these results are acceptable for the application at hand (here a micocell base
station) since the required bandwidth and operating frequency have been met.
The far-field E-plane and H-plane for co-polar and cross-polar radiation
patterns were measured at 1.9 GHz. The measured results are very similar to a
conventional microstrip patch antenna. The plots of E-plane and H-plane co-
polar are shown in Figure 3.4.4 and Figure 3.4.5, respectively. The cross-polar
levels in the E- and H-plane were 18 dB and 14 dB lower respectively than
the co-polar levels.
86
o
180
180
87
in the fundamental mode (TM Il ) the probe-fed printed ring has a high input
impedance at resonance. Thus there is a strong coupling between the feed and
the patch, that is, the radiator is overcoupled. This is why proximity coupling
a micros trip line to the ring is the preferred option, where by moving the
feedline away from the patch the coupling to the radiator can be reduced (and
hence the input impedance). The strong coupling between a probe feed and a
printed annular ring is extremely advantageous, however, when considering
stacked configurations. Typically, for any stacked patch configuration, the
lower resonator is deliberately over-coupled and the top patch is used to
effectively impedance match the entire configuration. Thus the high
impedance encountered for a probe-fed single layered ring is very suited to a
stacked version. ' Importantly too, the high impedance level is relatively
independent of substrate thickness. Thus thick dielectric layers below the
lower resonator can be used without the penalty of poorly behaved input
impedance due to the inductive nature of the feed. This is in direct contrast to
probe-fed circular and rectangular stacked patch configurations where the
lower substrate cannot be made too thick. Another advantage of the annular
ring over the circular patch version is the extra degree of freedom in design,
namely the inner radius. This can be varied independently of the outer radius
to provide fine tuning of the impedance variation of the antenna. The effect of
this extra degree of freedom is similar to that of varying the width of a
rectangular patch.
Figure 3.4.6 shows the predicted and measured input impedance plot
of a probe-fed stacked annular ring antenna (refer to the caption for the
relevant dimensions). As can be seen from this figure, very good agreement
between theory and experiment was achieved. The predicted and measured
10 dB return loss bandwidth was 21 % and 22 % respectively. The E- and H-
plane co-polar and cross-polar far-field radiation patterns were measured and
are very similar to a conventional patch radiation characteristics.
Comparisons of these measured values and the predicted results for the E-
plane and H-plane co-polar patterns are shown in Figure 3.4.7. The cross-
polarization levels were greater than 20 dB below these values in both planes.
The gain of the stacked configuration was measured as 8.2 dBi, compared to a
theoretical value of 8.5 dBi.
88
bandwidths of the three cases can readily be explained . It is now pertinent to
examine some design trends.
Figure 3.4.6 =2.2, tan 01 =0.001, dj =6.096 nun, aj =10.0 nun, bl =29.0 nun, Pp
Er l
=21.0 nun, ~ =0°, ro =0.325 nun, Er2 =1.07, tan 02 =0.001, dZ =8.0 nun , aZ =14.0 nun,
=
bZ 31.0 mm
180
Figure 3.4.7 Predicted and Measured Radiation Patterns of Stacked Probe-fed Annular
Rings
89
3.4-3 Design Trends
A. Substrates
A lot of research over the years has focused on what are the best
materials for the dielectric layers for microstrip patches. It has been shown
that thick, low dielectric constant laminates tend to give the largest bandwidth
responses and good surface wave efficiencies. Although it should be stated a
combination of high dielectric constant and low dielectric constant materials
can also yield good impedance behavior (we will investigate this further in
Chapters 4 and 6). Perhaps the easiest choice of materials to work with is a
combination of low dielectric constant material (say e, =2.2) and foam. This
tends to give the largest impedance bandwidth and also has the simplest
design procedure (We will examine this in a subsequent section in more
detail, when we optimize the substrate layers to give the best bandwidth
performance). Using two layers of foam does give better surface wave
efficiency, however at the expense of bandwidth reduction. We have found
that using the higher dielectric constant laminate for the substrate of the lower
patch gives enhanced results. This can be attributed to the coupling of the
modes on each patch. Unlike in [15] where the 'mutual coupling' terms are
the important factors in the design, here the resulting bandwidth is
predominantly governed by the current distribution on the lower patch. The
broadest bandwidths were achieved when the first order mode on the lower
patch is considerably greater in magnitude than the corresponding mode on
the top patch. In other words the top patch is 'loosely coupled' . To do so,
requires the lower dielectric layer to have a greater dielectric constant than the
upper layer. If the layers have the same dielectric constant or the upper layer
has a higher value, the modes are too strongly coupled yielding very tight
resonant loops and resulting in lower impedance bandwidths.
Stacked patch
Single layer
(a)
91
(b)
Figure 3.4.8 Effect of d l on Stacked Patch Design (frequen cy range: 6 - 11 GHz): (a)
parame ters: Erl =2.2, d l =1.524 mm, tan 0 1 =0.001, er2 =1.07, d2 =2.5 mm, tan ~ =0.001 , R 1
=7.0 rnm, R2 =7.4 mm, xp =5.4 rnm, ra =0.325 mm; (b) paramet ers: Erl =2.2, d l =2.524 rom,
tan O. =0.001, e,z =1.07, d2 =2.5 rom, Ian ~ =0.001, R. =7.0 rom, R2 =7.6 rnm, xp =5.5
mm, ra = 0.325 rom)
B. Starting Point
92
approximately 0.06 1.0 . These values were obtained using the combination of
the dielectric layers suggested previously.
Once the resonant loop has been formed, adjusting the dimensions of
the patches can control its location on the Smith chart. In this section, only
circular patches are presented, however the same trends are achieved when
varying the patch conductor sizes of rectangular patches (LI> L z) and annular
rings (bj , b z). Figure 3.4.9a shows the effect of varying R) on the impedance
locus while keeping the other parameters of the stacked patch constant. Note
that an arc can be drawn through the center of the resonant loops and the
position of the local minimum of this arc is dependent on the thickness of d).
As R) is decreased the center of the resonant loop shifts from a high real
component to a lower impedance. Figure 3.4.9b shows the effect of varying
Rz on the impedance behavior of the stacked patch configuration, once again
keeping all other parameters constant. Here increasing R z increases the real
component of the input impedance. Thus changing R) and Rz have contrary
effects on the overall behavior. This is very important when designing the
stacked configuration for a particular frequency band. For example, if the
resonant loop occurs at the desired frequency band, however the real part of
the impedance at the center of the resonant loop is low, then Rz can be
increased and R) decreased proportionally to shift the locus without affecting
the desired bandwidth. Similar loci to that shown in Figure 3.4.9 have been
93
achieved for square and annular ring patches when varying the outer
dimensions and so for the sake of brevity, these will not be presented here.
Adjusting the probe position has only a second order effect on the
overall impedance behavior of the stacked configuration. A trend similar to
that shown in Figure 3.4.9b can achieved when increasing xp, however the
relative movement in the center of the resonant loop is by no means as
extreme. Thus changing the position of the probe really only provides a fme
tuninz mechanism.
(a) (b)
Figure 3.4.9 Effects of Patch Dimensions on Impedance Locus (frequency range: 6 -11
GHz): parameters: Er. = 2.2, d) = 1.524 mm, tan SJ = 0.001, ea = 1.07, d2 = 2.5 mm, tan ~ =
0.001, xp = 5.4 mm, TO = 0.325 mm; (a) R J variation (R2 = 7.4 mm); (b) R2 variation (R J = 7.0
mm)
94
impedance control is reduced bandwidth, here typically by a few percent.
Simplistically this can be related to the size of the antenna. If we reduce the
volume of an antenna, for example reducing the width of a patch, the
bandwidth will decrease. This is indeed the case for the rectangular patches
where the width is reduced and for the annular rings where the inner radius is
increased. Varying the width of the top rectangular patch and the inner radius
of the annular ring has minor effects on the impedance loci of these stacked
patches.
3.4-4.1 Introduction
95
it will be evident that the derived figure of merit provides a good basis for the
design of micros trip patch antennas.
3.4-4.2 Theory
-
J(p,rp)=a - -{Xli')
a
P - 1 P
Ji'(XII'-)cosrp-a?-ll(XII'-)sinrp
a p a
(3.1)
Where 0 1 and G2 are the Tm and Tc Green's functions for the substrate
respectively and F I and F2 are the Fourier transforms of the currents given in
Equation 3.1 (please refer to [17] for more details of the formulation). The
power trapped in the surface wave modes is given by:
NSW( [ /16w+S ]]
Psw =~ - Re 4~ /16lf1 (P)F/ (P) +G2 (fJ)F/ (fJ)}pIp
(3.3)
From Equations (3.2) and (3.3) the power in the space wave can be
determined and therefore the surface wave efficiency can be evaluated.
96
The impedance of the mode can be defined as:
Z- -Prot
- 1 -- - R +]'X (3.4)
2//1
2
where I = P.d;.
It can be shown using simple circuit theory that the Q of the .mode is
given by:
Q= k o aX(ko)
(3.5)
2R(ko) ako
where the derivative, i)X/i)ko can be computed numerically and ko is the free
space wave number.
Thus the figure of merit for the circular microstrip can be computed
relatively easily as a function of the substrate parameters. It is important to
note that for this calculation the current coefficient of the mode is not required
and therefore a rigorous analysis is not needed.
97
stacked patch configuration (such as a probe or edge-fed) this assumption is
valid even for relatively broadband cases. Thus for the stacked configuration,
the impedance is written in terms of a simple two port network from which
the Q can be readily determined. The surface wave efficiency can be
determined from Equations (3.2) and (3.3) for the multi-layered case and thus
the figure of merit can be calculated.
3.4-4.3 Results
O. o.
o. Increasing lOr
mm
O. Decreasing Radius o. ~ d= 5.0
TJs.jQ 0.0
O.
."
• o.
~ 2.5 mm
0.01
~= 1.2 mm
o0 5 10 15 20 25 30 35 40 45 50 05 10 15 20 25 30 35 40 45 50 55
99
Case 5 in Table 3.4.1 presents the probe-fed circular stacked patch
configuration designed in the previous section. Here a combination of RT
Duriod 5880 (e, = 2.2) and Foam (e, = 1.07) was used to produce an antenna
with approximately 23 % bandwidth. As can be seen from the figure of merit
for this antenna, it is significantly better than for the single layered cases. As
for the single layered examples in Table 3.4.1, the dimensions of the stacked
patches were determined taking into consideration the properties of the
dielectric materials [19].
3.4-5.1 Introduction
100
likely be colored by the rules of thumb discovered previously. Instead we
employ numerical optimization to automatically synthesize high performance
antenna geometries under specified conditions and within given constraints.
Using this approach, the dielectric material values can be set and optimal
antennas using these materials can be synthesized. By comparing the
performance of the solutions resulting froin different material combinations,
design trends suggesting the best material combinations for efficient
broadband patch antennas may be identified.
The costfunction
Component weights
Keeping these rules of thumb' in mind, cases were run for different
Erl values leaving Er2 constant at 1.07, the value for hard foam. Values of Erl
simulated were: 2.2, 5.0, 7.5 and from 10 to 40 in steps of 5. All other
parameters in the probe-fed stacked patch configuration were set as variables.
Figure 3.4.13 presents the maximum impedance bandwidth achieved by the
optimizer for the stacked patch configuration as a function of the dielectric
constant of the lower layer. The open circles represent the optimum
bandwidth achieved when the procedure was constrained to produce solutions
exhibiting a minimum of 80 % in band surface wave efficiency, the filled
circles represent the bandwidths of optimum solutions with this constraint
relaxed.
40
l35
130
Qll
E
:>
.~ 25 -E-E-- ---
8
Bandwidlh (SWE > 80%) 0
20 Min. SWE (SWE > 80%) 0
Ban<lwidlh(relax SWE) V
Min. SWE (relax SWE) 0
15 0
o 5 10 15 20 25 30 35 40
Layer 1 Dielectric constant (l,)
Figure 3.4.14 presents the return loss and surface wave efficiency
frequency response for the optimal case with Erl =15. Note that the bands in
which high efficiency and good impedance bandwidths are achieved overlap,
although the maximum efficiency tends to be near the upper frequency edge
of the 10 dB return loss bandwidth. Similar frequency responses were
calculated for the range of values of Erl but are not shown for brevity. It is
interesting to note, however, that as Erl is reduced, the overlap between the
high efficiency band of frequencies and good impedance matching becomes
better and the efficiency is maintained over the matched impedance
bandwidth declining only slightly with reducing frequency. This implies that
the coupling between the 2 radiators improves the surface wave efficiency.
100
95
90
-5
85
l
~
e
80 s'"
75
'"'">
~
70
8
'"
't:
-15 cil
85
60
55
12 1~ lA 1~ 1~ 1.7 1.8
Frequency (GHz)
Figure 3.4.14 Return Loss Response and Surface Wave Efficiency for Optimized Patch
=
with Erl 15
105
The trend in Figure 3.4.13 shows that despite this coupling of the
surface wave mode to the upper patch, as the dielectric constant of the lower
layer is increased the efficiency starts to degrade. This can be explained by
having a closer look at the parameter values used to give the solutions in
Figure 3.4.13. As Erl is increased, good impedance bandwidth and efficiency
occurs when the height d, is decreased. Although intuitively, using well-
established trends for single layer patches (see Chapter 2), the lowering of the
height should to an extent counter the increase of Erl and therefore the overall
contribution to the surface wave may decrease, there is an important effect
that needs to be pointed out. As d, decreases, obviously the height of the
probe decreases and so the effect of the discontinuity associated with the
probe feed becomes less pronounced. It has been shown in the past that
surface waves couple well to discontinuities and that the presence of possible
discontinuities can determine whether the power is trapped within a surface
wave or radiated [21]. By reducing the height of d, the discontinuity
associated with the feed pin decreases and therefore the amount of power
coupled to the surface wave decreases.
106
for the optimum results presented in the previous section. Figure 3.4.15 also
presents the impedance locus for the optimized dimensions predicted by
Ensemble 6.1. For this simulation, the circular patch was divided into 36
segments and thus it is possible that discretization errors may account for the
slight discrepancy between the simulated results of Ensemble and the full
wave analysis used during optimization.
.. ..
7/'
~ ,
"
./
i
The theoretical and measured gain of the stacked antenna was 9.0 dBi
and 8.5 dBi at 1.4GHz, respectively. The discrepancy can be resolved by
considering the large backlobe in the H-plane and the accuracy of the antenna
range (± 0.5dB). These results verify that efficiency has not been
compromised to attain large bandwidth.
Figure 3.4.17 Cavitybacked stacked patch antenna(parameters: £'-. = 2.2, d l = 3.048 rom,
R1 =31.3 mm, xp =28 rom, yp =0, ro= 0.325 rom, fa = 1.07, dz = 10.0 rom, £'-3 = 2.2, d3 = 0.254
rom, Rz =34 rom, R3 =50 rom)
109
The simulated and experimental results of a cavity backed stacked
patch antenna designed for operation at 1.9 GHz are shown in Figure 3.4.18.
The parameters obtained from the simulation for this design are as follows: Erl
=2.2, dl =3.048 rnm, R, =31.3 rnm, xp =28 rnm, YP =0, ro =0.325 rnm, Ea =2.2,
d2 =0.254 rnm, R2 =34 mm, R3 =50 rnm. Foam with a thickness of 10 mm was
used between the lower patch and upper patch. It can be seen from this figure
that the theoretical and experimental impedance behavior are in very good
agreement. The 10 dB return loss bandwidths of 23% was obtained, the
largest ever reported for a cavity backed probe-fed stacked microstrip patch
antenna.
The measured gain of antenna was 7.35 dBi. The far-field radiation
patterns were measured at 1.9 GHz for E-plane and H-plane co-polar and
cross-polar levels. The plots of E-plane and H-plane co-polar are shown in
Figure 3.4.19 and Figure 3.4.20, respectively. The cross-polarization levels
were similar to a stacked circular patch reported in Section 3.4-3. Figure
3.4.21 shows a photograph of the cavity-backed antenna used in a printed
Antenna Remote Unit CARU) for optically distributed mobile communications
compliant with the 1.9 GHz peN specifications [25].
tOO
Figure 3.4.19 Measured Far-Field E-plane Co-polar Radiation Pattern of Stacked Patch
Antenna
110
'10
3.5-1 Introduction
112
Figure 3.5.2 Input Impedance of LargeSlot Excited Patch
113
As can be seen from Figure 3.5.3, the backward directed radiation
level is reasonably high with a front-to-back ratio (FIB) of 12 dB, compared
to a theoretical value of 15 dB. This sort of level can be an issue, in particular
for base stations for wireless systems. Typical base stations/radio hubs
provide a sectoral coverage area to increase the system capacity, for example
for LMDS (Local Multipoint Distribution Services) the 3600 azimuth plane is
usually split into four 900 sectors, whereas for a mobile base station it is
usually three 1200 sectors. Imperative for these sectoral systems is to
minimize the back radiation from each antenna to ensure interference from
adjoining sub cells is also minimized. Because a slot in a ground-plane
radiates equally in both the upper and lower half spaces (ignoring the affect of
the dielectric materials surrounding the ground-plane), the level of backward
radiation can become somewhat high. This leads to less power available in
the power budget of a link and more importantly for sectored wireless
communication systems, it can lead to increased levels of interference. The
predicted and measured gain of aperture coupled patch antenna was 9.0 dBi
and 8.8 dBi, respectively.
3.5-3.i introduction
Figure 3.5.4 Geometry of Aperture Coupled Microstrip Antenna with Reflector Element
115
patches
"
o
·10
..;r
'l\~~/I
~20 I
, ~
.;
::l
,
I
.3-30
~
E
::> ~! I-4-Theory
0-40 I -
a: ~ 1-'-'Expertment
(
· 50
I
-60
1.2 1.4 U U 2 2.2 2.4
Frequency (GHz)
116
''''
Figure 3.5.7 Measured Radiation Patternsof CavityBackedACP
0.5 - - - :- - 200
I "",/ I I I .....--.. ..... I I
.....)........ I I I I "'t-.__ ..1._ .._.
o w..J- 160
4 6 8 10 12 14 16 18 20
Reflector Length (mm)
Figure 3.5.8 Relative magnitude and phase of reflector and aperture far fields:
Aperture length =12.5 mm, aperture width = 1 mm, reflector width = 1 mm, reflector spacing =
5.635mm; fr = 9050Hz
Design Example
180 180
(a) (b)
Figure 3.5.9 Radiation patterns of microstripline fed aperture antenna with reflector
element at 9.5 GHz: (a) E-plane; (b) H-plane
Introduction
119
In this section three configurations of microstrip antennas based on
the aperture coupling feeding technique will be discussed. Each configuration
is capable of dual linear polarization, and is also capable of wideband
operation. Advantages and disadvantages of implementing each configuration
will be presented, with the main focus being on the potential of each
configuration to meet all of the performance specifications. Attention will
also be paid to ease of design, ease of fabrication, and overall material cost.
Dual OffsetApertures
By utilizing the fact that the aperture may be offset in the direction
perpendicular to resonance without suffering a drop in coupling, dual offset
apertures can be included in the design of an aperture coupled patch in order
to achieve dual linear polarization [30J. The geometry of this configuration
is shown in Figure 3.5.10. Here we have two perpendicular apertures feeding
a square patch, with each aperture feeding a separate polarization. This
geometry is attractive in that each aperture is coupled to a single feedline,
minimizing the design complexity of the feeding network. However, there are
two major drawbacks in implementing this geometry for this application. The
first drawback is the limitation of aperture size that is placed on the design -
by referring to the geometry in Figure 3.5.10 it is obvious that the size of each
aperture is restricted to less than half the patch length. This limitation on
aperture size forces the aperture to operate well below resonance, rendering it
extremely difficult to achieve broad bandwidth.
1/ Apertures
Pnrt 1 Coupling
"
II
U
= jf,
/'
Mlcrostrlp Patch
Port 2
Figure 3.5.10 Geometry of dual linear polarized patch antenna using offset apertures
120
The second, and most critical, drawback of this approach is the
inherent asymmetry introduced into the antenna. As mentioned previously,
this asymmetry decreases the amount of isolation that can be achieved
between the two polarizations. Practical limits on polarization isolation using
this technique are in the range of 20 to 25 dB. Also, the degradation of
polarization isolation is especially apparent when using thick substrates to
increase bandwidth [31], causing this option to be especially unattractive.
-,
Mlcrostrlp Patch
II Pnrt 1
,,;:¥
Crossed Slot,
V I)
In Ground Plane "-
""-- ,--/ Air Bridge
Port 2
Figure 3.5.11 Geometry Crossed Slot ACP with Balanced Feed Network
121
Dual Layer Crossed Slot with a Balanced Feed
I~
Port 1
Figure 3.5.12 Geometry of CrossedSlot Fed Patch Antenna with Dual Layer Balanced
Feed Network
3.6-1 Introduction
Ground Plano
123
From this enhanced bandwidth configuration, bandwidths (defined as
having a VSWR < 2:1) from 50 to 70% have been realized. This is the widest
instantaneous bandwidth yet achieved for a microstrip patch antenna. Section
3.6-2 provides a comparison of different methods for bandwidth
enhancement. Techniques for impedance matching are presented in Section
3.6-3. The impedance behavior is very sensitive to several physical
parameters of the antenna, and the effect of varying these parameters is
analyzed in Section 3.6-4. Variations on the design, such as incorporation of
additional patches, are discussed in Section 3.6-5. Details of an experimental
design are presented in Section 3.6-6. An example of an ASP designed at
millimeter-wave frequencies, where material selection is somewhat limited is
given in Section 3.6-7. Section 3.6-8 investigates techniques to reduce the
backward directed radiation of an ASP microstrip antenna, namely using
cavity backing and also a reflector patch.
124
At the expense of increased design complexity, back radiation levels
may be reduced through the use of stacked patches. The impedance locus for
a typical aperture coupled stacked patch design is shown in Figure 3.6.2b.
Here the overcoupled resonance of the lower patch is combined with a low-Q
mutual resonance between the patches, and an impedance locus similar to that
of Figure 3.6.2a results. The antenna substrates consist of foam with a total
thickness equal to that of the previous example. A 1.5:1 VSWR bandwidth of
20% is achieved, and the front-to-back ratio is improved to better than 18 dB
over the band.
(a) (b)
(c)
Figure 3.6.2 Impedance Loci of Wideband Aperture Coupled Microstrip Antennas:
(a) Broadband Patch using Thick Substrate and Near-Resonant Aperture;
(b) Aperture Coupled Stacked Patch; (c) ASP antenna
The use of a resonant aperture with stacked patches results in the ASP
antenna, with a typical impedance locus shown in Figure 3.6.2c. This locus
differs from those of Figures 2a and 2b, with two tight loops near the center of
the Smith chart. The locus is produced by the interaction of the individual
125
resonances - a low-Q mutual resonance produced by the interaction of the two
patches, and a mutual resonance produced by the interaction of the aperture
with the lower patch. The latter resonance is overcoupled in the sense that the
interaction between the aperture and patch is strong. In the absence of the
upper patch, the impedance locus resulting from this overcoupled resonance is
similar to Figure 3.6.2a, however exhibiting a larger loop that typically cannot
be matched to a VSWR<2:1. The mutual resonance between the aperture and
lower patch is associated with the lower frequency loop in the impedance
locus and the mutual resonance of the two patches produces the upper
frequency loop. These mutual resonances can also couple to each other,
changing the overall .behavior of the impedance locus. The addition of
another mutual resonance results in a dramatic increase in bandwidth
compared to the previous two configurations, with a 1.5:1 VSWR bandwidth
of 44% being achieved. The required antenna substrate thickness and
aperture size are also increased and this in tum produces higher back radiation
levels, with a front-to-back ratio of better than 10 dB over the band for this
example.
low Impedance
line 2d « 1000 line
~ o~ / '
../'
50nllne
50nllne ~
(a) (b)
Figure 3.6.3 Impedance Matching Techniques for ASP antenna : (a) wide centered
feedline; (b) dual offset feedlines
The dimensions of the prototype design are given in Figure 3.6.4, and
in each of the examples to follow, all parameters except for one are held
constant to the values given. To have an extremely rigorous theoretical model,
the currents on the feedline should be modeled for this wide feedline case.
127
However, the model chosen to generate the data was based on the reciprocity
method, as it is acceptable to show the trends as the individual parameters are
varied while being computationally efficient.
Figure 3.6.5 shows the effect of varying the aperture length on the
input impedance locus. It is seen that this has a pronounced effect on the
lower frequency loop while having a minor effect on the other loop. This
observation supports the previous statement that one loop is due to the
interaction of the aperture and lower patch, while the second is due mainly to
the interaction of the two patches. Lengthening the aperture has the effect of
increasing the coupling between the aperture and lower patch (therefore
increasing the size of the lower frequency loop), while at the same time
bringing the two loops closer together.
The effect of altering the size of the lower patch is shown in Figure
3.6.6. This is a critical parameter as the action of the lower patch contributes
to both mutual resonances, and therefore varying its dimensions has an effect
on both loops in the impedance locus. It is very interesting to compare the
plots of Figure 3.6.6 with those of Figure 3.6.7, which shows the effect of
varying the upper patch dimension. Note the similarities between Figures
3.6.6a and 3.6.7c, and Figures 3.6.6c and 3.6.7a. These similarities show that
128
it is not the absolute dimensions of each patch that govern the impedance
behavior, but the relative size of each patch to the other that is important.
(a) (b)
(c)
Figure 3.6.5 Input Impedance as Function of Aperture Length: (a) SL=8mm. (b)
SL=10mm; (c) SL=12mm
129
the corresponding loop in the impedance locus is smaller. The same effect
occurs when using a small aperture, and is evident by the similarities in
Figures 3.6.5a and 3.6.8c. In terms of design criteria, this signifies that when
aperture size is increased, the lower substrate thickness must also be increased
to maintain the same impedance behavior.
(a) (b)
(c)
130
(a) (b)
(c)
Figure 3.6.7 Input impedance as a function of upper patch size:. (a) P~= PW2=9mm;
(b) P~= PW2=1 Omm; (c) P~= PW2=12mm
(a) (b)
131
(c)
Figure 3.6.8 Input Impedance as Function of Layer #1 Dielectric Thickness: (a)
d t=2.5mm; (b) d 1=3mm; (a) d J=4mm
(a) (b)
(c)
132
3.6-5 Discussion
It was stated in Section 3.6-2 that the lower frequency loop in the
impedance locus was associated with the mutual resonance of the aperture and
lower patch while the upper frequency loop was due mainly to the interaction
of the two patches. While this statement was supported by the data presented
in Section 3.6-4 for the configuration under examination, it also raises two
important questions. First, is this true for any ASP configuration? And
second, can the incorporation of more than two stacked patches in the
multilayer geometry of the ASP produce a third mutual resonance and
enhance the bandwidth even further?
133
Feedparameters - Wr= 0.5mm, doff = 5.4mm, LSlUb= 5.8mm, Err = 2.2, dr =
0.635mm, tan or= 0.0009.
- ~ - ~ - ~ ~ - ~ -~-
I
'-'-1-
I J
--1--1-
I
0.4
200 _L _'-_L~_J_.1_.l _ _ .I_
180 - - - - - - --- - - - - - -
G.3 I I I I J I I I
(~ ~)
Figure 3.6.10 Ratio of DominantModePatch Currents, 12111: (a) Magnitude; (b) Phase
I I I I I I 1
V - '
iii ·10
I I I
- -r - - r - - , - - , - - ~
I
- r--
I 1 1 "'- , I I I
~ ·15 -- ~ -- ~ I - - I - 1 -7i1 ---1-- -~--
3 4 5 6 7 6 9 10 11 12
Frequency (GHz)
iii
~
c
ii
CJ
o ""'-'-'u-Wu-Wu..J..w...o...41I
3 4 5 6 7 6 9 10 11 12
Frequency (GHz)
3.6-7.1 Introduction
The predicted and measured return loss of the Ka-band ASP are
shown in Figure 3.6.13. The measurements were carried out using a 40 GHz
Wiltron 360 B vector network analyzer. As can be seen from this figure,
theory and experiment are in relatively good agreement. The dimensions for
the antenna are given in the figure caption. The discrepancies in the results
can be attributed to uncertainties in the dielectric constants of the materials
used (± 0.02), alignment errors of the layers and the uncertainty associated
with the presence of the small air gaps. Despite these, a 10 dB return loss
bandwidth covering the Ka-band was achieved for the fabricated mm-wave
antenna. The broadband nature of these printed antennas ensures that the
effects of design and fabrication uncertainties are minimal. The tightened
loops in the experimental impedance loci and hence the lower return loss over
a smaller bandwidth in Figure 3.6.13, are most likely due to an
137
underestimation in the thicknesses of the air gaps, which were assumed to be
0.1 mm. It should be noted that no modifications/reiterations in the design of
the ASP were undertaken as the specification was met in the first fabricated
antenna.
Figure 3.6.13 Measured and Predicted Impedance Loci of Ka-band ASP (Antenna
Parameters: Dielectric Layers : Er • = 1.0, d. = 0.1 rom, Ea = 2.2, d2 = 0.787 rom, Er3 = 1.07, d3 =
1.1 rom, Er4 = 2.2, d4 = 0.254 rom, Patches: P LI = PW1 = 2.5 rom, N = 2; Pu = P W2 = 2.4 rom, N
= 3, SL= 2.5 rom, Sw = 0.2 rom . Feed Parameters: Erf= 2.2, d r = 0.254 rom, Wr= 0.8 rom, L SlUb
=0.65 rom)
The E- and H-plane co- and cross-polar radiation patterns of the mm-
wave antenna were measured across the Ka-band. A typical measured
response for the co-polar E- and H-plane patterns are shown in Figure 3.6.14,
which were measured at 38 GHz. The evident ripple in the E-plane is typical
for an aperture-eoupled antenna with a finite ground-plane (for this antenna 5
x 3 em). Note also there is an apparent asymmetry in the E-plane pattern.
This is due to the presence of the K-type connector used to couple power to
and from the microstrip feed-line. The flange for the connector is relatively
large in comparison to the printed antenna and its shielding effect quite
pronounced for angles near the appropriate endfire (here e --7 + 90 0 ) . The
gain of the antenna was measured across the Ka-band and 'was greater than 6
dBi over this frequency range. The cross-polarization levels were more than
20 dB below the co-polar patterns in each plane. The front-to-back ratio was
greater than 12 dB across 26 - 40 GHz, approximately 4 dB greater than that
reported in Section 3.6.5 due to the reduced bandwidth requirement here and
hence reduced slot size of the antenna. The surface wave efficiency of the
mm-wave antenna was calculated as greater than 85 % across the Ka-band.
138
110
3.6-8.1 Introduction
-' .
40
·r·' :--'~'"
. ~
~ , ,, .'
35 ~-
.~-,- "~.....~
. .,
lil30
~25
o
~20
,,
,.
