close

Вход

Забыли?

вход по аккаунту

?

Inverted stripline technology for integrated active antenna and microwave integrated circuit applications

код для вставкиСкачать
INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UM I
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in typewriter face, while others may
be from any type o f computer printer.
The quality o f this reproduction is dependent upon the quality o f the
copy submitted. Broken or indistinct print, colored or poor quality
illustrations and photographs, print bleedthrough, substandard margins,
and improper alignment can adversely afreet reproduction.
In the unlikely event that the author did not send UMI a complete
manuscript and there are missing pages, these will be noted. Also, if
unauthorized copyright material had to be removed, a note will indicate
the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand com er and
continuing from left to right in equal sections with small overlaps. Each
original is also photographed in one exposure and is included in
reduced form at the back of the book.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6" x 9" black and white
photographic prints are available for any photographs or illustrations
appearing in this copy for an additional charge. Contact UMI directly
to order.
A Bell & Howell Information C o m p an y
3 0 0 N orth Z e e b Road. Ann Arbor. Ml 48106-1346 USA
313/761-4700 8 0 0 /5 2 1 -0 6 0 0
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
IN V E R T E D S T R IP L IN E T E C H N O L O G Y F O R IN T E G R A T E D
A C T IV E A N T E N N A S A N D M IC R O W A V E IN T E G R A T E D C IR C U IT
A P P L IC A T IO N S
A D issertation
by
JU LIO AN G EL N A V A R R O
Subm itted to th e Office o f G raduate Studies o f
Texas A & M U niversity
in partial fulfillment o f the requirem ents fo r the degree o f
D O C T O R O F P H IL O S O P H Y
Decem ber 1995
M ajor Subject:
Electrical Engineering
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
UMI N u m b er: 9 6 1 5 8 5 4
UMI Microform 9615854
Copyright 1996, by UMI Company. AH rights reserved.
This microform edition is protected against unauthorized
copying under Title 17, United States Code.
UMI
300 North Zeeb Road
Ann Arbor, MI 48103
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
IN V E R T E D S T R IP L IN E T E C H N O L O G Y F O R A C T IV E A N T E N N A
A N D M IC R O W A V E IN T E G R A T E D C IR C U IT A P P L IC A T IO N S
A D issertation
by
JU L IO A N G E L N A V A R R O
Subm itted to T exas A & M U niversity
in partial fulfillm ent o f the requirem ents
for the degree of
D O C T O R O F PH ILO SO PH Y
A pproved as to style and co n ten t by
K ai C hang
(C h air o f C om m ittee)
K aran W atson
(M em ber)
K rzy szto f A . M ichalski
(M em ber)
D onald P arker
(M em ber)
C iSj
C am N guyen
(M em ber)
Jianxin Zhou
(M em ber)
fi c cl:J
y
A . D .jP atton
(H ead o f D epartm ent)
D ecem ber 1995
M ajo r Subject:
Electrical Engineering
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
ABSTRACT
Inverted Stripline Technology fo r Integrated A ctive A ntenna and
M icrow ave In teg rated Circuit Applications. (D ecem ber 1995)
Julio A ngel N avarro, B.S.; M .S., Texas A & M U niversity
C hair o f A dvisory Com m ittee: D r. Kai C hang
T his d issertation presents studies and developm ents in passive an d activ e integrated
an tennas and beam steerable spatial pow er combiners.
functions
at
transm ission
n etw orks.
th e
antenna
term inals, integrated
line losses typically
B y in corporating th e com ponent
antennas
associated w ith
avoid
conventional
th e
transition and
pow er
distribution
In teg rated and active integrated antennas have th e potential o f reducing th e
size, w eight and cost o f conventional transm itter and receiver designs.
Passive devices
such as PIN , varacto r, o r Schottky-barrier diodes can b e in teg rated w ith antennas to
p ro v id e sw itching, tuning o r m ixing functions. Active devices such as G unn and IM P A T T
d io d es o r transistors can also b e integrated w ith antennas to co n v ert D C energy to R F
p o w e r and create radiating oscillators. A large num ber o f radiating oscillators o r active
an tennas can b e synchronized to operate coherently in p o w e r com bining applications.
P o w e r com bining techniques have been developed at the chip, circuit and spatial level to
o v erco m e p o w er deficiencies o f solid-state devices in the m illim eter and sub-m illim eter
bands.
Injection-locking techniques are used to ensure coherent o p e ratio n o f all active
a n ten n a so u rces and to o b tain high pow er combining efficiency.
C o n tro l over th e
frequencies o f each active antenna has also been show n to pro v id e beam steering
capability.
T his beam steering phenomena w ould present an inexpensive approach to
c re a te an electronic phased array. U p to 40 degrees o f steering has b een dem onstrated in
1 x 4 E - and H -plane arrays during this investigation.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
D E D IC A T IO N
/ d edicate this dissertation to my lovely w ife, Maggie.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
V
ACKNOW LEDGM ENTS
I ack now ledge th e U. S. A rm y R esearch Office, the N ational Science Foundation and
N A S A -L ew is R esearch C e n te r fo r th eir support o f this research as well as th e R ogers
C o rp o ra tio n fo r providing substrate m aterials. I thank m y advisor, Professor K ai Chang,
fo r guiding m e th ro u g h o u t my entire g raduate school tenure. I also thank the m em bers o f
m y com m ittee, K aran W atson, K rzy szto f Michalski, C am N guyen and D onald Parker, for
th eir k n o w ledge and assistance during this investigation.
I ack n o w led g e th e efforts o f students and m em bers o f th e electrom agnetics laboratory o f
T ex as A & M :
M ing Y i Li, Jam es M cSpadden, Jam es M cC leary and R obert Flynt.
Specifically, I th an k L u F a n fo r m aking this research a success.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
TABLE O F CONTENTS
Page
A B S T R A C T ...................................................................................................................................
iii
D E D IC A T IO N ..............................................................................................................................
iv
A C K N O W L E D G M E N T S .........................................................................................................
v
T A B L E O F C O N T E N T S ..........................................................................................................
vi
L IS T O F F IG U R E S .....................................................................................................................
viii
L IS T O F T A B L E S ......................................................................................................................
x
CHAPTER
I
IN T R O D U C T IO N TO M IC R O W A V E IN T E G R A T E D C IR C U IT S
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1
In tro d u ctio n ............................................................................................
M ic ro w a v e s............................................................................................
M IC Transm ission L in e s .....................................................................
D ielectric S u b stra tes.............................................................................
A ctive S olid-State D evices..................................................................
M icrow ave Intergrated C ircuits.........................................................
Integrated and A ctive Integrated A n ten n as....................................
System A pplications..............................................................................
1
1
5
9
15
16
19
22
II T H E P R O B E -F E D IN V E R T E D ST R IPL IN E A N T E N N A ...............................
25
2.1
2 .2
2.3
In tro d u ctio n .............................................................................................
T he Cavity M o d e l..................................................................................
M easured R esults...................................................................................
25
27
29
III D IO D E IN T E G R A T E D IN V E R T E D S T R IP L IN E A N T E N N A S .................
36
3.1
3.2
In tro d u ctio n .............................................................................................
D iode Integration...................................................................................
36
37
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER
Page
IV T H E O R Y A N D P R O P E R T IE S O F O S C IL L A T O R S ........................................
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
4 .1 0
48
In tro d u c tio n .............................................................................................
O scillation C o n d itio n s...........................................................................
V o ltag e C ontrolled O scillators...........................................................
D erivation o f O scillation C onditions.................................................
S -P aram eter Form ulation o f the O scillation C o n d itio n s
Q uality F a c to r..........................................................................................
S ta b ility .....................................................................................................
N o is e ..........................................................................................................
P ulling F ig u r e ..........................................................................................
O scillator S y n ch ro n izatio n ...................................................................
48
48
49
52
58
62
63
65
67
70
V G U N N & F E T -IN T E G R A T E D IN V E R T E D ST R IPL IN E A N T E N N A S ....
76
5.1
5.2
5.3
5.4
In tro d u c tio n ...............................................................................................
Single A ctive A ntenna.............................................................................
G u n n -In teg rated IS A s.............................................................................
F E T -In teg rated I S A s ..............................................................................
76
78
BO
85
V I SP A T IA L P O W E R C O M B IN IN G A ND B E A M S T E E R IN G .........................
91
6.1
6.2
6.3
6.4
6.5
In tro d u c tio n ...............................................................................................
In jection-L ocked O sc illa to rs................................................................
Spatial P o w e r C om bining......................................................................
Injection L ock ed P hased A rrays..........................................................
B eam Steering in A ctive A ntenna A rrays..........................................
91
93
95
101
102
V II C O N C L U S IO N S A N D R E C O M M E N D A T IO N S .............................................
112
R E F E R E N C E S ...............................................................................................................................
114
V IT A ..................................................................................................................................................
125
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
L IS T O F F IG U R E S
Figure
Page
1.1
C om m on bands in the frequency spectrum ...........................................................
2
1.2
A tm ospheric ab sorption as a function o f operating frequency........................
3
1.3
Coaxial and w aveguide transm ission lines............................................................
4
1.4
M icro w av e integrated circuit (M IC ) transm ission lin e s ...................................
6
1.5
Typical so lid -state device p a c k a g e s.......................................................................
17
1. 6
S olid-state device integration w ith various planar transm ission lines
19
1.7
C onventional and integrated antenna approaches for tran sm itte rs................
22
2.1
Inverted p a tc h configuration ....................................................................................
26
2.2
ISA con fig u ratio n dim ensions.................................................................................
30
2.3
IS A te st fix tu re .............................................................................................................
31
2 .4
ISA o p e ra tin g frequencies as a function o f patch d iam eter.............................
32
2.5
ISA in p u t im pedance a s a function o f probe position (/?)................................
33
2.6
P ro b e-fed IS A radiation p a tte rn s ...........................................................................
34
2.7
H alf-p o w er beam w idths and gain o f probe-fed IS A s ........................................
35
3.1
R ectan g u lar p atch antenna dimensions and transm ission line m o d el
38
3.2
C ircular p a tc h antenna configuration and coordinate s y s te m .........................
39
3.3
S o lid -state device package and integrated IS A co n fig u ra tio n .......................
41
3.4
P IN in te g ra te d IS A sw itch resu lts..........................................................................
42
3.5
F req u en cy agility o f a varactor-integrated IS A ...................................................
44
3.6
B an d w id th tunability an d corresponding V S W R versus varactor voltage..
45
3.7
P ercen tag e im pedance bandw idth versus frequency..........................................
46
4 .1
D io d e an d G unn o scillator equivalent circu it......................................................
52
4.2
O scillation conditions o f a typical circuit..............................................................
56
4.3
G unn dio d e curren t-v o ltag e relationship..............................................................
57
4.4
T w o -p o rt circuit defin itio n s.....................................................................................
58
4.5
T w o - and th ree-p o rt F E T equivalent circuit........................................................
60
4.6
S ynchronized oscillator n o tatio n .............................................................................
70
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
F ig u re
P age
4.7
Lead an d lag phase versu s frequency fo r a single-tuned o scillato r
71
4.8
C hange in b eat frequency due to in jectio n -lo ck in g ...........................................
74
5.1
C anonical oscillator c irc u it.......................................................................................
78
5.2
G unn-integrated I S A ..................................................................................................
82
5.3
T uning characteristics o f a G u nn-integrated ISA (D=10.4 mm) ...................
84
5.4
Typical X -band G unn-integrated IS A radiation p attern s.................................
85
5.5
F E T -in teg rated ISA co n fig u ratio n .........................................................................
86
5.6
F E T -in teg rated ISA radiation p a tte rn s .................................................................
87
5.7
IS A transceiver front-end co n fig u ratio n ...............................................................
89
5.8
IS A tran sceiv er front-end radiation p erform ance...............................................
89
5.9
T w o -w ay com m unication link u sing IS A transceiver fro n t-e n d ....................
90
.1
V ario u s feed structures u sed fo r antenna a rra y s ................................................
91
6.2
L ocking bandw idth results for a typical X -band Gunn IS A ............................
94
6.3
Linear n etw o rk o f o sc illa to rs...................................................................................
96
6.4
2x2 X -band active array co nfigurations................................................................
97
6.5
R adiation p attern s o f 2x2 square a rra y .................................................................
99
6 .6
R adiation p attern s o f 2x2 diam ond a rray.............................................................
100
6 .7
B eam steering p attern s o f th e 2x2 sq u are a rra y .................................................
108
6 .8
P a tte rn s o f 1x4 E- and H -plane b eam steering a rra y s......................................
Ill
6
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
L IS T O F T A B L E S
T able
1.1
1
Page
Transm ission line com parisons................................................................................
6
.2 Soft m icrow ave substrate p ro p e rtie s.....................................................................
11
1 .3
H ard m icrow ave substrate prop erties at 25 ° C ...................................................
12
1.4
H igh s, substrate p ro p e rtie s.....................................................................................
13
1.5
H igh thermal conductivity c e ra m ic s......................................................................
14
1 .6
M onolithic integrated circuit substrate properties..............................................
14
4.1
Typical com m ercial V CO specifications...............................................................
51
5.1
G u nn-integrated ISA operating frequencies, pow er vs. patch d iam ete r
83
O perating frequency, p o w er and efficiency o f each antenna in a rra y
98
6
.1
6 .2
6
B eam steering d em o n stratio n s................................................................................
103
.3 Frequency distribution in scanning dem onstration.............................................
107
6 .4
L inear array frequency distributions for beam steerin g ....................................
110
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
1
CH A PTER I
IN T R O D U C T IO N T O M IC R O W A V E IN T E G R A T E D C IR C U IT S
1.1
I n tr o d u c tio n .
active
[1
This d issertation introduces a novel an tenna configuration used for
] and integrated
arrays. [3 ]
[2
] antenna applications such as spatial p o w er com bining phased
T h e investigation in [3] represents th e first spatial p ow er com bining beam
steering d em o n stratio n using active antenna sources w hich are injection locked only
th ro u g h m u tu al coupling w ithout a n y external sources and steered w ith o u t conventional
p h ase shifters.
T h e scope o f this stu d y is very b road in th a t it involves m any areas o f
m icrow ave engineering such as circuits, solid-state devices, m aterials, planar antennas,
p o w e r com biners and phased arrays.
A short review o f each to p ic is included when
necessary.
A lthough th e con cep t o f active integ rated antennas has been around for several decades,
little p ro g ress has been m ade fo r practical applications.
Recently, m any possible
applications hav e stirred m uch in terest in the com m ercial m arketplace. T his is due to the
fact th a t activ e antennas are fully integ rated in very small, com p act packages and are ideal
d o p p ler sensors. A s devices and integration techniques m ature, perform ance and cost will
d ictate th eir u s e in com m ercial applications.
1.2
M ic ro w a v e s
M icrow ave circu its are an integral p art o f m odern society.
Initially
im plem ented in m ilitary radar, m icrow aves have been used in space, scientific and
com m ercial applications.