1/ ,.' ,,~
/ , ,, .' "
.'
Increase LR
\.
-,
~
u 15 ~
~ ' .' V
.,~
<, '\
0 10
";' -r ,
.' .' no_reflector
-,
C5 ---- 6.5mm
e0 /
IL
•..•..·· 6.1mm
- ·- ·6.9mm
"-
-5
i -"-" 5.5mm
7.5mm
-10
22 24 26 28 30 32 34 36 38 40 42 44
Frequency (GHz)
Figure 3.6.16 Effect of Length of Reflector Patch on Front-to-back Ratio
141
45
40
.... ......
' . ....... .. .
;J;'.
m35
:2- ~ ..., , .... ......... ...........
..../" ......
030
i ..,:>\ IncreasedR ....... ....... ..'
....,
~25 , .......'
u
~20 .".."'1
.~;;-
'. . ~~,
~15
c
o
.
.t10 f.it
~,.
.......V - no_reflector -, ......
V ····1.1mm
....····1.3mm '\.
"
5
- •• • O.9mm
o
~ ~ U ~ 30 ~ M ~ ~ ~ ~ ~
Frequency (GHz)
Figure 3.6.17 Effect of Spacer Thickness for Reflector Patch on Front-to-beck Ratio
r---.-r---.-r--.-.--.-.--.-.----,
40
..
..... ,~' :,.............
~30 1--t---"l~'_;jL-I---+-~t--'-t"""""'"'i--o.m--+---j
/,'11...;
o : ~#},
~ 25 F--++t-:L~,.___-b-"o.q...-+--1--+--P~+--f
~20 I-+~I-*---::~--+--t--+-+--+---"\~--j
X
~c·15
..
"
.'/
.l,~'· . / - r:
no_re IIec t or
~.
'.
E10 ~"""':"--1I---H"" O.3mm H--I--t--'<:I--;
IL
5 ....
~. ........ 0.2mm
- .-. O.4mm
·"···1.Omm
142
Figure 3.6.19 Simulated Input Impedance of ASP with and without Reflector Element
I--::::r----~---"ro:
~i
metallic cavity
144
Figure 3.6.21 shows the predicted and measured return loss of the designed
Ka-band cavity-backed ASP (refer to the figure captions for the dimensions of
the antenna). As was the case for the ASP, there is relatively good agreement
between theory and experiment with the bandwidth requirements clearly met.
Figure 3.6.22 shows the measured E- and H-plane co-polar radiation patterns
.of the cavity-backed antenna at 32 GHz. Once again the ripple and
asymmetry in the E-plane is evident in the plot. These features can be
attributed to both the finite ground-plane, which for this antenna is even
smaller than for the fundamental ASP of Section 3.6-7 hence leading to the
large ripple, and the K-type connector. For the cavity-backed ASP this
connector is located closer to the ASP and therefore its effect on the radiation
pattern is more dramatic than in Figure 3.6.14. The gain of the antenna was
measured as greater than 6 dBi across the 26 - 40 GHz frequency range and
as before the cross-polarization levels in each plane were at least 20 dB below
the co-polar patterns. The front-to-back ratio of the cavity backed antenna
was measured as greater than 20 dB across the Ka-band.
."
~
-5
~\
-m -: , , ,..--- '.
-10 ,,~~~
-
~
\, ,, ,
<,
-15
/ ,
'tJ ~
I~
,
(I)
(I) ·20
r\ /,l
.. ,
I
0
"
..J ~ I
f
c ·25
::J
l) -30
I
, I
I
\....,f
a: II
I
v
I--theory
·35 ,- - - - experiment I
-40 I
-45
25 27.5 30 32.5 35 37.5 40
Frequency (GHz)
Figure 3.6.21 Measured and Predicted Return Loss of Ka-band Cavity-backed ASP
(Antenna Parameters: Dielectric Layers: Ert = 2.2, d t = 0.125 nun, E,z = 1.0,d2 = 0.1 nun, Er3 =
2.2, d3 = 0.508nun, E,4 = 1.07, d, = I.l nun, E,4 = 2.2, d4 = 0.125 nun, Patches: PLI = PWt = 2.5
nun, N =3; Pu = PW2 = 2.3 nun, N = 4, SL = 2.45 nun, Sw = 0.2 nun. Feed Parameters: En =
2.2, de = 0.125 nun, We = 0.8 mm, L.,Ub = 0.75 nun, Ae = Be = 5 nun, De = 3 nun, Ere•• = 1.0)
145
3.6-9 Dual Polarized ASP
3.6-9.1 Introduction
~
c r8. h8
Patch 2 (W2. L2) - - -.L....",,- ~
cr7. h,
~
c r6. h6
~
Patch I (W I. LI) crs. hs
~
Feed 2 (W2F. cr4. h,
L2SruD)
~
cr3. h3
Er2. h2
~
ErI.hl
Reflector (W I REP Ll REF
Reflecto r (W2REP L2RE!F)c:====::t::
Figure 3.6.24 (a) and 3.6.24 (b) show the measured and calculated
return loss for Port 1 and Port 2 of the dual polarized ASP antenna,
respectively. As can be seen from these plots, there is good agreement
between theoretical and experimental results. Bandwidths of greater than 52%
(VSWR < 2:1) for both polarizations were obtained. The lower edge of the
matched band is slighter lower for Port 2 than Port 1. This can be attributed
to the height of upper feed substrate and so the coupling between slot and the
patch element for this polarization is stronger than between slot and the patch
element for Port 1 excitation. The shift is less than 1 %. Figure 3.6.25 shows
measured and calculated isolation between two ports. An isolation of more
than 39 dB over the bandwidth has been achieved. As mentioned before, the
two feed networks are separated by the ground plane, thereby eliminating any
cross coupling between the microstrip-lines in the two networks.
(a) (b)
Figure 3.6.24 Measured and Calculated RL of Dual polarized ASP Antenna : (a) Port I,
(b) Port 2 (------- calculated)
· 30
· 35
·40
en
:Eo
·45
I
I
i}j ·50 ,
I
,
I
I
·55 I
I
I
I
I
·60 \ I
I ,I'
·65
1.5 2 2.5 3
Frequency (GHz)
Figure 3.6.25 Measured and Calculated Isolation between Port 1 and Port 2 (------ calculated)
Figure 3.6.26 E-plane Radiation Patterns of Dual Polarized ASP Antenna: (a) port I, (b)
Port 2 (---- X-pol.)
149
Figure 3.6.27 H-plane Radiation Patterns of DualPolarized ASP Antenna: (a) Port I, (b)
Port2 (----X-pol.)
3.7-1 Introduction
Antenna 1: 1 - 2 GHz
Feed width, Wr 1.4 nun Feed thickness, hr 1.58 nun
Stub length, 1, 26 nun Laver 1 thickness, hi 19 nun
Offset, Lorr 30 nun Layer2 thickness, h2 3.175 nun
Aperturelength, L. 84 nun Layer 3 thickness, h3 9 nun
Aperturewidth, W. 10 nun Layer4 thickness, !4 1.58 nun
Patch 1 length,L1 64 nun Feed, En 2.2
Patch 1 width,WI 150 nun Layer 1 Erl 1.07
Patch 2 length, k 60 nun Laver 2, Er2 2.2
Patch 2 width,W2 150 nun Laver 3, Er3 1.07
Layer4, Er4 2.2
Antenna 2: 1 - 2 GHz
Feed width, Wr 1.4 nun Feed thickness, hr 1.58nun
Stub length, 1, 14 nun Layer I thickness, hi 8 nun
Offset, Lorr 15 nun Laver 2 thickness, h2 3.175 nun
Aperturelength, L. 41 nun Laver 3 thickness, h3 5 nun
Aperturewidth, W. 5 nun Laver4 thickness, !4 1.58 nun
Patch 1 length,L1 32 nun Feed,Err 2.2
Patch I width, WI 65 nun Laver 1 Erl 1.07
Patch 2 length, k 30 nun Layer2 ,Er2 2.2
Patch 2 width,W2 65 nun Laver3, Er3 1.07
Layer4, Er4 2.2
151
The calculated and measured Return Loss (RL) for the lower
frequency ASP are shown in Figure 3.7.1. The calculated bandwidth (RL < -
10 dB) is 1-2.3 GHz (75 %) and the measured impedance bandwidth (RL < -
10 dB) is 1-2.1 GHz (71 %). There is a discrepancy between the predicted and
measured results, with a more important impact at the higher frequency band
of the ASP. One possible cause of this narrowing of the acceptable return loss
bandwidth could be due to the construction of the multi-layered antenna. As
can be seen from Table 3.7.1, the first and third layers are required to be
19mm and 9mm of foam respectively. These layers were constructed using
IOmm, 5mm and 4mm sheets of Rohacell. Thus air-gaps could have been
present in the resulting ASP, reducing the overall bandwidth. However
despite this, there is a good agreement between measured and computed
return loss, so further iteration is not required.
10
iii
~ 15
oJ
a: 20
25
30 '----~---~---'------'
0.5 1.5 2 2.5
Frequency (GHz)
The E and H-plane radiation patterns (predicted and measured) for the
lower frequency ASP at 1.5 GHz are shown in Figure 3.7.2. The Front to
Back Ratio (FBR) is a bit higher than predicted which, is mainly due to the
use of a small ground-plane (the ground-plane of the antenna). The size of
ground-plane was 1.4 Ao x 0.7 Ao (at the center frequency). Using a larger
ground-plane would provide better FBR, however that increases the weight
and the cost of the ASP antenna.
EXP -
T>£O - •
(a) (b)
Figure 3.7.2 Radiation Patterns of ASPl at 1.5 GHz: (a) E-plane; (b) H-plane
TliEO-
10
•
f •
j
•
I
10 (
r-. i
1/
: : \ 0
00 ;
:: \,
25 II
I
0 0
,0
.f
0'
30
Frequency (GHz)
(a) (b)
Figure 3.7.5 Computed and Measured Radiation Patterns of ASP2 at 3 GHz: (a) E-
plane; (b) H-plane
The predicted and measured gain of the upper frequency ASP are
shown in Figure 3.7.6. This ASP has a measured gain of 8 ± 1.5dBi in the
frequency range of 2-4.1GHz. This gain response reduces the bandwidth of
the second antenna from 79 % to 70 % as the 3dB gain bandwidth is 2 -
4.15GHz. From Figure 3.7.8 it can be seen that the gain drops very rapidly at
4.10Hz. Once again the"radiation patterns of this ASP were very similar to
typical ASP antennas that are presented in Section 3.6. The FBR for this ASP
antenna over the 2-40Hz was no worse than 8dB.
154
El!J'-
THEO -
10
I u • u
'-flHol
The requirements for the filters compnsmg the diplexer for this
application are as follows. The filters must have pass bands at least equal to
the broadband antennas to which they are attached. The return loss within the
band of each filter must be as low as possible to ensure that the total RL of the
UBP antenna is less than -10dB across the entire band of operation. For this
application one filter must pass 1-2GHz, and reject 2-4GHz and the other
filter must pass 2-4GHz while rejecting 1-2GHz. It is imperative that each
filter has a sharp roll-off to ensure array-like distortions of the radiation
pattern of the UBP does not occur. For this case the interaction is at 2GHz.
Further, since the filters will be imperfect and are of limited order,
they will have finite impedance characteristics in their stop band. The result
of this is that when the two filters are connected, they will effectively load one
another and influence the respective band performances. In order to transfer
all the power to the designated port, the input admittance of each filter must
be made complementary. In normalized form this condition can be
represented as in [46]:
(3.6)
(3.7)
155
(3.8)
A final requirement, once all others have been met, is that filters must be
composed of realizable components.
When the realized filters were connected in parallel there was a dip
and then a sharp rise at the cross over frequency in the simulated results,
which implied that the imaginary part of the admittances were not canceling
each other. In order to reduce this distortion, the value of the first capacitor
for the high pass filter was increased and the value of the first inductor of the
low-pass filter decreased. The values of the other elements needed to be
altered in order to minimize the ripple in the pass band as well as reduce the
susceptance.
The final parameters and lumped element values for the diplexer are
shown in Table 3.7.3. It is necessary to point out that the width of microstrip
lines were calculated using equations in [47]. The width of the required
inductors were 0.33 mm which corresponds to 100 0), and the width of the
capacitors were 3.3 rom which equates to 25 0) using a substrate with a
dielectric constant of 3.05 and a height of 0.508mm.
iD
0 1--- - - _ . .-..-...--............-=---4
-" '- .
~ 10
C
o
.. 20
eu
:J
c
1lI 30
~
40
Frequency (GHz)
10
-"
U U 2 U 3 U 4 U
Frequency (GHz)
158
3.7-4 UBP Antenna Results
al,~ I al
---.... : --...
I
i
_1_ rn~~~ 1
bll~ i bl i
i i
l !
Figure 3.7.9 Block Diagram of Ultra-wideband Printed Antenna
Theoretical and measured results for the return loss of the UBP
antenna are shown in Figure 3.7.11. As it can be seen from Figure 3.7.11 the
measured return loss is not as low as the predicted results. At higher
frequencies (> 4GHz) the measured return loss approaches OdB, while the
159
predicted results approaches -5dB. This is mainly due to the difference in the
simulated and measured results of the diplexer alone. Referring to Figure
3.7.8 the measured return loss for the diplexer at higher frequencies (> 40Hz)
is around -5dB, while the simulated results are around -13dB. These
discrepancies were discussed in earlier. The fine discrepancies apparent in
Figure 3.7.11 between the experimental and theoretical results are due to no
de-embedding of the SMA conductors used in the UBP, as mentioned
previously.
iii 10
:2-
..J
II:
15
20
25
160
should be dominant; (2) the cross over' frequency where the mutual coupling
between the two radiators may be detrimental; and (3) the upper frequency
band where the high frequency ASP should be dominant.
Figure 3.7.11 shows the E-plane radiation pattern for ultra broadband
antenna at 1.50Hz, 2 OHz and 30Hz. At 1.5 OHz, the minimal effect of the
second sub-antenna (240Hz) is observed as in the E-plane the second sub-
antenna will not affect the performance of the UBP antenna because it is
located in the H-plane. The measured pattern is very similar to typical ASP
antennas. The ripples in the radiation patterns are due to diffraction of
aperture radiation at the edge of ground-plane. These ripples are within 2-3dB
which is in an acceptable range for most printed antennas.
(a) (b)
(c)
Figure 3.7.11 Measured E-planeRadiation PatternsofUBP Antenna: (a) 1.5 GHz; (b) 2
GHz; (c) 3 GHz
161
the diplexer at 2GHz (4dB) compensates for this variation in beamwidth
thereby yielding a constant gain antenna. For frequencies just beyond the
cross-over frequency (2GHz) less distortion in the H-plane pattern may have
been achieved, if the diplexer consisted of Chebyshev filters, with steeper
roll-offs than the used maximally flat filters. However, Chebyshev filters
would have compromised the return loss of the UBP. There is a distortion in
the right side of the radiation pattern of the UBP antenna at 3 GHz, due to the
electrically large size of the first sub-antenna.
(a) (b)
(c)
Figure 3.7.12 Measured H-plane Radiation Patterns ofUBP: (a) 1.5GHz; (b) 2.0 GHz
and (c) 3.0 GHz
Figure 3.7.13 shows the measured gain of the UBP. As can be seen
from this figure at least 6 dBi of gain has been achieved from 1 - 3.7 GHz.
3.8 Summary
162
have been presented and explored, each with their own merits and degree of
complexity.
10 , - - - - - , - - - - - - . - - - r - - - - , - - - r- - r -- -,-- - - ,
In Section 3.5 large slot excited patches were investigated and several
design examples presented. Techniques to overcome the backward directed
radiation were examined. In Section 3.6 design criteria for the ASP (aperture-
stacked patch) microstrip antenna, which utilizes a resonant aperture with
stacked patches, were presented. The characteristic action of the resonators in
this antenna produces a greatly enhanced bandwidth over that exhibited by
other aperture coupled microstrip elements. Impedance matching techniques
were discussed, and the effects of several key physical parameters of the
antenna were examined. Results of this parameter study provide a good
design guide for this antenna. A linearly polarized experimental design was
presented, from which an octave bandwidth was realized. A wideband dual-
163
polarized printed antenna based on a dual feed ASP configuration was also
presented. This microstrip antenna has a VSWR < 2:1 bandwidth of greater
than 50 % and isolation between the excitation ports of greater than 39 dB.
The decision of which of all the techniques presented to use can only
really be made taking into consideration other aspects of the antenna terminal.
A general 'rule of thumb' that applies here is to make the antenna as simple as
possible whilst stilI satisfying the performance goals. For this reason , the
progression of this chapter is probably the best plan of attack, when
considering an enhanced bandwidth problem. Further examples illustrating
this will be presented in subsequent chapters.
3.9 Bibliography
[1] T. Chio and D. H. Schaubert, "Effects of slotline cavity on dual-polarized tapered slot
antenna arrays", 1999 Antennas & Propagation International Symposium. Orlando
USA, pp. 130 -133, July 1999.
[2] 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, Orlando, FL, USA, pp. 924 - 927, July 1999.
[3] 1 M. Johnson and Y. Rahmat-Sarnii, "Genetic Algorithms and Method of Moments
(GAlMOM) for the design of integrated antennas," IEEE Trans. AntennasPropagat.,
Vol. AP-47, pp. 1606-1614, October 1999.
[4] H. F. Pues and A. R. Van de Capelle, "An Impedance Matching Technique for
Increasing the Bandwidth of Microstrip Antennas," IEEE Trans. AntennasPropagat.,
Vol. AP-37, pp. 1345 -1354, November 1989.
[5] l-S. Kuo and K. - L. Wong, "A dual-frequency L-shaped patch antenna," Microwave
& OpticalTechnology Letters, vol. 27, pp. 177-179, Nov. 2000.
[6] G. Kumar and K. C. Gupta, "Non radiating edges and four edges gap-coupled
multiple resonator broad-band microstrip antennas", IEEE Trans. Antennas
Propagat., Vol. AP-33, pp. 173-178, February 1985.
[7] R. B. Waterhouse, "Small microstrip patch antenna," Electronics Letters, vol. 31, pp.
604 - 605, April 1995.
[8] W. C. Chew, "A broadband Annular Ring Microstrip Antenna," IEEE Trans.
AntennasPropagat., vol. AP-30, pp. 918 - 922, May 1982.
[9] D. M. Kokotoff, 1 T. Aberle and R. B. Waterhouse, "Rigorous analysis of probe-fed
printed annular rings," IEEE Transactions Antennas & Propagation, vol. 47, pp. 384
- 388, Feb. 1999.
164
[10J S. A. Long and D. M. Walton, "A dual-frequency stacked circular-disc antenna,"
IEEE Trans. Antennas Propagat., Vol. AP-27, pp. 270-273, March 1979.
[IIJ 1. T. Aberle, D. M. Pozar and 1. Manges, "Phased arrays of probe-fed stacked
microstrip patches", IEEETrans. Ant. & Prop., vol. AP-42, pp. 920-927, July 1994.
[12J 1. T. Aberle and D. M. Pozar, "Analysis of infinite arrays of probe fed rectangular
microstrip patches using a rigorous feed model", Proc. lnst. Elec. Eng., Pt. H, vol.
136, pp. 110-119, April 1989.
[13J 1. S. Dahelle and K. F. Lee, "Characteristics of annular ring microstrip antenna",
Electron. Lett., Vol. 18, pp. 1051-1052, November 1982.
[14J Z. Nie, W.e. Chew and Y.T. Lo, "Analysis of the annular-ring-loaded circular disk
microstrip antenna," IEEE Trans. Antennas Propagat., vol. AP-38, no. 6, pp. 806-
813, June 1990.
[15J S. D. Targonski, R. B. Waterhouse, and D. M. Pozar, "A wideband aperture coupled
stacked patch antenna using thick substrates," Electronics Letters, vol. 32, pp. 1941 -
1942, October 1996.
[16J C. A. Balanis, AntennaTheory: Analysis and Design, Wiley, New York, 1997.
[17J J. T. Aberle and D. M. Pozar, "Accurate and versatile solutions for probe-fed
microstrip patch antennas and arrays", Electromagn., vol. 11, pp. 1-19, Jan. 1991.
[18J D. M. Pozar, "Microstrip antennas", Proc. IEEE, vol. 80, pp. 79-91, Jan. 1992
[19J J. R. James and P. S. Hall, Handbook of Microstrip Antennas, London, UK: Peter
Perergrinus, 1989.
[20J D. M. Kokotoff, R. B. Waterhouse, C. R. Birtcher and 1. T. Aberle, "Annular ring
coupled circular patch with enhanced performance,' Electronics Letters, vol. 33, pp.
2000 - 2001, Nov. 1997.
[21] R. B. Waterhouse, "The use of shorting posts to improve the scanning range of probe-
fed microstrip patch phased arrays," IEEE Transactions Antennas & Propagation,
vol. 44, pp. 302 - 309, March 1996.
[22] F. Croq and D. M. Pozar, "Millimeter wave design of wide-band aperture-coupled
stacked microstrip antennas", IEEE Transactions Antennas & Propagation, vol. AP-
39, pp. 1770-1776, December 1991.
[23] R. B. Waterhouse and N.V. Shuley, "Scan performance of infinite arrays of
microstrip patch elements loaded with varactor diodes," IEEE Trans. Antennas
Propagat., vol. AP-41, pp.l273 - 1280, Sept. 1993.
[24] 1. T. Aberle and F. Zavosh, "Analysis of Probe-fed Cicrular Micostrip Patches
Backed by Circular Cavities," Electromagnetics, Vol. 14, pp. 239 - 258, 1994.
[25] M. Lye, R. B. Waterhouse, D. Novak, F. Zavosh and J. T. Aberle, "Design and
development of a printed antenna remote unit for optically distributed mobile
communications," IEEEMicrowave and Guided Wave Letters, vol. 8, pp. 432 - 434,
December 1998.
[26] J.F. Zurcher, "The SSFIP: A global concept for high performance broadband
planar antennas", Electronics Letters, vol. 24, pp. 1433-1435, Nov. 1988.
[27] N.K. Das and D.M. Pozar, "A generalized spectral domain Green's function for
multilayer dielectric substrates with applications to multilayer transmission lines,"
IEEE Transactions on Microwave Theory and Techniques, vol. MTT-35, pp. 326-
335, Mar. 1987.
[28] S. D. Targonski and R. B. Waterhouse, "Reflector elements for aperture and aperture
coupled microstrip antennas," IEEE Antennas & Propagation Symposium, Montreal
Canada, pp. 1840 - 1843, July 1997.
[29] R. B. Waterhouse, M. Lye and S. D. Targonski, "Design of printed antennas for
mobile base station applications," Third Asia-Pacific Conference on
Communications, APCC '97, Sydney Australia, pp. 242 - 246, Dec. 1997.
165
[30] S.D. Targonski and D.M. Pozar, "Design of wideband circularly polarized
aperture coupled microstrip antennas" , IEEE Transactions on Antennas and
Propagation, Vol. 41, pp. 214-220 , February 1993.
[31] F. Croq and A. Papiemik, "Wideband aperture coupled microstrip antenna",
Electronics Letters, vol. 26, pp. 1293-1294, August 1990.
[32] 1.R. Sanford and A. Tengs , "A Two Substrate Dual Polarized Aperture Coupled
Patch ", IEEEAntennasand Propagation Symposium Digest, pp. 1544-1547,1996.
[33] D.M. Pozar, "A reciprocity method of analysis for printed slots and slot
coupled microstrip antennas ", IEEE Transactions on Antennas Propagat., Vol. AP-
34, pp. 1439-1446, December 1986.
[34] D. A. Gray, "Optimal cell deployment for LMDS systems at 28 GHz," Proc. Wireless
Broadband Conf., Washington DC, July 1996.
[35] H. Ogawa, D. Polifko, and S. Banba , "Millimeter-wave fiber optics systems for
personal radio communication," IEEE Trans. Microwave Theory & Techniques, vol.
40, pp. 2285 - 2292, Dec. 1992.
[36] Z. Ahmed, D. Novak , R. B. Waterhouse, and H. F. Liu, "37 GHz fiber-wireless
system for distribution of broadband signals," IEEE Trans. Microwave Theory &
Techniques, vol. 45, pp. 1431-1435, Aug. 1997.
[37] L. D. Westbrook and D. G. Moodie, "Simultaneous bi-directional analog fiber-optic
transmission using an electroabsorption modulator," Electronics Letters, vol. 32, pp.
1806 - 1807, Sept. 1996.
[38] K. Li, 1. X. Ge, T. Matsui, and M. Izutsu, "Millimeter-wave sub-carrier optical
modulation, photodetection and integration with antenna for optic fiber link system,"
Proc. 1999 IEEE MTT-S Int. Microwave Symp.• Anaheim, CA, USA, pp. 1015 -
1018, June 1999.
[39] A. Stohr, K. Kitayama , and D. Jager, "Full-duplex fiber-optic RF subcarrier
transmission using a dual-function modulator/photodetector," IEEE Trans.
Microwave Theory & Techniques, vol. 47, pp. 1338 -1341, July 1999.
[40] D. Novak , G. Smith, A. Nirmalathas, C. Lim and R. B. Waterhouse, (invited) "Fiber-
optic networks for millimeter-wave wireless communications," Proc. 1998 Asia
Pacific Microwave Conference, Yokohama, Japan, pp. 309-316, Dec. 1998.
[41] D. M. Pozar, S. D. Targonski and H. D. Syrigos, "Design of millimeter-wave
microstrip reflectarrays,' IEEE Trans. Antennas Propagat. , vol. 45, pp. 287 - 296,
February 1997.
[42] W. Menzel, D. Pilz and R. Leberer, "A 77 GHz FM/CW radar front-end with a low-
profile low-loss printed antenna," IEEE Trans. Microwave Theory Techn. , vol. 47,
pp. 2337 - 2241, December 1999.
[43] P. R. Haddad and D. M. Pozar, "Analysis of two aperture-coupled cavity-backed
antennas," IEEETrans. Antennas & Propagat., vol. 45, pp. 1717 -1726, Dec. 1997.
[44] 1.R. Sanford and A. Tengs, "A Two Substrate Dual Polarized Aperture Coupled
Patch",IEEETrans. Ant.& Prop. Symp. Digest, pp. 1544-1547, 1996.
[45] S.C Gao, L.W. u, P. Gardner, and P.S. Hall, "Wideband Dual-Polarized Microstrip
Patch Antenna", Electronics Letters, Vol. 37, pp. 1213-14, Sept. 2001.
[46] G. L. Matthaei, L. Young and E. M. T. Jones, Microwave Filters. Impedance-
Matching Networks and Coupling Structures, Artech House , Dedham, 1980.
[47] D. M. Pozar, Microwave Engineering - Second Edition, John Wiley and Sons lnc.,
New York, 1998.
166
Chapter 4 Improving the Efficiency of
Microstrip Patch Antennas
4.1 Introduction
168
n
fc = -4h--'j~;==&flo=~-r&=j.Jr=-=1 (4.1)
The quantity used to measure the impact of the excited surface wave
on the antenna performance is the surface wave efficiency defined as:
where Psw is the power trapped in the surface wave and Prad is the total
radiated power from the antenna.
4.3-1 Introduction
169
The principal of operation of these patches is based on the following:
there is some critical radius of the patch that minimizes the excitation of the
TMo surface wave mode:
~TMo Rcrit =X' In =1.8412,5.3314, .... (4.3)
where ~TMo is the propagation of the surface wave mode. Thus if the
radius of the patch is greater than Relit. the TM o surface wave mode will not be
excited.
~o --- ~--- ~- - -~ - ~-
I I I I I t I
.... - - -
E I I ,
Q::J ~ o
IL IL IL I
L I
L I
L _
a:: I I I I I I
I I I I I I
I I I I I I
·50 -- -r - - -r --- r- - - ---r---r - - -r - - -
I I I t t I
I I I I I I
. 60 L..-o...-J'----'----'-----'-----'-~--'-~__L.~--'-~....J_~_J
U U U 5 ~ U U U ~
Frequency (GHz)
The first design considered uses a probe and a shorting pin oppositely
placed to one another to increase the resonant frequency. This idea is similar
to the concept introduced in [3] (and also discussed in more detail in Chapter
6), which reduced the resonant frequency of a microstrip patch by placing a
shorting pin in close proximity to the probe feed. By positioning the shorting
pin away from the feed, one would expect the opposite affect on the resonant
frequency, simply because in this case the feed is seeing an inductive load.
The element and its dimensions are described in Figure 4.3.2.
Figure 4.3.2 CircularPatch Loaded withShorting Pin (R =16.25 rom, xpJ =5.7 rom, Xp2
= -13.5 rom)
Figure 4.3.3 Input Impedance Response of Circular Patch Loaded with Short Circuit
(fstan =4.9 GHz; f slop =5.3 GHz)
II.
Figure 4.3.4 E- and H-plane Co-polar Radiation Patterns of Circular Patch Loaded with
Short Circuit at 5.I GHz
172
constant Cro The coaxial probe feed of radius ro is located at (pp, (j)p) from the
center of the patch. The annular ring is of inner radius a and outer radius b
and is placed concentrically about the patch antenna.
d E,
Figure 4.3.6 Input ImpedanceResponseof Patch Loaded with Annular Ring (a =8.05
rom, b = 16.25 mm, R = 4.74 mm, xp =1 rom)
".
Figure 4.3.7 E- and H-plane RadiationPatterns Patch with Annular Ring at 5.1 GHz
174
Figure 4.3.8 shows the calculated surface wave efficiency of the
benchmark case, the circular patch loaded with a shorting pin and the circular
patch loaded with an annular ring as a function of frequency. It is interesting
to note that the maximum surface wave efficiency does not correspond to the
resonant frequency for any of the designs.
95.-----,-----r---..------.---.,....----,
90
4.3-3.1 Introduction
175
4.3-3.2 Design and Results
y z
Figure 4.3.9 Probe-fed Circular Patch with Parasitically Coupled Annular Ring
176
To investigate the multilayered configuration, a prototype was
fabricated on similar material used in Section 4.3-2 for operation at
approximately 5.1 GHz. A thin layer of RT-Duriod 6010.5 (e, = lOA and d =
0.254 mm) was used to etch the annular ring that was then adhered to the
substrate. Figure 4.3.10 shows the predicted and measured return loss
behavior of the printed antenna (refer to the figure captions for the
dimensions). As can be seen from these results very good agreement between
theory and experiment was achieved. The predicted 10 dB return loss
bandwidth is 3.0 % compared to a measured value of 3.3 %. Figure 4.3.11
shows the predicted and measured E- and H-plane co-polar radiation patterns
of the multilayered antenna. Once again good agreement between theory and
experiment was achieved. As can be seen from these plots there is very little
ripple in the patterns, similar to the antennas presented in the previous
subsection. The predicted and measured gain of the antenna was 6.5 dBi. A
conventional patch mounted on the same material has a bandwidth of 1.9 %
and a gain of 3.8 dBi. The cross-polarized levels of the multilayered antenna
in each plane were greater than 30 dB below the co-polar patterns for all
angles that is very useful for applications requiring dual or circular
polarization.
-.---\
o ......
....... ....
.......... L--
·5
, ~. -~# ••
n, Theory
• • •• Experiment
"'\ ,,
I'
/:
/"
.
,
-30 1"'1"
-35
~
4.8 4.85 4.9 4.95 5 5.05 5.1 5.15
Frequency (GHz)
Figure 4.3.10 Predicted and Measured Return Loss of Multilayered Annular Ring Loaded
Patch(parameters: £'-. =10.4, tan 0) =0.002, d. = 1.905 mm, Ea =10.4, tan ~ =0.002, d2 =
0.254 mm, R =4.57 mm, Pp = 1.13 mm, <l>p =0, ro =0.45 mm, a =8.5 mm, b = 16.65 mm)
Figure 4.3.11 Predicted and Measured E- and H-plane Radiation Patterns of Multilayered
Annular Ring Loaded Patch
4.4-1 Introduction
178
and importantly the surface wave efficiency over this band is high, greater
than 85 %. Comparisons with conventional patch antennas are also given.
4.4-3 Results
180
greater than 20 dB below the co-polar patterns for all angles that is very
useful for applications requiring dual or circular polarization. This will be
addressed in chapters 6 and 7.
Figure 4.4.2 Input Impedance Behavior of Hi-lo Patch (parameters: €r. = 10.4, tan Sl =
0.002, d 1 =1.905 rom, Ea = 1.07, tan ~ =0.002, d2 =4.0 rom, R 1 =4.74 rom, Pp =3.49 rom, %
=0, ro =0.325 rom, R2 =10.8 rom)
II.
Figure 4.4.3 Measured E- and H-plane Radiation Patterns of Hi-lo Stacked Patch
---
95
90
.>~
~85
1'\
c
~80
V \
IE
W 75
\
~ \
~70
365
\
\
-.
~
J:60
55
50
4 4.5 5 5.5 8 8.5 7
Frequency (GHz)
4.5-1.1 Introduction
182
These artificial composite structures (photonic bandgaps) have
attracted a lot of attention due to their unique and favorable properties and
their suitability to a variety of applications, namely: antenna gain
enhancement and radiation pattern shaping applications [6]-[8], filtering
applications including high Q filters, frequency selective surfaces (FSS) and
space filters, resonant cavities [9] and waveguide structures [10, 11].
lop view
(a)
183
I I
side view
top view
(b)
Figure 4.5.1 Sievenpiper's High Impedance GroundPlane [12]: (a) Two-layer design,
(b) Three-layerdesign
184
4.5-2 High Impedance Ground Plane Design
4.5-2.1 Introduction
185
because of the ease of integrating this high impedance ground plane with a
square or rectangular microstrip patch antenna.
~ COnduclin g Vias
"'rr~T----T""'--T---U_ Metal Layer - - - __
Side Vitw
Since the unit cells are small compared to the wavelength, the
electromagnetic properties can be described using lumped circuits.
Sievenpiper models his design as a parallel resonant LC circuit shown in
Figure 4.5.2. The capacitance in this two-layer model is dominated by the
fringing electric field between the adjacent elements, as shown in Figure
4.5.3. The fringing capacitance C is given by the expression:
186
C -- w(£! + £2)
Il
Cosh-
g
-l(a) (4.4)
#
--a;-fI
!J.OJ _
(4.8)
From Equation 4.8 the upper and lower frequency limits of the bandgap can
be predicted.
. -, -. •
.<,..."
\
\
'.
'.-, ' . .. ~ .
.. .. ",
"
Figure 4.5.5 HFSS simulation 5xl unit cell for HIGP <Parameters: s, =2.2, d =9.51
rnm, W = 9 mm, g = I mm, rOY = O.5mm)
5xl unit cell in Figure 4.5.5 for the HIOP structure was simulated
and the S21 transmission results plotted in Figure 4.5.6. A reference
simulation is also included in this figure, that is, a grounded dielectric slab
without the HIOP. As can be seen from the results the centre frequency of the
HIOP bandgap closely corresponds to the resonant frequency calculated using
the surface impedance model (refer to Figure 4.5.4). Furthermore the
predicted bandgap from 1.89 OHz to 3.15 OHz using the surface impedance
model compares well to the simulation results shown in Figure 4.5.6.
o .......... ............
~--
-5 ..".- I-"""
iii
:!:!. ·10
o
e
V / I...
~ Reference I
u; -HIGP
~ ·15 I
In
e
l!
I- .20
N
~
/V
I
<II
-25
-30
1 2 3 4 5 6
Frequency (GHz)
190
The results are shown below in Figure 4.5.8. A reference
measurement on a PEC was performed, to characterise the behaviour of the
TM probes and the environment the measurements were performed in. The
surface of the PEC was the same size as the mGP under test. A steep
decrease in transmission at 2.2 GHz signifies the TM band edge.
-15
...... #
.,- .#.,
-20 "
\L 1
f'\-...I r-\'
r
1.
.' . ~
-25
i ·30 ,"
•• ...1\ u \ rV\V
g -35
iii
A }/ \ r/v
~ -40
sc: -45 ~
'"
_ -50
~.,I" '- 6' I
I;j
-55
~ 1- •_. Reference ~
I TM Measurement
-60
-65
1 1.5 2 2.5 3 3.5 4
Frequency (GHz)
-30 r - - - , -- - ....----,-----,------,r-------,
-40 I----+----f----+----+---jl---~
...I!
N -80 OOU-J' -+-- - + -- --I-+--4fH-- - - jI--- ___i
(J)
4.6 Summary
4.7 Bibliography
[1] J.F. ZUrcher and F.E. Gardiol, BroadbandPatchAntennas,Artech House Inc, 1995.
[2] D. R. Jackson, J. T. Williams, A. K. Bhattacharyya, R. L. Smith, S. J. Buchheitt and
S. A. Long, "Microstrip Patch Designs That Do Not Excite Surface Waves," IEEE
Trans. Antennas Propagat., vol. AP-41, pp. 1026-1037, August 1993.
[3] R. B. Waterhouse, "Small Microstrip Patch Antenna," Electronics Letters, vol. 31,
pp. 604 - 605, Apr. 1995.
[4) 1. T. Aberle, D. M. Pozar and J. Manges, "Phased Arrays of Probe-fed Stacked
Microstrip Patches", IEEE Trans. Ant. & Prop., vol. AP-42, pp. 920-927, July 1994.
[5) T.K. Lo, Y. Hwang, E.K.W. Lam, B. Lee, "Miniature Aperture-Coupled Microstrip
Antenna of Very High Permittivity," Electron. Lett., vol. 33, pp. 9-10, Jan. 1997.
[6] J.D. Shumpert, W.J. Chappell and L.P.B. Katehi, "Parallel-Plate Mode Reduction in
Conductor-Backed Slots Using Electromagnetic Bandgap Substrates," IEEE Trans.
Microwave Theory and Tech., vol. 47, pp. 2099-2104, Nov. 1999.
[7] M. Thevenot, C. Cheype, A. Reineix and B. Jecko, "Directive Photonic-Bandgap
Antennas," IEEE Trans. Microwave Theory and Tech., vol. 47, pp. 2115-2122, Nov.
1999.
[8] T.H. Liu, W.X. Zhang, M. Zhang and K.F. Tsang, "Low Spiral Antenna with PBG
Substrate," Electron. Lett., vol. 36, pp. 779-780, Apr. 2000.
[9) M.M. Beaky, lB. Burk, H.O. Everitt, M.A. Haider and S. Venakides, "Two-
Dimensional Photonic Crystal Fabry-Perot Resonators with Lossy Dielectrics," IEEE
Trans. Microwave Theory and Tech., vol. 47, pp. 2085-2091, Nov. 1999.
[10) F. Gadot, A. Ammouche, A. de Lustrac, A. Chelnokov,F. Bouillault, P. Crozat, 1.M.
Lourtioz, "Photonic Band Gap Materials for Devices in the Microwave Domain,"
IEEE Trans. Magnetics; vol. 34, pp. 3028-3031, Sep. 1998.
[11) F. Yang, K. Ma, Y. Qian and T. Itoh, "A Novel TEM Waveguide Using Uniplanar
Compact Photonic-Bandgap (UC-PBG) Structure," IEEE Trans. Microwave Theory
and Tech., vol. 47, pp. 2092-2098, Nov. 1999.
[12] D. F. Sievenpiper, "High-Impedance Electromagnetic Surfaces," Ph.D. dissertation,
Dept. Elect. Eng., University of Califomia at Los Angeles, Los Angeles, CA, 1999.
194
[13] D. F. Sievenpiper, L. Zhang, R. F. Jimenez Broas, N. G. Alexopolous, E.
Yablonovitch, "High-Impedance Electromagnetic Surfaces with a Forbidden
Frequency Band," IEEE Trans. Microwave Theory and Tech., vol. 47, pp. 2059-
2074, Nov. 1999.
[14] R. Gonzalo, P. de Maagt, M. Sorolla, "Enhanced Patch-Antenna Performance by
Suppressing Surface Waves Using Photonic-Bandgap Substrates," IEEE Trans.
Microwave Theory Tech., vol. 47, pp. 2131-2138, Nov. 1999.
[I5] R. F. Jimenez Broas, D. F. Sievenpiper, E. Yablonovitch, "A High-Impedance
Ground Plane Applied to a Cellphone Handset Geometry," IEEE Trans. Microwave
Theory and Tech., vol. 49, pp. 1262-1265, Jul. 2001.
195
Chapter 5 Small Microstrip Patch Antennas
5.1 Introduction
The antennas for mobile units are fundamental parts of the overall
mobile system. The evolution of mobile systems requires portable antenna
elements that are small, low profile, lightweight and relatively inexpensive to
manufacture. As mentioned before, some of the desired features for mobile
antennas include omnidirectional radiation pattern in the horizontal plane,
broad bandwidth to accommodate spread-spectrum services, and ease of
construction. However, the fundamental problem is the size of the antenna
tends to be large for low microwave frequencies, typically less than 3 GHz.
As stated earlier, as the volume of the antenna reduced, the performance of
the antenna in terms of impedance and radiation patterns is sacrificed to
achieve reduced volume [7].
198
The dipole antenna is the simplest form of wire antenna. The basic
dipole antenna structure is a Y2A center-fed linear cylindrical antenna and it is
this unwieldly dimension that has seen its restricted use. Thick dipoles are
considered broadband while thin dipoles are more narrowband. For very thin
dipoles, small perturbations in the operating frequency will result in large
changes in its operational behavior. The bandwidth can be increased for thin
dipoles if the length to diameter ratio of the dipole is decreased.
Several methods have been reported in the literature to reduce the size
of the printed conductor of a microstrip patch. One such method of reduction
is to use high dielectric constant material [18]. This is a logical extension of
what was presented in chapter 2 on how to reduce the physical size of the
patch. The antenna in [18] was mounted on a high dielectric material and
loaded with superstrate material (ft. ~ 38). Only poor efficiency due to surface
wave excitation and narrow bandwidths have been achieved so far with this
form of patch. The availability of low cost and low loss high dielectric
constant material is another issue.
Replacing the shorting post with a low resistance chip resistor can
also demonstrate reduced antenna size [22]. A consequence of this action is
that the antenna bandwidth is also enhanced. However, the problem with
utilizing a chip resistor is that the efficiency and gain of the antenna are
dramatically reduced due to the power dissipated in the resistor that
significantly reduces the power budget.
It should be noted that throughout this chapter the term: shorted patch
is used to describe a patch incorporating several shorting posts. It does not
refer to the classical quarter-wave patch antenna where an edge of the radiator
is terminated with a short circuit plane.
5.2-1 Introduction
202
this section a parameter study showing the effects of the height of the
substrate, the dielectric constant of the material, the size of the feed and
shorting pins as well as the effect of a cover-layer on the electrical
performance and manufacturing ease of the shorted patch is provided. From
this study, valuable insight into the optimum design of shorted patches is
achieved; Section5.2.4 addresses the problemof the strong dependence of the
impedance behaviorof the shorted patch on the close proximity of the feed to
the pin. Here several simple techniques are summarized which overcome this
short fall. The advantages and disadvantages of each method are discussed in
this section.
y
y
x f ..
w
(Xp.Yp)
x
(xl'" »s)
~L~
(a) (b)
z
t,
5.2-2.1 Introduction
110
Figure 5.2.3 Predicted and Measured E-plane Ee Far-field Radiation Patterns of Circular
Probe-fed Shorted Patch at 1.9 GHz.
110
Figure 5.2.4 Predicted and Measured H-plane Be Far-field Radiation Patterns of Circular
Probe-fed Shorted Patch at 1.9 GHz
205
It should be noted that although the H-plane cross-polarization level
of this antenna (Figure 5.2.4) is relatively high, for the application of mobile
communications handsets this is not a major concern, as much of these fields
will diffract off the edges of the small, finite ground plane within the handset
[29]. The H-plane co-polarized field, E~, is very similar to that of a
conventional patch and therefore is not shown here. The predicted gain of the
antenna is 3.6 dBi. The location of the gain maximum is similar to a top
loaded monopole.
206
Another solution for the given specification can be achieved using the
rectangular shorted patch configuration presented in Figure 5.2.lb. The
= =
dimensions of this case are: L 17.3 mm, W 21.3 mm, xp 5.0 mm, ro = =
= =
0.325 mm, xps 7.4 mm, ros 0.6 mm. Once again, the impedance bandwidth
and radiation patterns are very similar to that presented in Figure 5.2.2-4 .
0.8 f>
0.8
~
0.4
1 '--"-tJr='-----'------'--'---'-----..---I.---'---''--o--J 0.2
o 0.03 0.04 0.05 0.08 0.07
Thickness , Ao
Figure 5.2.5 Performance Trends for Shorted Patches Mounted on Foam (e, =1.07)
207
= 10 mm and e, =2.2 has the following dimensions: R =7.8 mm and 11 =0.28
mm. Hence the fabrication tolerances become very stringent. Another very
important consequence of using a dielectric material is the significant
reduction in bandwidth. Although a decrease is expected when comparing to
conventional microstrip patch design, the degree of reduction is larger. For
the Er =2.2 case, the 10 dB return loss bandwidth is only 3.4 %. This can be
explained by considering the simple circuit theory outlined earlier. Recall
that the capacitive coupling between the probe feed and the shorting pin used
to counter the inductive nature of the patch below resonance is proportional to
the dielectric constant of the material. Thus, the rate of change of capacitance
with respect to frequency is also proportional to s, This being so, then the Q-
factor of the resonator will be very high and hence the bandwidth narrow.
For all cases considered so far, the radii of the feed (ro) and shorting
(rOs) pins have been kept constant. To obtain insight into the effect of each
radius on the overall size of the patch conductor and impedance
characteristics, several new cases were investigated. For the first series of
cases, the feed pin was kept constant (ro = 0.325 mm) and the radius of the
208
shorting pin varied. The general trend observed was the impedance locus of
the printed antenna remained similar, once the other dimensions were
modified to ensure 50 Q at 1.9 GHz. It was observed that as the radius of the
shorting post was reduced, the bandwidth did slightly decline, however only
by a fraction of a percent. The main effect of the size of the shorting pin was
on the overall size of the patch conductor as well as t:... The trend observed
was as the radius of the shorting pin is increased, the size of the patch
conductor increases, as does t:... For the previously incorporated substrate, an
increase in the radius of the shorting pin of 0.1 mm, is accompanied by an
increase in the radius of the patch by 0.4 mm and an increase in t:.. of 0.4 mm.
Hence a smaller patch size can be achieved with a thinner shorting pin
conductor, although the problem with accurate positioning of the pin with
respect to the feed becomes worse. Interestingly, the opposite trends occur
for the feed pin radius. As the radius of feed pin is decreased, while keeping
the shorting pin size constant, the radius of the patch conductor increases, as
does the separation distance between the probe and shorting pin required to
achieve 50 Q at 1.9 GHz. Similar incremental values for the radius of the
patch and t:.. were obtained as in the previous case. This phenomenon will be
examined in more detail in the next section.
As was seen in the previous section, the close proximity of the feed
and shorting pins is a major design flaw in the shorted patch antenna.
Although some of the procedures outlined in Section 5.2.3 can reduce the
stringent design tolerances on 11, more robust solutions have yet to be
provided. In this section, the issue of improving the manufacturing ease of a
shorted patch is addressed and two techniques are proposed.
__ -+_~x
I'--_ _~~x
(a) (b)
Figure 5.2.6 Top Viewof Probe-fed Patches Incorporating Multiple Shorting Posts: (a)
Two Posts; (b) ThreePosts
Consider the patch with .two shorting pins shown in Figure 5.2.6a.
Using the same substrate parameters and pin radii given earlier, a patch
configuration with the following dimensions was designed for 50 .Q input
210
impedance at 1.9 GHz: R = 13.2 mm; xp =4.95 mm, Xpst = xpsz = 11.08 mm
and CPt =-cpz = 10° (where cpj is the angle between the ilb shorting pin and the x-
axis). The input impedance locus of this shorted patch is shown in Figure
5.2.7. As can be seen from this figure, the 10 dB return loss bandwidth is
approximately 7.9 %. Thus the improvement in bandwidth and ~, here
approximately 6 mm, are achieved at the cost of increased patch real estate.
The predicted radiated Eo in the E- and H-plane are presented in Figure 5.2.8.
These results are similar to the patterns presented in Figures 5.2.3 and 5.2.4,
with the exception that there is less of a dip at broadside in the E-plane, by a
factor of 2 dB. Thus more power is being radiated with this configuration
than a patch incorporating a single shorting post. There is also a more
pronounced asymmetry in the E-plane pattern as a result of the presence of
more shorting pins.
,eo
211
As mentioned before, another degree of freedom with the multiple
shorting post antenna is the angular displacement of the posts. Consider the
patch with two shorting posts with the angular displacement of the pins
increased to ± 20°. For an input impedance of 50 Q at 1.9 GHz the
dimensions of this shorted patch are: R = 14.65 rom, xp = 3.65 rom and Xpsl =
X ps2 = 12.7 rom. The corresponding bandwidth of the shorted patch
configuration is 9.4 %. Thus to achieve greater bandwidth, the shorting posts
should be positioned further apart, once again at the expense of increased
patch area. An explanation of this phenomenon is that as 4> increases, the
coupling between the shorting pins decreases, as a function of cos 4>. To
counter this effect, stronger patch modes are required to maintain the
coupling. Thus the patch conductor size must be made larger.
212
Swap the Relative Location of the Feed with respect to the Post
___ t - - ~ X
shorting post
Figure 5.2.9 Schematic Diagram of Shorted Patch Configuration with Swapped Post and
Feed Pins
213
Figure 5.2.10 Predicted and Measured Input Impedance of Shorted Microstrip Patch
= = =
(parameters: Er 1.07, tan S 0.015, d 10.0 mm, R 12.1 rom, xp 10.1 mm, Yp O.xps = = = =
= =
5.2 rom, Yps 0, rO 0.325 rom, rOs 0.6 rom) =
As mentioned before, there are several degrees of freedom in the
design of a shorted patch of which the radius of the patch and the locations
shorting post and probe feed are the most critical in determining the
impedance characteristic. Many simulations were run which considered the
effect of fabrication errors on the impedance behavior of shorted patches. To
this end, we have assumed errors of ± 0.25 mm in each of the previously
mentioned parameters. Variations in each parameter were considered
accumulatively and Figure 5.2.11 displays the worse case scenarios in terms
of change in resonant frequency and return loss for both the traditional and the
modified shorted patches.
O.------,- - ---,.- - --r-- --r-- - --r-- --,
·5
iii' ·10
S
] -15
I
I
E
:J
-20 - _ 1-
I
_
I
~
a: -25
I
- - - - 1- - - - -
I
I I
-30 - - - - r - - - - - ,- - - - -
I I
I I
-35 ..................'"""--''-'-~ .........~~.....x. ~~.................................L..~~ ..J
1.75 1.8 1.85 1.9 1.95 2 2.05
Frequency (GHz)
Figure 5.2.11 Effect of Fabrication Errors on Return Loss Behavior of Traditional
Shorted Patch ( • ) and Modified Configuration ( • ). The solid lines are the original
cases and the dashed lines show the worse case scenarios
214
As can be seen from Figure 5.2.11, the changes are more dramatic for
the traditional case, once again highlighting the robustness of the modified
patch configuration. It should be noted that the cases considered here are
relatively broadband in nature, hence the effects of fabrication error are
somewhat reduced. For cases where thin substrates are used, these errors will
be more pronounced.
The previously discussed single shorting post design strategy can also
be applied to a printed annular ring to further reduce its surface area. Figure
5.2.12 shows a schematic of a shorted annular ring and the input impedance
variation of such an antenna using 10 mm 51 HF Rohacell foam as the
substrate. The dimensions of the structure are given in the figure caption. As
can be seen from the measured and predicted impedance loci, very good
agreement between the two is evident. The predicted and measured 10 dB
return loss bandwidths of the antenna are 5.8 % and 5.9 %, respectively. The
slight shift in resonant frequency apparent in Figure 5.2.12 can be attributed
to the thin layer of dielectric material required to etch the patch conductor.
This layer (height of 0.254 rom, e, = 2.2) is not included in the analysis. For
the theoretical results displayed in Figure 5.2.12, two attachment modes and
twenty-five entire domain basis functions were used. The radiation patterns
of the shorted ring were measured and are very similar to those presented in
Figure 5.2.3 and 4 of a shorted circular patch. For the sake of brevity these
patterns will not be given here. The gain of the antenna was measured as 3.5
dBi, compared to a theoretical value of 3.6 dBi.
y
t:
Figure 5.2.12 (a) Schematic Diagram of Shorted Annular Ring; (b) Input Impedance
Behavior of Shorted Annular Ring (parameters : d = 10.0 rom, E,. = 1.07, tan 0 = 0.01, a = 6.0
rom, b = 10.3 mm, Pp = 7.0 rom. $p = 0, ro= 0.325 rom, PI" = 9.3 rom, $ps = 0, ros = 0.6 rom)
215
It is interesting to compare the properties of the shorted annular ring
with the previously presented shorted patches. Using the same dielectric
materials, the outer radius of the shorted ring is slightly smaller than the
radius of a circular shorted patch, by approximately 5%. The bandwidth and
the gain of the shorted ring are once again slightly less than the circular patch
version, by 0.5 % and 0.2 dBi, respectively. These results can be attributed to
the smaller area of the ring-based antenna. Despite these relative
shortcomings in the electrical performance of the shorted annular ring there is
a distinct advantage. The gap between the shorting and feed pins is greater,
by 14 %. This allows for easier manufacturing of the antenna and makes it
less susceptible to errors in the positioning of these pins. As was shown in
Section 5.2.2, the coupling between the feed, the shorting post and the closest
radiating edge plays a significant role in the impedance behavior and hence
the size of the patch. Like for the conventional stacked patch case considered
earlier, the annular ring has another degree of freedom that contributes to the
impedance behavior, namely, the inner conductor edge. Thus the strong
capacitive coupling required for a shorted patch to operate can be distributed
amongst all these variables. It was shown in Section 5.2.2 that for minimum
area the shorting pin must be located closest to one of the radiating edges of
the patch. However, it was also shown that for maximum spacing between
the feed and shorting pins, the feed should be located nearest to this edge.
Hence for a circular or rectangular patch configuration, a compromise must be
made. However, by using an annular ring both criteria can be achieved. Here
the feed pin is located nearest the inner conductor edge and the shorting pin
located near the outer edge, providing a shorted patch of minimal area as well
as acceptable spacing between the pins.
5.3-1 Introduction
216
Throughout this investigation, edge-fed rectangular shorted patches
are examined, although the findings can be applied to any other form of the
shorted patch. To accurately analyze this style of shorted patch, the full-wave
software package, Ensemble 5.1 [31] was utilized. Figure 5.3.1 shows a
schematic of the proposed shorted patch antenna. As can be seen from the
top-view of the antenna, there are two 'notches' each located on
symmetrically opposite sides of the feed, near the shorting pin.
Yt
z
i
L ~
L~
E,
i
• w
I
r
1
Shorting pin (rOo)
SMA feed
Figure 5.3.1 Schematic of Edge-fed Shorted Patch: (a) Top View; (b) Side View
217
Simulations have shown that the length of the first pad (L 1 in Figure
5.3.1) should be set to the diameter of the shorting pin to yield the smallest
size patch conductor. The width of the first pad (WI) plays a key role in the
overall size of the printed antenna and its input impedance nature. If WI is set
to the width of feed insert (Wfi in Figure 5.3.1) then the antenna behaves as a
standard short patch. As WI is reduced, the overall patch conductor size
decreases, due to the phenomenon described earlier associated with the fringe
capacitance. When the width is reduced to that of the shorting pin, the
maximum size reduction is achieved. However, as WI is decreased, the input
impedance is also reduced until this point where of course the antenna appears
almost as a short circuit at the feed port. Thus a compromise must be made
between maximum size reduction and input impedance behavior. This control
of the input impedance response alleviates to an extent the required very
accurate positioning of the shorting pin with respect to the feed of a standard
shorted patch [21].
The presence of the other two pads once again increases the fringing
capacitance around the post/feed thereby reducing the size of the patch
conductor. These additional pads have a secondary effect on the impedance
response.
-5
, ~'i ;:
::
'i
~10
I
~
III
,i
III I
,3.15
c::
:i t - ••• Standard S-P
Q).20 ~ ••••••.. 1 Pad Sop
a: 2 Notches Sop
~i - • _. Measured
-25
I
-30
0.5 0.55 0.6 0.65 0.7 0.75 0.8
Frequency (GHz)
Figure 5.3.2 Return Loss of Various Shorted Patches (Substrate Parameters: Er =2.2, d =
1.59 rom. Shorted Patch Dimensions: L = 44.5 rom, W = 48.75 mID, 4= 16 rom, Wf = 1.0 rom,
Ln = 11.5 mm, Wfi= 7 mm, L. = 1.0 mm, WI = 3.0 mm, ~ = 2.0 mm, Wz = 3.0 rom, xps = 0.5
mm,ros = 1.0 mm)
o
110
5.4-1 Introduction
t
4 W, ..
t, +- 1..--" w.
t -.. y
L, + i-
t .-i- x,.
+
L,
T
t
L,
~.-
....vr,- .-
x,
(a)
(b)
Figure 5.4.2 shows the predicted and measured input impedance for a
wing antenna designed for operation at a frequency of 2.7 GHz mounted on
221
10 mm foam (refer to Figure 5.4.2 captions for dimensions). The predicted
and measured 10-dB return loss bandwidth was 7.4 % and 6.9 % respectively.
This impedance bandwidth is comparable with a standard shorted patch
presented earlier in this chapter. The measured far-field radiation
characteristics of the shorted winged patch are shown in Figure 5.4.3 (note:
broadside is at 1800 in this figure). For the measurement, the ground plane for
the antenna was very small, extending 15 mm each side of the printed
conductor. The radiation pattern is similar to that of the shorted patches
presented in this chapter, however, the cross-polarization level for this
antenna is somewhat higher, particularly in the H- plane, due to the small
sized ground plane. As was shown in [34], the size of the handset antenna
affects the radiation properties. The smaller the ground-plane size, the more
the fields diffract around the ground-plane edges. Of course, the size of the
ground plane cannot be too large as it will not be able to be accommodated for
on the handset. The 6ain of the antenna was measured as 1 dBi.
e -15
-20
2.3 2.4 2.5 2.6 2.7 2.8 2.9
frequency (GHz)
Figure 5.4.2 Predicted and Measured Return Loss of Winged Shorted Patch
(Parameters: E, =1.07, tan &=0.015, d =10 nun, WI =19 nun, L I =6 nun, Wn =4 nun, L.. =5
mm, Lz =2 mm, Lr=7 nun, Wf = 1.2 mm, xp = I nun, Xps =1 mm, r p =0.6 mm, rps =0.325
m~
''0
Figure 5.4.3 Measured Far-field Radiation Patterns of Winged Shorted Patch at 2.7 GHz
222
To put these results into perspective, a standard edge-fed rectangular
shorted patch designed to operate at this frequency using the same dielectric
material has the dimensions of 66 mm x 46 mm. Thus the new antenna is
smaller by a factor of 26 comparing the conductor area. The bandwidth of the
wing antenna is slightly smaller, 7 % compared to 10 %.
r ,.
Probe
Figure 5.4.6 shows the predicted and measured return loss of a wing
antenna on the developed handset designed for operation at 1.9 GHz. The
slight discrepancy between the experimental and calculated results can be
attributed to the handset not being taken into consideration during the
simulation. The measured bandwidth of the antenna on the handset was 5.7 %,
which is compliant with most cordless systems (for example the DECT [37]).
To improve the bandwidth performance to be consistent with PCS
requirements, stacking the printed antenna as in [38, 39] or using a resistive
termination as in [22] could be done, although this latter method should be
avoided if possible and this will be addressed later. The measured far-field
radiation pattern is shown in Figure 5.4.7 . The ripples observed are commonly
224
associated with the interaction of the handset and the antenna [33, 34]. The
measured gain of this configuration was 0.6 dBi.
o
" .... --~-~
--10
m
"C
';;-20
en
..2
E-30
::J
li
"'-40
-50
1.75 1.8 1.85 1.9 1.95 2
frequency (GHz)
Figure 5.4.7 Measured Far-field Radiation Patterns of Winged ShortedPatch at 1.9 GHz
on MobileHandset
225
method here. Although the volume of the PIFA in [14] is smaller, it is
substantially more difficult to manufacture than the proposed printed antenna
here. One of the key issues associated with any low cost, high volume
antenna is the ease of manufacturing and its susceptibility to manufacturing
errors. There are several very important properties of the proposed antenna
compared to that in [14] related to its ease of construction. Firstly, the
antenna here is a single layer structure (as opposed to the design presented in
[14]) and therefore there are no associated alignment problems. Also the
antenna here requires less solder joints and therefore is more robust. The
antenna in [14] also has a higher component account and therefore is more
expensive to manufacture. It should be pointed out that the size reduction
proposed here can be applied to each of [14], to give even further size
reduction, as these papers have not attempted to optimize the 'patch'
conductor shape, as done here.
The next higher order resonance for the two designed antennas
presented here are weakly coupled and occur at 8 GHz and 9.5 GHz,
respectively. These frequencies are somewhat higher than expected for a
small antenna and can be attributed to the non-conically shaped conductor
making it difficult for higher order resonances to be impedance matched.
This property of the antenna is very advantageous ; especially as tighter
emission requirements by government agencies are enforced for these
systems.
5.5-4 Introduction
226
reduction recorded in the literature. The two spiral shorted patches are
investigated theoretically and experimentally with their input impedance and
radiation characteristics given.