The term microwave refers to th a t part o f the frequency
spectru m fro m 300 M H z to 300 G H z. M icrow aves have been im plem ented for point-topoint com m unication links, wireless com m unication, m otion sensing fo r security systems,
rem o te sensing and im aging, selective heating o f tu m o rs in m edicine, spectroscopy for
m aterial identification, and police radar. The use o f "m icrow ave" ovens at 2.45 G H z has
m ade th e term a household word. T h e frequency spectrum is divided into m any different
bands. T hese ban d s an d som e com m on applications are show n in F igure 1.1. T h e federal
co m m unications com m ission(FC C ) allocates frequency ranges and specifications for
different
ap plications
in
the
U n ited
States
including
television,
radio,
satellite
com m unications, cellular phone, police radar, burglar alarm s, navigation beacons, etc.
P erfo rm an ce o f each application is strongly affected by th e atm ospheric ab sorption curve
sh o w n in F ig u re 1.2. F o r example, a secure local area netw o rk w ould be ideal at 60 G H z
du e th e high atte n u a tio n caused by th e O 2 resonance.
The journal model for this dissertation is IEEE Transactions on Microwave Theory and Techniques.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2
G am ma Rays
X-Rays
Ultra-Violet
Visible
Infra-Red
--~~
~~
~~
3 00 GHz
mm
40 GHz
3 0 0 MHz-
Ka
30 GHz
26.5 GHz
30 MHz
20 GHz
18 GH:
Ku
3 MHz
10 GHz
GHz
0.3 MHz-
GHz
6 GHz
GHz
3 GHz
30 KHz
GHz
3 KHz
GHz
0.5 GHz
30 Hz
DC
DC
IEEE R a d a r
Bands
New
Designations
Figure 1.1. C om m on bands in the frequency spectrum .
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
3
FREQUENCY (GHz)
150
300
100
•Q
2
O
1.0
0.10 - S e a Level
HjO
<
4 km
0.01
HjO
0.001
W AVELENGTH (millimeter*)
F igure 1.2. A tm ospheric absorption as a function o f operating frequency.
A s m ore applications spring-up, overcrow ding and interference at low er frequency bands
pushes applications tow ard higher operating frequencies. H igher frequency o p era tio n has
several advantages including:
1.
L arger instantaneous bandw idth for g reater transfer o f inform ation.
2.
H igher resolution for radar and m ore detailed im aging and sensing.
3.
R educed dim ensions fo r resonant antennas.
4.
L ess interference from nearby applications.
A t higher frequencies, how ever, devices perform poorly and require m o re expensive
m aterials w ith com plicated fabrication m ethods.
Furtherm ore, high frequency circuit
m odels tend to be over-sim plified and inaccurate. As these technologies m ature, operating
frequencies will increase w ith o u t com prom ising system perform ance.
F rom th e beginning, m ost m icrow ave com ponents w ere designed using the w aveguides
and coaxial cables show n in Figure 1.3.
T oday coaxial cables are still used as
interconnects du e to their flexibility and good perform ance.
Circular and rectangular
w aveguides, o n the other hand, a re b est suited for high pow er and/or low -loss applications
b ut w o rk ov er a specified bandw idth w ith a low cu t-o ff frequency. W aveguides are also
heavy, bulky a n d require costly high-tolerance machining. In spite o f g o o d perform ance.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
4
w aveguides and coaxial configurations are not suited fo r im proving reproducibility,
reducing size and w eight, and low ering overall system costs. F urtherm ore, m odem trends
have pushed m icrow ave circuits to w ard m iniaturization, im proved reliability and greater
functionality. Such im provem ents require a com plete change in to p o lo g y and m ethods o f
fabrication.
Starting from conventionally assem bled coaxial and w aveguide system s, hybrid m icrow ave
in teg rated circuits(M IC ) evolved.
M ethods o f integration include hybrid M IC s and
m iniature hybrid M ICs. T hese hybrid techniques have since evolved to m onolithic M ICs
(M M IC ). T hese m ethods achieve g reater functionality and reproducibility by using p h o to ­
lithographic techniques for circuit fabrication. T hese approaches, how ever, require planar
transm ission lines to replace w aveguides and coaxial cables. Planar transm ission lines are,
for the m ost part, open structures w hich tend to have low er Q -factors and higher crosstalk
betw een lines. These transm ission lines depend on th e type o f substrate used and th e type
o f solid -state devices available.
R ectangular
C ircular
Elliptical
Coaxial Cable
Figure 1.3. Coaxial and w aveguide transm ission lines.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
5
T h e overall characteristics o f th e integrated circuit design depends on th e several m ajor
facto rs including:
1.
2.
3.
1.3
M IC transm ission line configurations:
a. M o d e o f propagation (i. e. Z o, Xg, dispersion, etc.)
b. R adiation and conductor losses.
c. In teg ratio n o f series/shunt devices and biasing n e tw o rk designs.
S ubstrate M aterials:
a. D ielectric losses and P ropagation effects.
b. M echanical strength and heat expansion.
c. M eth o d s o f fabrication.
Solid-S tate D evices:
a. M icro w av e com ponent function (i. e. sw itching, tuning, mixing,
am plifying, oscillating etc.).
b. O p eratio n Frequency, pow er and conversion efficiency.
c. M aterials and m ethods o f fabrication available.
M I C T ra n s m is s io n L in e s
M IC transm ission lines consist o f m etallized circuit
p attern s supported by a dielectric substrate as show n in Figure 1.4. T h ese planar lines can
be
fabricated
using
photolithographic
techniques.
P hoto lith o g rap h y
reproducibility and allow s m ass-production o f integrated circuits.
im proves
T he circuit pattern
layout used to tran sp o rt energy th ro u g h out a circuit determ ines m any characteristics o f the
transm ission line.
U nlike th e T E M m o d es o f coaxial lines, planar M IC lines are often non-T E M .
The
characteristics o f som e M IC lines, how ever, can be approxim ated by assum ing quasi-T E M
propagation.
Q uasi-T E M lines include m icrostrip, inverted m icrostrip, tra p p e d inverted
m icrostrip, suspended stripline, coplanar w aveguide and coplanar strips.
N o n -T E M M IC lines include slotline, finline and imageline. W aveguides are non-T E M in
n a tu re but unlike w aveguides, M IC lines suffer from dielectric and radiation losses and
cro ss-talk betw een lines in a d ense circuit. T hese transm ission lines can n o t handle highp o w e r like w aveguides, b u t th ey are m ore th an suitable fo r low -to -m ed iu m p o w er
applications.
The redu ctio n in size, w eight and overall cost offsets th e slight drop in
perform ance. A qualitative com parison o f these transm ission lines is given in T able 1.1.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
6
T able 1.1. Transm ission L ine C om parisons. [4 ]
Transmission Line
Useful Freq.
Range
(GHz)
Impedance
Cross-Sectional
R ange(Q )
Dimensions
Rectangular Waveguide
<300
1 0 0 -5 0 0
Coaxial Line
<50
1 0 -1 0 0
< 10
1 0 -1 0 0
< = 100
1 0 -1 0 0
Strip Line
Microstrip Line
VI
2 0 -1 5 0
Fin Line
II
V
V~l
20 - 400
o
il
V
o
Suspended StripLine
Q Factor
Power
Active Device
Potential for LowCost
Production
Rating
Mounting
Moderate to Large
High
High
Easy
Poor
Moderate
Moderate
Moderate
Fair
Poor
Moderate
Low
Low
Fair
Good
Small
Low
Low
Easy
Good
Small
Moderate
Low
Easy
Fair
Moderate
Moderate
Low
Easy
Fair
Slot Line
<=60
60 - 200
Small
Low
Low
Fair
Good
Coplanar Waveguide
< = 60
4 0 -1 5 0
Small
Low
Low
Fair
Good
Image Guide
<300
2 0 -3 0
Moderate
High
Low
Poor
Good
Dielectric Line
<300
2 0 -5 0
Moderate
High
Low
Poor
Fair
CPW
Microstrip
CPS
HP
Suspended
Stripline
Sl o t l i ne
Inverted
Microstrip
T r a p p e d Invert ed
Mi c r o s t ri p
I m a g e Li ne
T r a p p e d Ima g e Line
Fin Line
Di e l e c t ri c Gui de
I n s u l a t e d I m a g e Line
F igure 1.4. M icro w av e integrated circuit (M IC ) transm ission lines.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
7
T h e microstrip line [5
M M IC s.
,6
] has been the m ost used planar transm ission line in M IC s and
Synthesis and analysis form ulas are well docum ented and m any discontinuities
have b een characterized.
C om m ercial program s are available to use th ese m odels for
circuit prediction and optim ization.
D evices can be integrated in series using planar
p ack ag es but shunt connections require drilling through the substrate in M IC s and via-hole
p ro cessin g in M M IC s.
C om pared to the other open structures, m icrostrip is a proven
tech n o lo gy w hich has higher po w er handling, lower loss and is easily packaged. [7 ]
In spite o f m icrostrip's advantages, o ther transmission lines offer viable alternatives to
conventional m icrostrip integration. A derivative o f microstrip, inverted microstrip (IM )
[8
,9 ] line uses a m etallized circuit pattern suspended with spacers o v er a g ro u n d plane
su p p o rt. U nlike m icrostrip, th e ground plane has been rem oved from th e back-side o f the
su b strate.
Since m ost o f th e transm ission occurs in air below the strip, th e guided
w av elen g th and characteristic im pedance o f inverted m icrostrip are larg er than a sim ilar
line usin g conventional m icrostrip. Series devices are connected ju st as in m icrostrip w hile
shunt devices do n o t require drilling through the substrate. This characteristic allow s n o n ­
destru ctiv e device testing an d position optim ization in inverted m icrostrip.
H ow ever,
in v erted m icrostrip is very prone to surface m ode excitations w hich w o u ld cause
considerable crosstalk in a d en se circuit. Thinner substrates and red u ced strip -to -g ro u n d
height reduces th e excitation o f surface modes but limits its usefulness.
Smaller
dim ensions place higher to lerances on spacers used to separate su b strate and ground
su p p o rt.
A further m odification to m icrostrip, the trapped inverted microstrip (TEM) [10 ] line uses
a channel to cho k e o u t surface w ave m odes which may otherw ise p ro p ag ate along the
inv erted m icrostrip configuration. This m odification raises the characteristic im pedance o f
T IM and relaxes th e tolerance requirem ent o f the spacers used in inverted m icrostrip. T he
en clo su re allow s th e use o f thicker substrates and larger g ro u n d -to -strip separations
w ith o u t surface w av e effects. Shunt and series devices are connected ju s t as in inverted
m icrostrip.
T he channel allow s the use o f larger strip-to-ground spacing and provides
m o re m etal fo r im proved heat dissipation in active applications.
Suspended stripline (SS) [11 ] is the last o f the m icrostrip derivatives. Sim ilar to IM, the
b ack -sid e ground plane is rem oved from the substrate. The strip and su b strate are totally
enclosed and oriented along the H -plane o f a rectangular w aveguide.
O peration
frequencies are lim ited to avoid excitation o f waveguide m odes. A lthough the operation
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
bandw idth is lim ited, this variation o f m icrostrip achieves low er losses than all o th e r tw oco n d u cto r M IC lines.
A lthough series device integration is straightforw ard, shunt
insertion is sim ilar to EM. T he characteristic im pedance and guided wavelength depend on
th e w aveguide dim ensions, su bstrate dielectric constant, substrate thickness, strip width
and position in th e waveguide.
Coplanar Waveguide (C PW ) [ 1 2 ] is quite different from microstrip and its derivatives.
R eferred to as a uniplanar line, C P W has both the co nductor and ground plane o n a single
side o f th e substrate.
Drilling is n o t required for shunt connections and both shunt and
series devices can be integrated using planar packages.
slo ts etched o n a single-sided substrate.
Propagation is achieved via tw o
The im pedance and guided w avelength
characteristics depend on th e dielectric constant, substrate height, slot separation and slot
w idth.
The stru ctu re supports even and odd-m ode propagation when th e g round planes
o n eith er side o f th e cond u cto r are at different potentials. Even mode propagation can be
cho k ed out by m aintaining these ground planes at the sam e potential using bridges o r
bond-w ires. F ailu re to com pletely choke o u t the even m ode causes an inductive effect on
stan d ard o d d -m o d e operation.
Unlike m icrostrip, packaging often causes parallel plate
m o d e problem s fo r C PW w hen th e package walls are to o close to th e circuit.
This
problem is relieved som ew hat w ith a grounded version o f CPW . A further m odification o f
C P W is the channelized version w hich parallels w hat T IM does for IM.
Coplanar Strips (C P S ) [13 ] consists o f tw o lines etched on the dielectric substrate. CPS
is uniplanar and c a n be integrated w ith planar devices in either series or shunt connections.
T his line suffers from higher radiation losses.
The characteristic impedance and guided
w avelength d e p en d o n th e strip widths, separation, substrate thickness and dielectric
constant. Similar to o ther uniplanar lines, enclosure proxim ity affects perform ance.
Slotline [14 ] is a non-T E M uniplanar M IC line using a single slot etched on a dielectricsu p p o rted layer o f metal w ith o u t a backside ground plane. T he characteristic im pedance
and guided w avelength depend on th e dielectric constant, substrate height and slot width.
Series and shunt devices can be incorporated for integration. Shunt m ounting is straight­
fo rw ard while series insertion requires some modifications.
L ow values o f characteristic
im pedance (< 6 0 Q ) are often inpractical. Low values o f Z0 require very narrow slots on
thin substrates w ith high dielectric constants..
F or circuit interconnections, slotline also
suffers from high radiation losses and packaging difficulties.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
9
Finline [15 , 1 6 ] is a com bination o f w aveguide and slotline.
A slot is etched on a
dielectric su p p o rted lay er o f metal and oriented along th e E-plane o f a rectangular
w aveguide. T h e w aveg u id e encloses either side o f the slotline. This structure com bines
the high perform ance o f w aveguides with the integration capabilities o f uniplanar lines.
H ow ever, th e stru c tu re ten d s to be ju st as bulky and heavy as standard w aveguide.
A
sh o rt is needed at th e ed g es o f the substrate which is o ften disturbed for device biasing.
Finlines generally have ab o u t a third o f the loss o f conventional m icrostrip and are useful
up to 100 GHz.
Dielectric guide [17 ], imageline [18 ] and insulated imageline [19 ] use a dielectric for
w av e p ropagation.
Im agelines use only a dielectric o v e r a ground plane.
This is in
co n trast to all previously m entioned transm ission lines which require som e metallic
co n d u cto r fo r propag atio n . P o w er transm ission occurs through th e dielectric. It exhibits
low loss and is extrem ely useful at frequencies above 100 G H z.
H ow ever, integration
w ith devices an d biasing m ust be addressed as well as high radiation at discontinuities.
C rosstalk w ith o th e r lines is another issue which m ust b e overcom e before it can be used
for M ICs.
An enclosed version o f this line, trapped imageline, attem pts to co rrect
crosstalk problem s in d en se circuits.
T h e p roperties o f th e M IC
lines described above are greatly dependent on the
characteristics o f th e supporting dielectric substrate.
T he substrate greatly affects
electrom agnetic w ave p ropagation and its associated transm ission losses.
The substrate
m aterial also determ ines th e process o f fabrication and several o th er characteristics critical
for M IC s and M M IC s. D ifferent dielectrics and their properties are discussed below.