5.5-5 Configurations
227
Shorting pin <x,.. y.,l
I .
\
Probe <x,. Y~
--- - L - - --.
Probe (x,.. Y,) Shorting pin (x,.. Y,.)
(a) (b)
(e)
Figure 5.5.4 shows the measured return loss of the two proposed
shorted patches . As can be seen from Figure 5.5.4, configuration 1 has a
resonant frequency at 715 MHz and configuration 2 has a resonant frequency
at 440 MHz. The simulated resonant frequencies were at 690 MHz and 460
MHz, respectively. These frequencies are significantly lower than a standard
microstrip patch antenna of the same area (f, = 4.5 GHz) as a well as a
conventional shorted patch (f, = 1.9 GHz). The reductions in resonant
228
frequency correspond to area size reductions of the order of 111 and 261,
respectively. Of course the 10 dB return loss bandwidths of the two
configurations shown in Figure 5.5.4 are very narrow, however this is only
due to the material thickness not being scaled for the new operating
frequencies 450 MHz & 735 MHz, respectively. There does exist a
compromise between size reduction and achievable bandwidth, which is
consistent with fundamental antenna theory [10].
229
0
~
. r#,
~~ I
·5 .• !
iii !:l
~
III ·10
:!
III
~u
.
0
..J u
··
C u
:2 -15
Gi
a:
···
- • - conllgur.Uon 1
.. .. . .. conllguflUon 2
·20 Theory(conllg.2)
••• •Theory(conlllJ.l)
-25
0.3 0.4 0.5 0.6 0.7 0.6 0.9
Frequency (GHz)
,.
Figure 5.5.5 Measured Radiation Patterns of Configuration 1
230
5.6 Improving the Performances of Small Microstrip Patches
5.6-1 Introduction
231
5.6-2 Stacked Shorted Patches
5.6-2.1 Introduction
" - - -1.. ~
233
shown in Section 5.2. The asymmetry in the E-plane co-polar plot can be
attributed to the presence of the shorting post. A gain of 3.8 dBi was
measured using a ScientificAtlanta 12-1.7 standard gain hom.
Figure 5.6.2 Predicted and Measured Input Impedance of Stacked Shorted Microstrip
Patch
110
Figure 5.6.3 Measured E-plane and H-plane Co-polar and Cross-polar Radiation
Patterns
234
As stated in Chapter 3, the key to obtain a wide-band impedance
characteristic, is to ensure that a 'mutual resonance' has occur. This
phenomenon yields a loop in the impedance locus when plotted on a Smith
chart. A hint to why this was not obtained in the previous subsection can be
found by examining the Smith chart of the results (Figure 5.6.2). As can be
seen from this figure, the shorted patches are over-coupled, producing a
somewhat large 'coupling loop', similar to a conventional stacked patch when
the patches are located to close together. To achieve a wide-band impedance
response the shorted patches must be located further apart.
235
thinner the upper layer must be made if good control of the impedance locus
is desired. Thus there is a compromise between bandwidth and impedance
control. The size of the patches in the initial design phase is not too important
as the relative dimensions will be modified through the course of the design to
ensure the loop in the impedance locus is center around 50 Q. A typical
starting point is a length and width for each patch of Arj8.
Once the resonant loop has been formed in the impedance response as
a result of the choice of materials used, its location on the Smith chart can be
controlled by changing the relative dimensions of the patches. Making the
bottom patch larger than the top patch lowers the real part of the input
impedance. The opposite effect can be achieved by reversing the size trend.
The position of the feed probe and the second shorting pin provide means of
fine tuning the impedance locus.
Figure 5.6.5 Predicted and Measured Input Impedance Response (Parameters: e,.1 =ea =
1.07, d l = 2.0 mm, d2 = 18.0 mm, L I = 36.0 mm, WI = 23.0 mm, ~ = W2 = 23.0 mm, xp = Xpa2
= 15.0 mm, ro = r0a2 = 0.325 nun, Xpsl = 1.0 rom, rOsI = 0.6 mm)
236
The measured E- and H-plane Eo radiation patterns are shown in
Figure 5.6.6 and are similar to those presented in the previous sub-section.
The gain of the antenna was measured as 3.5 dBi across the relevant band of
frequencies. It should be noted that for all measurements presented in this
paper, the ground-plane was large, 50 x 50 em. A further improvement in the
impedance behavior may be achieved similar to the cases for a monopole and
PIFA [12] if the handset was included in this study. The overall dimensions
of the microstrip antenna presented here are 4 x 4 x 2 em, of which can be
reduced by using the technique outlined in Chapter 3 at the slight expense of
bandwidth.
,.
Figure 5.6.6 Measured Eo Radiation Patterns at 1.9 GHz
237
proximity to the feed pin to ensure that the size of the conductor is reduced as
mentioned before. The two spirals are located close to each other with the gap
spacing Wg between the radiators being 0.3 mm. It is interesting to note that
by decreasing We, the resonant frequency of the antenna decreases; however,
the limit of this reduction is set by the diameter of either the shorting pin or
the feed pin.
+- L -..
I
r
( ~-r.,
( • p.'
w -~1
~ I .+
~y )1
~ (,d.y~)-
1 ~I w
t
1. w,
T
Figure 5.6.7 Interleaved Shorted Printed Spiral Antenna
238
material (RT/Duroid 6010, e, = 10.2) was used for the feed network to
maintain the small size of the antenna. A photograph of the manufactured
dual spiral printed antenna is shown in Figure 5.6.8.
t,
239
!g"0
.
o
-- r-,
\
I
I
/
,;
,2·15 \
E
\v
"
e".20
·25
-30
1.8 1.9 2.1 2.2 2.3
V
2.4 2.5
frequency , GHz
Figure 5.6.9 Measured Return Loss of Shorted Dual Spiral Printed Antenna
(parameters: e, =1.07, d = 10 nun, L =10 nun, W =9.5 nun, WI =0.6 nun, wg =0.3 nun, TO =
0.3 nun, ro. =0.15 nun, xpJ =-2 nun, YpJ =0 nun, X p2 =2 nun, Yp2 =0 nun, xpsJ = 1 nun, YpsJ =1,
x psl = -1 111Ip, Yps2 = -1)
110
Figure 5.6.10 Measured Radiation Patterns of Dual Spiral Printed Antenna at 2.3 GHz
I<--_-+-~x
Figure 5.6.13 shows the predicted and measured E- and H-plane co-
polarized radiation patterns for the balanced shorting post patch configuration,
at 1.9 GHz. To provide the 1800 phase difference between the two feeds, a
Norsal 4114 coupler was used. As can be seen from this graph good
agreement between theory and experiment was achieved. The measured
cross-polarization levels were greater than 20 dB below the co-polar fields in
both planes. Theoretically, the cross-polarized fields should be more than 100
242
dB lower, however, this discrepancy could be due to diffraction from the
finite ground plane (in this case 7cm2) . It should be noted that the measured
values in the H-plane are still more than 20 dB below that for a typical shorted
patch. The theoretical gain of the balanced configuration is 5.0 dBi compared
to 3.5 dBi for a typical shorted patch mounted on the same material.
Figure 5.6.12 Predicted and Measured SI1 of a Balanced Shorted Microstrip Patch
(parameters: e, = 1.07, d = Smm, R = 20.6mm, Xpi = lS.8mm, Xpsi = 18.6mm, Xp2 = -lS.8mm,
Xps2 = -18.6mm, rOi = r02 = 0.32Smm, rOsi = rOs2 = O.6mm)
o
270 1---r--~~~_.c::+:-:----l-- -l 80
1110
Figure 5.6.13 Predicted and Measured Radiated Fields from the Balanced Shorted
Microstrip Patch at 1.9 GHz
243
eighth of a wavelength for a conventional shorted patch using a single
shorting pin to approximately a quarter-wavelength. The increase in antenna
real estate was attributed to each excited probe 'seeing' the shorting post used
for the other feed. This 'inductive' loading somewhat cancels the 'capacitive'
effect of the closeness of the excited probe and its associated shorting post.
The more shorting pins located on the patch conductor the greater the size of
the patch. Therefore a CP version utilizing 4 shorting pins and feeds
(required to ensure good cancellation of cross-polarized fields) would give
very little size reduction compared to a conventional microstrip patch antenna.
A key problem here is that the shorting pins must be electrically isolated if
more than one is to be used on one patch conductor.
244
y
0..
2
h. III
X
d
Er2 Er l
245
for both the balanced feed configuration and the CP version of the shorted
concentric rings shared this common procedure.
1 /
Port 2
The measured return loss bandwidth of the proposed balanced dual concentric
ring antenna is shown in Figure 5.6.17 (refer to the figure captions for the
locations of these probes as well as the appropriate shorting pins and the
246
dimensions of the rings). The 10 dB return loss bandwidth of the antenna is
8.9%. It should be noted that the simulated results for the inner ring alone is
4.4% whilst for the outer ring a bandwidth of 3.6% was simulated. The
proposed antenna has a higher bandwidth due to the concept based upon the
principal of any dual resonance antenna, namely that if a mutual resonance
can be formed, the overall impedance variation of the antenna will be
minimized, enhancing the bandwidth of the antenna. Importantly too, the
proposed antenna has a larger bandwidth than that presented before, namely
4.4%.
\
-5
,/
<> \
~10
rii
III
.915
\
E
.a
!!20
-25
-30
1.B 1.9 2 2.1 2.2
frequency, GHz
Figure 5.6.17 Measured Return Loss of Balanced Dual Concentric Shorted Ring Antenna
(Parameters: al =10 mID, az =7 mID, £,.. =2.2, d = 10.3 mID, bl = 14 mID, bz =11 mID, E,.z =
1.07, d = 10 mID, rol =0.2 mID, rosl =0.3 mID, roz =0.325 mID, fosZ =0.6 mID, Xpl =7.5 mID, Ypl
=0 mID, Xp2 =- 11.5 mm, Yp2 =Omm, Xpsl =9.5 mm, Ypsl =0 mm, Xps2 =-13.3 mm, Yps2 =0
mm)
247
Figure 5.6.18 shows the radiation patterns for the balanced dual
concentric shorted ring antenna measured at 2.1 GHz. In Figure 5.6.18, 0°
corresponds to broadside. The measured cross polarization levels were
greater than 15 dB below the co- polarized fields in both E- (tP = O~ and H-
planes (tP = 90~. A significantly lower cross polarization level was achieved
for the H-plane measurement, and this discrepancy may be due to the
presence of the connectors and cables used to record the radiation pattern not
being calibrated out of the measurement. These cables will have a more
adverse effect when scanning in the E-plane. As can be seen in Figure 5.6.18,
the patterns are fairly symmetrical around broadside. It should be noted that a
greater degree of cross-polarized field reduction could have been achieved if a
larger ground plane was used; here the ground-plane was only 7 em x 5 em.
The measured gain of the antenna was ~. O dBi.
180
5.6-4.1 Introduction
249
To ensure the overall size of the antenna is small, the 4 shorted
patches are located in close proximity to each other, with the gap between the
patches approximately 1 rom. It should be noted that the 4 patches must not
be in direct contact as this will substantially increase the size of the antenna
for a set operation frequency due to the inductive loading effect of each
additional shorting pin. To maximize the impedance bandwidth of the shorted
patches a low dielectric constant, thick substrate (Rohacell: Er = 1.07, d = 9
rom) was utilized. A thin layer of RT/Duroid 5880 (e, = 2.2, d = 0.254 rom)
was used to etch the patch conductors.
Figure 5.6.20 shows a schematic diagram of the feed network for the
synchronous subarray. As can be seen in this diagram, additional lengths of
microstrip line are used to provide the required 900,1800 and 2700 phase shifts
at 3 of the excitation ports (labeled as 2 - 4 in the layout). Also apparent in
Figure 5.6.19 are the impedance transformers needed to ensure the antenna is
well matched. To minimize the size of the feed network, a high dielectric
constant substrate was used (RT/Duroid 6010, Er = 10.2). A relatively thick
feed substrate was chosen (d = 1.9 rom) to ensure the widths of the feedlines
required were not too small, thereby improving the robustness of the proposed
antenna. To connect the patch conductors to the feed network, pins of radius
0.45 rom were used. Similar pins were soldered to the 4 patch conductors and
the shared ground-plane between the feed and the printed antenna to provide
the required short circuits. Finally a SMA connector was soldered to the input
port shown in Figure 5.6.19. Both antenna and feed substrates were truncated
such that the area of the CP antenna was 4.5 x 4.5 em. A photograph of the
manufactured 2 x 2 synchronous subarray is show in Figure 5.6.21.
2
3
1
4
Input port
Figure 5.6.20 Layout of feed network for 4 Element Synchronous Subarray of Shorted
Patches
250
Figure 5.6.21 Photograph of 4 Element Synchronous Subarray of Shorted Patches
--- -, V- r---
10
9
I
·5 8
m
~ \ -, 7
I
-10 6
:l
.9 ~
~
5
E 4 0"
·15
a:
\j
::>
iii ~
a: 3 ~
-20 2
~ 0
1.1 1.15 1.2 1.25 1.3 ' 1.35 1.4 1.45 1.5
Frequency (GHz)
Figure 5.6.22 Return Loss and Axial Ratio of 4 Element Synchronous Sub array of
Shorted Patches
251
It should be noted that there are techniques that can be incorporated to
further reduce the size of the proposed antenna including placing 'slits' in the
printed conductors (as in [47]) and also deforming the printed conductor
around the handset (as in [48]). If the symmetry of these configurations can
be maintained, then similar results to those presented in Figure 5.6.22 should
be expected.
I
I. <x.,3· YpJ)
'-'J .K
(xpo!'
\
Ypa)
';..
L
2
I <J,yps3) ( t ~
IA~ ~'YP2 : ~
(~I·Ypol)
~ «,r- Yp , )
.K
- - - L 1- - -
x
Figure 5.6.23 Schematic Diagram of 3 Element Synchronous Subarray of Shorted
Triangular Patches
252
Here, each of the 3 equilateral triangle patches of length, L is fed by a
probe of radius ro located at (Xph ypi) relative to the center of the ilb patch (i =
1, 2, 3). A shorting pin of radius ro s is positioned (Xpsh ypsi) from the center of
each of the triangular shaped patch conductors. The subarray is mounted on a
grounded foam substrate (e, = 1.07) to maximize the bandwidth of the shorted
patch configuration [25]. The height of the substrate (d) is 10 mm, A thin
layer of Taconic TLC30 (e, = 3.0, d = 0.5 mm) is used to etch the patch
conductors. The patches are sequentially positioned and fed in a 0, 120 and
2400 phasing arrangement and so the entire antenna demonstrates symmetry
about boresight. A small gap (1 mm) exists between the apexes of the
triangles. The feed network is similar to that presented for the 4 element
synchronous subarray, although here a 3-way power splitter was developed to
ensure the correct magnitude distribution to each port. As for the 4 element
synchronous subarray, to achieve the required phasing to the second and third
patches, additional lengths of microstrip line were used. To minimize the size
of the feed network, a high dielectric substrate was used (RT/Duroid 6010, e,
= 10). A relatively thick substrate (2.5 mm) was chosen to ensure the widths
of the feedlines are relatively thick, thus making the antenna robust. In a
similar arrangement to that given for the 4 element synchronous subarray,
pins of radius 0.6 mm were used to connect the short patches to the feed
network. Shorting pins (ros = 0.325 mm) were soldered between the patch
conductors and the common ground-plane. An SMA connector is the input to
the feed network and the substrates of both the antenna elements and the feed
were truncated to give an antenna with the dimensions of 4.5 em x 4.5 ern. A
photograph of the manufactured 3-element antenna is presented in Figure
5.6.24.
lr- B
! e!
:
-10
v-.. r-, /
.3 -15
I
E
... I
~ ·20
IX:
II 2
-25
~ 0
1 1.2 1.4 U 1.8 2 2.2 2.4 2.8 2.B 3
Frequency (GHz)
Figure 5.6.25 Return Loss and Axial Ratioof 3 Element Synchronous Subarray
(Parameters: ~ = 21 mm, Xpl = 0, Ypl = -8.1 mm, Xpsl = 0, Ypsl = -6.1 mm,Xp2 = 7.0 mm, Yp2 =
4.1 mm,Xps2 = 5.3 mm, Yps2 = 3.1 mm, xpJ = -1.0 mm, YpJ = 4.1 mm, xpsJ = -5.3 mm,YpsJ = 3.1
mm)
254
110
Thus far in this section we have shown that a small printed antenna
with good impedance and axial ratio characteristics can be developed using a
subarray of shorted patches with sequential rotation. The synchronous
subarray procedure presented earlier was applied to both 4-element and 3-
element shorted patch subarrays and yield 10 dB return loss and 3 dB axial
ratio bandwidths in excess of 15 %. Although good impedance and radiation
characteristics were obtained for these configurations, the overall size of the
antenna was still significantly larger than the individual shorted patch due to
the need for the other elements in the subarray. It should be noted that a 2-
element synchronous subarray has not been considered using conventional or
shorted patches due to the lack of spatial symmetry that is necessary to
achieve good CP qualities.
/
Port 2
-10 i\ I
ID
'tJ
"15
\ / /
.2
520
\ / 1\, I
~
I
",
f·25
·30
-35
1.6 1.7 1.8 1.9 2 2.1
frequency, GHz
Figure 5.6.29 Measured Return Loss of CP Dual Concentric Shorted Ring Antenna
= = = = =
(Parameters: a, 10 mm, az 7 mm, £'1 2.2, d 10.3 mm, b J 14 mm, b z 11 mm, E,z = =
= = = = =
1.07, d 10 mm, rol 0.2 mm, rosl 0.3 mm, r o2 0.325 mm, ros2:: 0.6 mm, Xpl 7.5 mm, Ypl
= = = = = =
0 mm, Xp2 0 rnm, Yp2 11.5 mm, Xpsl 9.5 mm, Ypsl 0 mm, Xps2 0 mm, Yps2 13.3 mm) =
The CP version of the shorted concentric ring antenna is significantly
smaller in size than the other printed subarray antennas mounted on similar
substrates and operating at the same frequency band. Table 5.6.1 gives a
comparison between the three configurations, in terms of size, impedance
behavior and radiation characteristics.
257
110 110
(a) (b)
110
(c)
Figure 5.6.30 Spin Linear Radiation Pattern of CP Printed Antenna: (a) Measured at 1.7
GHz; (b) Measured at 1.8 GHz; (c) Measured at 1.9 GHz
5.6-5.i introduction
258
microwave laminates. However, using such a material (foam) jeopardizes the
ease of manufacturing and the robustness of the antenna, contrary to the often-
touted advantages of printed circuit technology.
Figure 5.6.31 shows a schematic diagram for the two shorted patch
antennas. The antenna consists of a coaxial probe of radius r, located at (xp,
yp) and two shorting posts of radius ros located at (Xpsh YPls) and (Xps2' yps2)
from the center of the patch conductor. The shorted patch antenna has been
constructed with two shorting posts as it reduces the fabrication tolerance
between the probe and shorting pins, as was shown earlier this chapter. The
patch conductor has a length L I and width WI shown in Figure 5.6.31a. In
order to investigate the effectiveness of PBG structures on shorted patch
antennas a second antenna shown Figure 5.6.31b was constructed. This
shorted patch antenna has identical dimensions as the previously mentioned
259
shorted patch only it is surrounded by a PBG structure. Both shorted patch
antennas are mounted on a grounded substrate of thickness d and dielectric
constant Er (not shown in Figure 5.6.31). The PBG structure surrounding the
shorted patch antenna consists of periodic square metallic pads of length L 3•
width W3 and a lattice constant a, located on the same surface as the patch
conductor. Each metallic pad has a via in its center, with a radius rOY
connecting the pad to the ground plane. The presence of the nearby metallic
pads and vias tend to raise the resonant frequency of the shorted patch antenna
since the effective cavity volume is reduced [50]. Given that it is important
the resonant frequency of the antenna remain unchanged in order that the
antenna operates within the band gap of the structure, a small ring of bare
substrate around the patch conductor was incorporated [51]. Thus the metallic
pads are separated from the patch conductor by a distance gx (r-direction) and
gy (y-direction) as shown in Figure 5.52b. It is important to note that the
overall dimensions (L, x W2 X d) for both shorted patch antennas are the
same. For more detail on the PBG design refer to Chapter 4.
)' 1
(a) (b)
Figure 5.6.31 Schematic diagram of (a) Conventional Microstrip ShortedPatch Antenna
and (b) proposed paG shortedpatch.(Configuration 1 parameters: e,= 2.2, d = 9.51 mm,LJ =
12 mm,WJ =6 mm, fa = 0.6 mm, f oo =0.6 mm, Xp = 3 mm, Yp = 0 mm, XpaJ= 5 mm, YpsJ =1.5
mm, Xpa2 = 5 mm, Yp02 = -1.5mm, Lz = 69mm, W2 = 69mm) (Configuration 2 parameters: e, =
2.2, d = 9.51 mm, L = 12 mm, W = 6 mm,r, = 0.6 mm, ro• = 0.6 mm, x p = 3 mm, Yp = 0 mm,
xpsJ = 5 mm, YpoJ = 1.5mm, Xps2 = 5 mm, Yps2 = -1.5 mm, Lz = 69mm, W2 = 69mm, L, = 9mm,
W3 =9mm, a = 10 mm, rOY =0.5mm, gx =2.5 mm, gy =3.5mm)
Figure 5.6.32a shows the measured return loss (SJI) of the probe-fed
shorted patch antenna on an ordinary metallic ground plane and the other with
the PBG structure. Although the presence of the PBG has raised the resonant
frequency of the antenna slightly, it still remained within the bandgap and so
260
no measures to compensate for the increase in resonance was undertaken. The
measured 10 dB return loss bandwidth for the antenna surrounded by the PBG
structure is narrower compared to the antenna on the conventional ground
plane. The bandwidths are 3.1% and 5.9% for the shorted patch with the PBG
ground plane and the conventional shorted patch, respectively. This is a
characteristic common to an antenna with higher efficiency as surface waves
(or any loss for that matter) tend to mask/improve the return loss of the
antenna. On the other hand, a significant improvement in the broadside gain
and gain bandwidth were observed as shown in Figure 5.6.32b. The maximum
gain for the ordinary shorted patch is 1.2dBi at 2.97 GHz with a IdB gain
bandwidth of 1.3 %, while the maximum gain for the shorted patch antenna
with the PBG ground plane is 4.4 dBi at 2.9 GHz with a 1 dB gain bandwidth
of 4.5 %. This gain enhancement is equivalent to a 109 % increase in effective
radiated power (ERP) at broadside.
o
.. ,
. . .
,
. . ('
1\
,
,
H ~"- '==WIUtPBO I v VV\
,
,
, y
,,
JV , I~
... ..
'. ~1
... ,.tett .5 0 i '
.~-:.I .t I.
• 8hor1ed
~ hloh wlh PSO
~ I. ,, ':
"t\:
'!. V\(
·1
~
.
I~· ·"",,~, "
"
t-
-2
";:'.:\ fJ' -. ".:
-3
-<lO -4
\f"
2.8 2.7 2.8 U
Frequency (GHz)
, 3.2 :1.8 2.7 2.8 2.' ' .1
Frequency (GHz)
(a) (b)
Figure 5.6.31 Comparison of Performance of Conventional Shorted Patch and PBG
Based Shorted Patch: (a) Return Loss; (b) Gain
180 180
180 180
AnI E-plo"" (erOM-polar) - Ant H-pI_ (ero_lat)
••• •Ant&PBG E-pI.... (_.o-polor)
- - - . Ant & PBG Hop..... (ero_lar)
(c) (d)
Figure 5.6.33 Measured Radiation Patterns of Conventional Shorted Patch and PBG
Shorted Patch (E-plane: x-z plane in Figure I , H-plane: y-z plane in Figure 1)
5.7-1 Introduction
e, cr(S/m) p
(kg/nr')
cartilage cartilage 40.22 1.29 1000
muscle muscle 53.55 1.34 1020
eye sclera, cortex, nucleus, vitreous 50.58 1.49 1000
humour,
nerve, brain cerebellum, cerebro spinal 50.11 1.85 1050
fluid, white matter
skin skin (dry) 38 .87 1.19 1000
fat, bone averaged infiltrated fat and 11.40 0.23 1200
cortical bone
blood blood 59.37 2.04 1000
Table 5.7.1 FDTD Mesh Tissue Group ings and Corresponding Tissues Averaged from
[51] to Provide Dielectric Properties
x~
Figure 5.7.1 Perspective View Showing Effective Orientation of Handset Relative to
Head and Coordinate System
The resampling of the head model for 2.5 rom voxels and the
rotations were done on the original head mesh to minimize anatomical errors,
however any distortions are expected to be within the range of human
variation. The major anatomical features were cross-checked with an atlas of
sectional anatomy before and after the resampling and rotation so that gross
distortions would be detected. The block model of the hand wraps around
three sides of the lower half of the handset and consists of a core of bone
surrounded with a layer of muscle similar to that used in. Details of the hand
model can be seen in Figure 5.7.2.
Muscle
Figure 5.7.2 View of the handset showing the location of the shorted patch antenna and
the dimensions of the block model hand
265
The resulting mesh has dimensions of 147x161x148 (= 3.4million) cells
requiring approximately 105 Mbytes of RAM. The analysis was run on a HP
715/100 with 128 Mbytes of RAM and took about 8 hours for a transient or
steady state excitation. The results for the antennas on the handsets in
isolation used a space of 65x70x120 cells and took just over 1 hour to
calculate.
266
with the head present are 'corrected' for the rotation of the head so that they
represent the antenna patterns with the handset in the intended position of use.
The antenna radiation patterns for the antennas on the handset in isolation are
not transformed as it was observed that handsets are generally positioned in
an upright orientation when not in use.
Table 5.7.2 Dimensions of Single Layered and Stacked Shorted Patch Antennas
~ \. ~
~f#" """
tti,
~ ... ~ l:t:e: ~ 't ~ :\ 1"" '\ J!
.D' fU" r'\
,t """ ·'0 ':'; .i .,.
r~ "'Jlf- IJ '11.."
I 1\
"
· 10
...il \ If) j ~
."
II \
\ r, j
-20
'J
·20
v. ~
u
.-.. ."
ri\ r"\
I I)
"'
· 10 ·30
o ~ ~ ~ ~ 00 1m 2~ ~ m D m ~ o » ~ ~ ~ ~ 1m 2W * m m m ~
Df~U DfCl'!'rJ
Pp~po ~ -O-Mcu_ _ po~
(a) (b)
267
./,~ '\
,
0
~ \ r\ )tj~
4 v:
1. -
I ~ \f",
·5
"r
~r<1
(j \If rJ
· '0
T; ~ ~ lJ~ r..,
0;
-e
· 15
-:lll
l~
ft"" ~ \ i
18
-25
r-.t
"V If
· 30
l
o 30 00 so l :lll ' ''' 1110 2'0 24/l 270 3CIl 3JO 360
D.S"'''
Ephi-Ja:lDOpo1e -o- B.bcta-cro. opo1c
- 0 - Epbk .a: 1ep~b -0- &betH.,1e, .ceil
____ Epbktltbd p.tcb -*- Bbcta.. ,.bd p-.h
(c)
Figure 5.7.3 Azimuth (9 =90°) Patterns for Monopole, SingleLayered and Sacked
ShortedPatches on: (a) Vertical handsetin isolation; (b) Handset Adjacent to Head Rotated to
CENELEC normal Position of Use;(c) Including BlockModelHand
268
J
IP'" _I-
~
"I Ij ~ ?f-~ If
..,;,;
V ~l -
"'" ~
~ l.P ~ r-... I ...
'l:. II
.J
I' t-,~ \J f- 1-
>...:
~~ If
~
·'0 II( · 10
1\
~
il
-e T I\..J ~ ~ I~ ~
~ 1'\h WII'(J
.1$ . 1$
·lll
~
~ iK\
I ~
·1O
loA" r-... ~ l1 r0o- F::: . J)
o ]() so :wo xo :m
... .
o 30 60 to 120 l'O 110 210 )10 l?D )CO )JO )60 90 120 ISO 110 210 VO )00
DC&ITU Dcpul
~_aopo lll -o- E1lltla-nc>MJlClII: 1f'W-~ -o- BJKt.· goaopok
~.I p. kiI -o- E'a' la-JM... ". td ....o- r"bH.... ' .l(b -o- Bbd•.•• ' RP-CdI
-r-~.ucbol .. Ie' ~ .lH tac bcl "d ...... 1l'...... lKkDcI ,.tch _abCII' -ltad,,,4p.r.c.b
(a) , (b)
_h<
-
'rF'
l..il: W ~ ~b.. l,., ~
....
"V
·'0
~ \ II Ir -
.....
II \
.....
·U
v r
;rl-
h ~
r1 M'I'I
f--
·lll
v,\
. 1j
·lO
o
f--
~
)0 ro 90 120 1.50 110 210 NO
\ ,II
no 300 3JO )6()
Del~U
flo"--.op:>}, -1UwU·OPft09Ok
-'- lVW-... ,1o p4 l o;b ~Bheta"'''. 10 '''cb
-+- fl>bl-.uc b4 p' lch Bbeta-atac.Lt.ip.lcb
(c)
Figure 5.7.4 Elevation <Ill =00 ) Patternsfor Monopole, SingleLayeredand Sacked
ShortedPatches on: (a) Vertical handsetin isolation; (b) Handset Adjacent to Head Rotatedto
CENELEC normalPositionof Use; (c) Including BlockModel Hand
~ . IS f--=...~-~_t_---:~+--___j
!a. -20 t - - --I-- - ="ki ---+-- - --j !a. ·20 t -- - -I--"<-- --+-&d=---t-:;r-- ---j
-25 t -- - - I -- --t------''''r- - --j ·25 . f - -- 4-- .:s,-__1_- +-t-- - -j
. )1) t----I----t----+---"O;~'-i . )1) f---_t_- -1r/--f-- -t-- --j
.lS ~~~+____~~.........
~~........~~~
·'0 .f-----"-,.--1-~.......-+___pl~..+"'--_j
~ ·'1 t---'f""=~~+----+----;
(c)
Figure 5.7.5 Isul plotsfor Monopole. Single Layered and Sacked ShortedPatches on:
(a) Vertical handset in isolation; (b) Handset Adjacent to HeadRotated to CENELEC normal
Position of Use; (c) Including BlockModel Hand
Table 5.7.3 and Figure 5.7.6 compare the SAR distributions for each
of the antenna types. A previous study which modeled a ')../4 monopole at
1900 MHz close to an upright or 300 forward tilted head found peak 1 g
SARs of 8.88 and 8.64 Wlkg respectively for 1 W of time averaged radiated
power. The lower SARs in our study are explained by the finite return loss of
270
the antennas and the rotation of the head placing the antenna further from the
surface of the head. The peak spatial SARs for the shorted patch antennas are
about one third of those for the monopole. This is better than the reduction
seen at 900 MHz with back mounted planar inverted-F antennas and can be
attributed to the better directivity achievable at higher frequencies. For each
of the antennas the inclusion of the hand in the present position reduces the
efficiency by about 10%. However if the hand is moved further up the
handset the efficiencies of the shorted patch antennas is expected to degrade
faster than that of the monopole due to current flows and possible physical
obstruction of the antenna. The SAR contours of Figure 5.7.6 show the
monopole to be exposing a wide area of the side of the head whereas the
shorted patch antennas show distinct volumes of absorption in the area of the
ear and the cheek. The shorted patch antennas produce higher SARs in the
hand than does the monopole due to the greater current flows on the handset
case.