1.4
Dielectric Substrates
The dielectric substrate supports each o f the M IC lines
discussed above and its prop erties affect the overall perform ance o f th e ICs.
each
ty p e
of
transm ission
line
configuration,
each
substrate
m aterial
Just as in
posseses
characteristics w hich m ay m ake it b etter suited for an application. F or instance, antennas
require lo w er dielectric co nstants to ensure good radiation w hile higher dielectric constant
m aterials w ould en sure sm aller circuit size.
T he cost, frequency o f operation, loss-
tangent, m echanical stren g th and surface finish o f a m aterial are all im portant for hybrid
M IC designs. F o r hybrid M IC s, there is a choice b etw een organic and inorganic materials.
O rganic m aterials include a w ide range plastics o r soft substrates.
Inorganic or hard
substrates include ceram ics, m onocrystalline m aterials, ferrom agnetic m aterials and
sem iconductors.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
10
In general, th e e r an d p r should be hom ogenous (independent o f position) and isotropic
(independent o f p ro p ag atio n direction). Furtherm ore, th ese p a ram ete rs should have very
small v ariation w ith tem p eratu re to ensure circuit stability.
T h e substrate therm al
conductivity sho u ld be high enough to ensure efficient rem oval o f heat from p ow er
transistors, a tte n u a to rs and loads in high p o w er applications. In hig h -p o w er applications,
a high b reak d o w n voltage is also desirable.
The therm al expansion coefficient o f the
su b strate m aterial should b e similar to that o f the deposited co n d u cto rs and housing to
w ith stan d te m p e ra tu re fluctuations and im prove reliability.
T he m aterial m ust allow
drilling, cuttin g a n d m achining for easy w orkability and low er p ro d u c tio n costs.
Also
im p o rtan t is a g o o d surface finish (o.os to o.i nm) to ensure good c o n d u c to r adhesion and
red u ce co n d u c to r loss. T hese substrates should be available in sh eets o f various sizes to
acco m o d ate large circuit layouts and integration o f many com ponent functions.
P lastics o ften hav e low loss tangents and lo w er. Som e typical so ft substrate m aterials
include
p o lystyrene,
polyolefin
and
polytetrafluroethylene(P T F E )
substrates.
To
o v erco m e m echanical instability and cold flow problems, these su b strate s are reinforced
w ith either glass fibers o r ceram ic particles. T he form er m aintains th e dielectric constant
but ad d loss w h ile th e latter tends to increase the dielectric constant.
T w o com m on
com binations fo r lo w er include w oven PTFE/fiberglass and m icrofiber P T FE /fiberglass
w hile a high er so ft substrate is ceramic PTFE.
T he choice depends on electrical,
m echanical, chem ical and therm al conditions encountered during fabrication as w ell as its
intended w o rk in g environm ent.
c o n stru ctio n m ethods.
The m aterial m ust also be am enable to low -cost
H ow ever, soft substrates exhibit som e undesirable characteristics
including creep u n d e r stress (solid-state deform ation), high coefficients o f therm al
expansion and lo w therm al conductivity.
L o w therm al conductivity coefficients create
therm al m anag em en t problem s in high p o w e r applications w hile su b strate deform ation
p resents p roblem s fo r clam ping o r other stresses on the substrate. H ig h therm al expansion
coefficients cau se fatig u e at jo in ts betw een the substrate and housing.
R einforcing glass
fibers can red u ce d efo rm atio n and thermal expansion. Typically, how ever, this reduction
o ccu rs in th e x-y p lane o f th e substrate w hile the r-direction's therm al expansion rem ains
unchanged cau sin g circuit reliability and durability concerns.
ch aracteristics o f w id e range o f organic substrates.
C eram ic
su b stra te s
include
aluminum
oxide
[2 0
T able
1.2 lists the
]
(AI2 O 3 ),
sapphire,
oxide(T iO x), beryllium oxide (B eO ) and alum inum nitride(A IN ).
quartz,
titanium
AI 2O 3 o r alum ina is an
inexpensive m aterial used fo r thin and thick-film hybrid IC s. T he g ra d e o f purity (0.95 to
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
11
0 .9 9 5 ) determ ines th e dielectric constant (9.5 to 10.5), low loss tangent (tan S) (0.0001 to
0 .0 0 0 5 ) and surface finish (
to
Alumina exhibits high er>, low deform ation, g o o d
therm al expansion coefficient (i.e. m atches base metals and GaAs) and high insulation
resistance.
T able 1.2. Soft M icrow ave Substrate Properties. [2 0 ]
Laminate/Substrate
Dielectric Constant
X-band
Loss Tangent
X-band
Dimensional
Stability
Chemical
Resistance
Temperature
Range
PTEE Unreinforced
2.1
0.0004
Poor
Excellent
-27 to +260
PTFE Glass Woven Web
2 .1 7 -2 .5 5
0.0009 0.0022
Excellent
Excellent
-27 to +260
PTFE Glass Random Fiber
2 .1 7 -2 .3 5
0.0009 - 0.0015
Fair
Excellent
-27 to -2 6 0
PTFE Q uartz Reinforced
2.47
0.0006
Excellent
Excellent
-27 to +260
Ceramic PTFE Composite
10.2
0.002
Excellent
Good
-15 to +170
Cross Linked Polystrene
2.54
0.0005
Good
Good
-27 t o +110
Cross Linked Polystrene/
2.62
0.001
Good
Good
-27 t o +110
Glass Reinforced Cross Linked Polystrene
Q uartz
Cross Linked Polystrene/ Woven Quartz
2.6
0.0005
Good
Good
-27 t o +110
2.65
0.0005
Good
Good
-27 to + 110
Cross Linked Polystrene/ Ceramic Powder
Filled
Teflon/Unreinforced (unclad)
3 -1 5
0.0005 - 0.0015
Fair to Good
Fair
-27 t o +110
2.1
0.0004
Poor
Excellent
-27 to +260
Teflon/Glass Reinforced
2.55
0.0015
Good
Excellent
-27 to -2 6 0
Teflon/Ceramic Reinforced
2.3
0.001
Fair to Good
Excellent
-27 t o -2 6 0
Teflon/Quartz Reinforced
2.47
0.0006
Good
Excellent
-27 to +260
Tenon/Ceramic Filled
10.3
0.002
Good
Excellent
-27 to -2 6 0
Polyphenylene Oxide (PPO)
2.55
0.0016
Good
Poor
-27 t o -1 9 3
Irradiated Polyolefin
2.32
0.0005
Poor
Excellent
-27 to -1 0 0
Irradiated Polyolefin/
2.42
0.001
Fair
Excellent
-27 t o -1 0 0
Glass Reinforced Polyolefm/Ceramic
Powder-Filled
3 -1 0
0.001
Poor
Excellent
-27 to *100
Sapphire (aka corundum o r alpha alum ina) is the single crystal form o f A]20
3
and it also
exhibits low loss tangent, high er, good surface finish, high insulation resistance and a
therm al expansion coefficient w hich m atches well w ith GaAs.
H ow ever, sapphire is
anisotropic in perm ittivity w hich varies from 9.4 to 11.5 for different directions o f
propagation. Fused quartz offers low loss tangent, relatively low sr, g o o d surface finish,
dielectric stren g th and repeatability.
th ro u g h and costly.
H ow ever, fused quartz is brittle, difficult to drill
T ab le 1.3 lists the properties o f these three popular m icrow ave
substrates m aterials.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
12
Table 1.3. H ard M icrow ave Substrate Properties at 25 °C.[ 20 ]
Property
Alumina
Sapphire
Fused Quartz
Dielectric Constant @ 1 0 GHz
9.8
11.5
3.78
Loss Tangent @ 10 GHz
0.0002
<0.0001
0.0001
Resistivity (/>cm )
10“
10“
10“
100
Dielectric Strength (kV/mm)
7.9
48
Temp. Cocff. o f e, (ppm/°C)
+136
■
-
Coefficient o f Thermal Expansion (ppnVC)
6.7
5.3
0.55
Thermal Conductivity (li'm K )
37
46
1
Melting Point (°Q
2030
2053
-
Density (gm/cm’)
3.9
3.97
2.203
Bending Strength (N/mm2)
245
-
-
Modulus o f Elasticity (kN/tnm2)
340
413
72
260
8
Tensile Strength (kg/mm2)
21
Rockwell Hardness (1 5N scale)
97
-
-
Flexural Strength (k ^ m m 2)
34
64
7
Surface Finish p a n )
<0.1
<0.1
-
Grain Size (/an)
< 1.5
-
■
Ceram ics w hich have very high Sr values (15 to 240) allow m iniaturization o f circuits
(specially at lo w frequencies) and often provide low er loss and b etter tem p eratu re stability.
P roperties o f several high Sr m aterials are listed in Table 1.4. M aterials developed
specifically fo r dielectric reso n ato rs (D R ) are generally based on titanates. T h ese titanates
provide very high quality reso n ato rs w ith low therm al expansion and tem perature
coefficients.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
13
T able 1.4. H igh er S ubstrate Properties. [ 20,2 1 ]
Dielectric Designation Code
CF
CB
CD
CO
NR
Dielectric Constants
21.6 ± 0 .6
29 ± 0 .7
37 ± 1
67.5 ± 2
152 ± 5
19.3 + 0.6
29 ± 0 .7
37 ± I
67.5 ± 2
1 5 2 -5
M aximum Loss Tangent
0.0003
0.0004
0.0004
0.0008
0.001
Temp. Coefficient (ppm/K)
-1 7 + 5
-10 ± 2
-30 ± 3
-23 ± 3
-1700 ± 120
Minimum Resistivity (Q-cm)
10“
10‘“
10“
10“
10“
Thermal Expansion (ppm/K)
7.8
6.3
5.8
9
10
Flexural Strength (kgf7cm‘)
1492
1590
1535
1910
1548
Density (g'cm 1)
3.89
4.32
4.75
5.56
3.91
Heat Conductivity (W/cm-K)
0.06
0.033
0.018
0.02
0.047
Specific Heat (J/g-K)
0.8
0.66
0.58
0.48
0.77
Maximum W ater Absorption (%)
0.01
0.01
0.01
0.01
0.01
F o r high therm al conductivity, beryllium oxide and alum inum nitride are available. B eO is
very toxic an d co stly which has limited its use b u t provides low er, good dielectric
strength, and high b u lk resistivity. Aluminum nitride co m p ares favorably w ith beryllia plus
it maintains a m o re constant therm al conductivity coefficient at higher tem peratures.
H ow ever, large single crystals o f A IN are not readily available.
A IN pow der m ust be
processed in an oxidizing atm osphere w ith com pounds such as yttrium oxide, yttrium
flouride and calcium carbide.
This processing low ers th e therm al conductivity but
m etalization is difficult and long term reliability is questionable.
overcom e som e m etalization problem s encountered.
M ulti-layer ceram ics
B eO and A IN characteristics are
listed in T able 1.5.
Overall, hard su b stra te s allow precise dim ensional c o n tro l o f circuit layout p atterns and
tend to have lo w e r therm al expansion coefficients and higher coefficients o f conductivity
th an soft su bstrates.
W elding o r device soldering is done easily on ceram ic substrates.
T hese p ro p erties m ake ceram ics m ore useful th an
tem perature an d high-stress applications.
plastics fo r high-pow er,
high-
H ow ever, unlike plastics, ceram ics are
expensive and not easily am enable to the construction o f intricate circuit periphery, holes,
slots, m achined depressions, etc.
Ceram ics are available in m uch sm aller substrate sizes
than plastics w hich m akes integration o f several functions m ore difficult.
In the end, the
choice b etw een h ard and soft substrates in hybrid M IC s depends on co st and perform ance
requirem ents discu ssed previously.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
14
T able 1.5. H igh Thermal C onductivity C eram ics.[20]
Property
Dil coin c Constant
® 1 MHz
@ 10 GHz
Loss Tangent
® 1 MHz
@ 10 GHz
Resistivity (Q-cm)
BeO
AIN
6.5
6.7
8.9
0.0004
0.004
10"
0.0005
>5
—
—
10"
Dielectric Strength (kV/mm)
9.6
Coefficient Thermal Expansion (ppm/°C)
9
4.4
Thermal Conductivity (W/mK)
260
1 40-230
Melting Point (°C)
2330
2400
Density (gm/cm!)
2.9
3.26
Modulus o f Elasticity (kN/mm3)
345
3 0 0 -3 1 0
315
Bending Strength (N/mm3)
200
Rockwell Strength (15N scale)
91
94
Flexural Strength (MPa)
241
28 -320
Surface Finish (mm)
< 0.4
< 0.5
Grain Size (mm)
9 -1 6
5 -1 0
M onolithic M IC s require substrates from w hich devices and circuits can b e 'grow n'
th ro u g h fabricatio n processes. Different m aterials available for M M IC s include: G allium A rsenide(G aA s), Indium -Phosphide(InP) and Silicon(Si). T he properties o f each different
m aterials u sed are listed in T able 1.6.
T ab le 1.6. M onolithic Integrated C ircuit Substrate Properties. [19]
Property
Silicon
Silicon-on-Sapphire
GaAs
InP
Semi-Insulating
No
Yes
Yes
Yes
Resisitivity (Q-cm)
10! - 105
> 1014
107 - 10’
- 107
Dielectric Constant
11.7
11.6
12.9
14
Electrical Mobility1(cm:/V-scc)
700
700
4300
3000
Saturation Electrical Velocity (cm/sec)
9 x 10s
9 x 104
1.3 x 107
1.9 x 107
Radiation Hardness
Poor
Poor
Very Good
Good
Density (g e m 3)
2.3
3.9
5.3
4.8
Thermal Conductivity (W/ctn-°C)
1.45
0.46
0.46
0.68
O perating Temperature (°C)
250
250
350
300
Handling
Very Good
Excellent
Good
Poor
Cost (2’ diameter)
S15
S50 (base sapphire)
S100
S3 00
At 10' 'em * doping level
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
15
G aA s is th e m ost prom inent high-frequency substrate used in M M ICs. E lectrons in G aA s
have high saturation velocities and mobilities achieved w ith relatively low electric fields.
T h e small transit tim es for electrons flow ing across a device allow high-frequency
operation. G aA s can be prepared in a sem i-insulating form w ith a resistivity o f ~ 1 0 9n-cm
and ar = 1 2 .9 . Sem i-insulating GaAs is used for m icrostrip line interconnections allowing
th e fabrication o f very densely-packed m onolithic circuits. A lthough, silicon has saturation
velocities low er than G aA s it is a very m ature fabrication technology used exclusively at
lo w er frequencies. U nlike sem i-insulationg G aA s, semi-insulating Silicon is to o lossy for
high frequency applications. O ne solution has been to use the Silicon-on-Sapphire(SO S)
app ro ach to overcom e th e sem i-insulating loss problem s encountered w ith silicon.
As
listed earlier, sapphire has m any attractive qualities for circuits including surface finish, low
lo ss and g o o d therm al conductivity. I f one accounts fo r its anisotropic nature, Si circuits
can aligned along th e sapphire substrate C -axis to obtain an sr = 11.6 and a loss tan g en t o f
0 .0 0 2 . This com bination has regenerated interest in silicon bipolar tran sisto r technology
fo r low er m icrow ave frequencies.