(a) (b)
271
(c) (d)
(e) (0
272
proximity to the shorting post was then presented and verified. The resulting
antenna was 75% smaller in surface area than a conventional microstrip patch.
Bibliography
[1] J. E. Padgett, C. G. Guntherand T. Hattori, "Overview of wireless personal
communications," IEEECommunications Magazine, pp. 28-41,Jan. 1995.
[2] T. S. Rappaport, "Future trendsof mobileand personal communications," (Invited
paper) SBMOIIEEE MIT-S IMOC'95 Proceedings, pp. 387-395.
[3] R. Prasad,"Overview of wireless personal communications: Microwave
Perspective," IEEE Communication Magazine, pp. 104-108, April 1997.
[4] J. L. Finol and lG Mielke,"Past and futuredirections in cellar telephony," (Invited
paper) IEEEAP-S, pp. 7-13,1998.
[5] T. Ojanpera and R. Prasad, "An overview of third-generation wireless personal
communications: A European Perspective," IEEEPersonal Communications, pp. 59-
65, Dec. 1998.
[6] D. M. Pozar, "An overview of wireless systems and antennas," IEEEProc. AP-S, pp.
566-569,2000.
[7] K. Hirasawa and M. Haneishi, Analysis. design. and measurement of small and low-
profileantennas, ArtechHouse, Inc., 1992.
[8] K. Fujimotoand l R. James,MobileAntennaSystems Handbook. Norwood, MA:
ArtechHouse, 1994.
[9] C. A. Balanis, AntennaTheol)': Analysis and Design, secondedition,John Wiley &
Sons Inc., 1997.
[10] R. C. Johnsonand H. Jasik, AntennaEngineering Handbook, 2nd Edition,McGraw-
HilI,1984.
274
[11] Z. D. Liu, P. S. Hall and D. Wake, "Dual-frequency planar inverted-F antenna," IEEE
Trans. Antennas & Propagat., vol. 45, pp. 1451-1458, Oct. 1997.
[12] M. A. Jensen and Y. Rahmat-Samii, "Performance analysis of antennas for hand-held
transceivers using FOID," IEEETrans. Antennas & Propagat., vol. AP-42, (8), pp.
1106-1113, Aug. 1994.
[13] P. Salonen, M. Keskilarnmi and M. Kivikoski, "Single-feed dual-band planar
inverted-F antenna using V-shaped slot," IEEETrans. Antennas & Propagat. , vol.
AP-48, (8), pp. 1262-1264, Aug. 2000.
[14] C. R. Rowell and R. D. Murch, "A capacitively loaded PIFA for compact mobile
telephone handset," IEEETrans. Antennas & Propagat., vol. 45, pp. 837-843,1997.
[15] R. E. Munson, "Conformal microstrip antennas and microstrip phased arrays," IEEE
Trans. Antennas & Propagat., vol. AP-42, (1), pp, 74-78, Jan. 1974.
[16] 1. W. Howell, "Microstrip antennas," IEEE Trans. Antennas & Propagat., vol. AP-
23, (1), pp. 90-93, Jan. 1975.
[17] K. R. Carver and 1. W. Mink, "Microstrip antenna technology," IEEE Trans.
Antennas & Propagat. , vol. AP-29, (I), pp. 2-24, Jan. 1981.
[18] T. K. Lo, Y. Hwang, E. K. W. Lam and B. Lee, "Miniature aperture-coupled
microstrip antenna of very high permittivity," Electron. Len; vol 33, pp, 9-10, Jan.
1997.
[19] G. A. Kyriacou and J. N. Sahalos, "Analysis of a probe-fed short-circuited microstrip
antenna," IEEETrans. Vehicular Tech., vol. 45, (3), pp. 427-430, Aug. 1996.
[20] A. Boag, Y. Shimony and R. Mittra, "Dual band cavity-backed quarter-wave patch
antenna," IEEEAP-S Digest, vol. 4, pp. 2124-2127, 1995.
[21] R. B. Waterhouse, "Small microstrip patch antenna," Electron. Lett., vol. 31, pp. 604-
605, April 1995.
[22] K. -L. Wong and Y. -F. Lin, "Small broadband rectangular microstrip antenna with
chip-resistor loading," Electron. Lett., pp. 1593-1594, Sept. 1997.
[23] 1. -H. Lu, "Single-feed dual-frequency rectangular microstrip antennas with pair of
step-slots," Electron. Lett ., vol. 35, (5), pp. 354-355, Mar. 1999.
[24] D. M. Pozar, Microwave Engineering - Second Edition, John Wiley and Sons Inc.,
New York, 1998.
[25] 1.T. Aberle, D.M. Pozar and C.R. Birtcher, "Evaluation of input impedance and radar
cross section of probe fed microstrip patch elements using an accurate feed model",
IEEETrans. Antennas Propagat., AP-39, pp. 1691-1697, December 1991.
[26] M. Sanad, "Effect of the shorting posts on short circuit microstrip antennas", Proc.
IEEEAP-Symp., pp. 794-797, June 1994.
[27] I. Park and R. Mittra, "Aperture-coupled small microstrip antenna", Electron. Lett.,
Vol. 32, pp. 1741-1742, Sept. 1996.
[28] R. B. Waterhouse and S. D. Targonski, "Performance of microstrip patches
incorporating a single shorting post", Proc. IEEEAP-Symp., pp. 29-32, July 1996.
[29] 1. Huang, Personal Communication.
[30] D. H. Schaubert, "A review of some microstrip antenna characteristics", Microstrip
Antenna Design, IEEE Press, pp.59-67, 1995.
[31] Ensemble 5.1, Ansoft, 1998.
[32] R. B. Waterhouse, S. D. Targonski and D. M. Kokotoff, "Design and performance of
small printed Antennas", IEEE Trans. Antennas & Prop., vol. AP-46, pp. 1629-1633,
Nov.1998.
[33] J. T Rowley and R. B. Waterhouse, "Performance of shorted microstrip patch
antennas for mobile communications handset at 1800 MHz," IEEE Trans. Antennas
& Prop. Vol. AP-47, pp. 815-822, May. 1999.
[34] M. A. Jensen and Y. Rahmat-Samii, "EM interaction of handset antennas and a
human in personal communications", Proc. IEEE, vol. 83, pp. 7 -17, Jan. 1995.
275
[35] D'Inzeo, "Proposal for numerical canonical models in mobile communications" in
Biomedical Effects of Electromagnetic Fields - Reference Models in Mobile
Communications, D. Simunic, Ed. Rome, Italy: COST244, pp. 1-7, 1994.
[36] R. B. Waterhouse, "Small printed antenna easily integrated into a mobile handset
terminal," Electron. Lett., Vol. 34, pp. 1629-1631, Aug. 1998.
[37] F. Ali and J. B. Horton, "Introduction to special issue on emerging commercial and
consumer circuits, systems, and their applications," IEEE Trans. Microwave Theory
Tech., Vol. 43, pp. 1633-1638, July 1995.
[38] R. B. Waterhouse, "Small printed antenna easily integrated into a mobile handset
terminal," Electronics Letters, vol. 34, pp. 1629 -1631, Aug. 1998.
[39] R. B. Waterhouse, 1. T. Rowley and K. H. Joyner, "A stacked shorted patch,"
Electronics Letters, vol. 34, pp, 612 - 614, April 1998.
[40] 1. Rashed, and C.-T. Tai, "A new class of resonant antennas", IEEE Trans. Antennas
& Propagat, vol. 39, pp. 1428 - 1430, Sept. 1991.
[41] P. S. Hall, Private Communication.
[42] XFDTD (1997), User's manual for XFDTD the X-Window Finite Difference Time
Domain Graphical User Interface for Electromagnetic Calculations. Version 4.03,
Remcom Inc., June 1997.
[43] R. B. Waterhouse, "Printed antenna suitable for mobile communication handsets",
Electron. Lett.• Vol. 33, pp. 1831-1832, Oct. 1997.
[44] Zealand Software Inc., IE3D™ Version 6.01.
[45] D. M. Kokotoff, 1. T. Aberle and R.B. Waterhouse, "Rigorous analysis of probe-fed
printed annular rings," IEEE Trans. Antennas & Prop., vol. AP-47, Feb. 1999.
[46] J. Huang, "A technique for an array to generate circular polarization with linearly
polarized elements," IEEE Trans. Antennas & Propagat., vol. 34, pp. 1113-1124,
Sept. 1986.
[47] C. -Yo Huang, 1. -YoWu and K. -L. Wong, "Broadband circularly polarized square
microstrip antenna using chip-resistor loading," lEE Proc. Micro. Antennas &
Propagat., vol. 146, pp. 94-96, Feb. 1999.
[48] M. A. Jensen and Y. Rahmat-Sarnii, "Performance of circularly polarized patch
antennas for personal satellite communications including biological effects," Proc.
AP-S, Newport Beach, Califoma , USA, pp. 1112-1115, June 1995.
[49] P. S. Hall, 1. S. Dahele and J. R. James, "Design principles of sequentially fed, wide
bandwidth, circularly polarized microstrip antennas," lEE Proc. H, vol. 136, pp. 381-
389, Oct. 1989.
[50] D. F. Sievenpiper, L. Zhang, R. F. Jimenez Broas, N. G. Alex6polous, E.
Yablonovitch , "High-Impedance Electromagnetic Surfaces with a Forbidden
Frequency Band," IEEE Trans. Microwave Theory and Tech., vol. 47, pp. 2059-
2074, Nov. 1999.
[51] R. Coccoli, F. Yang, K. Ma, T. Itoh, "Aperture-Coupled Patch Antenna on UC-PBG
Substrate," IEEE Trans. Microwave Theory Tech., vol. 47, pp, 2123-2130, Nov.
1999.
[52] C. Gabriel, "Compilation of the dielectric properties of body tissues at RF and
microwave frequencies", Brooks Air Force Base, report no. AUOE-TR-1996-0037,
1996.
276
Chapter 6 Direct Integration of Microstrip
Antennas
palch
ar1ema
280
6.1-3 Single Layer Printed Antennas on High Dielectric Constant
Material
Edge feeding
Patch conductor
-
E
~
-20
CD
a: ·25
-30
-35
3.5 4 4.5 5
Frequency (GHz)
281
Figure 6.1.5 show the E- and H-plane radiation patterns for the
benchmark antenna. A wide beamwidth in both planes is evident, which is a
typical trait of printed microstrip patch antennas. The backfire radiation level
is low, exhibiting a front to back ratio of 24 dB. Figure 6.1.6 plots the
magnitude of the E-fIeld for a two dimensional E-plane cut through the
benchmark antenna. A propagating wave is evident on the microstrip line,
showing field concentrations in the high dielectric constant substrate between
the strip and ground plane. Also note the high intensity field in the high
permittivity substrate under the patch conductor, implying that a large portion
of the field is trapped within the substrate.
o
90
- H-plane
E-plane
180
•
Max
_".,"C--'"
1 High
•
pennittivity
T substrate
Min
Ground plane
282
6.2 Slot-coupled Procedures and Solutions
~-Z~~~~;~~~~~-COUPl i ng aperture
Feed substra te
284
when the aperture is mutually coupled to parasitic patch elements, the overall
aperture stacked patch antenna bandwidth is diminished.
70
:.........
65
.........
r-,
---
60
::::l!
0 ~ '\..
55
J:
'C
<,
<,
50
j
'C
<,
C 45
lU
lD
40
35
-,
~
30
2 3 4 5 6 7 8 9 10 11
Relative Permittivity
'.
I'll
'0
8. 300 \--- f -- H I -+---:r+--+'tt--t---tf-t---l
.5
,,
'5 200 f - - I ' t - -+++-+--i++---f..I.& --- t - ---lf--l--i
Q.
c
,,
.1
5 10
Figure 6.2.4 Input Impedance Comparison for High and Low Feed Substrate
Permittivity ApertureAntennas
The peak gain within the impedance bandwidth of the antenna is also
diminished with increasing permittivity, although it still remains at a typical
level for microstrip patch antennas. This is shown in Figure 6.2.5. The
minimum gain remains relatively constant. It should also be noted that the
width of the 100 .Q sections of microstrip line in the reactive power divider
became extremely thin (80J.lIIl) when the permittivity of the feed material
reached 10.2. The level of back radiation remained constant.
285
Due to the very thin microstrip lines required to realize the reactive
power divider of the antenna structure of Figure 6.2.2 when a high
permittivity feed substrate is employed, a single feed line aperture stacked
patch antenna was developed. A schematic of the single feed line aperture
stacked patch antenna with a feed substrate dielectric constant of 10.2 is given
in Figure 6.2.6. For OEICIMMIC integration purposes, the thickness of the
high permittivity feed substrate (hi) is excessive. A theoretical study of the
consequences of decreasing hi was undertaken using [10]. Once again, the
structure was optimized for maximum operational bandwidth after each
iteration of hi.
1.5
:....
...... 1'\..
. . . Maximum Cain
1-. - MInimum Cain
r-
I
........
~
~
iii 7.5
........... Ie
:!:!.
c 7
'iii
-- .. .. .- -- -. . . .. .. .. ..
C) 1.5
I
:-- -
~
5.5
5
2 4 5 I 7 I a 10 11
Relative Permittivity
Parasitic patch
element I (I" wr)
_ High E, feed substrate (h i)
286
aperture is mutually coupled to parasitic patch elements, the overall aperture
stacked patch antenna bandwidth is diminished with decreasing hi.
42
40 »:
38 /
~
..
~
.t: 38
V
·"i 34
./
"III
C 32
./
ca
30
»"
28
... /
26
0.2 0.4 0.6 0.8 1.2 1.4 1.6
Thickness (mm)
As can be seen from Figure 6.2.8, the peak gain within the impedance
bandwidth of the antenna is relatively consistent as hi is varied, with slight
fluctuations around 8 dBi. The minimum gain displayed some minor
variations, but maintained a level above 6 dBi. The front to back ratio
exhibited improvement for lower values of hi.
8.5
I,'"
....- -
Mlllmum G.ln
----... ~ ,.
- • • Minimum G.ln . /
~ 7.5
~ 7
Thickness (mm)
287
6.2-3 CPW fed Aperture Stacked Patch on high dielectric material
60 In(8d +~)
F: W 4d
for W
d
~1
Zom = 120n
ohms (6.1)
e, is the effective relative permittivity
d is the thickness of the dielectric substrate
w is the microstrip line width
ac=--
Rs
Np/m where R. =~CDIl'
ZomW 20-
(6.2)
R, is the surface resistivity of the conductor
Z = 30n K'(k)
where k =
ohms S
ocpw ~£'2+1 K(k) S+2W
(6.3)
S is the CPW center conductor width
W is the CPW gap width
K(k) is the complete elliptic integral of the first kind and
K'(k) = K(k')
kl=~(1-k2)
288
As shown in Equation 6.1, the impedance of a microstrip line (Zom> is
dependant on the thickness of the substrate. Hence, for a given substrate
thickness and permittivity, only a single value of conductor width can achieve
a desired Zorn' For OEICIMMIC integration applications where the substrates
materials are generally thin and of high dielectric constant, very fine
microstrip lines are required. Aside from elevating the complexity of
fabrication, decreasing the width of a microstrip line increases the conductor
loss. This is especially pertinent at millimeter-wave frequencies, as the
attenuation due to conductor loss is also escalated at higher frequency (refer
to Equation 6.2).
CPW also has other advantages over microstrip transmission lines for
OEIC/MMIC integration. CPW circuits only require processing on a single
side of an OEICIMMIC wafer. Aside from the ease of connecting series and
shunt devices as previously mentioned, single sided processing requires a
lower number of fabrication steps than microstrip based circuits, reducing
manufacturing expenses. Having the center conductor and ground electrodes
of the CPW in the same plane provides a greater level of isolation between
adjacent transmission lines. Therefore, it may be possible to achieve smaller
circuit/chips sizes if CPW is used as the transmission medium, rather than
microstrip lines.
- - - Palch substrate
Parusitic patch
elemen t I
Figure 6.2.11 compares the simulated and measured far field radiation
patterns. Reasonable similarity in the shape of the radiation patterns is
observed, and front to back ratios of 13 dB and 11 dB were obtained for the
theoretical and experimental results, respectively. This high level of back
radiation is a consequence of the resonant aperture in the ground plane. A
resonant aperture element in isolation will radiate to an almost equal level at
broadside and backfire. Hence, a trade-off exists between the enhanced
290
bandwidth that can be achieved with a resonant aperture element in a CPW
fed ASP, and the level of back radiation. Techniques to reduce the back
radiation of CPW fed ASP antennas are given in Section 6.2.4. Both patterns
in Figure 6.2.11 exhibit a side lobe approximately 5 dB down from the main
lobe in the E-plane, which is a consequence of spurious radiation from the
CPW feed line.
(a) (b)
Figure 6.2.10 Input Impedance of CPW fed ASP: (a) Predicted; (b) Measured
The theoretical gain of the CPW fed ASP was in excess of 8 dBi for
the majority of the impedance bandwidth of the antenna,remains above 7.5
dBi. The gain remained extremely close to the directivity of the antenna,
differing by no more than 0.2 dB. This implies that the antenna is
exceptionally efficient, and its performance is not being degraded by the
excitation of surface waves.
o
-90 I-w-t--+~iE-+--h41--1 90
180
1:l0
1- II·pIane
-E-pIoDt
I tOO
(a) (b)
1 High
permittivity
T substrate
Groond pione
(a)
l Uigh" .
PCmutUVl ty
T substrate
(b)
Figure 6.2.12 E-fieldIntensity Plot of (a) conventional microstrip patch antenna; (b)
CPW fed ASP antenna
292
properties for OEICIMMIC integration. Using a substrate structure consisting
of a high permittivity feed substrate and foam parasitic patch dielectric, low
surface wave excitation and an ample impedance bandwidth can be achieved.
This single parasitic patch element antenna has a lower profile than the CPW
fed ASP configuration, but has a reduced, although still generous bandwidth
(approximately 20 - 25%) due to the decrease in the number of resonators.
The use of a shielding plane separated from the rear of the antenna by
a foam dielectric can reduce the backward radiated power. This shielding
plane configuration is shown in Figure 6.2.13, attached to a single parasitic
patch element CPW fed ASP. However, the presence of this shielding plane
and the CPW ground plane can promote the generation of parallel plate
modes, reducing the efficiency/gain of the antenna. Parallel plate mode
excitation can also lead to impedance abnormalities, and pattern degradation
due to diffraction at the edges of the parallel plates.
iii ·10
\
~ \ f"'\ I'U
I ,
,~
Gl · 15
"C
:::l I
~ ·20
l:
CI Due to parallel
C'II ·25
== plate mode
·30 excita tion
·35
-40
4 5 e 7 e
Frequency (GHz)
Figure 6.2.14 Measured Return Loss of CPW fed ASP with Shielding Plane
Another concern unique to using a cavity for a CPW fed ASP is that
the lateral metallic bounds of the cavity can form a section of conductor
backing in the CPW feed line. The conductor backed CPW section will then
have a contrasting characteristic impedance to the ordinary CPW line,
resulting in an impedance mismatch.
294
6.2-4.3 Reflector Patch element
In this section, the use of patch elements at the rear of a CPW fed
ASP on a finite size ground plane to reduce the level of back radiation is
experimentally analyzed. A comparison between a standard CPW fed ASP
with and without a reflector patch is presented. The effect of varying some of
the parameters of the reflector patch to tailor the rear hemisphere radiation
pattern is also considered.
Figure 6.2.15 CPW fed ASP with reflectorpatchelement (lp = wp = 16 rom, I. = 11 rom,
w.= 0.2 mm.L« 27.3 rom, wr= 0.2 rom)
295
As in Chapter 3, the reflector element is designed to operate well
above it's first resonance, producing a radiated field approximately 1800 out
of phase with the field radiated in the backward direction from the aperture.
With the appropriate selection of reflector dimensions and substrate thickness,
the magnitude of these radiated fields can be made almost equal. This results
in the cancellation of the total radiated field in the backward direction. In
Chapter 3, an infmite ground plane was used in the analysis, producing almost
ideal cancellation of the backfire fields. In a more realistic scenario, the work
here uses a small, finite ground-plane, hence the cancellation effects are also
influenced by edge diffraction. The reflector dimensions used were quite
different when compared with a frequency scaled version of the reflector used
in Chapter 3, highlighting the impact of the finite ground plane.
Figure 6.2.16 depicts the return loss for the antenna, with and without
the reflector element. A 10 dB return loss bandwidth of 23% is achieved. As
the thickness of the foam reflector substrate is relatively large, the reflector
element has little effect on the input impedance of the antenna. The E and H
plane radiation patterns shown in Figures 6.2.17, 6.2.18 and 6.2.19 highlight
the reduction in field radiation in the rear half plane of the antenna when the
reflector element is added to the CPW fed ASP. The antenna and reflector
dimensions are the same as those given in Figure 6.2.15. Reduction of the
total power radiated into the rear hemisphere of the antenna is evident across
the entire impedance bandwidth, emphasizing the broadband ability of this
technique. The major field cancellation occurs in the region around backfire
(180°), which is particularly noticeable in the E-plane. In the forward
hemisphere of the H-plane the total radiated power has increased, and the
indentations in the pattern are smoothed over. Minor alignment errors were
experienced in the measurement of the radiated fields.
o
"\ ./
,
·5
/
--8
iii'
'C
·10
·15
1
r <, I
=C · 20
::I
~,f I
Cl
III
== ·25
r I
·30
~ - without reflector
• • •• with reflector
II
·35
5 5.5 8 8.5 7 7.5
Frequency (GHz)
Figure 6.2.16 Return Loss of CPW fed ASP with and without Reflector Patch Element
296
(a) (b)
Figure 6.2.17 Measured Far Field Radiation Patterns of CPW fed ASP with and without
Reflector Patch Element at 6.1 GHz: (a) E-plane; (b) H-Plane
(a) (b)
Figure 6.2.18 Measured Far Field Radiation Patterns of CPW fed ASP with and without
Reflector Patch Element at 6.7 GHz: (a) E-plane; (b) H-Plane
o
110 100
(a) (b )
Figure 6.2.19 Measured Far Field Radiation Patterns of CPW fed ASP with and without
Reflector Patch Element at 7.3 GHz: (a) E-plane; (b) H-Plane
297
A 2 nun offset of the reflector element in the positive direction along
the y-axis produces a reduction in the rear directed radiation in the H-plane, as
seen in Figure 6.2.20. This effect is particularly evident in the sidelobes away
from backfire. The E-plane pattern remains relatively unchanged.
1&0 1&0
(a) (b)
- H-plane
••••• E-plane
1&0
(c)
Figure 6.2.20 Measured Radiation Patterns of CPWfed ASP with + 2 mmy-axis Offset:
(a) at 6.1 GHz; (b) at 6.7 GHz; (c) at 7.3 GHz
298
Theoretical trends infer that an increase in the width of the reflector patch
shifts the E-plane dips away from backfire.
o
110 110
(a) (b)
_ H-plane
110 ••••• E-plane
(c)
Figure 6.2.21 Measured Radiation Patterns of CPW fed ASP with 5 rom Wide Reflector
Element: (a) at 6.1 GHz; (b) at 6.7 GHz; (c) at 7.3 GHz
6.2-5.1 Introduction
There has been the occasional dual polarized antenna design reported
that has the desired characteristics for integration with high permittivity
MMIC or OEIC substrates. Hettak et al. have proposed CPW fed aperture
coupled antenna designs on a single layer of high dielectric constant material,
for both simultaneous orthogonal linear polarization [16] and polarization
diversity [17] applications. However, these structures display a very narrow
impedance bandwidth (approximately 2%) that could be a limiting factor to a
communications system design. Stotz et al. demonstrated the use of silicon
nitride membranes to suspend a parasitic patch above microstrip line fed
coupling apertures on a Gallium Arsenide substrate for both single and dual
polarizations [18]. These configurations also had a narrow impedance
bandwidth, and a complex fabrication procedure. None of the aforementioned
designs made reference to the gain or efficiency of the antenna to determine
the influence of surface wave excitation.
The simulated and measured results for the return loss of the SFDP
antenna are given in Figure 6.2.23. The measured results show a slight shift
in the lower frequency resonance. This shift could be caused by tolerances in
the substrate permittivity (to.25 for Duroid 6010.2), or small air gaps
between substrate layers in the fabricated antenna. Otherwise, reasonable
agreement between theory and experiment is achieved. A measured
bandwidth (defined as >10 dB return loss) of 19% was observed.
o
--- .. .... .
............
I\. "
~~
·5
-
in 'Y
~ ~
\ "
·10
~
.3 V
, 'I,'
· 15
c
:; ,
OJ
. L
·20
VI
a: I , ' ~
••••
Simula led
Measured
·25
I
It
·30
i
5 5.5 I 1.5 7 7.5 I
Frequency (GHz)
The co- and cross-polarized radiated fields for the SFDP antenna in
the ep =0° , 90° and 45° planes at 6.5 GHz are represented in Figure 6.2.24a, b
301
and c, respectively. For most angles within ±45° from broadside, the co- and
cross-polarized fields are of approximately equal magnitude for all three
planes depicted. This implies that the antenna is capable of receiving
arbitrary polarization angles at a relatively constant level. Such a
characteristic can alleviate polarization loss concerns at a mobile cellular base
station, where the incoming radiation from the handset may vary in
polarization due to the orientation of the handset with respect to the base
station antenna. Minor alignment errors may have been incurred during
measurement. Also note the relatively low front to back ratio of less than 10
dB, common to resonant aperture antennas (see Chapter 3). The radiation
characteristics presented in Figure 6.2.24 remained fairly constant over the
operational bandwidth of the antenna. The theoretical antenna gain is above 7
dBi across the entire return loss bandwidth, and the antenna exhibits high
efficiency.
'OIl 'OIl
(a) (b)
-90 1-lr+-+-+-?lE-+-f-j-t;~
Co -polarization
Cross-polarization
110
(c)
Figure 6.2.24 Measured Radiation Patterns of SFDPAntenna at 6.5 GHz: (a) cjl = 00
plane; (b) cjl = 90° plane; (c) cjl = 45° plane
·5
iii'
...
~
0
· 10
-J
C ·15
~CD
a:
·20
Figure 6.2.25 Return Loss of stacked SFDP antenna (aperture lengths = Ilmm, aperture
widths = lmm, CPW strip width =O.255mm, CPW gap width =O.l7mm, open circuit stub
length = l.lmm, foam thickness =3mm, patch size = 16.6mm2, additional foam thickness =
=
3mm, additional patch size 14.6mm2)
303
dimensions of the apertures are identical. The magnitude of 5 21 demonstrates
the high level of isolation achieved by arranging the apertures in the 'T'
configuration. The measured 5-parameter results shown in Figure 6.2.27b
confirm the predicted isolation, remaining below -29 dB across the region of
interest. The simulated and measured 5 11 results show excellent agreement in
resonant frequencies and bandwidth. However, the 5 22 results show a reduced
bandwidth due to the influence of a strong parasitic effect at approximately
7.25 GHz caused by the interaction of the aperture and the edges of the small
ground plane used. This parasitic effect was predicted by HF55 at around 7.6
GHz, but it did not impact on the impedance bandwidth performance of the
antenna. A similar effect can also be seen in the measured 5 11 results at 7.9
GHz. It was found both analytically and experimentally that varying the size
of the ground plane and feed substrate can alter the position and severity of
the parasitic effect. The unaltered configuration exhibited a measured
bandwidth of 14% common to both ports of the DFDP antenna.
PorI 2
Port I
Figure 6.2.26
-
Dual Polarization Dual Feed Configuration
~ /' ~
...
. , .......... ,..
V
•
\
~
\ .
"
, - ".'
.",
\
..
I
~' .
' .
a ~5 I
Frequency (GHz)
(a)
~5 7 7.S
'8]"•••••••• Su
_ . - S21
iii'
~ -10
~
~I r<-. "
V· ,, .
\.,....-: ~
,.
.§ ...
~
... ...
IV
" " .
~
~ ... .
.' ~,\., ; .v,"
E .IIj, ..
:""
:J
;! -so
... I ~5 U J U
Frequency (GHz)
(b)
Figure 6.2.27 Simulated and Measured S-parameters of DFDP Antenna
304
The co- and cross-polarized radiated fields for port 1 of the DFDP
antenna in the H- and E-planes are represented in Figure 6.2.28a and b,
respectively. Port 2 results are given in Figure 6.2.28c and d. Cross-
polarization levels at broadside are more than 20 dB below the co-polarized
levels, giving a high level of polarization isolation . This is not only attractive
for polarization diversity applications, but also the generation of circular
polarization could be easily achieved by attaching a 900 hybrid to the two
input ports. Away from broadside the cross-polarization level increased to
approximately 13 dB down from the co-polarized level, implying that the
antenna is more suited to point-to-point links in its current state. As
mentioned previously, minor alignment errors may have been present in the
measurement of the radiation patterns . Relatively low front to back ratios are
again observed as a result of the resonant apertures. The cross-polarization at
broadside remains very low over the bandwidth common to both ports of the
DFDP antenna . The theoretical antenna gain is in excess of 7 dBi across
almost the entire common return loss bandwidth with high efficiency. The
distorted patterns observed at port 2 of the DFDP antenna at 7 GHz are a
consequence of the parasitic impedance effects seen in Figure 6.2.27, caused
by the aperture/ground plane interaction. The marginally higher cross-
polarization level at broadside for port 1 of the DFDP antenna at 7 GHz is of
little concern, as this frequency point is outside the return loss bandwidth
common to both ports of the antenna.
o o
(a) (b)
305
o o
180 180
(c) (d)
Figure 6.2.28 Measured Radiation Patterns ofDFD Antenna at 6.5 GHz:(a) H-plane (Port
1); (b) E-plane (port 1); (c) H-plane (port 2);(d) E-plane (port 2)
The low front to back ratio of the SFDP and DFDP antennas could
pose a problem in applications that use sectorial coverage, or where other
circuitry or support structures are mounted behind the antenna. Microstrip
reflector elements can be attached to the rear of antennas that use resonant
aperture coupling to reduce the level of power radiated into the rear
hemisphere (refer to Chapter 3). This method of back radiation reduction does
not promote the propagation of parallel plate modes or having any major
manufacturing problems, as can be the case with other methods. In this
section, the concepts for cancellation of rear directed radiation outlined before
are applied to dual polarized antennas.
SFDP ReflectorElement
Figures 6.2.29a, b, c and d compare the measured far field patterns for
the stacked SFDP antenna with and without the crossed reflector element.
The power radiated into the rear hemisphere is significantly reduced with the
addition of the reflector, demonstrating a similar order of improvement to that
achieved in [22]. The results shown are from the middle of the return loss
bandwidth, at 6.7 GHz. Similar improvement is observed across the entire
33% bandwidth, exemplifying the broadband ability of the technique.
o o
180 180
180 180
(c) (d)
Figure 6.29 Comparison of Radiated Fields of SFDP Antenna with and without
Reflector Element at 6.7 GHz: (a) $ = 00 plane (co-pol); (b) $ = 00 plane (cross-poL); (c) $
=900 plane (co-pol.): (d) $ =900 plane (cross-pol.)
307
DFDPReflectorElement
As the DFDP antenna has apertures that are separated from each
other, individual reflector elements of length 27.3 mm and width 0.2 mm are
placed directly beneath, but perpendicular to the apertures, as shown in Figure
6.2.30. The foam thickness is 5 mm, and the two reflector elements reside on
opposite sides of a 0.254 mm thick sheet of RT Duroid 5880. Again, the
reflector element has minimal influence on the input impedance of the
antenna.
l.
r
On reverse side of
~~:~~~~:~fl~~rto.,?
elements
Figure 6.2.30 Reflector Element Configuration for DFDP Antenna
The effect of adding the reflector elements on the measured far field
patterns for the DFDP antenna is apparent in Figure 6.2.31a-d. Once again
the rear directed radiation is suppressed when the reflectors are attached. The
results given are at 6 GHz, which is at the lower end of the common
impedance bandwidth. Field patterns at higher frequencies show a reduction
in the level of improvement, when compared to the DFDP antenna without the
reflectors. The decrease in rear field cancellation is a consequence of the
parasitic impedance effects seen in Figure 6.2.31b, which causes the field
radiated by the apertures to change due to their interaction with the ground
plane. Hence the reflector element no longer satisfies the radiated field
criteria of similar magnitude to the apertures with approximately 1800 phase
shift. Further investigation must be undertaken to optimize performance and
the reflector elements for the DFDP antenna.
o
110 110
(a) (b)
308
- Withoutreflector
••••• With reflector
180
(c) (d)
Figure 6.29 Comparison of Radiated Fields of DFDP Antenna with and without
Reflector Element at 6.7 GHz: (a) ep =0° plane (co-pol.); (b) ep =0° plane (cross-pol.); (c) ep
= 90° plane (co-pol.); (d) ep = 900 plane (cross-pol.)
6.3-1 Introduction
309
surface thereby reducing its effect on the overall radiation performance of the
antenna.
310
comparable to a stacked patch incorporating low dielectric constant material
combinations (see Chapter 3), inferring that the surface wave efficiency is
relatively high.
Low dielectric
conslant material Patch conductor
Figure 6.3.2 Input Impedance of Rectangular hi-to Stacked Patch Antenna (parameters:
=10.4, d, =1.925 mm, tan c~h =0.001, ea =1.07, d2 =4.5 mm, tan ~ =0.001, L I =9.4 mm,
Erl
WI =8.8 mm, ~ =21.4 mm, W2 =20.4 mm, xp =4.3 mm, ro =0.325 mm)
311
configuration consists of a stacked rectangular patch combination and the
second is for a stacked annular ring scenario. As for the circular patch case
presented in [7], Tlsw is greater than 85 % over the 10 dB return loss band of
frequencies (ring: 0.895 r, to 1.105 fc; rectangular: 0.885 fc to 1.115 fc) . Note
the peak in efficiency for the rectangular case is approximately 93 % and is
near the upper edge of the 10 dB return loss bandwidth. After this peak Tlsw
gradually decreases until the two patches no longer strongly couple resulting
in the rapid decay of Tlsw. Eventually Tlsw approaches values for the case of a
single layered geometry etched on the high dielectric constant material (refer
to Chapter 2). For the annular ring case note the maximum efficiency is
slightly higher than the rectangular patches at the expense of reduced
bandwidth where the efficiency is high (above 85 %). Indeed the impedance
bandwidth of the annular ring configuration is slightly lower too,
approximately 21 % compared to 23 % for the rectangular hi-lo case. Thus
there appears to be a trade-off between maximum surface efficiency and both
return loss and Tlsw bandwidth, as for the case for a conventional stacked patch
configuration. Once again to put these results into perspective, a patch etched
on the lower dielectric material alone has a surface wave efficiency of 66 %.
It is interesting to note that for all the cases considered; including a circular
patch hi-lo structure, the maximum peak in efficiency does not match the
lowest return loss value. This is similar to the cases presented in Chapter 3,
although there is a better correlation here.
tS Ic!- _
-
ltS .-.- v ........
10
......
f:
'\.\
\.
!l70 ~
312
configuration utilizing a hi-lo combination of materials (Er! = 10.4 and Erz =
1.07). All combinations have been designed for the same center frequency
and the thicknesses of lower dielectric layers are also the same (0.03 Ao). For
the stacked cases the upper dielectric layer has a thickness of 0.06 Ao. From
Figure 6.3.4, there are several important findings. Before exploring these, it is
important to note that the main source of cross-polarized fields is the
discontinuity associated with the feed (here a probe) and the driven patch
conductor. This is why direct contact fed patches have greater cross-
polarization levels than non-contact printed antennas (such as aperture
coupled patches).
single layer (10) __--.-__
.00
The first trend to observe from Figure 6.3.4 is that the cross-polarized
field level for the 10-10 stacked patch combination is less than for the single
layered patch utilizing a low dielectric constant substrate (here by 5 dB). This
observation can be attributed to the presence of the parasitically coupled
patch. As there are no discontinuities associated with this second patch and
the centers of each patch are aligned, as well as the second element being
reasonably far from the discontinuity associated with the driven patch and the
probe, it can be expected that the cross-polarized radiated power will be
lower.
From Figure 6.3.4 it can be seen that the single layer patch using a
high dielectric constant substrate has significantly less cross-polarization than
the patch with e, =2.2. This is a result of the cross-polar fields being strongly
bound to the material [27]. The hi-lo stacked patch has slightly higher cross-
polarization levels (5 dB) than the single layer scenario due to the
parasitically coupled element drawing power from the lower element. Thus
there is one minor trade-off of the new configuration, the hi-lo patch has far
313
superior llsw and impedance bandwidths compared to the single layer
geometry at the expense of slightly higher cross-polarization levels. It should
be noted that the cross-polarized fields of the hi-lo stacked patch is
significantly lower (15 dB) than the conventional stacked configuration. This
makes the hi-lo structure a prime candidate for circular polarization
generation, where the cross-polarization level governs the quality of the axial
ratio (AR).
High
permittivity
substrate
Foam
die lect ric Gro und plane
. _.....
re
'. I
~ ....
...........,1
~
. "
'
. ., . ., , :
~
,,
.,
,
:
(a) (b)
Figure 6.3.6 Schematic Diagram and Axial Ratio for Single Feed hi-lo Stacked Patch
= = =
Configuration (parameters: £,.\ 10.4, d\ 1.524 mm, tan 0\ 0.001, Cr2 1.07, d2 5.0 mm, = =
= = = = =
tan ~ 0.001, £"3 2.2, d\ 0.254 mm, tan 03 0.001, L. 9.0 mm, W\ 7.8 mm, ~ 20.7 = =
mm, W2 = 15.5 mm, Wf= 0.25 mm)
315
As mentioned in Chapter 2, probably the most common means of
obtaining CP from a micros trip patch antenna is to incorporate two orthogonal
feeds excited 90 0 out of phase. Once again this method has had limited
success when using conventional dielectric structures due to the spurious
fields radiated from the feed network, in particular when direct contact feeds
are used. Figure 6.3.7a shows a schematic representation of an edge-fed hi-lo
stacked patch version of this CP microstrip antenna. Here the patch is fed by
a Wilkinson power divider (refer to [1] for design details). To achieve the 900
0
phase shift between the two orthogonal feeds a 90 delay line was
incorporated into the microstrip feed network. To model this configuration,
Ensemble 5.1 was utilized, although the Wilkinson power divider was not
modeled in the simulation. Figure 6.3.7b shows the predicted and measured
axial ratio of this configuration. As can be seen from Figure 6.3.7b the
agreement between experiment and prediction is good, with the measured 3
dB axial ratio bandwidth is in excess of 32 %. The discrepancies between the
results can be attributed to simulation assuming an equal magnitude excitation
at both ports (l and 2 in Figure 6.3.7a) and a 90 0 phase shift between these
ports across the frequency band. The 10 dB return loss bandwidth was of
similar order to the AR bandwidth as a consequence of the resistor (100 Q) in
the power divider.
..S
I Q - ... r
exp.,lrnenl
~ ~
~ /
\.~/ <, /v
.
1.5
-."
I
• U 5 s.s
FNqUllncy (GHz)
Figure 6.3.7 Schematic Diagramand Axial Ratio for DualFeed hi-lo StackedPatch
Configuration (parameters: £"1 = 10.4, dl = 1.925 mm,tan SI = 0.001, ea = 1.07, dz =4.5 mm,
tan Oz = 0.001, £"3 =2.2, dl =0.254 rom, tan ~ =0.001, L1 =9.4 mm, WI = 9.4 mm, Lz =21.4
mm, Wz = 21.4 mm, Wf= 1.766 rom)
316
6.3-4 Bandwidth Enhancement
' - /'oam 2
..-----,~£..----Pal ch conductor (I". w,,)
. . - - Foam I
......~'---- Patch conductor (1,2'W,2)
Figure 6.3.8 Geometry ofhi-Io-Io Stacked Patch Antenna (dimensions: Ipl =9.9mm, Wpl
= 6.4mm, Ip2= 20.4mm, Wp2 = 24.7mm, Ip3= 19.4mm, w p3 = 11.2mm)
317
During the design of the hi-lo-Io printed antenna, it was found that the
dimension variations (such as patch length or substrate thickness) had similar
effects on the impedance locus as the parameter study given in Chapter 3 for
microstrip fed ASP's. This is considering that alteration in aperture length in
[9] equates to modifying the patch element size on the feed substrate. The hi-
10-10 patch was found to be slightly more sensitive to changes in dimensions,
especially if the parameter being adjusted is physically near the high dielectric
feed substrate.
(~ ~)
Figure 6.3.9 InputImpedance of Hi-Io-Io Stacked PatchAntenna: (a) Predicted; (b)
Measured
-90
ISO
6.3-5. I Introduction
319
43 o o
(6.1)
o 28 o
o o 43
43 o o
(6.2)
o 43 o
o o 28
There have been very few reported explorations into printed antennas
on LiNb03, as the anisotropic nature and high permittivity of the material
make antenna design extremely complex. The single layered approach in
Chapter 2 suffers the predictable narrow impedance bandwidth and surface
wave effects due to the high permittivity substrate. In this section two hi-lo
stacked patch antenna elements that employ LiNb03 as the feed substrate, for
both z-cut and x-cut crystal orientations are presented. An experimental
investigation into the performance of these LiNb03 hi-lo printed antennas is
given, including the impedance and radiation characteristics.
320
_ Foa m dielect ric (hI)
_---,HO-.-- Parasi tic patch O.,.w,,)
Figure 6.3.12a shows the measured return loss behavior of the z-cut
LiNb03 hi-fa stacked patch configuration of Figure 6.3.11, incorporating
rectangular patch elements. The measured bandwidth, defined as being
greater than 10 dB return loss, is 12.4%. This bandwidth is sufficient for most
wireless systems. A conventional single patch element etched on a similar
high permittivity material typically has a useable bandwidth of less than one
percent. If required, stacking an additional parasitic patch mounted on its
own foam substrate can further increase the useable bandwidth, resulting in
--
the hi-lo-lo configuration, as was shown in the previous section.
iii •
-.....
_'0 .'0 '\ /
r-
8 · 11
\ I
I
oJ
E ... \
v
:>
Ci -os
II:
I
.,.
II 'U I. IU II tU It t~1
Frequency (GHz)
(a) (b)
Figure 6.3.12 (a) Return loss of z-cutLiNb03 hi-lo Stacked Patch Antenna (~ =
O.254mm, hf =2mm,hi =O.5mm, lpl = 1.6mm, wpl =3.2mm, Ip2 =7.5mm, W p2 =O.2mm (b)
Radiation patterns for the z-cut LiNb03 hi-lo stacked patch antenna
_ _ _ H-plane
• •••••• E-plane
321
Figure 6.3.12b shows the measured E- and H-plane co-polar radiation
patterns the z-eut LiNb03 hi-to stacked patch antenna. The LiNb03 wafer was
cropped to 20mm x 20mm (approximately one wavelength square), with the
brass mounting block (which helps form the ground plane) extending only a
few millimeters beyond this. The slight ripple in the patterns can be attributed
to diffraction from the edges of the small ground plane and substrate, and
experimental error in the measurement set-up. The FIB ratio is approximately
13 dB. A wide 3 dB beamwidth in both principle planes is observed, which is
a common characteristic of patch antenna elements. This coverage area is
compatible with a variety of picocellular base station applications. The cross-
polarization level in each plane of the z-eut LiNb03 hi-lo stacked patch was
more than 20 dB below the co-polarization level at broadside. The gain of the
z-eut LiNb03 hi-to antenna was around 7.5 dBi. This value is comparable to
a traditional stacked patch structure that utilizes only low dielectric constant
substrate layers, inferring that the surface wave excitation is very low.
Figure 6.3.13a displays the measured return loss characteristic for the
x-eut version of the LiNb03 hi-to antenna. A measured bandwidth of 9.2%
was observed, which is slightly lower than the x-eut antenna, but still
adequate for most wireless communication systems. The far field radiation
patterns for the x-eut LiNb03 hi-to antenna are given in Figure 6.3.13b. As
the LiNb03 wafer and the brass mounting block were approximately the same
dimensions as in the z-eut case, the slight ripple in the patterns is still evident.
The FIB ratio was measured to be approximately 26 dB, and a wide 3 dB
beamwidth is again observed. The cross-polarization level in each plane of
the z-eut LiNb03 hi-lo stacked patch was more than 30 dB below the co-
--
polarization level at broadside. The gain of the z-eut LiNb03 hi-to antenna
was approximately 7 dBi.
-----
0
~
III
. r-, I
!!..
"'\
· tI
:l I
, '\.,
·11
o
..J ... I
.aE ... /
\
~
...... \I
... ,u
1
,u
1'-1
" " "
Frequency (GHz)
Figure 6.3.13 (a) Return Loss of x-cut UNb03 hi-lo Stacked Patch Antenna <hd =
O.127mm, hr =2mm, hI =O.5mm, Ipl = = = =
1.6mm, Wpl 3.2mm, Ip2 8mm, Wp2 5mm) (b)
Radiation patterns for the x-cut UNb03 hi-lo stacked patch antenna
_ _ _ H-plane • •••••• E-plane
322
6.4 Summary
Section 6.2-5 introduced two variants of the CPW fed ASP antenna,
which produced broadband dual polarized structures suitable for microwave
and optical device integration. The SFDP antenna utilized a single diagonal
CPW feed line and a crossed aperture to create a simultaneous dual linear
323
polarized antenna element with an impedance bandwidth of up to 33%. The
DFDP antenna displayed independent control of orthogonal linear
polarizations by using separated apertures fed by their own individual CPW
line. Cross-polar isolation in excess of 20dB at broadside was achieved over
a bandwidth common to both feed ports of 14%. Due to the high level of
back radiation that is inherent to antennas with resonant coupling apertures,
dual polarized reflector patch elements were attached to the rear of the dual
polarized antennas. Both the SFDP and DFDP exhibited a decrease in the
total field radiated into the rear hemisphere when the reflector elements were
attached. As is the case with other methods of back radiation reduction, the
reflector elements do not promote parallel plate modes, and there are no major
machining/fabrication concerns.
6.5 Bibliography
[I] D. M. Pozar, Microwave Engineering - Second Edition, John Wiley and Sons Inc.,
New York, 1998.
[2] G. H. Smith, R. B. Waterhouse, A. Nirmalathas, D. Novak, C. Lim, and O. Sevimli,
"A broadband integrated photonic-antenna interface for multiservice millimeter-wave
324
fiber-wireless applications," MWP 2001, Long Beach California, pp. 173 - 176,
October 200 1.
[3] R. B. Waterhouse, "Stacked patches using high and low dielectric constant material
combination," IEEE Transactions Antennas & Propagation, vol. 47, pp. 1767 -
1771, December 1999.
[4] Z. Zhang and C. P. Wong, "Assembly of Lead Free Bumped Flip-Chip with No-Flow
Underfills," IEEE Trans. Electronics Packaing Manufacturing, Vol. 25, pp. 113 -
119, April 2002.
[5) L. Harle and L. P. Katehi, A Vertically Integrated Micromachined Filter," IEEE
Trans. Microwaves Theory Techniques, Vol. 50, pp. 263 - 2068, September 2002.
(6) Y. Furuhama, "Research and developments of millimeter-wave technologies for
advanced communications" , 3Td RIEC Symp. Novel Techns & Appls. of Millimeter-
Waves, Sendai, Japan, pp. 1- 6, December 1998.
[7] Y. Qian and T. Itoh, "Progress in Active Integrated Antennas and Their
Applications", IEEE Trans. Microwave Theory and Techniques, 1998, vol. 46, no.
11, pp. 1891-1900.
[8) High Frequency Structure Simulator- Version 5.5, Agilent Technologies, 2000.
[9] S. D. Targonski, R. B. Waterhouse and D. M. Pozar, "Design of Wide-Band
Aperture-Stacked Patch Microstrip Antennas", IEEE Trans. Antennas and
Propagation, 1998, vol. 46, no. 9, pp. 1245-1251.
(10) Ensemble 6.0, Ansoft, 1999.
(11) K. C. Gupta, R. Garg and I. J. Bahl, Microstrip Lines and Slotlines, Artech House,
1979.
[12] M. Riaziat, R. Majidi-ahy and I-J. Feng, "Propagation modes and dispersion
characteristics of coplanar waveguides" , IEEE Trans. Microwave Theory and
Techniques, March 1990, vol. 38, pp. 245-251.
[13] W. Menzel and W. Grabherr, "A microstrip patch antenna with coplanar feed line",
IEEE Microwave and Guided Wave Letters, Nov. 1991, vol. I, pp. 340-342.
[14] L. Giauffret, 1. M. Laheurte and A. Papiemik, "Experimental and theoretical
investigations of a new compact large bandwidth aperture-coupled microstrip
antenna", Electronics Letters, 1995, vol. 31, no. 25, pp, 2139-2140 .
[15] L. Giauffret and 1. M. Laheurte, "Theoretical and experimental characterisation of
CPW-fed microstrip antennas", lEE Proc. Microw. Antennas Propag., 1996, vol.
143. no. l.pp. 13-17.
[16] K. Hettak and G. Delisle, "A novel antenna configuration for millimetre wave
communication systems," IEEE Antennas Propag. Symp. Dig., 1998, pp, 2092 -
2095.
[17] K. Hettak, G. Y. Delisle and M. G. Stubbs, "A novel variant of dual polarized CPW
fed patch antenna for broadband wireless communications", IEEE Antennas Propag.
Symp. Dig., Utah USA, July 2000, pp, 286 - 289.
[18] M. Stotz, G. Gottwald, H. Haspeklo and J. Wenger, "Planar single- and dual-
polarized aperture coupled E-band antennas on GaAs using SiN.-membranes", IEEE
AntennasPropag. Symp. Dig., 1996, pp. 1540 - 1543.
[19] K. Siwiak, Radiowave Propagation and Antennas for Personal Communications,
Artech House, Boston, 1998.
[20] G. Rosol, "Environmental factors contribute to antenna selection", Microwaves &
RF, August 1995, pp. 117 -123.
[21) B. G. Porter, L. L. Rauth, 1. R. Mura, and S. S. Gearhart, "Dual-Polarized Slot-
Coupled Patch Antennas on Duroid with Teflon Lenses for 76.5-GHz Automotive
Radar Systems", IEEE Trans. Antennas and Propagation, 1999, vol. 47, no. 12, pp
1836-1842.
325
[22] S. D. Targonski and R. B. Waterhouse, 'Reflector elements for aperture and aperture
coupled microstrip antennas', IEEE Antennas Propag. Symp. Dig., 1997, vol. 3, pp.
1840-1843.
[23] D. M. Pozar, "Analysis of an infinite phased array of aperture coupled microstrip
patches," IEEE Trans. Antenna Propagat., vol. 37, pp. 418 -424, April 1989.
[24] D. R. Jackson,1. T. Williams, A. K. Bhattacharyya, R. L. Smith, S. 1. Buchheitt and
S. A. Long, "Microstrip Patch Designs That Do Not Excite Surface Waves," IEEE
Trans. AntennasPropagat., vol. AP-41,pp. 1026-1037, August 1993.
[25] D. M. Kokotoff, R. B. Waterhouse, C. R. Birtcher and J. T. Aberle, "Annular ring
coupled circular patch with enhanced performance," Electronics Letters, vol. 33, pp.
2000-2001, Nov. 1997.
[26] D. M. Kokotoff, R. B. Waterhouse and 1. T. Aberle,"On the use of attachment modes
in the analysis of printed antennas," Progress in Electromagnetics Research, Nice
France, July 1998.
[27] R. B. Waterhouse, "Improving the scan performance of probe-fed microstrip patch
arrays," IEEE Transactions Antennas & Propagation, vol. 43, pp. 705 - 712, July
1995.
326
Chapter 7 Microstrip Patch Arrays
7.1 Introduction
7.2-1 Introduction
One of the first realizations of a microstrip patch array was the series
fed array (see for example, [1]). Here each element of the array is connected
in series via an arrangement of transmission lines. Figure 7.2.1 shows a
schematic diagram of an 8-element series fed array consisting of edge-fed
patches. The array is fed from the left and is classified as a standing wave
array. Series fed arrays have been developed in waveguide realizations for
decades, however microstrip forms have much more flexibility. This is
mainly due to the fact that it is easy to change the impedance of the microstrip
feed lines between the radiating elements to give the desired amplitude taper.
The patch width can also be varied to give this same effect.
i
L=lOmm L=IOmm L=IO.55mm L=IO.45mm L=IO.6mm
W=1.5mm W=6.4mm W=IO.4mm W=IO.4mm W=6.4mm
The return loss response and the E-plane radiation performance of the
array across the 10 dB return loss bandwidth of 1 % is shown in Figures
7.2.3a and b, respectively. As can be seen from Figure 3b the radiation
pattern remains generally constant across the matched impedance bandwidth,
unlike some of the early cases of series fed microstrip patch arrays. There is a
slight asymmetry in the radiation pattern that is due to the presence of the feed
network. The H-plane pattern is similar to a conventional edge-fed microstrip
patch. The cross-polarization level in both principal planes (E- and H-plane)
was less than 40 dB below the co-polar fields. The relatively constant
radiation performance of the antenna can be attributed to using a full-wave
simulator to synthesize the antenna, software tools that were not available in
the early days of microstrip patch technology development or were simply too
329
slow for design purposes. Utilizing such tools as well as enhanced bandwidth
elements (such as stacked patches) should see an improvement in the
performance of series fed arrays, but probably not to the same degree as
corporate (or parallel) fed microstrip arrays.
r-, )---
1\ I
\
· 25
...•~ u U I~ U ' .7 U U 10
Frequency (GHz)
II.
(a) (b)
Figure 7.23 Response of 4 Element Series Fed Array : (a) Return Loss; (b)
Radiation Patterns
Looking through the literature, there does not appear to be any case of
planar series fed arrays. This is intuitive, for as the length of the array
increases, or in the planar case, the area, the elements at the end of the array
are less likely to receive any power from the source. Thus these elements in a
series feed array are redundant. It is perhaps possible to develop small (say 8
element) 2-dimensional arrays of series fed microstrip patch elements,
however, the shortcomings highlighted previously for the series fed linear
array would still hold for this case. There is reported in the literature a case of
several linear series arrays connected in parallel, using the impedance
matching procedure summarized before [2].
7.3-1 Introduction
Parallel or corporate fed microstrip patch arrays are the most common
type of array using microstrip patch technology. Here, unlike the series fed
array each element has its own excitation transmission line, which can be
made independent of the feed lines of the other elements as well as the other
elements of the array. Figure 7.3.1 shows a schematic diagram of an 8-
element corporate feed array of edge-fed microstrip patches. As can be seen
330
from the figure, each element has its own excitation transmission line. Each
of these transmission lines is then connected together via a series of 2-way
power combiners, although 3-way dividers are commonly used if an odd
number of elements are used in the array. The power combiners can either be
reactive, such as shown in Figure 7.3.1, or based on Wilkinson dividers. The
Wilkinson divider gives broader band isolation between the elements at the
expense of increased complexity and also loss. It should be noted that most
microstrip patches have impedance bandwidths smaller than that of a reactive
power divider.
7.3-1.1 Introduction
Technical Specifications
The technical requirements for the linearly polarized low profile array
are summarized in Table 7.3.1.
Characteristic Requirement or Target
Passband(MHz) 850-960
Polarisation Linear, vertical
RF Powerinput (W) 50
Dimensions (mm,not to exceed) 390w x 244h x 15d (Design Goal)
480w x 330h x 25d (Not to be exceeded)
Input Impedance 50 ohm
Beamwidth (mainlobe) As narrowas possible
Beam width (side lobe level) 20 dB below main lobe at +/- 30° (goal)
Beamheight (mainlobe) narrow
Beamheight (side lobe level) secondarvimpOrtance
Physical MountingArrangements a. Perimetermounting points are to be within the
overalldimensions given.
b. Antennawill be attachedto the face of a metal
frame.
c. Othermetallicstructures will be mountedbehindthe
antennawith no more than 10mmclearance.
AntennaConfiguration
332
Upper Patch
.>
~ Matching Structure
Probe Feed - - - - -
The return loss of the stacked array is shown in Figure 7.3.3. As can
be seen from this figure the required 10 dB return loss bandwidth (850 - 960
MHz) has been achieved. In fact the 10 dB return loss bandwidth is from 844
- 966 MHz (13 %), giving the antenna some extra degree of freedom .
333
CH 3 - S11
REF . PLANE
5.087... ell
811 F~ REFLECTION
~ NARK£R 1:
LOO MAG . 10 .000.lll/DIU 0 .832000 GHz
- 8 . 766 d8
: i
. _·· ··T··.. ··· ~ ! . . . tIRRI<I:R TO I1flX
MARK£R TO nIN
·····..···i'·_······;··········r········f·········r 2 0.873600 GHz
I -10.2-11 J8
,,.
0 .700000 GHz s. .socccc
,.
Figure 7.3.4 Measured E-plane Co-polarRadiation Pattern
'10
Figure 7.3.5 Measured H-planeCo-polarRadiation Pattern
334
7.3-1.3 Narrowband circularly polarized patch array
Technical Specifications
Antenna Configuration
335
To achieve the required bandwidth, a combination of microwave
laminates was used. For the lower grounded substrate, 3.175 mm Rogers RT
Duroid 5880 (Er = 2.2) was incorporated. For the parasitically coupled
patches, a 6 mm layer of Rohacell form (e, = 1.07) was used to separate these
patches from the driven elements. A thin layer of TLY Taconic (e, = 3.0) was
utilized to etch the parasitically coupled patches. The top dielectric layer also
protects the top patch conductors from the environment. Ensemble 6.1 was
used to design the stacked patch array .
•• op Patches
The return loss of the stacked patch array is shown in Figure 7.3.7.
As can be seen from this figure the required 10 dB return loss bandwidth
(1750 ± 50 MHz) has been achieved. In fact the 10 dB return loss bandwidth
is from 1650 - 1900 MHz (14 %), giving the antenna some extra degree of
freedom.
The axial ratio of the antenna was measured across 1650 - 1900 MHz
using a rotating, linearly polarized source antenna. The axial ratio across this
band was less than 3 dB as required. The measured H-plane far-field axial
336
ratio patterns at 1750 MHz is shown in Figure 7.3.8. The 3 dB beamwidth in
the H-plane is 45°. The 3 dB beamwidth in the E-plane, not shown here, is
73°. As can be seen from Figure 7.3.8, the sidelobe levels are very low. The
measured 3 dB axial angle in the Hsolane is 65° whereas in the E-plane it is
82°.
l~:'::~::;~/~l~
mP'SIlt t !::H..r., .I.: :
.C .I . .-r u
1.~lJz
_lAW'
.ISPm nu
.... IIfT.
nelli_
n_
T~t~~I~•
I.n,fat 1M, eM
- .' .. ...,.,
• ~~S(~>
Technical Specifications
337
Characteristic Requirement or Target
Antenna Configuration
(a)
339
(b)
(c)
Figure 7.3.10 Schematic diagram of Triple Stacked Patch Elements : (a) Driven patches;
(b) First parasitic patches; (c) Second parasitic patches
340
To achieve the required bandwidth, a combination of microwave
laminates was used. For the feed network, 1.5 mm Rogers RT Duroid 5880
(e, =2.2) was incorporated. Once again to isolate the feed from the mounting
structure, a 3 mm layer of Rohacell form (e, = 1.07) and a thin layer of TLY
Taconic (Er = 3.0) were used. Using a rigorous full-wave analysis to design
the stacked patch configuration the following material selection was required:
3 mm of Rogers RT Duroid 5880 for the lower substrate; 3 mm of Rohacell
foam for the middle dielectric layer; and 6 mm of Rohacell foam for the upper
dielectric layer. Thin layers of TLY Taconic were used to fabricate the middle
and top patches. The top dielectric layer also protects the top patch
conductors from the environment.
The return loss of the stacked array is shown in Figure 7.3.11. As can
be seen from this figure the required 10 dB return loss bandwidth (850 - 960
MHz) has been achieved. In fact the 10 dB return loss bandwidth is from 829
-1012 MHz (20 %), giving the antenna some extra degree of freedom.
CH3 -S11
REF . PlfViE
S11 FORlJARO REFLECTION 0 .0000 PI~
H "lRKER 1
o. 6 19000 GHz
................;.....'1" .....:
The axial ratio of the antenna was measured across the band of
frequencies using a rotating, linearly polarized source antenna. The axial ratio
of the antenna was less than 3 dB as required. The measured E- and H-plane
far-field axial ratio patterns at 900 MHz are shown in Figures 7.3.12 and
7.3.13, respectively. The 3 dB beamwidth in the H-plane is 45°. The 3 dB
beamwidth in the E-plane is 70°. As can be seen from these patterns, the
sidelobe levels in each plane are very low. The measured 3 dB axial angle in
the H-plane is 65° whereas in the E-plane it is 80°.