T h e re are m any different techniques for deposition o f metallic an d /o r dielectric layers.
R egardless o f w hether on e u ses organic o r inorganic substrates o n e m ust u se one o r
several o f the following fo r th e conductor: silver, copper, gold, aluminum, etc. D ielectric
films used include SiO, SiC ^, S ij N ^ Ta 2 C>5 , etc. Resistive films u sed include N iC r, Ta,
Ti, TaN , C erm et, G aA s, etc.
All o f these materials are used w ith different fabrication
p ro cesses to provide an arsenal w ith which to develop solid-state circuits and systems.
C learly, hybrid integration is not limited to a particular technology because it can use
previously fabricated devices. H ow ever, if th e materials, devices and fabrication m ethods
a re suitable fo r m onolithic im plem entation, great im provem ents in size and w eight can be
m ade over a hybrid circuit.
Furtherm ore, M M IC s avoid many parasitics w hich limit
hybrids as w ell as costly p o st fabrication tuning to substantially reduce co sts in mass
production.
1.5 Active Solid-State Devices A lthough limited in pow er ou tp u t, solid-state device
technology has m atured to provide repeatability, low -cost and g o o d perform ance. There
a re tw o - and three-term inal types o f solid-state devices. Typical packages for solid-state
d evices used in hybrid M IC s are show n in Figure 1.5. T w o-term inal devices available for
in tegration include G unn, IM P ATT, varactor, P IN and schottky-barrier diodes.
Three-
term inal transistor devices to consider include bipolar junction transistors(B JT ), metalsem iconductor field-effect transistors(M E SFE T ), hetero-junction bipolar transistors(H B T )
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
16
and high electro n m obility transistors(H E M T ).
These devices form the basic building
blocks for m ost circuit applications.
The diodes and tran sisto rs listed above can be used to provide voltage-controlled
reactance, negative-resistance and current-controlled reactance.
V oltage- o r current-
control o v er reactance o r negative-resistance provides the m echanism fo r developing
circuits w hich tune, sw itch, mix, am plify o r oscillate. The choice o f diodes o r transistors
determ ines th e cost, p o w e r output, o p erating frequency, D C -to-R F conversion efficiency,
stability, noise and biasing requirem ents. This choice is not alw ays clear cut o r obvious.
D iodes a re tw o-term inal devices th at require less complex loading and biasing schemes
than three-term inal transistors.
H ow ever, transistors can be used to pro v id e many
different functions such as sw itching,tuning, amplifying, etc. which w ould req u ire several
different typ es o f diodes.
D iodes reach higher operating frequencies w ith higher RF
o u tp u t p o w er but tran sisto rs have higher D C -to-R F conversion efficiencies and o p e ra te at
low er D C input levels.
F or in teg rated circuits, th e device chosen determ ines if hybrid o r m onolithic technology
best suits th e application.
A thorough understanding o f the difficulties in developing an
integrated circuit com ponent can shed light on the best possible road to follow .
For
integrated antennas, th e choice is further clouded by the introduction o f m ore constraints
related to th e antenna perform ance.
T he problem is m ore difficult w ith integrated
antennas b ecause biasing schem es, loading and matching circuits can adversely affect the
perform ance o f th e antenna. Furtherm ore, circuits attem pt to keep fields from radiating
aw ay w hile antennas try to optim ize radiation. W hen choosing betw een m onolithic and
hybrid
im plem entation, th e circuit o r the antenna perform ance m ay
com prom ised in the integration.
have to
be
T his has been a contributing factor in m aintaining
conventional approaches to transceiver designs and it m ust be overcom e befo re integrated
and active integrated antennas can be used widely.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
17
S c r e w Di ode
P a c k a g e Ty pe
Planar
Chip S t y l e
T ra ns ist o r
Figure 1.5. Typical solid-state device packages.
1.6
Microwave Integrated Circuits
By com bining solid-state devices, dielectric
m aterial and M IC lines, high-perform ance circuits for m any applications can be developed.
E ach M IC line provides a starting point tow ards basic elem ent designs fo r use in many
com ponents.
T hese com ponents include filters, couplers, transform ers, etc.
A basic
design elem ent for th e design o f these com ponents is a resonator. A t low er frequencies
( < 3 0 0MHz), lum ped inductor/capacitor (LC ) elem ents are suitable resonators.
As the
frequencies increase into th e m icrow ave range, such resonant elem ents are not easily
constructed and ten d to lose pow er due to radiation.
At m icrow ave frequencies, these resonant elem ents are developed using distributed
transm ission lines, sh o rt sections o f w aveguides o r dielectric cavities. Shorted o r opened
sections o f transm ission lines can serve as either inductors o r capacitors at a frequency
point in a circuit. O n ce the resonator configuration is established, several resonators can
be cascaded to develop a filter. Similarly, oscillators, amplifiers, active antennas and other
com ponents can be designed by imbedding the solid-state devices discussed in section 1.4
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
18
in to these reso n an t elements.
M ethods o f hybrid integration can be classified into hybrid
M IC s and m iniature hybrid M ICs.
H ybrid M IC s
integrate
pre-fabricated solid-state devices,
cap acito rs w ith the circuit layout pattern.
resistors,
inductors
and
M iniature hybrid M IC s fabricate inductors,
capacitors and resistors w ithin th e circuit layout pattern and later integrate solid-state
devices.
H ybrid M IC s require system assem bly and tend to be labor intensive in
p roduction.
H ybrid M IC s are often lim ited at higher frequencies by the effects from
device in tegration discontinuities and p ackage parasitics. M iniature hybrid M IC s alleviate
so m e o f th e problem s o f hybrid M IC s by reducing the circuit com ponents w hich require
m anual integration.
In M M IC s, assembly a n d integration discontinuities are avoided by
fabricating solid -state devices as well as capacitors, inductors and resistors w ithin the
circuit layout.
T h e M IC transm ission line an d device package plays a m ajor role in the
su ccess o f hybrid technology circuits.
F igure 1.6 show s schem atic differences in tw o-
term inal device in tegration b etw een several com m on planar transm ission lines.
M M IC s o v erco m e hybrid M IC integration problem s by sim ultaneously fabricating solidsta te devices an d circuits o n a single substrate. Since th e device and circuit share th e sam e
substrate, th ere a re n o bonding, hybrid integration discontinuities o r package parasitics to
ta k e into acco u n t.
H ow ever, th e developm ent cost o f M M IC s is usually very high
requiring a c c u ra te m odeling softw are and stringent fabrication standards. D ue to the small
dim ensions o f th e devices, special testing a n d troubleshooting techniques are needed to
determ ine w hich com ponents m eet specifications. D ue to small device-to-chip area, yield
(i.e. co m p o n en ts which m eet specifications) depends mainly o n th e success o f th e device
fabrication.
E rro rs in m odeling o r fabrication tolerances affect circuit perform ance and
increase system developm ent tim es.
com pensate fo r these errors.
Som e clever tuning schem es have been devised to
H ow ever, post-fabrication tu n in g not only defeats the
p u rp o se o f m onolithic im plem entation (i.e. avoid post-assem bly tuning) but also raises
c o sts and u ses u p valuable w afer space.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
F ig u re 1.6. S olid-state device integration with various planar transm ission lines.
T here are many consid eratio n s for monolithic fabrication o f circuits[22 ] w hich are
discussed w ith m icrostrip line in mind. A s stated previously, reducing circuit size, weight,
cost a re very im portant.
H ow ever, many problems arise as a result o f m iniaturization.
C oupling o r cro ssta lk b etw een adjacent lines limits the packing density o f m iniaturized
circuits. W ider spacings betw een adjacent lines increases circuit size w hile enclosed lines
increase costs.
etc.
O th e r problem s include packaging, circuit losses, device reproducibility,
F o r active applications (i.e. amplifiers, oscillators, etc.), energy which is not
co nverted to R F p o w e r is lo st to heat.
H eat generated m ust be rem oved efficiently to
m aintain perfo rm an ce and reliability.
Thermal conductivity is a m easure for heat
cond u ctio n w hich increases fo r thinner substrates, but thinner substrates require thinner
m icrostrip lines to m aintain th e sam e system im pedance level.
T hinner m icrostrip lines
suffer from higher resistive losses and greater parasitic capacitance to ground.
resistive losses and parasitic capacitances result in low er quality factors.
These
Such low Q-
values prevent the realization o f low -loss, narrow -band filters and low -noise oscillators.
O ther com plications arise from th e fabrication o f via holes fo r shunt co nnections which
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
20
lo w er yield. Finally, M M IC designs carry an expensive initial developm ent cost w hich can
only be reco v ered from large production quantities.
Depending on th e fabrication
m ethods, M M IC s can produce high-volum es and high-yields w ithout labor-intensive
tw eaking. T his is th e essence o f M M IC s using m icrostrip.
A lternative M IC configurations alleviate som e o f the problem s m entioned above b ut
introduce others.
In M M IC s, th e transm ission line, device and com ponent perform ance
m ust be accom plished thro u g h the properties o f a particular material and fabrication
m ethod. This constraint often limits the com ponent, device o r transm ission line optim um
perform ance.
F o r m ost circuit applications, M M IC s serve to provide very small reliable
circuits w ith im proved reproducibility.
H ow ever, w afer costs, substrate characteristics,
small p roduction quantities and the large real estate occupied by passive elem ents can
m ake the use o f hybrid M IC s m ore attractive for som e applications. O n e such application
m ay be in tegrated and active integrated antennas.
1.7 I n te g r a te d a n d A ctiv e I n te g ra te d A n te n n a s Similar to ICs, integrated antennas a re
th o se radiators directly integrated w ith solid-state devices to provide so m e m icrow ave
com ponent function.
Integrated antenna applications require substrates and antenna
configurations w hich allow efficient radiation.
Packaged devices cre ate d in Si o r G aA s
can be conveniently integrated within the antenna on substrates w hich optim ize the
antenna radiation perform ance. In hybrid M IC s, special low -loss substrates can be used to
m inim ize dielectric losses an d optim ize antenna efficiencies.
Packaging m ethods can be
im proved to reduce parasitics and integration discontinuities for higher frequencies o f
operation.
C om plete m onolithic im plem entation o f integrated antennas m ay not yield
optim um results because M M IC substrates tend to be o f higher er w hich reduces radiation
efficiency. A lso, antennas a re typically m uch larger than th e integrated so lid-state device
w hich w ould leave a large am o u n t o f valuable w afer real state un-used.
Integrated an tennas are th o se radiators integrated directly w ith solid-state devices such as
P IN , v aractor, o r Schottky-barrier diodes to
functions.
provide switching, tuning o r mixing
A ctive integrated antennas are th o se integrated antennas w hich use active
devices (G u n n diodes, FETs, etc.) to convert D C energy to RF p o w e r (i.e. oscillators,
am plifiers, etc.). Integrated and active integrated antennas have the potential o f reducing
the size, w eight and co st o f conventional transm itter, receiver and tran sceiver designs by
incorporating th e circuit com ponent functions at th e antenna terminals.
C onventional tranm itter, receiver o r transceiver system s use several different circuit
com ponents connected to an antenna via a transm ission line.
At th e ju n ctio n betw een
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
21
circuit and transm ission line, a transition attem pts to provide a sm ooth change in m ode o f
propagation.
Sim ilarly, th e transition from the transm ission line to the antenna transfers
the guided w ave to th e radiating m ode o f the radiator.
A lthough this m ethod allow s
separate optim ization o f antenna, transm ission line and circuit com ponent, it also limits
certain im portant factors.
T he transitions at the com ponent and antenna to transm ission
line increase th e size, weight and cost o f the overall system.
T he transitions also add
circuit com plexity an d introduce discontinuities which cause losses and limit frequency o f
operation. T hese lo sses add o n to the losses incurred in th e transm ission line w hich often
w orsens w ith increasing frequency.
C onventional application o f M IC s and M M IC s to
m iniaturize the circu it com ponents, transm ission lines and antenna only reduce these
losses.
In contrast, in tegrated antennas attem pt to eliminate these losses alto g eth er by
integrating solid -state devices at the antenna.
This integration, how ever, affects the
perform ance o f the a n ten n a and th e com ponent function. O ne obvious exam ple o f these
trade-offs is the use o f active integrated antennas for RF transm itter applications.
Unlike conventional transm itters, an active antenna integrates active devices directly at the
antenna term inals.
B y designing the antenna and oscillator on a single substrate, one
avoids transition/transm ission line losses from pow er distribution netw orks.
show s conventional and integrated antenna approaches fo r transm itters.
F igure 1.7
Ideal active
integrated antennas w o u ld provide goo d com ponent characteristics w ithout com prom ising
antenna perform ance. A lthough the concept is straightforw ard, successful im plem entation
o f this m ethod have b een difficult to achieve for several reasons.
A ctive integrated an ten n a design requires know ledge in several areas o f m icrow ave
engineering including solid-state devices, circuits and antennas.
E x p erts in circuits and
oscillator design seldom have antenna experience and vice-versa. Som e obvious trade-offs
o f th e active in teg rated antenna approach is that optim ization o f the oscillator and antenna
m ust b e m ade sim ultaneously o n th e same structure. This is som etim es self-defeating since
characteristics
perform ance.
w hich
im prove
circuit
oscillators
often
degrade
antenna
radiation
A ctive integrated antennas can be realized using tw o and three terminal
devices for RF p o w e r generation, low -cost sensors, decoys, m odulators, amplifiers and
sources for p o w er com bining.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
TRANSITION
transitio n
H
2 .\
!
r
1
I
ANTENNA
l
i
i
1
|
“
1
1
COMPONENT
TRANSMISSION LINE
( a)
ANTENNA
COMPONENT
( b)
I n t e g r a t e d An t e n n a A p p r o a c h
F igure 1.7. C o nventional and integrated antenna approaches fo r transm itters.
1.8 S y ste m A p p lic a tio n s A ctive and integrated antennas can b e used as an alternative to
conventional app ro ach es. Functions which w ould normally o c c u r in the circuit aw ay from
th e an te n n a (d etection, tuning, m odulation, mixing, amplifying, etc.) can be included with
th e a n ten n a to reduce size, w eight and cost. I f the com ponent and antenna pro v id e nearly
th e sam e perform ance a s that using a conventional approach, the reduction o f size and cost
w ould greatly enhance th e system ’s usefulness.
PIN in te g ra te d an ten n as can serve as m odulators a t the antenna term inals.
Such a
m o d u lato r w ould av o id long R F line losses. The am plitude m odulator(A M ) could b e set
up to transm itt inform ation using digital coding o r analog transm ission techniques.
Similarly, v aracto r-lo ad ed antennas can be used as frequency m odulators(F M ). Again, this
m eth o d w ould avoid lo n g RF lines to reduce carrier losses. F rom a receiver standpoint,
P IN s c a n be used a s frequency sw itches or as a squelch setting. V aractors ca n b e used as
a front-end filter fo r band o r channel selection as well as to increase a narrow band
an te n n a ’s operating bandw idth. In array applications, these solid-state devices can be used
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
23
to in crease th e usable bandw idth and/or provide am plitude o r phase tapering mechanism.
In p h ased-arrays, th e usable scanning range can be increased by tuning (o r de-tuning)
individual antenna elements.