341
o
180
180
Technical Specifications
The technical requirements for the linearly polarized low profile array
are summarized in Table 7.3.4.
342
Characteristic Requirement or Target
Antenna Configuration
l - - - -_ _ l __ ----.J
(a)
(b)
344
Taconic were used to fabricate the middle and top patches. The top dielectric
layer also protects the top patch conductors from the environment.
• (b)
•
• (c)
•
Figure 7..3.l5 Schematic Diagram of Aperture TripleStacked Patch Elements: (a) lower
patches; (b) middlepatches; (c) top patches
345
Impedance Behavior and Radiation Characteristics
The return loss of the aperture stacked patch array is shown in Figure
7.3.16. As can be seen from this figure the required IO dB return loss
bandwidth for the three bands, namely, 1525 - 1560 MHz, 1805 - 1990 MHz
and 2480 - 2500 MHz has been easily achieved. This makes the antenna less
susceptible to fabrication errors.
ii CH 3 - 911
REF . f'l AI'!E
Sii f"ORUffiO R£Fl.ECTI OI1 . I O.OOOCr ""
1 .8OSf:tl<1 (ii1z
-1Q·~5 d!l
1 .~W4l)u 0Hz
-1~ .SQfJ dB
2 .18QOO() GI1z
-Hi·OOl} o:ltt
2 .601600 0Hz
1.41)0000 GHz a.ooocoo : I - :17 . (l:~O dB
I
346
II I ' U) rn.,
U' ,In.•• •..n l' u rul'
11m
'.'4''''GIt ,
II G1
o;• .I,-J41U'lIJ!r
SU CUlIU' /I'."
un U TI.t UIIJU )
. .....UIlIlJ
SIIm'l
III '1 .s, 1001 0.
I .H.... GII..
M IlIi[l a ll~r'
OIIC11I1 IILL
NOlO . unu
rulUIDI
lurun u.
0111,...
.r"l~'
• IU[U ."...,
P ~I .1 !)WIll"
I."'U.. ,"I til
SZ I UDl un
ltI 'II U Il! ..
t l' l:r--$1."" • .1.- ..,." SWff' IU"
If""
I . tu&&t A.I:
II"
.. 1"'1.......
III U.,U~..
te .• . Me ..
1. ' UItt AI
MIld. . ."
.11CI(f[ f lU
. . . .tr..
rllltTlli
IU},.I....
~r'::Wt
. :rtu .,11'
t."..... lb01
t .• • U D[
""'1[1 SIIIU'
onunE nu
1I0Ultlno.
ru.UIDII
.~Yt,"'~. LI
'~I <li lt ..
u V,V
6a'.tL,t.C,'crr
347
.. -
"' l"'~
.,.t.
lOW"""
...
1...... -
., .W- It "
...."1'-- -
.I ttt''''''' .. I U ....' _
I,,''''' "I''UI
,:i::'J'- 'r:.::: ::"'11
"1"1 'l~
.... 1.
ncn ..
1'\\,"',. '
z:r:r:,:;,
-"-="-----------'==
(a) (b)
,. ,. .
Itl , . ,.... ..
"
· - It.ln •
..","'"
. ...........
..\1..... •..
."MI
I"' ""'''III1U
......n ...
...........
UtliU li
I . .. .... II
m uUl m.L
.' 1111"
' III(U"
111-'.1 1......
::«:~JE:.:
-=~ -==o:.J !"••
(c)
Figure 7.3.20 H-planeRadiation Patterns: (a) 1.54GHz; (b) 1.9GHz; (c) 2.45 GHz
348
Patches
f..----
\(1\ 1/
<,
\v/
V
\/ ,
)J )4
F_y(QHI)
:N :N .
'10
(a) (b)
Figure 7.3.22 Performance of Aperture Coupled Patch Corporate Array (a) Return Loss;
(b) H-plane Co-polar Radiation Pattern
Figure 7.3.23 Photograph ofH-plane Aperture Coupled Array and Feed Network
349
7.3-3.2 Eight element hi-Io patch array
Due to its low surface wave excitation and wide bandwidth, the hi-lo
stacked patch is an ideal candidate for use in antenna arrays. This sub-section
presents an 8-element corporate fed array of hi-lo stacked patch components,
suitable for integration applications that require fan-beam coverage.
Measured impedance and radiation pattern results are given.
The array consists of the driven patch elements (1.6 x 1.25 mm) and
corporate feed network, seen in Figure 7.3.24a, formed on a communal 0.254
mm thick RT Duroid 6010.2 substrate (e, = 10.2). As with the hi-lo stacked
patch elements of Section 5.3, the high permittivity feed material is used to
emulate the materials used in MMIC/OEIC design. A 0.5 mm thick layer of
foam dielectric (s, = 1.07) resides over the feed substrate, with a 0.127 mm RT
Duroid 5880 (e, =2.2) laminate supporting the parasitic patch elements (3.2 x
4.4 mm) of Figure 7.3.24b stacked on top. The thin RT Duroid 5880 layer
also acts as a radome, shielding the metallic patch elements from the
environment. The array elements are spaced by 6.6 mm (approximately 0.62
A at the center frequency). The overall ground plane and substrate size of the
antenna arra is 23 x 60 mID.
--------
(a)
(b)
Figure 7.3.24 Layoutof Hi-IoPatch Array: (a) Feed Network and LowerPatches; (b)
ParasiticPatches
350
Figure 7.3.25 represents the measured input impedance of the hi-lo
stacked patch array. The results show a 10 dB return loss bandwidth (shown
between marker 1 and 2) of 15%, centered at approximately 28 GHz. This
bandwidth is lower than that achieved for the single hi-lo stacked patch
element in Section 5.3, due to the limited selection of commercially available
substrate thicknesses. At frequencies outside the return loss bandwidth,
considerable surface wave excitation in the high permittivity feed material is
evident. As shown in Section 5.3, the surface wave efficiency decreases
significantly when the patch elements of the hi-lo stacked patch antenna are
not strongly coupled. Hence, a residual surface wave loss is observed.
r-"--r"-'T---r--"-r--'r--l . ·-1---.--···:1
t : ;
.L~·~:.·:·.:I: .::~..:.~ :::.'::::\
I::'.: : ·~::: . ·: : !· ·:· ..: : · 1 ·: ·: · ~::1 : : :·:.:.1::..
i : ! ,
iii'
~
"~" 0
;;'" ·10
·20
·30
L- ~ _ _....._.l....•••_..t......_.I....... _~ ._.~_ _._ I--..•_ •._.__ J
22 24 26 28 30 32
Frequency (GHz)
The measured far field co-polar patterns of the hi-lo stacked patch
array are portrayed in Figure 7.3.26. The apparent pattern rippling may be
caused by the small ground plane, or experimental error in the measurement
set-up. The 3 dB H-plane beamwidth is approximately 10 degrees. Although
not shown, the back radiation level was low, with a front to back ratio of 21
dB. Cross-polar levels were greater than 18 dB below the co-polar level at
broadside in the H-plane, and in excess of 17 dB down in the E-plane.
•
- H-plane
••••• E·p lane
.
,
Figure 7.3.26 Radiation Patterns of Hi-lo Patch rom-wave Array
351
7.3-3.3 Eight element ASP array
Spacer
352
The input impedance response and the radiation patterns of the array
were then measured. The return loss of the array with the reflector elements
was identical to the case with the reflector elements absent and is shown in
Figure 7.3.28. The H-plane patterns of the array with and without the reflector
elements at 28 GHz and 38 GHz are shown in Figure 7.3.29a and b,
respectively. From both these figures it can be seen that the backlobe has
been reduced by a further 15 dB when the reflector elements are utilized.
Similar H-plane radiation patterns were obtained across the Ka-band. A slight
increase in the gain of the array of approximately 2 dB across the entire Ka-
band was also observed, due to the lower level of power radiated in the back
plane. The cross-polarization levels in both E- and H-planes were measured
to be as at least 20 dB below the co-polar values for all frequencies across the
Ka-band.
Figure 7.3.28 Input Impedance Response of Array with and without Reflector Elements
353
of the slot/directive patch edges, then the diffraction phenomenon can alter
the radiation patterns and the design trends shown in Section 3.5 no longer are
valid. For these situations an analysis technique that takes into consideration
the size of the ground-plane should be used to design the reflector patches,
such as the Finite Element Method, as presented in Chapter 6. In that
chapter, for small/truncated ground-planes we have found that the reflector
patch concept is still very effective in reducing the backward directed
radiation when appropriately designed.
'00
...
(a) (b)
Figure 7.3.29 H-plane Co-polar Radiation Patterns of Array at: (a) 28 GHz; (b) 38 GHz
7.3-4.1 Introduction
Figure 7.3.30 shows the unit cell of the 8 element linear array of hi-lo
stacked patches. Here the lower patch is etched on a grounded RT Duriod
6010 (e, = 10.2). A foam spacer resides between the stacked patches and a
thin layer of RT Duriod 5880 (Er = 2.2) is used to etch the top patch, which is
positioned above the foam. The input impedance and radiation performance
of an individual hi-lo stacked patch was simulated using Ensemble 6.1. The
design procedure given in Chapter 5 was followed to ensure good impedance
and efficiency properties were achieved for the hi-lo stacked patch. The
predicted 10 dB return loss bandwidth was approximately 20 %, centered at
an operation frequency of 6.35 GHz (refer to Figure 7.3.30 caption for
dimensions of the stacked patch). The predicted gain of this printed antenna
was 8 - 9 dBi across this frequency band, ensuring the printed antenna was
highly efficient. The cross-polarization levels in both principal planes (E- and
H-plane) were at least 30 dB below the co-polar fields.
355
Bottom patch with
branchlinc couplerfeed
Low .. ..b'trat.
Figure 7.3.30 Schematic Diagram of Unit Cell consisting of Hi-Io Stacked Patch Capable
of Generating CP <dimensions of antenna: layer 1 - Err =10.4, d l =1.27 mm; layer 2 - fa =
1.07, d2 =3 mm; layer 3 - Er3 = 2.2, d3 =0.254 mm; bottom patch: 7.15 mm2 ; top patch : 16.4
mm 2,feed width: 1.2 mm)
• Figure 7.3.31
• •• •
Schematic diagram of the array layout
~ 0.2
ti
cl: ~i
0.1
0
5 5.5 8 8.5 7 7.5 8
Frequency (GHz)
358
The radiation patterns and axial ratio of the scannable printed array
were measured at a variety of frequencies and scan angles. A sample of the
results is presented in Figures 7.3.34, 7.3.35 and 7.3.36. As can be seen from
these figures the array can be readily scanned to ± 45° whilst maintaining an
axial ratio of less than 3 dB. The gain across the scanned range of angles and
frequencies is approximately 3 dB lower than the expected/predicted values
due to the insertion loss of the phasing module discussed previously. There
was no amplitude taper on the array, although despite this, the side-lobe levels
were at least 10 dB below the main beam for all cases examined. This level
could be enhanced by incorporating sidelobe reduction techniques such as
those outlined in [3], although it should be noted that the array elements here
are inherently broader in bandwidth than those used in [3] and therefore the
element excitation amplitude and phase errors should be less problematic
here.
li I' 11 1--
I
'5
11 I: 'J~
'0 I{i" II 1\= II -
I
II , t hl l
I\P 1;1 i :r: ill
.\. ;; 1 1111 : '
f s ); j \ I ,II j} \ -
h.l \..·1·..·'\•..-
5 \ :
I,
1 0
'\ j ( ----
I ..
-
~ I
~-5 -
Figure 7.3.34 Radiation Pattern and Axial Ratio when scanned to 15° at 5.7 GHz
Figure 7.3.35 Radiation Pattern and Axial Ratio when scanned to 30° at 6.3 GHz
359
"
." . ,- . --!
~--L _L _~
..
.to - - -
I I
Figure 7.3.36 Radiation Pattern and Axial Ratio when scanned to 40" at 6.9 GHz
By far the most common type of fixed beam planar microstrip patch
array is based on the corporate feed [6]. These planar arrays are utilized in
applications such as millimeter-wave collision avoidance radar for vehicles,
local-multipoint distribution services and imaging. A schematic diagram of a
256-element corporate fed patch array is shown in Figure 7.3.37 [7]. The
design of these arrays can be somewhat complicated, not so much in terms of
the antenna element design, but the feed network layout.
360
A good rule of thumb to minimize spurious radiation from feed lines
is to keep the structure as symmetrical as possible, which tends to minimize
cross-polarization levels and to use thin transmission lines. An excellent
paper on the effect of the feed network on the overall performance of a
corporate fed microstrip patch array is given in [7]. In this paper, it was
shown that as the array gets larger, the loss associated with the feed network
gets more and more until it can be substantial. For a 32 x 32 element array,
the loss was more than 7.5 dB. Table 7.3.5 shows a comparison of planar
arrays of microstrip patches versus reflectors with efficiencies of 50 % [7].
The table highlights the issues related to large patch arrays. We can see from
this table that although the directivity increases as the number of elements
increase and microstrip technology can yield similar gains to that of a
reflector for array sizes less than about 1000 elements, the feed related losses
(radiation, dielectric and ohmic) become significant.
Number of Elements 16 64 256 1024 4096
Directivity without 20.9 27.0 33.0 39.2 45.1
network
Radiation loss 0.8 1.0 1.3 1.9 2.6
Surface wave loss 0.3 0.3 0.2 0.2 0.1
Dielectric loss 0.1 0.3 0.5 1.0 2.1
Ohmic loss 0.1 0.3 0.6 1.2 2.4
Calculated aaln 19.5 25 30 34.5 37.5
Gain of reflector 18 24 30 36 42
7.4-1 Introduction
/
Zin (patch)
200n
=200n
200n
loon
200n
67Q
200n
son
, Zin
362
Figure 7.4.3 Radiation Performance of 8-element Combination FeedLinearArray: (a)
H-plane; (b) E-plane
363
- -"
"
'..J - - •
• • • • • • 1527
" g . III Ell'r.';
• " , ., ¢.h ~ . ~
;_:t'..~.·,'~:.·_-~.L',:-;;.""
.. .•.• ,.,:".~·, ,'i~._.,·.~:.", _~.>:,."·:,·_':;-. :;r
.. ..-, ~';£~~"\~"<;-~-~~. , . ,
.., .
Figure 7.4.5 Radiation Performance of 32-element Combination Feed Planar Array : (a)
H-plane; (b) E-plane
7.5-1 Introduction
364
There is one important issue related to microstrip patch antenna
technology that was discussed in Chapter 4 and this is surface wave
excitation. As mentioned in Chapter 4, surface waves (sometime referred to
as leaky waves) are 'trapped' waves excited due to the presence of the
substrate or dielectric layers associated with the microstrip antenna. Because
the energy is generally trapped within the material and not radiated surface
waves are classified as a loss mechanism. The presence of a surface wave can
cause increases in cross-polarization levels due to the trapped wave refracting
off the finite edges of the ground-plane of the antenna. Surface waves can
also cause unwanted coupling between the antenna and any active devices.
For a single layer microstrip patch antenna, the thicker the material
used the larger the power lost to the surface wave. Also the higher the
dielectric constant the less efficient the antenna becomes due to surface wave
excitation. For large arrays of microstrip patches, the resonance of modes
associated with these surface waves can severely limit the scan performance
of the array by inducing a phenomenon known as a scan blindness. For a scan
blindness all (or at least most) of the power is coupled back to the source and
subsequently is not radiated. A common means to examine the scanning
potential of a large array of microstrip patches is to consider the theoretical
active reflection coefficient of the array, which is defined as the reflection
coefficient of an element in the array as a function of scan angle [10]. Figure
7.5.1 shows the active reflection coefficient of a large array of probe-fed
patches when scanning in the E-, H- and D-planes. As can be seen here, in
the E-plane, the active reflection becomes larger as the scan angle increases
due to mutual coupling until a point where it levels off and then it increases to
unity at endfire. The scan angle where it is becomes large (approximately
75°) is the scan blindness. Note the degree of blindness depends on what
element is used. For example, if aperture-eoupled patches were used in this
array, the active reflection coefficient at the scan blindness would approach
one. The magnitude at the scan blindness is dependent on the level of
spurious radiation from the antenna.
1.2 , - - - -- - - -- - - - - - - ---,
---2 :I~R
- a - E· planc
.......-H.planc
-o-. D-planc
o ~q:R.~+-t--j-_t-I__+___+__~
o 9 18 27 36 45 54 63 72 81 90
Theta (d e g re es)
..0.9
V\
c
.!!!D.8
I
10 / \ A
VI
~
u
~.7 - ---- 1.1510
········0.8510
{ V
Bo.6
I.
I,
/' .........
c ,I ,
/ ..... '" :
t·
,I
5
t } I'
,,
V·
:;:0.4
III
';0.3
> ,'/... .... ........ ,. ......
..
......
.~ '
,Y- ......
~.2
'
<
...,.
... R
0.1
o
o 10 20 30 40 50 60 70 80 90
Theta (degrees)
366
The methods effectively couple power from the surface wave and radiate this
power. We also investigate the scanning potential of broadband (stacked) and
wideband (aperture stacked) patches. It will be shown that these arrays of
bandwidth enhanced elements do not suffer the same scanning issues as the
single layer geometries. Throughout the investigations presented, the general
scanning performance of conventional microstrip patch arrays will be given,
to benchmark the other configurations.
7.5-3.1 Introduction
368
resulting bandwidths and therefore yields very simple antenna structures and
avoids potential blindnesses associated with these networks [20]. All the
arrays examined were analyzed using the full-wave Spectral Domain infinite
array analysis described earlier.
Figure 7.5.3 shows the unit cell of the probe-fed stacked patch array.
As can be seen from Figure 7.5.3 the rectangular stacked patch element is fed
by a coaxial pin located at (pp, %) from the center of the bottom patch. The
lower patch resides on a dielectric layer with a dielectric constant of Ert and a
thickness d., The parasitic coupled patch is mounted on a laminate with a
dielectric constant Er2 and a thickness d2• The unit cell size is a in the x-
direction and b in the y-direction for the array. To investigate the
performance of the probe-fed stacked patch configuration, the full-wave
analysis presented in [7] was used. This analysis is based on the Spectral
Domain Integral Equation (SDIE) technique with boundary conditions
enforced using a Galerkin moment method. Also, the probe feed is accurately
modeled with an attachment mode [22]. For scanning in the E-plane (cj> = 0°
in Figure 7.5.3), 10 Entire Domain Basis Functions per patch were used in
addition to the attachment mode. For scanning in the H-plane (cj> = 90°) the
complimentary set of orthogonal modes were incorporated (refer to [22] for
details). It should be noted that the unit cell given in Figure 7.5.3 was also
used to examine the performance of stacked patches utilizing a combination
of high and low dielectric constant laminates. Although these hi-lo
configurations would typically incorporate edge feeding of the driven
element, it has been shown that edge and probe feeding these structures give
similar responses. Therefore, for the ease of analysis, only probe-fed
configurations are presented here.
Top View
Side View
It has been shown in the past [12], that the presence of grating lobes
can severely limit the useful scanning range of probe-fed patch arrays. Thus
to maximize the scanning range of a probe-fed stacked patch array over the
broadest band of frequencies, the element spacing should be near IJ2 at the
maximum operating frequency of the array. The need to avoid grating lobes
can be attributed to the excited surface wave before the onset of the grating
lobe and hence the large impedance mismatch and resulting blindness at this
scan angle [10]. Of course by having the elements relatively close together as
required by this constraint, means that the overall thickness of the dielectric
layers has to be limited or the impedance nature of the stacked antenna will be
too inductive. Despite this limit, good bandwidths can be achieved.
370
Figure 7.5.4a shows the impedance response as a function of
frequency of an infinite array of probe fed rectangular stacked patches at
broadside (9 = 0°). Here Erl = 2.2, d, = 0.02 Ac and Er2 = 1.07 and d2 = 0.025
Ac, where Ac is the center frequency of the band of interest (refer to the figure
caption for the other dimensions). The 10 dB return loss bandwidth of the
antenna is approximately 27 % and the response resembles that for a single
stacked patch (approximately 30 %) using this combination of dielectric
materials in Chapter 3. The design procedure is in fact fairly similar to
Section 3.4 in terms of dielectric constant selection; starting points for the
conductor sizes and how to control the impedance response and so for the
sake of brevity will not be repeated here. The only exceptions to this are the
thicknesses of the materials. For the infinite array cases, d, and d2 are thinner
by close to factors of 2 and 3 respectively. Once again, this can be attributed
to the mutual coupling within the array. Also shown in this figure is the
scanning performance (or spiders [24]) in the E- and H-planes over the
scanning range of 9 = 0° - 60°. As can be seen from Figure 7.5.4a the array
has very limited scanning potential in particular near the center of the
operation band, only ± 15° in each principle plane. How to improve the
scanning range will be addressed later.
t. '.O.llIe
2. f. 0.17fe
3. f _1.03lc
~. f .1.ottc
5. ' . 1.121'c
~~
(a) (b)
372
return loss of 27 % over a scanning range of ± 45° in both planes compared to
17 % over ± 35°. This can be attributed yet again to the greater mutual
coupling between elements when the lower dielectric constant laminates are
used, which not only gives a more rapid variation of impedance as a function
of frequency but also scan angle.
(a) (b)
Probe-fed stacked patch arrays with erl = 12, 14 and 16 were also
designed and results similar to Figure 7.5.5b were achieved using these
laminates, namely bandwidths greater than 27 % over a scanning range of ±
45° in each plane. Importantly too, the surface wave efficiency for all these
configurations were high, in excess of 90%. The only modification to the
design presented in Figure 7.5.5b was the reduction in the conductor size of
the lower patch as the dielectric constant was increased and a slight
modification to the upper patch conductor dimensions to ensure the resonant
loop was located near 50 .0. These results illustrate that very good radiation,
bandwidth and scan performance can be achieved for patch arrays developed
373
on the same materials as the active devices are grown (such as GaAs) if a
second layer consisting of a low dielectric constant laminate is used for the
upper patches. Such configurations are very applicable to millimeter-wave
systems including collision avoidance radar [25] or base stations for
microcellular communications [26]. Perhaps the only perceivable drawback
of using the higher dielectric constant materials (such as Erl = 16) is the lower
radiator does indeed become smaller and thus requires more accurate
fabrication procedures.
-
l.a
orC5.0(_1)
• ••• " '.5.0(111\89) ,1 ..-........ 1--•••• ••1.5.0('001)
• ••• or1 _5.O(Imog)
iJ
100 ··· ··· .".10A(_ 1)
··_ ···· . r1_10.4{_ 1)
- .- . orCl0A(11I\89) ~ ~.
.. -.......
.._. ort _l0A(lmog)
.. .
- .. _.. or1-2-2(....) ~~
- -_.. "1..2.2(_1)
....... .
.r1_2.2(11I\89) ~ ~
.. .. . ::::::'"
..' - .,1 _2.2(lmog)
:l'ITlnll flW U"':
.......
...... ~-~'Ie.e.l Il ln . erl
._... .. . ..... ~ ........
I-.... ~
...... . .-....
. . ........... . .... . . -
~ ......... :/ _0""
lY_ r-;
~
"7 t-- <; . .... . ~ ••• .J
I R r:-::.:: . . . . . . ....... ........ . ... . .
· 10
·100
o ~ ~ ~ .a ~ ~ ~ ~ ~ .~
Theta (degreel)
o 10 ~ 30 .0 ~ eo 70 10 80
Theta (dot/ .ee.)
(a) (b )
374
o
.. '
.., . .
iii · 10
.._u_ . 1-'
'
...... ,.•..•.. .., ..........,,. .
::E. .20
III
~ ...... ....
....... .....
-. ..."...,.
·30 !>to'Pl'" ~/
.'
~.
,.....
~.,
---
o .' .Nt' I--c',.:"~
--- -.......
~ ·50 ! <;J V-
~
i .;p'
i ·60
.......
i>" ./
8.~ ·70
-er1_2.2
•••• er1_5 f---
e
o
·60
.90
/ .... .... er1_10.4
_._. er1_14 f--
_"_00 er1 2.2(flg.28) f - -
·100
o 10 20 30 40 50 60 70 80 90
Theta (degrees)
7.5-4.i introduction
375
In this section we investigate the impedance response and scanning
performance of a large array of ASPs. As was shown in Chapter 3, this form
of microstrip antenna can have impedance and radiation bandwidths in excess
of an octave as well as all the other features associated with microstrip
technology given earlier. We explore techniques to optimize the VSWR < 2:1
bandwidth over the maximum scanning range. It will be shown that
bandwidths in excess of an octave can be achieved over a scanning range of ±
45° in the principal planes making this technology a possible alternative to
TSA phased array solutions.
;f~
+ .-::n:~:~:~ (~.~~.~) 12
Figure 7.5.8 Schematic Diagram of Unit Cell of Infinite Array of Aperture Stacked
Patches ( a : element spacing in the x-direction: b : element spacing in the y-direction)
As was the case in Chapter 3, the aperture in Figure 7.5.8 is not only
used to couple the microstrip feed to the patches, but also as a radiator. To
determine the performance of the ASP in an infinite array environment, the
376
analysis used in Chapter 3 was modified to a similar form presented in [30].
The analysis used here is based on the Spectral Domain Integral Equation
technique with boundary conditions enforced using the Galerkin method and
incorporates multilayered Greens functions as given in [31]. The interaction
between the slot and the microstrip line is modeled using the reciprocity
method outlined in [30, 32]. The ASP configuration analysis uses five Entire
Domain Basis functions on each patch and the slot for scanning in the E-plane
(cjl = O~ and the orthogonal set of modes on the patches with scanning in the
H-plane (\p= 90°).
A. Array Design
377
back ratio and reduced surface wave efficiency. Since the array is to be
scanned, surface wave effects were deemed important and therefore the
material selection was considered satisfactory.
(a) (b)
Figure 7.5.9 Predicted input impedance of some infinite arrays of ASPs (a) Parameters:
£"1= = = = = = = =
2.2, d l 3.175 mm, Ea 1.07, d z 3 mm, L I WI 9.1 mm, Lz Wz 10 mm, S. 10 =
= = = = = =
mm, S, I mm, En 2.33, dr = 1.6 mm, Wr 4.75 mm, L. 3.4 mm, a b 13 nun ; (b)
= = =
Parameters: same as in part (a) except d l 1.7 nun , dz 2.1 nun and L. 5 mm
378
Figure 7.5.10 Predicted Input Impedance of Infinite Arrayof ASPs (wI =Ao)
(parameters: Erl = 2.2, d. = 2.45 rom, £(l = 1.07, d2= 1.45rom, L. = WJ = 9.1 rom, ~ = W2=
10rom, SI = 10 rom, S, = I rom, En= 2.33,df = 1.6rom, Wf= 7 rom, Ls = 3.5 rom, a:::b = 17
rom)
B. Scan Performance
Figure 7.s.11 Predicted Input Impedance of Infinite Array of ASPs (parameters: Erl = 2.2,
d l =2 mm, fa= 1.07, d2= 1.4mm, LI = WI =9.1 mm, ~=W2= 10mm, SI= 10mm, Sw= 1
mm, En = 2.33, dc= 1.6 nun, Wr = 4 mm, L. = 4 mm, a = b = 12.5 nun; frequency increment =
0.125 GHz; scan increment = 5°)
0.' I A
C
.!!! o.a ~
........ E·p1ana: 10.5GHz
-"'p1ana: 10.5GHz
Jli
u . . . .n E.pIana: 5.25GHz /;
~ - ..... If.pIana: 5.25GHz
o
0.7
o.a ' - ..- ... . H-plana:
E-plana: 7.785GH .. .....J. f·;:.I ·:•
..
,
7.a75GH .'
~u ' Ilf
.'
0.5 .'
.
~
a:
0,4
.... ~
.~
,/ I
..
~
~ 0.3
n 0.2
I>'" '
~.~ ." -, I
ol(
0.1
..--- .......,....
.n
A
.:~.
n'
\ IJ
o
o ~ ~ ~ ~ ~ ~ ro ~ ~
(a) (b)
Figure 7.5.12 (a) Predicted broadside input impedance of an infinite arrayof ASPs
(parameters: e, = 2.2, d, = 1.7 mm, Ea = 1.07,d2= 2.1 mm, L, = W, = 9.1 mm, ~ = W2 = 10
mm, SJ = 10.5 mm, S, = 1 mm, En= 2.33,dr= 1.6 mm, Wr= 4 mm, L, = 5 mm, a = b = 12.5
mm; frequency increment = 0.125GHz; scan increment = 5°);(b) E- and H-planescan
performance of the infinitearray
Many other simulations were run and from these and the cases
presented here, it appears that having the broadside impedance locus tightly
bound to the center of the Smith chart gives the best scanning performance.
This reflects the complicated impedance variation of the scanning
381
performance of the array as a function of frequency. It should be noted that
the radiation patterns of such an array are another important quantity to
ascertain the overall performance of the antenna, however, it is not possible to
determine the far-field patterns from an infmite array. Although the active
element gain [5] can give a similar insight into the scan performance as the
active reflection coefficient can, it is a normalized value and therefore it is not
that useful in determining quantities such as the front-to-back ratio and the
antenna efficiency. To get a good understanding of the radiation patterns, the
single element (or a finite array) needs to be examined. It should be noted
that the gain and radiation patterns of an ASP are relatively constant over the
VSWR < 2:1 bandwidth as was shown in Chapter 3 and therefore it can be
extrapolated that the ASP will behave in a similar manner in an infinite array
environment. To obtain some of the radiation properties of the ASP array, a
single element was evaluated using the materials given in the Figure 7.5.12
caption. For this antenna, the front-to-back ratio was greater than 12 dB for
all frequencies and scan angles where the VSWR is less than 2:1. The surface
wave efficiency was greater than 84 % for the antenna.
It should be noted that a large array of ASPs does not avoid surface
wave problems associated with other printed antenna arrays. If anything, the
array of ASPs will suffer more from surface wave effects due to its
significantly larger bandwidth. As mentioned previously, the key to avoid
these problems is the element spacing. Of course this philosophy can be
applied to printed dipoles and conventional patches too, although these
printed antennas have less degrees of freedom as the ASP and therefore it
could be difficult to obtain good impedance matching over a broad range of
frequencies.
7.6-1 Introduction
382
arrays of microstrip patches. These antennas can yield gains greater than 50
dBi, although the bandwidths are typically small, to date a couple of percent.
Figure 7.6.2 shows a photograph of a millimeter-wave reflectarray [34]. Of
course these printed antennas have many of the features of microstrip patch
arrays, however, the conformal nature of the entire antenna is no longer a
feature.