A ctive antennas as oscillators m ake very inexpensive m icrow ave sources.
Given a DC
p o w e r supply and a m odulating circuit, such an inexpensive source may becom e a useful
decoy for electronic w arfare and jam m ing applications o r for FM com m unication links.
D ecoy system s have been incorporated using m onolithic techniques and conventional
tran sceiv er approaches.
F urther im provem ents could b e accom plished using an active
integ rated antenna approach. Com m unication applications norm ally require a transm itter
w ith low oscillator noise and high stability. A ctive antennas require further developm ent
to m eet th ese requirem ents
A ctive antennas are inherently go o d doppler sensors because radiating sources are
sensitive to d o p p ler return from moving objects. The antenna serves as a self-oscillating
m ixer to m ix in th e the local oscillating frequency w ith th e low -level frequency reflection
from th e m oving object.
detected.
B y extracting this doppler return, relative m otion can be
Such an application could greatly reduce the current size, w eight and cost o f
radar, R F identification system s, autom otive collision w arning system s, autom atic d o o r
openers, perim eter m onitoring system s and burglar alarms.
S ch o ttk y -d etecto r diodes can be integrated with the antenna for R F detection.
The
anten n a selectivity determ ines th e RF band o r channel being m onitored. This detector can
be u se d fo r m any different m onitoring and im aging applications.
This approach also
applies directly to rectifying antenna w o rk for m icrow ave p ow er transm ission.
To
illustrate, p o w e r lines w ould n o t be an effective m ethod o f distributing p o w er in space. It
w o u ld also be to o costly fo r each satellite to have its o w n generating plant. H ow ever, a
centrallized p o w e r source could distribute p ow er to
m icrow aves using large antennas.
its neighboring satellites via
Beam s w ould focus o n th e receiveing satellite which
w o u ld have a large rectenna array to convert the RF to DC.
G ranted positioning
equipm ent and com puter m onitoring system s w ould be required to m aintain beam taper
and phasing b u t such com plex autom ated system s already keep F-16s up in the air.
As
device conversion efficiencies im prove and fabrication co sts are reduced such a system
m ay b eco m e a reality. A sim ilar application involves spatial p o w er com bining.
A ctive antennas are ideally suited for spatial p ow er com bining sources in the millimeter
and sub-m illim eter bands. A t these high frequencies, solid-state devices produce very little
p o w e r and tend to be very inefficient. T o com pensate, p o w e r com bining techniques have
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
24
been developed.
spatial level.
P o w er com bining can be accomplished at either the chip, circuit or
Spatial pow er com bining techniques are specifically su ited for active
antennas because they incorporate many low -pow er radiating sources.
T echniques are
applied to ensure coherent operation o f each source to obtain a large am ount o f com biner
RF pow er.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
25
CH A PTER H
T H E P R O B E -F E D IN V E R T E D S T R IP L IN E A N T E N N A
2.1
I n tr o d u c tio n
O f th e transm ission lines discussed in chapter I, m icrostrip is, by far,
th e m ost widely used for circuits and antennas.
It has m any desirable qualities such as
com pact packaging, low -loss and higher pow er capability than o ther plan ar transm ission
lines. H ow ever, it has som e draw backs. Figure 1.6 show s how tw o-term inal devices are
integ rated in series and shunt connections for various planar transm ission lines. A s shown,
m icrostrip requires drilling or via hole processing for shunt connected devices.
This
deficiency can be overcom e by u sing either coplanar w aveguide (C PW ), coplanar strips
(C P S ) or slotline. H ow ever, o th er com plications arise from the use o f th ese alternative
lines.
F o r instance, slotline and C P S tend to b e overly lossy at higher frequencies and
C P W requires air bridges to m aintain the ground potential on either side o f the center
co nductor.
O n e w ay to ta k e advantage o f m icrostrip’s g o o d qualities while circum venting its
deficiencies is through th e use o f inverted m icrostrip.
U nlike m icrostrip, inverted
m icrostrip rem oves the ground plane from th e substrate backside and inverts the
c o n d u c to r over a ground plane su p p o rt as shown in Figure 2.1. T he electric and m agnetic
fields are prim arily concentrated in th e air betw een the patch and the g ro u n d low ering the
effective dielectric co n sta n t(% ^) o f this configuration. A low er eej j m eans that inverted
m icrostrip dem onstrates a longer guided w avelength and higher characteristic im pedance
o v e r a com parable line in m icrostrip. Insertion o f series o r shunt devices d oes not require
drilling thro u g h
the substrate w hich often introduces unw anted
discontinuities in
m icrostrip.
T h e use o f air fo r transm ission m akes inverted m icrostrip less lossy than m icrostrip but the
spacing betw een the conductor and baseplate ground plane m ust be controlled accurately.
S pacers are often used w ith som e success up through K u-band. A s operatio n frequencies
increase, the inverted m icrostrip configuration is prone to exciting surface w ave m odes
w hich cause considerable cross-talk in densely packed circuits (or high m utual coupling in
anten n a arrays).
These surface m odes can render an M IC useless.
F o r antennas, these
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
26
m o d e s reduce rad iatio n efficiency and distort the overall pattern.
In order to eliminate
u n w an ted m odes and reduce cross-talk, shorting pins o r electric w alls on either side o f the
c o n d u c to r can be used. T he use o f metallic walls results in w hat is often called trapped
in v erted m icrostrip. M etallic channels o r grooves in the baseplate provide accurate and
rep eatab le spacing betw een conductor and ground plane as well as isolation b etw een lines.
T ra p p e d inverted m icrostrip is useful at higher o p erating frequencies and provides strong
m echanical su p p o rt as well as heat sinking capability fo r active devices.
aftclh C o m d y e t o r
pac® rs
.t® @r@ymicfl
F igure 2.1. Inverted patch configuration.
In v e rte d m icrostrip is attractive for hybrid M IC s because shunt connections (i.e. diodes,
tra n sisto rs o r pro b es) do n o t require drilling th rough the circuit substrate as in m icrostrip.
This characteristic can be exploited to allow non-destructive m echanical optim ization o f a
d ev ice o r a p ro b e input. F o r integrated antennas, the inverted m icrostrip provides a builtin rad o m e for protection. By choosing the substrate and m etal support carefully, quality
h erm etic seals can be achieved to protect solid-state devices and im prove system reliability
and durability. T hese qualities m ake this transm ission line an ideal candidate for integrated
an ten n a applications.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
27
This ch ap ter describes th e design o f cavity enclosed inverted stripline antennas ( / SA).
A lthough a m ore descriptive name w ould b e trap p ed inverted m icrostrip antennas, the
stru ctu re falls under the m ore general stripline type geom etry. T he operating frequencies
o f these an ten n as are found using a standard cavity m odel and com pared to m easured
results.
V ario u s im portant param eters are varied to dem onstrate the usefulness o f the
model.
2.2 T h e Cavity Model T h e simplicity o f th e cavity m odel is due to the assum ption that
th e sep aratio n betw een th e patch co n d u cto r and the g ro u n d plane is m uch less than the
o p eratio n w avelength (i.e. h « X ) . In such a case, the electric field com ponent only has a
single com p o n en t in the z-direction and, consequently, the m agnetic field only has
transverse com ponents in th e volum e betw een the c o n d u cto r and ground plane (i.e. E:, Hx
and Hy fo r rectangular coordinates an d E:, H+ and Hp fo r cylindrical coordinates).
T h ere
are no n o rm al com ponents o f the electric field a t the edge o f th e patch w hich, in turn,
renders th e m agnetic fields at th e ed g e to b e very small. T hese assum ptions c re a te a cavity
bounded o n th e to p and b o tto m by perfect electric conductors (PEC) and o n the side by
perfect m agnetic conductors (PMC).
F o r an arbitrary-shaped p atch antenna excited b y a current J at a frequency co on a
su bstrate o f thickness h, perm eability n and perm ittivity e, Maxwell's equations are
w ritten as [23 ]
V x E = -jcopB.
V x H = j <vsE + J
v
.e = £
(2,)
E
V- H = 0
T h e exciting current J can be introduced w ith a m icrostrip line feed o r a probe-feed. T he
m icrostrip
line approach requires q u arter-w ave im pedance transform ers
introduces unw anted radiation.
and
often
H ow ever, a large array can be photolithographically
rep ro d u ced in m ass-quantities once th ese effects have been accounted for. T h e probe-feed
is desirable in the sense th at it is placed below the antenna minimizing any coupling to the
radiating edges. T he p roper probe location can be used to provide very g o o d m atch to the
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
coaxial feedline. A lthough it is not readily used for large arrays since it is labor intensive,
it can provide very g o o d perform ance in small sub-arrays.
W hen m odelling th e probe
feed, an equivalent w id th o f line o r ribbon can b e used at the probe position.
J = z - Sp
(2.2)
%
elsewhere
I0
I f th e height o f the su b stra te is electrically thin, the z-directed current is assum ed to be
constant(i.e. V ' J = - j o / ) * 0 ) w hich reduces G auss’ law in Eqn. (2.1) to zero (i.e.
V ■E = 0 ). T h ese equatio n s allow us to obtain
V 2E + «y2^ £ E = y'rUyuJ
(2.3)
w h ere th e electric field and electric current is z-directed and the w avenum ber in the
dielectic substrate is defined by k = co^[fie. PM C boundaries on the sides o f th e cavity
en su res th at
=
Eqn. (2.3) can be solved first as a hom ogeneous equation by
setting th e R H S (o r c u rre n t ex citation) to zero.
V 2E + k 2E = 0
(2.4)
T he eigenfunctions an d eigenvalues o f the hom ogenous equation are given by y/ma, and
kmn, respectively. I f th e eigen functions are orthogonal, then th e com plete solution to eqn.
(2 .3 ) is given by
1
E,=jeo/iZZ.
2 ------- r~7 ~Vmn
” " k - k mn
(2 5)
T h e reso nan t freq u en cies for th e structure are found by setting the denom inator o f the first
term in th e double sum m ation o f eqn. (2.5) to zero (i.e. k 1 - k 2
mn = 0).
T he resonant
frequencies a re given by
F0=
* " " ■
2 x-yjp e
(2.6)
I f th e patch is rectan g u lar w ith w id th W and length L, the resonant frequencies can be
found using
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
29
(2.7)
W here the dom inant m ode is obtained by setting m=0 and n=J.
On the other hand, the
resonant frequencies o f a circular patch antenna with a diam eter D is given by[24 ]
W here a„„ is th e n-th zero o f the derivative o f the Bessel function o f order m.
dom inant m ode for a circular p atch is obtained by setting am„=1.84118.
The
For an actual
resonator, th e electric field near th e edges is not only in th e z-direction since it fringes over
th e edges.
Fringing fields lo w er the observed resonant frequencies below that o f
calculated values.
T he difference can be consolidated by using an effective diam eter
instead o f th e physical diam eter.
2.3 M e a s u re d R e su lts In this study, the antennas u sed are circular w ith diam eter D as
show n in F igure 2.2.
T h e patch es are enclosed by a cavity o f diam eter C and depth
(A +B'). A is the substrate thickness and B is the center con d u cto r separation to ground. Si
is typically air (i.e. £t =l) and e2 is the dielectric constant o f th e substrate, shows the ISA
configuration and the critical dim ensions. The m icrostrip cavity model can also be used to
calculate the resonant frequency o f the inverted structure by using the structure’s effective
perm ittivity
and an effective diam eter (Deg). T he effective diam eter is the observed
electrical size w hich is larg er th an th e physical diam eter an d accounts for the fringing fields
o f the radiating structure. T he resonant frequency can be w ritten as
(2.9a)
w here c is the speed o f light and am„ is the mn-th m ode coefficient. Deff is the effective
patch diam eter du e to fringing fields which is given by[25 ]
( l n ( £ ) + 1.77 + 1 .4 1 ^ + £ ( 0 2 6 8 * ^ + 1 .6 5 ) )
(2.9b)
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
30
In the above equation D is th e physical patch diam eter and B is the conductor to ground
separation. T he effective perm ittivity is a value very close to one since m ost o f the fields
are concentrated in air. Typical empirically determ ined values o f 1.1 to 1.2 provide good
agreem ent w ith m easured results.
P atch
A ntenna
___ .P
M icrom eter
Cap is n o t s o l d e r e d to
P a t c h A n te n n a
P r o b e Feed
F igure 2.2. ISA configuration dimensions.
T he novel test fixture show n in Figure 2.3 w as designed to non-destructively characterize
th e input im pedance as a function o f probe position.
T he probe can be moved
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
31
m echanically w ith a m icro m eter in o rd er to accurately o p tim ize th e retu rn loss at the
o p eratin g frequency. T h is trait also aids in the integration o f so lid -state devices w hich will
b e d e m o n strated in later chapters.
Figure 2.3. ISA test fixture.
A w ide ran g e o f p atch diam eters and tw o types o f substrates w e re m easured during this
investigation. T he fo llo w in g param eters are used during th e se m easurem ents: A -1 .5 mm,
B ~ 1.5 m m , C = 62 mm, Si=1.0, £7=2.2 or 10.2, and D = 10,15,•••,60 mm.
F igure 2.4
sh o w s th e calcu lated an d m easured operating frequencies as a function o f p atch diam eters
fo r th e tw o su bstrates.
A s show n, the m easured results follow th e calculated values
closely.
Since th e p ro b e can be m oved along th e antenna, th e return loss at th e o p eratin g frequency
w as o p tim ized for a m inim um value equal to less than 35 dB (i.e. RL < + 3 5 JB) w hich
c o rresp o n d s to a VSWR » 1.0006. T h e calculations are ac cu ra te to w ithin a few percent
o f th e m easu red results fo r the param eters given above. E rro rs in calculations m ay be due
to th e p a tc h -to -g ro u n d separation, probe contact to the patch and p robe inductive effects.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
32
13
N
X
o
12
calculated
e = 2 .2
11
u>> 10
B
9
<u
S
cr
u
•U.
oxi
c
«V.
o
a.
8
7
6=10.5
6
5
m easured
«
O
5
10
15
20
25
30
35
40
P a tc h D ia m e te r (m m )
45
50
55
60
F ig u re 2.4. ISA operating frequencies as a function o f patch diam eter.
T he p ro b e can b e m oved easily along th e patch to non-destructively obtain im pedance
inform ation ab o u t th is antenna structure.
An H P-8510B N etw ork A nalyzer w as used to
m easure th e in p u t im pedance b y calibrating to the ground plane o f the inverted m icrostrip
patch.
T h e te st fixture used allow s accurate non-destructive positioning o f th e probe
along the anten n a radius.
T h e input im pedance o f a 30 mm diam eter circular patch
enclosed b y a 62 mm diam eter cavity vs. probe position is show n in F igure 2.5.
O ther
dim ensions are as given before w ith er=2.2.
F o r circular p atch es, th e electrom agnetic fields are given by the follow ing equations
E (x, y ) = z E ^ J j k p ) cos {n<£)
H (x ,y ) = ^ - E m
COjU
~k
{kp) s in (n ^ ) - 4»— J'm(ikp) cos(«<z5)
P
( 2 . 10)
T he electric c u rre n ts ( J ) w hich flow on th e patch co nductor can b e calculated and used to
determ ine th e radiatio n perform ance o f the m icrostrip antenna.
J=nxH=
j
COfJ.
P ^ m( V ) c o s ( « 0 ) - <j>- J'm(kp)wL{n<f)
( 2 .11)
P
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
33
o<?
p
F ig u re 2.5. IS A input im pedance as a function o f probe position ( p ).
Similarly, th e m agnetic current is given by
M = E x n = <jj£mnJ m{kp) cos (n<f)
(2.12)
I f th e m ag n etic lo o p cu rren t is used, th e problem can be sim plified by im aging this current
o v e r th e g ro u n d plane.
In this fashion, the ground plane is rem oved and a m uch m ore
sim pler p roblem rem ains.
T he only difference is that th e im age essentially doubles the
quqntity. F ro m this, th e electric v ector potential is found u sing
=
7
J EmnJ m{ka)cos{n<j>)—
*=Q;=0
r
(2.13)
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
34
The resulting fields in the Fraunhofer region o r far field are then given by
E=
(2.14)
' -jk B co iB
jk cos 9
H = jcoF^
(2.15)
jkB coiO
jk cos 6
A lthough effects o f the w all proxim ity are not directly noticeable on th e p atc h input
im pedance, they can b e readily seen on the radiation patterns. T he radiation p a ttern s at
the o p eratin g frequency o f 5 G H z for the 30 mm diam eter antenna on a substrate o f er=2.2
are show n in F igure 2.6.
T he E - and H -plane half-pow er beam w idths(H PB W s) are 51°
and 61°, respectively. C ross-polarization levels are 19.3 dB below the m easured m axim um
gain level o f 10.5 dBi.
105
255
90
270
75
105
285
255
(a) H -plane and cross-pol
90
270
75
285
(B ) E-plane and cross-pol
Figure 2.6. Probe-fed ISA radiation patterns.
R adiation p atterns for o th er p atch diam eters vary w ith respect to the H PB W because it
depends o n the patch-to-cavity ratio and frequency o f operation.
T he inverted antenna
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
35
and enclosure com bination sharpens th e radiation p a ttern o f this antenna. F ig u re 2.7 shows
th e H P B W s o f several antennas te ste d in the 62 mm diam eter cavity.
O th e r param eters
rem ain co n stan t as listed previously w ith an er=2.2 and th e variation in th e p a tc h diam eter
(D).
N oticeable b e a m sharpening p eak s at ~ 0 .5 patch-to-cavity d iam eter ratio and
broadens to either side.
T hese antennas dem o n strate a useful alternative for conventional m icrostrip patches.
In
som e cases such as in integrated o r active integrated antennas, the IS A d em onstrates a
clear advantage since device integration is straightforw ard. Gunn diodes an d F E T s can be
integrated for o scillato r applications.
sw itching an d tun in g functions.
P IN and v a ra c to r diodes can b e integrated for
O ther devices can be integrated to achieve mixing,
am plification o r m ultiplication. B iasing circuits can be etched on the su b strate o r com e up
through th e g ro u n d plane.
The integration is com pact, low -cost and exhibits good
radiation p erfo rm an ce as will be show n in th e follow ing chapters.
90
Gain
85
-10
80
-9
75
H-plane
70
-7
65
60
55
-3
50
--2
45
62 mm Diameter Cavity
40
10.0
15.0
20.0
25.0
30.0
35.0
40.0
Patch Diameter (mm)
45.0
50.0
55.0
60.0
F igure 2.7. H alf-pow er b eam w idths and gain o f probe-fed ISA s.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
36
CHAPTER HI
D IO D E IN T E G R A T E D IN V E R T E D S T R IP L IN E A N T E N N A S
3.1.
I n tr o d u c tio n
O ver th e last few years, m icrow ave integrated circuits (M IC s) and
m onolithic circuits (M M IC s) have received a g rea t deal o f attention for m any system
applications. T hey are intended to m ake the R F front-end easier to reproduce and m ore
reliable
w ith
im proved
perform ance
at
low er
costs.
Im proved
perform ance,
reproducibility an d reliability is due to strict control on every line dim ension and device
used in th e circuit w hile lo w er co st per unit o ccurs only in large quantities. T o w ard s this
goal, in teg rated antennas a tte m p t to place RF front-end com ponents right a t the antenna
term inals.
O scillators, mixers, amplifiers, sw itches, tuners have been dem o n strated while
th e ultim ate goal is to com pletely integrate the entire front-end w ithin the antenna volum e.
This w o u ld p ro v id e a com pact, reproducible p ro d u ct w ith potential fo r low cost in m ass
quantities.
In teg rated an ten n as pose an interesting problem because they require som e understanding
o f several different areas o f m icrow ave engineering (i.e. devices, antennas, circuits,
m aterials, co m p o n en t functions, etc.).
S olid-state devices and their operation as well as
the co m p o n en t fun ctio n (i.e. sw itching, tuning, mixing, amplification, oscillation, etc.) are
critical in th e system design.
effects a re also im portant.
p ro p erties w hile th e
A ntenna design considerations and device com bination
D evice biasing circuits tend to disturb antenna radiation
antenna configuration
may degrade the com ponent
function
characteristics. T h ese difficulties have kept perform ance o f such integrations belo w useful
specifications and o u t o f m ilitary and com m ercial applications.
T here are m any ty p es o f antennas that are are well suited for device integration.
By far,
m ost in teg rated antennas have used m icrostrip patches b u t other configurations have also
been in teg rated w ith solid-state devices.
Viable alternatives to m icrostrip fo r integration
include tap e re d n otches, slotlines and dipoles.
T he antenna m ust be easily accessible to
solid-state devices w ith respect to size, input im pedance and biasing netw orks.
The
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
37
inverted stripline antennas described in C hapter II are ideally suited for integration with
tw o-term inal devices.
3 .2
D io d e I n te g r a tio n
An im portant m ilestone in integrated antennas occurred in the
early 1980s w hen B hartia and Bahl introduced solid-state varactor diodes on the radiating
edges o f a rectangular m icrostrip patch antenna [26 ] to provide resonance tunability. This
frequency agility allow s a narrow band antenna to operate o v er a w ide range o f
frequencies.
M o re im portantly, it brought solid-state devices (w hich w ere previously
confined to guided w ave circuits) together w ith planar antennas. T he approach treats the
anten n a as ju st an o th er circuit com ponent w hose properties can also be electronically
controlled.
A lth o u g h varactors w ere used in this integration for frequency tuning, many o th er devices
an d functions can also be realized. The addition o f an electronically tuned reactance to the
rectan g u lar patch antenna can b e analyzed using a sim ple transm ission line approach. The
v aracto rs are variable reactive com ponents w hich load the transm ission line [27 ,28 ].
Fringing fields are accounted fo r with an equivalent length o f line and radiation is m odeled
by a resistor. A lthough this m ethod ignores package parasitics, integration discontinuities
and cavity perturbations, it g o es a long w ay in predicting the antenna frequency response.
A m o re intensive full w ave approach using m ethod o f m om ents (M O M ), finite difference,
o r finite difference tim e dom ain (FD TD ) can provide an improved m odel which accounts
fo r radiation effects due to th e interaction o f the solid-state package and the antenna.
H ow ever, these num erical approaches also incur m ore tim e and com putational costs.
T h e following eq uations describe m odels u sed to determ ine characteristics o f m icrostrip
p atch antennas. A s far as these circuit and cavity m odels are concerned, the differences
betw een a m icrostrip p atch and an inverted stripline antenna(ISA ) can consolidated by a
sim ple scaling factor. F o r the ISA, th e relative dielectric constant (<sy) is replaced by an
effective dielectric constant ( ^ ) which is determ ined from the inverted configuration.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
F ig u re 3.1 show s th e antenna configuration dimensions and approxim ate transm ission line
m odel o f a rectan g u lar m icrostrip patch antenna. The transm ission line m odel can b e used
to determ ine th e input adm ittance at a probe feed w hich is given by
Ytn = G + j(B + Bv) + Yg
G + j(B + Bv + Y0 lm j3 L)
Y0 + j(G + j( B + B v)) tan ft L
(3.1)
w h e re G is th e radiation conductance given by [29 ]
W2
90%
G=
W
[\ 20%
W {{%
(3.2)
w ))%
and th e suscep tan ce B o f the open end is approxim ated by
B=
*0A/V %
(3.3)
w h e re Y0 is th e characteristic adm ittance o f the transm ission line, P is th e transm ission line
p ro p ag atio n constant, Al is th e extension which accounts for fringing fields, L is the
physical reso n an t length and Bv is an electronically-variable susceptance provided by the
so lid -state diode.
Patch antenna
YV'///// /
y/Y Y„ y Baseplate
Figure 3.1. R ectangular patch antenna dim ensions and transm ission line m odel.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
39
C ircular p atch antennas can also be used for integration and analyzed in a similar manner.
T he cavity m odel assum es that the circular p atch antenna has perfect electric conductors
and a perfect m agnetic c o n d u c to r w hich encloses the patch co n d u cto r and ground plane.
F igure 3.2 show s the circular p atch configuration and coordinate system used. The fields
w ithin th e cavity m ust satisfy th ese boundary conditions. I f the cavity is very thin, the zdirected electric fields are approxim ately given by
Ez = k 2J x( kp ) cos(0)
(3.4)
and th e corresponding f-directed m agnetic fields are given by
H^= jco£kJl(kp)cos(<f)
(3.5)
A
7
ila n n e te r (D)
E le ctric Field Lines
P a t c h a n te n n a
''V'V///
'//fy Baseplate! ■/,
Figure 3.2. C ircular p atch antenna configuration and co o rd in ate system.
T he field distribution fo r a given m o de indicate places w here a solid-state device may
provide th e m ost effect o n the frequency o r radiating characteristics th e patch. First, the
resonant frequency o f a given circular patch can b e calculated using
P _
&mn_____________________________ ^
^
y/l + ( % * . „ Xln<&> + 177 +14le« + o (0-268 V +1-65))
W here c is th e speed o f light, a™, refers to the m ode num ber (i.e. 1.8411 for th e dom inant
m ode) and D is the physical p atch diam eter.
%-
is the effective perm ittivity o f the
inverted configuration.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
40
O nce th e p atch dim ensions have been finalized to ensure dom inant m ode operation at F0,
th e im pedance ch aracteristics w ithin th e antenna cavity can be found. A lthough several
assum ptions have been m ade, these equations give insight into m ethods and places to best
integrate passive and active devices. D uring integration, m atching becom es critical and, to
this end, th e input im pedance at a position p from the center can be approxim ated with
{
\
J im
(3.7)
2 J
w h ere R0 is th e resistance at resonance fo r p = ^
■
This equation is useful in determ ining a position w here passive and active com ponents
an d /o r devices can be placed. This includes probes, diodes, transistors and o ther circuits.
W hen this device is a P IN diode, a switching o r m odulating antenna com ponent can be
created.
E q u atio n (3.7) is u sed to m atch probe feeds o r devices to the antenna.
In teg rated devices can be m odeled as lum ped loads which affect the resonant frequencies
o f th e circular p atch antenna. Since the m icrostrip patch is a resonant cavity, there are a
large num ber m odes w hich can be excited w ithin the antenna due to discontinuities or
p erturbations.
E x citation o f these m odes often degrades th e circuit and radiation
perform ance o f the p atch antenna.
m odes. [3 0 ,3 1 ,3 2 ]
Shorting pins can be used to suppress unw anted
E xternal elem ents such as capacitors and inductors can also be
strategically placed to enhance o r suppress different m odes.
In th e same manner, PIN
d iodes can provide electronic control over the radiation efficiency o f a m icrostrip patch.
T h e zero and fo rw ard bias turn s th e antenna on and off, respectively. The integration can
serve as a m icrow ave sw itch and /o r m odulator. [33 ,34 ]
D io d es are easily connected across the patch to g round w ith a low pass filter to provide
th e necessary biasing voltage.
By varying the diode position u nder the patch, the
im pedance presented to th e diode term inals can be changed.
p osition determ ines th e d io d e’s effect on the antenna cavity.
A lternatively, the diode
Since the electric field is a
m axim um at the radiating edges o f the antenna, PEN and v aracto r diodes will have the
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
41
m ost effect a t these positions. D iodes placed near the radiating edges can easily tune the
resonance o v er 30% . This is a significant range since a typical m icrostrip patch has less
than 3% instantaneous im pedance bandw idth.
An inverted m icrostrip antenna has been used extensively to integrate solid-state devices.
The stru ctu re has dem onstrated g o o d radiation perform ance and design flexibility. Figure
3.3 show s a circular IS A integrated w ith tw o packaged diodes. T he diodes can either be
P IN o r varactors.
In th e case o f tw o P IN diodes, the integrated antenna will behave
according to the voltag e used to bias the devices. N o voltage represents an open circuit
state w hile 1 V tu rn s th e P IN s into a short circuit.
C
C
R
C ' 0.3
C ‘ 0.3
' 0.4 raH
' 2 Ohms
PIN Diode
L
0.3 - 1.4 pF
0.3 pF
2.31 Ohms
0.4 nH
R
2 Ohms
Varactor Diode
Figure 3.3. Solid-state device package and integrated ISA configuration.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
42
Pow er Input
VSWR
VSWR: 1.014 to 22.87
VSWR=2 line
0
0.1
02
0.3
0.4
05
0.6
0.7
08
09
I
PIN Bias Voltage (Volts)
(a)
0.7 V, 6.61 mA
1.2 V, 100 mA
-5 -0 .6 V, 0.64 mA
-10
- -
0Q
T3
w
C/5
</)
-15 -
O
-30 --
0 V, 0 mA
VSWR: 1.014 to 22.87
-35
3.75
3.85
3.95
4.05
4.15
4.25
4.35
4.45
O perating F req u en cy (GHz)
(b)
F ig u re 3.4. PIN integrated ISA switch results.
T he follow ing param eters a re u sed during these measurem ents: A -1 .5 mm, B ~ 1.5 mm,
C =62 mm, £i=I.O, £2=2.2 or 10.2, and D=30 mm.
F o r the given p aram eters, the
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
43
structure o p erates a t 4.97 GHz. A t zero voltage bias (O N state), the open circuited PIN s
are reactances w hich detune the ISA operating frequency about 20% .
The O N state
allow s the antenna to o perate at 4.125 GHz, w ith a V S W R o f 1.013 and a 2:1 im pedance
bandw idth o f 130 M H z (3.15% ). W hen the P IN diodes a re biased on, they provide very
g o o d short at th e operating frequency. For a 1.2 V bias level (100 mA), the OFF state
show s a V S W R o f 22.87. Figure 3.4 shows the perform ance o f the P IN integrated ISA.
G eneral p u rp o se P IN diodes used in this investigation are housed in the case style shown
in Figure 3.3. T he probe-fed antenna can also be am plitude m odulated by varying the bias
continuously from 0 to 100 m A (0 to 1.2 V).
In a similar m anner, P IN diodes can be replaced w ith varactors to provide the frequency
agility show n in Figure 3.5.
As show n, the operating frequency is shifted or tuned w ith
resp ect to th e v a ra c to r bias voltage.