.,/ 1--
-
383
7.6-2 Reflectarrays
384
(a) (b)
Figure 7.6.3 Reflectarray Elements: (a) Patch with stub; (b) Variablesize Patch
385
applications; a bandwidth of 10 % or more is needed. The two main factors
limiting the bandwidth of the reflectarray are differential spatial phase delay
and the bandwidth of the individual microstrip elements. Bandwidth
limitations caused by differential spatial phase delay are a major factor only in
very large reflectarrays and can be minimized by designing the reflectarray
with a small fID ratio. To increase the bandwidth of the individual microstrip
elements, several techniques used in conventional microstrip array designs
may be implemented. A straightforward method is to use a thick substrate
with a low dielectric constant, and bandwidths of around 10% could be
achieved in this manner. However, use of thick substrates in the reflectarray
application increases phase errors inherent in the design, limiting the
effectiveness of this technique.
270 I--+--'---;'=:
110
110
386
7.6-3 Lens Coupled Printed Antennas - Proximity coupled patch
version
7.6-3.1 Introduction
The use of the millimeter-wave spectrum (26 - 140 GHz) has been
experiencing a resurgence of recent as many systems have been allocated or
are proposing to use frequencies within this operating band. Such
applications include collision avoidance radar [39], high capacity
microcellular communication networks [40], Local-Multipoint Distribution
Services (LMDS) [41] and simple radio links between Base Station
Controllers (BSC) and Base Transceiver Stations (BTS) for present day
Personal Communication Systems (peS). Associated with these applications
is the need for high gain, low cost and unobtrusive antennas.
Figure 7.6.8 shows the predicted and measured return loss of the
proximity-coupled lens antenna. The measurements were done using a
Wiltron 360B 40 GHz Vector Network Analyzer. As can be seen from these
results, good agreement between theory and experiment was achieved. The
predicted and measured 10 dB return loss bandwidths are 7.0 % and 7.2 %,
respectively. The slight shift in resonant frequency of 0.3 % can be attributed
to uncertainties in the dielectric properties of the materials (~E,. ± 0.02) as well
389
as the possible presence of small air gaps (fractions of millimeters) in the
multilayered structure or even slight misalignment of the layers.
i
o
-5
·10
·15
- -.::::.: e-,
r-..
.~
\t
ff,
i
;"" --
...\ I!
"\.-, I I~."..".. ~"=.,.nt ~
\I
\11
-45
~ ~ M ~ ~ ~ H ~ ~ U ~
Frequency (GHz)
270 1-++-+-1-
lao
270 eo
'00
Figure 7.6.11 Co-polar Radiation Pattern at 36.5 GHz
391
210 1-+-1-+-+-+-+
180
7.6-4.1 Introduction
392
----_.:._;:....._....._._...._-- ..
H
diJi§§~§~.i.d
~ L" En f
\ Microstrip feed-line
K-typeconnector
Figure 7.6.14 shows the predicted and measured return loss of the
lens-coupled ASP and once again it is evident that the specification of a 10 dB
return loss bandwidth across the entire Ka-band has been met. Figure 7.6.15a
shows the measured E- and H-plane co-polar radiation patterns of the antenna
at 37 GHz. The 3 dB beamwidths in the E- and H-planes were 4.r and 4.4°,
respectively. Figure 7.6.15b shows the measured co-polar radiation patterns
for the principal planes at 28 GHz. Here the 3 dB beamwidths are 5.2° and
4.8° in the E- and H-planes, respectively. The measured gain of the antenna
was 32.0 dBi at 37 GHz and 31.1 dBi at 28 GHz. The cross-polarization
levels were at least 20 dB below the co-polar patterns for all frequencies
measured and the front-to-back ratio was greater than 16 dB across the entire
frequency span.
393
~
.. ~ 1:- ThOOry
....... Experiment
I
Iii'
\\
·1 0
:!!. ........
:
.3
E
·15
·20
\ \ .. ,,, 1;/ -, .\ .
"
'ii
a: .25
\ ! \'
\/
V
-35
20 25 30 35
Frequency (GHz)
100 100
(a) (b)
7.7-1 Introduction
For this design we decided, however, not to use the continuous patch.
We choose to use a uniform microstrip patch array of which the 'laid out'
version of this is shown in Figure 7.7.2. The uniform microstrip patch array
has several advantages over the "long patch" antenna: the use of discrete
elements allows for the input impedance to be nominally independent of the
395
number of feed points; the patch width can be adjusted to control the input
impedance of each patch; and the use of the insert fed patches allow another
degree of freedom in controlling the input impedance. These design freedoms
are very useful for producing an omni-directional radiator.
1111111111111111
. . . . , ,
396
,,,
....
, ," ,
.... , , " ,,
., >
\
\
,,
... \
.
".. ,
"
....0' .. ~
.70 to
.70 , , I I
10
" , ;r.
I
,
I
\
I
~ , ,I
> I
~
',>'
, ,
110
110
0
(a) (b)
.~
315
,, ,
....
, , <\
~ ,", \
,,
,
' ) / \
~ ;' '' \
,l6' .. ,.. ',04
, , " eo
", , ,, I ,I
I
~
" , I
I
"'''-'</
, ,
(c)
Figure 7.7.3 Effectof Number of Elements on Ripple in H-plane: (a) 4-elements; (b) 6-
elements; (c) 8-elements
o
270 l-l-e_.....+..,...*+-!-....._6~-l to
180
397
The following design procedure was incorporated for this work. A
single element of the array was to be designed, fabricated and tested. An 8
element uniform array with a corporate feed network would then be designed
based on the single element design. Here impedance matching would be
incorporated at the junctions. This array would then be fabricated and tested.
A matching network to compensate for the mutual coupling and microstrip
discontinuity effects would then be designed using the measured data from the
previous step. The new array would then be fabricated and tested.
Figure 7.7.5 gives the initial layout ofthe patch elements to be used in
the uniform array. The edge fed patch was designed using a simple cavity
model approach and the insert length was determined using the approximation
given in [9]. The input impedance of the patch was designed to be 116 Q at
resonance and this value was controlled using the insert dimensions.
48.75 mm
44.00 mm
10.00 mmI
398
7.7-2.3 Design Iterations
Figure 7.7.7 shows the return loss performance of the first design
iteration of the uniform array. As can be seen from this figure, the
performance is unacceptable. The mutual coupling and microstrip
discontinuity effects have caused the array to be not impedance matched at
the design frequency.
Y '-"-'-';--
Or--....,...-..,.---,.--r---,r----,--~__,_~==
t I I J,...~....._, - - - 1- - I --
-5 - - L __ .1 __ J _ _: :_ L~~ __ i __ i __ ~ _
I I I I I I I I
iii' I
I
I
I
I
I I
I
I
I
I
I
I
I
I
I
I
:Eo -10 --r- -~ --,--~
I I I I
-rI - - Ir - - rI - - 7--
I
'I - - -
1/1
1/1 I I I I I I I I I
.3 ·15
__ L __
I
~
I
__ ~
I
__ ~
I I
L __ L __ L __
I I
~
I
__ ~
I
_
E I I I I I I I I I
.aCIl . 20
I I I
--r--7 --,---r---r--r --r--7--' ---
I I I I I I
I I I I I I I I I
II: I I I I I I I I I
__ 1 __ 1 __ __ L __ L __ 1 __ 1 __ _
-25
~ ~ ~
I I I I I I I I I
I I I I I I I I I
In the next phase two changes were made in the design. Firstly, an
impedance transformer was incorporated to provide impedance matching at
the appropriate frequency band. Also the spacing between the microstrip
feedlines was increased to negate the feedline coupling which can possibly
degrade the radiation pattern. The increased spacing could affect the input
impedance slightly. These changes are shown schematically in Figure 7.7.8.
The return loss performance of the array is shown in Figure 7.7.9. As can be
seen from this figure, although the response has been improved, the return
loss bandwidth is not appropriate for the application at hand.
399
11,11!"IIII,I!,II,II
Figure 7.7.8 Changes madeto Array Layout for Second Iteration
_
. : -:- :
O l""::'==,---..,...-...------,.-----r-..,...- =="",...,,=l"'~=!
:---'~\: :4::-:~
-- - : - --
I I I I I I I I I
-5 --'---T-
I I
-r- -,--- r- ' - - - T- - - r - - ' - - -
I I I I I I I
iii'
~ -10 ~--- ~ -
I I I
/J\-~ --:---~ --- ~ ~---
I I I I I I
Ul I I I I I I I I
Ul
.3
.
I I I I I I I I
-15 -- '--- T--- r-- , - --r- - - I - - -~ - - - r - -, - - -
I I I I I I I I I
e I I I I I I I I I
.a
ell
·20
I I I I
-- 1 - - - T - - -r - - , - --r- - -I- - - ~ - - - ~ - - ' - - -
I I I I I
I I
I
I
I
I
I
I
I
a: t I I I I I I I I
I I I I I I I I I
45 --~-- -T --- I- - -' - - -~ - - -I - - - ~-- -~--~---
: : : : : : : : :
1 I I I I I 1 I I
2 2.05 2.1 2.15 2.2 2.25 2.3 2.35 2.4 2.45 2.5
Frequency (GHz)
400
Or:::::'= = ...-----.- -.-----,,..---,--- = = = .,.....--,
I I
I , I I I I I I
-5 -- ~-- ~- ~- - ~- - ~ - - ~ -- + -- +--~- -
I I I ' I I I I
iil I I I I I I I I
~ ·10
III
III
-- 1--.--' -T--'--'- -'--' --'--
1
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
.3 -15
I
-- r--~--~ -
I I I
I
I
I I
rr I I
- T-- T - - ~ - - ~ --
I I I I
I I
E
:l
I I I I
__ L __ L _ _ L __ L _ _
I I I
_ L _ _
I I
__ _ _ _ _
Qj ·20
~ ~ ~
I I 1 I I I I I I
a: I I I I I I I I
I I I I I I I I
·25 -- r-- r- - r - - r - - --r--r --r--1- -
I I I I I I I I
I I I I I I
-30 L...........J..............J..............L.u.........J.......... ...u..........JL.u..u...L..............L............:.......-.u
2 2.05 2.1 2.15 2.2 2.25 2.3 2.35 2.4 2.45 2.5
Frequency (GHz)
180
401
Design Iteration Number4
In this phase the feed insert was shorted by 2 mID on the end patches
as shown in Figure 7.7.13. This was done to counter the effects of the
mounting screws on the radiation pattern characteristics. The return loss
performance of this design iteration is shown in Figure 7.7.14. As can be seen
here, there is only a small change in performance from that given for the
previous design phase. Because of this, no additional iterations were deemed
necessary for impedance matching of the antenna. Figure 7.7.15 shows the H-
plane radiation pattern. Here the maximum ripple is 3.1 dB, 0.9 dB below the
requirement. Also shown in this figure is the third design iteration pattern for
comparison purposes . The E-plane radiation pattern is shown in Figure 7.7.16
and once again the requirement for a 10 degree angular width for less than -
10 dBi gain was met.
11,11.11,11.11,11,11,11
.
Figure 7.7.13 Layout of Fourth Iteration
I I ,....- j"
I I , / I
I
I
I
I
I
--I
I
-r --' ---r--,--- r---
I
I
I
I
I
I
I
I
I
I
1/1
1/1 I I I I I I I I
...c I
I
I
I
I
t
I
I
I
I
I
I
I
I
EQ) -20
I
- - -. - - - r - - - ,-
I I I
1 - -
I
-1- - - , - - -
, I
r --, ---r ---
I I
a: I
I
I
f
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
,
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I I I I I I I I I
.30 ....................................................I--I...o.....u....L..I...I.........,;u................L............i.J....l...........L.J...J............u.............
2 2.05 2.1 2.15 2.2 2.25 2.3 2.35 2.4 2.45 2.5
Frequency (GHz)
402
90
180 t-i-l---r-----+----!--.....l----10
270
- - Design Iteration #3
- - DesIgn iteration #4
180
7.7-3.1 Introduction
,-,-,-:-,-,-, I: 1'02~
467mm
../" e
screw\ e _
Fastening I
e
/
SMA Connector
The wraparound antenna was designed for operation at 870 MHz and
so the size of each shorted patch is 30 rnm x 48.75 rnm. To achieve a 1oo.Q
resonance, the shorting pin on each patch (represented by a white circle in
Figure 7.7.18) is located 1 rnrn from the insert microstrip feed. Here the insert
is 10.5 rnrn from the patch edge. As can be seen from Figure 7.7.18, the two
end shorted patches of the array are narrower than the other elements (32.75
rnrn compared to 48.75 rnrn). This was done to compensate for the effect on
the input impedance behavior of these shorted patches of the fastening screws
needed to secure the antenna.
As the diameter of the cylinder is 150 mm, the necessary spacing for
an eight element array is 58.75 rnrn (center to center). The length of the
mounting cylinder was 340 rnrn and a 2 rnrn grove is present in the cylinder
such that the antenna is flush with the external surface. The array is fed via an
SMA through the ground-plane (fuselage) of the array. To curve the array
405
into the required shape, standard bend techniques outlined in [54] were
utilized.
7.7-3.3 Results
8-10 \ I
-I I
E'12
%14
a:·18
-18
·20
-22
800 820 840 880 880 900
Frequency (MHz)
Figure 7.7.19 Measured Return Loss of Wraparound Array of Shorted Patches
o
110
Figure 7.7.20 Measured Radiation Patterns of Small Wraparound Antenna
406
7.8 Summary
Corporate fed arrays are probably the most versatile with the largest
bandwidth, although these arrays do suffer from higher feed loss than series
and combination arrays. Combination arrays can provide a relatively simple
design procedure and also good radiation and bandwidth results. Planar
fixed beam and scanned arrays utilizing microstrip patches were also
investigated. Once again corporate feeding is probably the easiest means of
forming a planar array, especially if scanning the beam is required.
It was shown in this chapter that surface waves associated with the
dielectric materials of the antenna can have detrimental effects on the
scanning performance of a large array of patches, although several methods
have been established to overcome this inherent problem. One such example
incorporated the use of shorting pins.
In this chapter it was seen that the scan/materials trends for single
layer geometries do not necessarily hold for more complicated, broader
bandwidth printed structures, which is very fortuitous. The bandwidth
performance of probe-fed stacked patch antenna arrays over a useful scanning
range has been investigated. It was found that bandwidths in excess of 27%·
over a scanning range of ± 45° could be achieved if high dielectric constant
407
materials were used for the lower laminates of the stacked configuration. The
design and scan performance of large arrays of ASP antennas was also
presented. It was shown that VSWR < 2:1 bandwidths in excess of an octave
for scanning in the E- and H-plane up to ± 45° and ± 55° respectively, can be
readily achieved. The design procedure for these arrays is relatively
straightforward and is based on the previously reported work for single ASPs
(see Chapter 3) with modifications required to the substrate thicknesses to
take into consideration the mutual coupling of the array. It has been shown
that to maximize the scanning range of the array the impedance locus at
broadside should be tightly bound to 50 Q . The results presented herein
show that microstrip patch technology can be applied to situations where it
previously would have not been considered, such as radio astronomy, due to
the large bandwidths and good scanning performance presented herein.
7.9 Bibliography
[lJ 1. R. James, P. S. Hall and C. Wood, Microstrip Antenna Theory and Design,
London, Peter Peregrinus, 1981.
[2J 1. Huang, "A parallel-series-fed microstrip array with high efficiency and low cross-
polarization," Microwave Optical Technology Letts., vol. 5, pp. 230 - 233, May
1992.
[3J D. M. Pozar and B. Kaufman. "Design considerations for low sidelobe microstrip
arrays," IEEE Trans . Antennas & Propagation, vol. 38, pp. 1176 - 1185, Aug. 1990.
[4J Ensemble 6.1, Ansoft, 1999.
408
[5] D. M. Pozar, "The active element gain," IEEE Trans. Antennas & Propagation, vol.
42, pp. 1176 - 1178, Aug. 1994.
[6] R. J. Mailloux , 1. F. McIlvenna and N. P. Kemweis, "Microstrip array technology,"
IEEE Trans. Antennas & Propagation, vol. 29, pp. 25 - 37, Jan. 1981.
[7] E. Levine, G. Malamud, S. Shtrikman and D. Treves, "A study of microstrip array
antennas with the feed network ," IEEE Trans. Antennas & Propagation, vol. 37, pp.
426 - 434, April 1989.
[8] D. M. Pozar and D. H. Schaubert, "Comparison of three series fed microstrip array
geometries," IEEEAntennas & Propagation Symposium, Ann Arbor USA, pp. 728 -
731, July 1993.
[9] C. A. Balanis, Antenna Theory: Analysis and Design, 2nd Edition , Wiley, New York,
1996.
[10] D. M. Pozar and D. H. Schaubert, "Scan blindness in infinite arrays of printed
dipoles," IEEETrans. Antennas & Propagation, vol. 32, pp. 602 - 610, June 1984.
[11] D. M. Pozar, "Scanning characteristics of infinite arrays of printed antenna
subarrays," IEEE Trans. Antennas & Propagation, vol. 40, pp. 666 - 674, June 1992.
[12] D. Novak and R. B. Waterhouse, "Impedance behaviour and scan performance of
microstrip patch arrays configurations suitable for optical beamforming networks,"
IEEE Trans. Antennas & Propagation, vol. 42, pp. 432 - 435, Mar. 1994.
[13) R. B. Waterhouse and N.V. Shuley, "Scan performance of infinite arrays of
microstrip patch elements loaded with varactor diodes ," IEEE Transactions Antennas
& Propagation, vol. 41, pp.l273 - 1280, Sept. 1993.
[14] F. Zavosh and J.T. Aberle, "Infinite phased arrays of cavity-backed patches", IEEE
Trans. Antenn. Propagat., vol. AP - 42, pp. 390-398 , March 1994.
[15] R. B. Waterhouse, "Improving the scan performance of probe-fed microstrip patch
arrays," IEEE Transactions Antennas & Propagation, vol. 43, pp. 705 - 712, July
1995.
[16] R. B. Waterhouse, "The use of shorting posts to improve the scanning range of
probe-fed microstrip patch phased arrays," IEEE Transactions Antennas &
Propagation, vol. 44, pp. 302 - 309, March 1996.
[17) A. Klouche-Djedid and M. Fujita, "Adaptive array sensor processing applications for
mobile telephone communications", IEEE Trans. Vehicular Technology, Vol. VT-45,
Mar. 1996.
[18] 1. J. Schuss , 1. Upton, B. Myers, T. Sikina,A. Rohwer, P. Makridakas, R. Francois ,
L. Wardle and R. Smith, "The IRIDIUM main mission antenna concept", IEEE
Trans. Antennas & Propagation, vol. AP-47, pp. 416-425, Mar. 1999.
[19) J. S. Herd, "Full wave analysis of proximity coupled rectangular microstrip antenna
arrays", Electromagnetics, Vol. 11, pp. 21 - 46, Jan. 1991.
[20) 1. 1. Schuss , "Numerical design of patch radiator arrays", Electromagnetics, Vol. 11,
pp. 47 - 68, Jan. 1991.
[21] Y. Lubin and A. Hessel, "Wide-band, wide-angle microstrip stacked-patch-element
phased arrays", IEEE Trans. Antennas & Propagation, vol. AP-39, pp. 1062 - 1070,
Aug. 1991.
[22) J. T. Aberle, D. M. Pozar and 1. Manges, "Phased arrays of probe-fed stacked
microstrip patches", IEEE Trans. Antennas & Propagation, vol. AP-42, pp. 920-927,
July 1994.
[23] R. B. Waterhouse, "A novel technique for increasing the scanning range of infinite
arrays of microstrip patches," IEEE Microwave & Guided Wave Letters, vol. 3,
pp.450 - 452, Dec. 1993.
[24] 1. S. Herd, privatecommunication.
409
[25] W. Menzel, D. Pilz and R. Leberer, "A 77 GHz FMlCW radar frontend with a low-
profile, low-loss printed antenna", 1999 IEEE MTI-S Intemational Microwave
Symp., Anaheim, pp. 1485-1488, June 1999.
[26] Z. Ahmed, D. Novak, R. B. Waterhouse and H. F. Liu, "37 GHz Fiber-Wireless
System for Distribution of Broadband Signals", IEEE Trans. Microwave Theory &
Techniques, pp. 1431-1435, Aug. 1997.
[27] P. 1. Hall, A. S. Mohan and R. S. Soretz, "Interference characterization at Australian
astronomy sites", IkT Intemational Technical Workshop, Sydney Australia, p. 37,
Dec. 1997.
[28] B. MacThomas, "An evolutionary approach to the development of the 'Square
Kilometre Array', and related generalised antenna layouts and concepts", ATNF
Technical Document 39.31087,22 Jan. 1999.
[29] T. Chio and D. H. Schaubert, "Effectsof slotline cavityon dual-polarized tapered slot
antenna arrays," 1999 Antennas & Propagation International Symp., Orlando USA,
pp. 130-133, July 1999.
[30] D. M. Pozar, "Analysis of an infinite phased array of aperture coupled microstrip
patches," IEEE Trans. Antenna Propagat., vol. 37, pp. 418 - 424, April 1989.
[31] N. K. Das and D. M. Pozar, A generalized spectral domain Green's function for
multilayerdielectricsubstrates with applications to multilayered transmission lines,"
IEEE Trans. Microwave Theory Tech., vol. 35, pp, 326- 335, Mar. 1987.
[32] D. M. Pozar, A reciprocity method of analysis for printed slot coupled microstrip
antennas," IEEE Trans. AntennaPropagat., vol. 34, pp. 1439-1446, Dec. 1986.
[33] L . Mall and R. B. Waterhouse, "Millimeter-wave proximity-coupled microstrip
antenna on an extended hemispherical dielectric lens," IEEE Transactions on
Antennas & Propagation, vol. 49, pp. 1769-1772, Dec. 2001.
[34] D. M. Pozar, S. D. Targonski and H. D. Syrigos, "Design of millimeter-wave
microstrip reflectarrays," IEEE Trans. Antennas & Propagation, vol. 45, pp, 287 -
296, Feb. 1997.
[35] C.S. Malagisi, "Microstrip Disc Element Reflectarray", Electronics and Aerospace
SystemsConvention, Sep. 1978.
[36] T.A. Metzler, "Stub Loaded Microstrip Reflectarray", IEEE AP-S International
Symposium Digest,pp. 574·577, 1995.
[37] A. Kelkar,"FLAPS: Conformal Phased Reflecting Surfaces", Proc. of IEEE National
RadarConference, pp. 58-62, March 1991.
[38] S.D. Targonski and D.M. Pozar, "Analysis and Design of a Microstrip Reflectarray
Using Patches of Variable Size", IEEE AP-S International Symposium Digest, pp.
1820-1823, 1994.
[39] S. Ohshima, Y. Asano, T. Harada, N. Yamada, M. Usui, H. Hayashi,T. Watanbeand
H. Iizuka, "Phase-comparison monopulse radar with switched transmit beams for
automotive application", 1999 IEEE MTI-S Intemational Microwave Symp.,
Anaheim, pp. 1493-1496, June 1999.
[40] A. Nirmalathas, C. Lim, D. Novak and R. B. Waterhouse, "Progress in millimeter-
wave fiber-radio access networks," (invited) Annals of Telecommunications, vol. 56,
pp. 27 - 38, Jan./Feb. 2001.
[41] D. A. Gray, "Optimalcell deployment for LMDS systems at 28 GHz", Proc. Wireless
Broadband Conf., Washington DC, July 1996.
[42] D. F. Filipovic, S. G. Gearhart and G. M. Rebeiz, "Double-slot antennas on extended
hemispherical and elliptical Silicon dielectric lenses", IEEE Trans. Microwave
Theory & Techniques, vol. MIT - 41, pp. 1738-1749, Oct. 1993.
[43] D. F. Filipovic and G. M. Rebeiz, "Double-slot antennas on extended hemispherical
and elliptical Quartz dielectric lenses", Intemational Journal of Infrared and
Millimeter Waves, vol. 14, pp. 1905-1924, 1993.
410
[44] 1. Zmuidzinas and H. G. leDuc, "Quasi-optical slot antenna SIS mixers," IEEE
Trans. Microwave Theory & Techniques, vol. MIT - 40, pp. 1797-1804, Sept. 1992.
[45] D. F. Filipovic, G. P. Gauthier, S. Raman and G. M. Rebeiz, "Off-axis properties of
silicon and quartz dieleclric lens antennas", IEEE Trans. Antennas & Propagation,
vol. AP - 45, pp. 760 - 766, May 1997.
[46] G. V. Eleftheriades, Y. Brand, J-F Zurcher and J. R. Mosig, "ALPSS: A millimetre-
wave aperture-coupled patch antenna on a substrate lens", Electronics Letters, vol.
33,pp. 169-170, Jan. 1997.
[47] X. Wu, G. V. Eleftheriades and E. Van Deventer, "Design and characterization of
single and multiple beam mm-wave circularly polarized lens antennas for wireless
communications", 1999 IEEE AP-S International Antennas & Propagation Symp.,
Orlando, pp:1200-1204, July 1999.
[48] P. Otero, G. V. Eleftheriades and 1. R. Mosig, "Integrated modified rectangular loop
slot antenna on substrate lenses for millimeter- and submillimeter-wave frequencies
mixer applications", IEEE Trans. Antennas & Propagation, vol. AP - 46, pp. 1489-
1497, Oct. 1998.
[49] D. M. Pozar and S. M. Voda, "A rigorous analysis of a microslripline-fed patch
antenna", IEEE Trans. Antennas & Propagation, vol. AP - 35, pp. 1343-1349, Dec.
1987.
[50] D. M. Pozar, "Radiation and scattering from a microslrip patch on an uniaxial
substrate", IEEE Trans. Antennas & Propagation, vol. AP - 35, pp. 613 - 621, June
1987.
[51] R..C. Johnson and H. Jasik, Antenna Engineering Handbook, 2nd Edition, Chapter 7,
McGraw-Hill,1984.
[52] J. R. James and P. S. Hall, Handbook of Microstrip Antennas, Vol. 2, Chapter 20,
Peter Peregrinus , 1989.
[53] D. I. Wu, "Omnidirectional circularly-polarized conformal microslrip array for
telemetry applications", Proc. IEEE Antennas & Propagat. Symp., pp. 998 - 1001,
June 1995.
[54] Rogers Corporation : http:/www.rogers-corp .com/mwullitinbl.htm (Fabrication
Information)
411
Chapter 8 Summary
8.1 Overview
Material Optimization
AntennaOptimization
Integration
8.3 Bibliography
[1] A. Nirmalathas, C. Lim, D. Novak and R. B. Waterhouse, "Progress in millimeter-
wave fiber-radio access networks, " (invited) Annals of Telecommunications, vol. 56,
pp. 27 - 38, Jan./Feb. 2001.
[2] D. H. Schuabert, F. D. Farrar, A. Sindoris and S. T. Hayes, "Microstrip Antennas
with Frequency and Polarization Diversity," IEEE Transactions Antennas &
Propagation, vol. 29, pp. 118 - 123, January 1981..
414
[3J R. B. Waterhouse, 'The Use of Shorting Posts to Improve the Scanning Range of
Probe-fed Microstrip Patch Phased Arrays," IEEE Transactions Antennas &
Propagation, vol. 44, pp. 302 - 309, March 1996.
[4] G. M. Rebeiz, G.-L. Tan and 1. S. Hayden, "RF MEMS Phase Shifters: Design and
Applications", IEEE Microwave Magazine, vol. 3, pp. 72 - 81, June 2002.
415
Index
Balanced fed
Conventional patches 121-122, 147,343
Shorted patches 240-257
Branchline coupler 357
c
Cavity backed
Aperture coupled patches 114-116,294
Aperture stacked patches 143 -145
Stacked patches 109- 11Q, 367
Cavity model 3,5,17,23,25,38,41 ,78,398
Circular patches 7, 17,27 - 29,46 - 48,71 ,73,77,79,84,96
- 99, 169, 170, 171 - 177, 193, 212, 252,
312,314
Cost function 101-103
Coverlayer/superstrate 1,80,200,208
Critical radius 170, 171
I,J, K
Indirectintegration 278-280
Integral Equations 4,5,24,25,26,39,47,51,54,59,77,81,84,
109,124,173,176,179,369,377,388
Integrated patches 414
Inverted F antennas 6,200,225,271
418
L
MEMs 2,414
Millimeter-waves 11,18 ,39,124,136,137,139,277,280,288 ,
289, 322, 323, 324, 348, 360, 374, 383, 384,
387,390
Mobilehandsets 10, 15, 16, 26, 80, 194, 197, 198, 199, 200,
204, 206, 208, 223 - 225, 262, 264, 273
MonolithicMicrowave Integrated Circuits 5,33,34, 100, 180, 193,280,281,283,286,
289, 290, 292, 293, 295, 300, 317, 323, 350,
355
Monopoles 72, 190, 191, 198, 199, 200, 205, 206, 223,
226,237,259,263,266 - 271 ,274
Mutualcoupling 47,73 ,90 ,115,147,156, 159,161,251,292,
314, 328, 357, 358, 365, 370, 371 , 372, 373,
377,396,398,399,408
o
Optimization 6, 28, 43, 70, 90, 100- 108, 278, 281 , 372,
413
Opto-electronic Integrated Circuits 33, 34, 100, 180, 193, 279, 289, 281, 283,
286, 289, 290, 292, 293, 300, 310, 317, 319,
320,323,324,350,355
Rectangular patches 1, 7, 26 - 28,29, 31, 35, 38, 39, 70, 75, 84,
85, 88, 93, 94, 95, 112, 134, 144, 151, 186,
202, 203, 206, 207, 216, 217, 220, 223, 227,
228, 232, 233, 239, 249, 252, 266, 273, 310,
311,312,321,357,369,371 ,376,388
Reflectarrays 139, 382, 383, 384 - 386, 392
Reflecting patches 114, 115, 117, 119, 124, 139 - 143, 295 -
299,323,325
Ring patches 31,84,94,217,246
s
Scan blindness 4,365,366,367,368,372,381
Shielding planes 114, 138, 293 - 294
Shorted patches 201-272
Single feed dual polarizedpatches 301
Spatial domain integralequationtechnique 24,25
SpecificAbsorption Rate 202,223 ,263,271
Spectraldomainintegral equation technique 24, 25, 28, 39, 47, 51, 54, 59, 77, 81, 84,
109, 112, 116, 117, 124, 140, 173, 176, 179,
203,242,278,310,369,377,388
Spiral patches 201,226 - 228,231,237 - 240
Stackedpatches 4,7,17,18,19,28,49,70,75,76,83 -111,
123 - 162, 178 - 182, 216, 234 - 235,266-
270, 283 - 289, 310 - 322, 330, 332 - 343,
350,351,355 - 357,369 - 382
Surfacewaves 9,10,33,39,55,96,106,109,136,143,167,
168 - 169, 184, 185, 187, 189 - 193, 259,
261,280,291,309,365 - 367,388
Synchronous subarrays 17, 60, 62 - 64, 244, 249 - 258, 273
420
T
Taperedslots 1,69,375
Transmission line model 3,17,23,24,25,36,38
Triangular patches 29 - 30,95,217 ,249,252,253
W,X, Y,Z
421