This change can b e accom plished electronically
w ithin a couple o f nanoseco n d s w ith very little current drain. The concept can be used for
channelized receiver reception, frequency m odulation, etc.
T he voltage-variable reactance w hich loads th e integrated antenna can b e m odeled as a
lum ped load on th e resonant circuit. Device parasitics w hich depend on case styles o f the
pack ag es will affect the antenna cavity m ore than a typical circuit.
D epending on the
frequency o f o p eratio n and package style, th e integration may introduce higher cross­
polarization and a ch an g e in the principal planes. H ow ever, for a simple lum ped model,
th e variable ju n c tio n capacitance o f the v aracto r can provide a g o o d description o f the
frequency tuning characteristics o f the integrated antenna.
T he v aracto r ju n c tio n capacitance varies as a function o f bias voltage. T he bias voltage
(V ) is described by
(3 8)
w h ere C(0) is th e capacitance at 0 Volts, Vbl is the built-in potential (G aA s=1.3 V ) and / is
0.5 for abrupt junctions.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
44
0
Return
Loss (d B )
•5
-10
-15
-20
-25
-30
-35
2.7
2.8
2.9
3
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
4
3.7
3.8
3.9
4
O peration Frequency (G H z)
(a)
R eflection
C o e ffic ie n t (dB)
12
- -
14 -16 --
20
- -
22
- -
24 - -
-28 --30
2 .7
2.8
2.9
3
3.1
3.2
3.3
3.4
3.5
3.6
O peration Frequency (G H z)
(b)
F ig u re 3.5. Frequency agility o f a varactor-integrated ISA.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
45
V aracto r diodes will have m axim um effect on th e antenna if they are placed at electric field
maxima.
A t the radiating edges, the v aracto rs couple very strongly with the radiating
electric field loading the antenna and low ering th e operating frequency. W hen biased, the
v aracto rs can quickly tune o v er a range o f frequencies.
V aractor integrated IS A s have
exhibited a w ide operating tuning range o f 31 % centered at 3.4 GH z. A s show n in Figure
3.6, at 30 V the operating frequency is at 3.88 G H z with a VSW R o f 1.0105. A t 0 V the
operatin g frequency is at 2.84 G H z with a V S W R o f 1.0053.
1.1
3.95
2.84 to 3.88 GHz
3.85
109
3.75
1.08
•S 3.65
1.07
't
>,
3.55
O
106
|
,,s
I
g)
1.04
=3
>
O
'w'
55 3.45
&
°
5
^
rj)
1.03
3.05
1.02
2.95
SC
o
>
1.01
2.85
1
0
5
10
15
20
25
30
Varactor Bias Voltage (Volts)
F igure 3.6. Bandw idth tunability and corresponding V SW R versus varactor voltage.
T he instantaneous im pedance bandw idth rem ains nearly constant at ab o u t 2 % . F igure 3.7
sh o w s th e IS A instantaneous 2:1 im pedance bandw idth vs. operating frequency.
As
show n, th e V S W R remains belo w 1.02 while th e im pedance bandw idth varies from 1.8 to
2.5% . T his integrated antenna can be used to rapidly scan over its tuning bandw idth for
w ideband receiver or transm itter applications. T h e small instantaneous bandw idth (~ 2 %)
can red u ce th e noise pick-up in a wideband channelized receiver system.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
46
A s in m ost integrated antennas, there m ay be a degradation in radiation perform ance due
to antenna m odifications and device perturbation o f th e antenna fields. F o r example, the
cross-polarization fo r the varacto r integrated patch is -1 0 dB com pared to -19 dB for the
non-integrated patch
Sm aller devices, im proved biasing schem es and optim ized diode
positions reduce these adverse effects o n the radiation pattern.
In identical fashion,
d e te c to rs and m ixers can be integrated within the confines o f th e antenna to provides
vario u s circuit functions.
3
2.8
2.6
1.1
V aractor Tunable ISA
1.09
C=62 mm, D=30 mm,
A=B=1.5 mm
06
£
1.08 00
>
e 2.4
1.07
■5
<3
•o
S 2.2
•o
§
OQ 2
uso
3 1.8
1.05 £
3 1.6
1.03
1.4
1.02
1.2
1.01
1.06 as
O
>
(3
00
G
1.04 ■3
s
so
3
o
>
1
2.8
2.9
3
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
Operating Frequency (GHz)
F igure 3.7. Percentage im pedance bandw idth versus frequency.
A different integrated antenna which used P IN diodes to electronically affect the frequency
characterisitcs o f a patch antenna was dem onstrated by D aryoush et. al. in 1986 [35 ]. In
th is investigation, a stub w as loosely coupled to the radiating edge o f a patch antenna. A
P IN d io d e is placed from the antenna to th e probe. F orw ard and reverse biasing o f the
P IN diode provides tw o different operating frequencies for the antenna at the input port.
T h e zero bias operating frequency is 3.285 GH z while the forw ard bias operating
frequency is 3.207 GHz. T he radiation pattern s show 70 degree H -plane beam width and 7
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
47
dB gain.
In this case, th e PIN can also be optically controlled.
W hite light illumination
varies th e d io d e ’s reactan ce en ough to tune the operating frequency 15 M H z at 3.270
GH z. O ptical control o f th e antenna can provide a very useful m ethod o f m odulation.
A d raw back in th e use o f integrated antennas is th a t they still require som e ty p e o f a feed
n etw ork to distribute th e RF signal to the term inals o f each integrated antenna.
C onsequently, a large array o f antennas integrated w ith varactors o r PIN diodes w ould still
need a lossy R F netw ork.
T he array aperture, how ever, w ould becom e agile either in
frequency, p h ase or am plitude according to the biasing o f each device at ea ch antenna.
In teg rated and active integrated antennas require the ability to com bine different types o f
devices easily. Specifically, m icrostrip requires th at shunt connections be drilled in hybrid
M IC s w hile in m onolithics via holes m ust be processed.
Some alternatives to m icrostrip
w h ich d o n o t require drilling include coplanar w aveguide, slotline, coplanar strips and
inverted m icrostrip.
Inverted
m icrostrip also alleviates hybrid integration discontinuities en co untered in
m icrostrip. Inverted m icrostrip is n o t a true uniplanar line but it has several advantages.
Since th e m ajority o f th e fields propagate through air, it has potential for low er
transm ission loss.
It also offers som e integration advantages over a conventional
m icrostrip line. Inverted m icrostrip does not require drilling fo r shunt connections which
allow s n o n-destructive experim ental testing as well as position optim ization o f diodes and
coaxial p ro b e inputs.
applications.
T his trait m akes inverted m icrostrip attractive fo r m any hybrid
W hen used fo r integrated antennas, inverted m icrostrip provides a built-in
rad o m e fo r p rotection.
A s previously shown by P IN and v aracto r integrated patches,
th e se integrations provide very g o od circuit perform ance and m o derate
radiation
perform ance. T he IS A stru ctu re has been extensively integrated w ith passive, active and
com binations o f devices to m ake trasm itters, receivers, detectors and transceivers.
T he
follow ing ch ap ter review s oscillators and their characteristics. This is necessary for G unnand F E T -in teg rated antennas which follow.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
48
C H A P T E R IV
T H E O R Y A N D P R O P E R T IE S O F O S C IL L A T O R S
4.1 I n tr o d u c tio n
F o r th e m ost part, integrated antennas have em phasized active diode
and tra n sisto r integrations for transm itters and p o w er com biners. T he com biners typically
involve oscillators fo r spatial and quasi-optical p o w er combining. T he follow ing sections
review im p o rtan t o scillato r th eo ry and characteristics.
T hese characteristics include
derivation o f th e oscillation conditions and load-pulling. Also included are descriptions o f
spectral p ro p e rtie s such as quality-factor, stability and noise.
4.2
O sc illa tio n C o n d itio n s
Solid-state sinusoidal oscillators are resonant circuits
integrated w ith active solid-state devices w hich convert DC energy to R F pow er.
The
resonant circuit o r re so n a to r can take m any shapes and form s and, because o f its
im portance, usually determ ines o f nam e o f the resulting oscillator (i.e. w aveguide and
m icrostrip oscillators).
R eso n ato rs include short lengths o f transm ission lines, rings,
dielectric disks, rectan g u lar and circular w aveguide cavities and yttrium iron garn et (Y IG )
spheres. In active in teg rated antennas, the antenna serves as th e resonant circuit a s w ell as
the rad iato r. A D C b ias line delivers the p roper voltage and current for th e active device
to g e n erate electro m ag n etic energy. The reso n ato r circuit provides th e necessary reactive
storage w hich com p en sates fo r th e active device reactance.
O scillations o c cu r a t the
frequency w h ere th e overall circuit reactances cancel out.
O scillation start-u p o ccurs due to noise o r D C bias transients. T he p o w e r level builds
up until th e active device is saturated. The device achieves a steady-state equilibrium by
continually resto rin g th e p o w er delivered to and dissipated in th e circuit. Since th e device
im pedance is a function o f th e R F current, th e oscillation frequency m ay change w hile the
device reach es its steady-state.
A t equilibrium, the oscillation frequency and am plitude
rem ain unchanged. F o r a free-running, steady-state oscillator, th e sum o f th e circuit ( Z c )
and device ( Z d ) im pedances m ust be zero at th e device's operating point as show n by
Z d ( V d c J d c , a > n , I r f J , - - - ) + Z c {con) = 0
( 4 . 1)
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
49
w h e re the circuit im pedance varies only with frequency and the device im pedance is a
fu n ctio n o f its D C operatin g point iydc, I dc), operating frequency com ponents {con), RF
c u rre n t am plitude ( / ^ ) an d tem perature (T). T h e subscript n refers to the harmonic
frequency w hich can take o n values o f n = 1,2,— .
T h ere are several different types o f sinusoidal oscillators including sub-harm onic,
fund am en tal and harm onic oscillators. O scillators provide p o w er at each o f its harmonics.
I f th e em bedding circuit is designed to operate at th e first harm onic and suppress others, it
is called a fundam ental oscillator. Similarly, sub-harm onic oscillators operate at fractional
m ultiples below the fundam ental, and harm onic oscillators operate at integer multiples
a b o v e the fundam ental.
Since the im pedance characteristics o f the devices change with
D C operating point, the bias voltage can be used to tune th e oscillation frequency. This
ch aracteristic is called bias tuning (o r pushing) w hich is used for frequency m odulation
(F M ) and has c re a te d a large com m ercial m arket for the voltage-controlled oscillator
(V C O ).
4 .3
V o ltag e C o n tro lle d O scillato rs
The V C O uses voltage to control the output
freq u en cy o f th e oscillator. This characteristic is used for frequency m odulation (FM ) in
com m unication links.
A change in voltage can change th e device or circuit reactance.
O n e can change the D C operating point ( Vdc,I dc) o f th e active device w hich changes the
d ev ice im pedance 2d and p u shes the circuit to the new operating frequency o f eqn. (4.1).
O n e draw back o f bias tuning is that the m odulating circuit m ust be able to drive the DC
p o w e r drained in the active device which may be several w atts. Also, since the operating
p o in t is altered, th e co rresponding change in the negative resistance may lead to wide RF
p o w e r deviations.
A n o th e r m ethod relies on a second solid-state device called a v ara cto r (variable
c a p a c ito r) to alter th e reactance o f the circuit. T he varactor is strategically placed to alter
th e circuit reactance Xc(con).
Unlike bias tuning, the active device operating point
rem ain s th e sam e and the o u tp u t pow er remains m ore constant over a w id er frequency
I
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
50
tuning range. Furtherm ore, th e v aracto r consum es little pow er which puts few constraints
o n the m odulating circuit.
H ow ever, a second device and another control voltage
increases circuit com plexity and
system
cost.
D epending on
the cost allow ed,
perform ance requirem ents, space available, materials provided and pow er allocated to the
R F section in a system, one m ethod may be better suited than the other.
M icrow ave
designers evaluate all o f these trade-offs and attem pt to achieve the best possible design
fo r a particular set o f specifications.
V C O s form an integral part o f many military and com m ercial m icrow ave system s for
local oscillator and transm itter applications.
designers th e ability to
Oscillator specifications give system
categorize oscillators for different applications.
Several
ch aracteristics used to describe oscillators are defined as
1.
Stability: Ability o f an oscillator to return to the original operating point after
experiencing a slight electrical or mechanical disturbance.
2.
N o ise
• A m plitude m odulation (A M ) noise: A m plitude variations o f th e o u tp u t
signal.
• FM noise: U nw anted frequency variations.
• Phase noise: Phase variations.
3.
Q uality (0 1 factor
• U nloaded: A ccounts for resonator losses
4.
only.
only and assum es R{oss = 0.
•
External: A ccounts fo r th e load resistor
•
Loaded: A ccounts fo r both resonator losses and external loading.
Frequency
• -Jum ping: D iscontinuous change in oscillator frequency due to nonlinearities in th e device impedance.
• -Pulling: C hange in oscillator frequency vs. a specified load m ism atch over
•
360° o f phase variation.
-Pushing: C hange in oscillator frequency vs. D C bias point variation.
5.
S purious signals:
oscillation carrier.
O u tp u t signals at frequencies o th e r than th e desired
6.
Post-T uning drift: Frequency and pow er drift o f a steady-state oscillator due
to heating o f the solid-state device.
7.
Therm al stability: C hange in output pow er and frequency vs. tem perature.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
51
Table 4.1 gives d ata for typical com m ercial G unn V C O s op eratin g at 35 and 9 4 G H z.
T hese tw o frequencies have m any com m ercial and military uses.
At 35 G H z, the
oscillators can be u sed for police radars.
T able 4.1. Typical C om m ercial V C O Specifications.
F requency (fD)
P o w e r (P0)
B ias P ushing R ange
(typical)
V a ra c to r T uning R ange
F requency D rift over
T em perature
P o w e r D ro p over
T em perature
Q ext
H arm onics Level
M o d u latio n B andw idth
M od u latio n Sensitivity
F M N oise
(ql 100 K H z offset
A M N oise
@ 100 K H z offset
35 G H z
250 m W
94 G H z
50 m W
50 MH z/ v
im M n z /y
± 250 M H z
, M H z/
+250 M H z
~L
7°C
- ^ A
c
- 0.03 dB/ o c
- 0.03 dB/ o c
800 - 1000
- 2 0 dB
D C - 50 M H z
2 5 -5 0
800 - 1000
-2 0 dB
D C - 50 M H z
2 5 -5 0
-* > dBC/K H z
• 80 ‘B c/k H z
- 155
' 150 dBC/K H z
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
52
4.4
D erivation o f O scillation C on d ition s T here are several w ays to study oscillators
including u sing scattering-param eters, stability regions and th e N yquist th e o ry applied to
feedback am plifiers and a negative resistance network.
K urokaw a used a negative
resistance in a tw o-term inal n etw o rk to describe oscillator characteristics. [3 5 ]
T he
follow ing derivation follow s his w o rk closely to lead to oscillation conditions, stability,
noise an d Q -facto r form ulations. W e begin w ith the one-port equivalent circuit o f Figure
4.1. T h e forcing function e(l) is u sed to model noise for oscillation sta rt-u p o r an external
injection locking signal u sed fo r synchronization.
r
r
Imbedding
Ci r c ui t
Z c = R C+ j X c
Zd. ~ Rd +jXd
L
J
I(t )
L
Figure 4.1. D iode and Gunn oscillator equivalent circuit.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
53
In th e circuit, an R F cu rren t flow s through the device. T he RF current has m agnitude
An(t), frequency con, and phase $ ,( /) .
/ ( 0 = 2 A ( 0 e ;('U"'+*"(' ))
« = 1 ,2 ,3 ,-
(4.2)
n =1
T h e m a g n itu d e /^ (/) and p hase
are functions o f tim e because they can be
subjected to noise a n d D C bias transients. These transients a re often present at very low
frequencies so that 4 j ( 0 and <
j>n(t) are slowly varying w ith respect to th e carrier frequency
<y„. E a c h frequency harm onic contrib u tes to th e overall current tim e function / ( / ) .
E ach o f th e harm onics w h ere oscillations are occurring m ust satisfy eqn. (4.1). U sing
K irc h h o ffs v o ltag e la w for th e loo p , w e can w rite equations for each harm onic as follow s
e , ( 0 = 4 ( /) •e ](a^
,))[Zd{Al,(ol) + Zc(a>,)]
^ ( 0 = 4 ( 0 - e J(^ +^ (0)[ Z 4 4 ^ 2 ) + 4 ( « 2 ) ]
i
(4.3)
eN{t) = AN{ t ) - e ^ ,+Mt)\ Z d{AN,coN) + Zc{coN)]
w h ere th e device and circuit im pedances are
4 4 4 , . ^ ) = 4 / ( 4 , + jX d(An,o)n)
Zc(con) = X d(con) + jX c(o„)
T he device im pedance Zd(An,eon ) is a function o f both the m agnitude o f th e R F current
and oscillating frequency w hile th e circuit im pedance Zc(con) only depends on the
frequency.
T he eqns. (4.3) im ply that the sum o f the circuit and device v o ltages m ust
equal th e forcing function (<?„(/)) at each harm onic and can be re-w ritten in a m ore
co m p act form .
e„(0 = An(l)ej(^ ^ Z
d(An,con) + Zc(co„))
n = 1, 2, 3 -
(4.5)
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
54
I f th e change in frequency So)n is small w ith respect to the carrier co„, w e expand the
circuit im pedance Zc(a>n) in a T aylor series about (a>„ + 8 co„). We also expand the device
im pedance Zd(A„,(On) about {An + 8An,a>n + Ston) with a tw o-variable expansion.
Zc( <o„ + 8 co„ ) = Zc( con) + 8(0 „— Zc
da>
da}m
ml
(4.6)
c
d_
+ SA„Z-Zd
n dA
Zd(An + SA„,6)n + 5con) = Zd(A„,a)n) + 8 ( o — Zd
0(0
(4.7)
, (54n)m (T v
ml
i.5a>nY 0*
8Am “
ml
dco"
4 ,.® »
F o r small perturbations, w e can neglect higher pow ers o f the Taylor series and p ro d u cts o f
small quantities.
T h e result is a com pact approxim ation fo r the perturbed circuit and
device im pedances.
Zc(£»„ + Sa>n) ~ Z c( 6)n) + S(on— Zc
Z d (An + &4„,con + S(on) ~ Z d(An,(on) + Scon— Zd
(4.8)
SA
4 ,.®,
+
d
-7
n~M
dA d
(4.9)
4 ,.® .
= c o s ((ont + <
f>n) + jsin((ont + </>„)) and include
W e replace th e exponential (i.e.
th e real and im aginary p arts o f th e im pedances. Keeping only the real part o f the resulting
v o ltag e drops a ro u n d the current lo o p o f eqn. (4.5), results in th e follow ing equation.
e„{t) = A„(l)cos(a>„/ + $ > „ (/)X ^ (4 n fl»„) + / ^ (<»„))
(4.10)
- 4 , ( 0 - sin(<y„/ + <pn(ty iX d(An, a>„) + X e(a>„))
W e th en replace th e real part o f th e circuit and device im pedance w ith th e appropriate
approxim ations o f eqns. (4.8) and (4.9). T he time dependence o f en(t), 4 i ( 0 a° d ^ „ ( 0 *s
im plied. T h e notation' ZCn stands fo r the localized circuit im pedance at th e n-th harm onic
frequency £y„. Similarly, the n o tatio n
denotes the localized device im pedance at the
frequency <on and a current m agnitude An.
The forcing function en(t) can be used to
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
55
m odel noise flu ctu atio n during oscillator start-up o r an external injection-locking signal.
Eqn. (4 .1 0 ) becom es
W e m ultiply th e result above by cos(<y„/ +<j>n) and sin (a>nt + <
f>n) and integrate over a
single period o f oscillation. D u e to the orthogonal pro p erties o f sine and cosine functions,
tw o equations are form ed.
+ R cn + 5 ( 0 n
+ O
+ 5A n
= J ^ n(0 co s(£ y nt + <fin)dt = ecn
(4.11)
^ x dH+X Cn + f a n^ { X eH+ x J + 5 A „ ^ { x j y j e nm n(a> „t + 4n) = em (4.12)
w here the right hand side o f th e equations above are the cosine and sine transform ations o f
th e forcing function fo r each o f the N harmonics. T o conserve space, the cosine and sine
transform ations w ill b e d enoted by eCn and eSn, respectively.
E quations (4 .1 1 ) and (4 .1 2 ) represent general oscillation conditions accounting for the
device im pedance's frequency and current m agnitude dependence a s well as th e circuit
im pedance's d ep en d en ce on frequency. T hese equations also assum e that N harm onics are
presen t in the circuit w hich can be super-im posed o v e r each o th e r in the frequency
dom ain.
F o r a free-running, steady-state oscillator, the o p erating frequency is stable and the
o u tp u t p ow er is co n stan t (i.e. ecn = esn = 0, San = 0 and SA^ = 0 ) . The equations above
then red u ce to th e follow ing conditions within a m ultiplicative constant.
o
■{xdn+ x Cn) = o
(4.13)
(4.14)
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
56
T hese eq u atio n s sh o w th a t each o f th e harm onics must satisfy eqn. (4.1). The re so n a to r
characteristics
determ ine th e higher o rd er harm onics w hich satisfy the oscillation
conditions.
F o r a free-running, stead y -state oscillator to operate, the m agnitude o f the active
device's n eg ative-resistance m ust be equal to th e circu it resistance presented to its
term inals. O scillation frequency o ccu rs w here the reactances cancel out. T hese conditions
are sum m arized belo w and show n in Figure 4.2 at th e oscillation frequency as an exam ple
(4 .1 5 )
F u rth erm ore, th e sta rt-u p is m ore likely to occur i f th e m agnitude o f the negative
resistance at least 2 0 % g reater than th e circuit resistance (i.e.
> 1.2-f^.)
Impedance
R e [ Zd ]
0
Frequency
Figure 4.2. O scillation conditions o f a typical circuit.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
57
F igure 4.3 show s the G unn diode I-V characteristics including th e threshhold voltage,
o p eratin g current and o p erating voltage. Im pressed on th e graph is an RF variation which
sim ulates oscillation conditions.
Uc
op
0
Ko p
dc
Figure 4.3. G unn diode current-voltage relationship.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
58
4.5
S-Parameter Formulation of the Oscillation Conditions
A t m icrow ave
frequencies an d above, it b ecom es inpractical to m easure cu rren ts and voltages b ut relative
p o w er and scatterin g (S ) param eters can be m easured accurately on netw o rk analyzers,
spectrum analyzers and p o w er m eters.
T hese test m easurem ent system s m easuring S-
param eters are very useful fo r M IC designs. The equations above are th erefo re modified
to u se S -p aram eter definitions. First, th e circuit and device im pedances are u sed to obtain
th e co rresp o n d in g reflection coefficients (T ).
r ,W
p M
= z- k > ' z - h >
Zc K ) + 2 > „ ,
(4 i6 )
v
Zd (4 ,^ r ,) + Zo M
W h ere ZQ represents th e characteristic im pedance o f the transm ission line used w hich may
be a function o f frequency. T he reflection coefficients in eqn. (4 .1 6 ) are equivalent to onep o rt S -p aram eters denoted by iSM. T he equivalent condition to eqn. (4 .1 5 ) fo r the onep o rt negative-resistance oscillator becomes
r , K , < u „ ) . r cK ) = i
(4 .i7 )
This eq u atio n can be derived by using definitions for th e o n e -p o rt S -param eter
(4.1 6 ), Tc - S c - %
is the ratio o f the reflected signal bc to a transm itted
F igure 4.4). Sim ilarly, Yd = Sd = %
in eqns.
signal ac (see
fo r the device. Solving fo r bc and bd leads to
bc = Sc -ac
and
bd = Sd -ad
(4.18)
T wo
Port
9 ftw © rk
Figure 4.4. T w o-port circuit definitions.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
59
W hen th e device and the circuit p o rts are connected such that ac = bd and ad = bc, w e can
solve fo r the transm itted signal ac.
(4.19)
R e-arranging eqn. (4.19), the S -param eter condition for oscillations is [36 ,37 ,38 ,39 ]
(s a - iK
(4.20a)
= o
Since ac * 0, the oscillation frequency will exist only if {SdSc - 1 = 0) is zero o r
S,Sr = 1
(4.20b)
N o te th a t eqn. (4.20b) is equivalent to eqn. (4.17).
N ote that the basic form o f the
oscillation condition in eqns. (4 .2 0 a and 4.20b) remains unchanged. This is tru e even for
an N -p o rt n etw ork w here the quantity in parentheses becomes an NxN m atrix equation.
T he oscillation condition is given by th e determ inant o f the matrix.
F o r transistors, proper D C biasing may not be enough to guarantee a negative
resistance region.
A negative resistance region can be induced by p ro p er feedback and
term ination elem ents. An FE T circuit in a com m on gate configuration form s th e tw o -p o rt
netw ork
show n
in Figure 4.5a.
Similarly,
common source and
com m on
drain
configurations are possible. T h e tw o -p o rt S-param eter oscillation condition is no longer a
single m ultiplier as in th e o n e-p o rt case. The device and load netw ork param eters are now
defined by 2x2 m atrices.
A n elem ent coefficient Sv o f the m atrix is the ratio o f the
resulting voltage a t port / due to the induced voltage at port j. T he oscillation condition is
th e determ inant o f th e 2x2 S-m atrix.
=0
(4.21)
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
60
in
171
o u t'
o u t'
(a)
(b)
F ig u re 4.5. T w o- and three-port F E T equivalent circuit.
The solution o f eq n .(4 .2 1 ) is the determ inant o f the m atrix below
?d
Sa
n 'Scu + 5,1 •5j, -1
Cd
J 21
nc
'
, ad
J 11 i " L>2 2
■
5,1 •5,c2 + 5,1 •5 !
ac
ad
°21
° 2 1 ' J 12 i _ ° 2 2 ' ° 2 2
ac
, ad
ac
,
=
0
1
(5,1 •5,c, + 5,1 •52e, - l) •(5! •5,1 + 521 ■5|2- 1)- (sd
2l •5,1 + 521 •52c,) •(5,1 •5,1 + 5,1 •S c22) - 0
For
sim ple
loads,
th ere
is no coupling o r feedback b etw een
5,1
= 5 2, = 0 ). T h e equation above reduces to
( ^ n ' 5 , i _ l) • (5 22 ■5 22 - 1) - ( 5 ,1 ■5 22 • 5,1 • S2\ ) = 0
circuit
p o rts (i.e.
(4.22)
U sing co nventional notatio n , set 5,e, = T ,, 5 22 = T2 and rem oving superscripts from device
param eters resu lts in tw o well know n oscillation conditions for th e tw o -p o rt netw ork.
5,, +
^12 ^2 1 ^ 2
1
1- 522T2 r,
(4.23 a)
5 +
5,2521r,
i-5,,r,
(4.23b)
1
r2
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
70
4 .1 0 O s c illa to r S y n c h ro n iz a tio n A dler [44 ] derived the oscillator phase w ith respect to
tim e form ulation giving relations betw een th e oscillator and th e external injected signal.
T h e form ulation w as instrum ental in describing th e injection locking process as well as th e
p ro d u c tio n o f a d isto rted beat note.
Figure 4.6 show s the notation used fo r the
synchronized oscillator described below . G iven the oscillator voltage, E, and the injected
v o ltag e, Ei, th e resu ltan t voltage, Eg, is th e vector sum o f E and Ej. The corresponding
an g u lar frequencies fo r E, E0 and Eg are co0, <y, and co. T he phase difference betw een the
in d u ced voltage an d th e im pressed v o ltag e is given by
<t>= - ^ - s i n ( a )
E
(4 .6 3 )
da
Figure 4.6. Synchronized oscillator notation.
F o r th e typical single tu n ed oscillator show n in Figure 4 .7 , w e can linearly approxim ate the
ch an g e in phase v ersu s frequency by th e slope, A =
as long as it is close to tu0.
T h erefo re, th e frequency difference betw een the Eg and E is used to calculate the phase by
0 =
a
( co -
co0 )
(4.64)
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
71
lag
lead
F ig u re 4.7. Lead and lag phase versus frequency fo r a single-tuned oscillator.
Since th e injected signal is a t a frequency different from th e oscillator signal, there is an
instan tan eo u s b eat frequency generated, Aco. Aco is described by the ch a n g e in the angle
b e tw e e n E and E , versus tim e, ^ a/ i t From eq u atio n s
da
—Ei
Atu = — = —
(4 .6 3 ) and (4 .6 4 ), w e have
s i n ( a ) + A<y0
(4.65)
w here A (o=co-(oi and A co0 =o)0 - co,.
Eqn. (4 .6 5 ) can be rew ritten to show that th e instantaneous frequency,
th e free running frequency, coo, by an am ount proportional to the
<o, is shifted from
sine o f th e phase angle,
a , and th e relative am plitude o f th e injected to oscillator signal, E'/E, existing at th e locking
instant.
-E ,
a - — —s i n ( a ) + <o0
EA
(4 .66)
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
72
F o r a single tu n e d circuit operated near resonance,
e q u a tio n (4 .6 4 ) w ith
This eq u ation is sim ilar in form to
j
ta n ( ^ ) = 2 Q
and f o r small angles, <
f>* 2
.
(0-Q>o
<on
F or frequencies near cd0,
2Q
A - * L - .
~ dt ~
(4.67)
2Q „
fo r small angles.
con
=s —
4 Q2[a)-a)oy
03 r
(4.68)
<on
S u b stitu tin g fo r A in eqn. (4 .6 5 ) and using the approxim ation fo r small an g les in eqn.
(4 .6 8 ) results in an expression relating the relative am plitude o f th e voltages, Q -fa cto r and
d ifferences b etw een th e free running and injected frequencies (i.e. the u n d isturbed beat
frequency).
-E , Q)0 . , x
da
A
d ‘0 = ' 5 ' = T 5 e s,n (“ ) + i W »
(4.69)
E qn. (4 .6 9 ) m u st equal to ze ro fo r a system w hich has reached steady state. W e can then
so lv e fo r th e stationary p hase angle b etw een the oscillator and injected signal.
s in (a ) = 2 0 —
0)n
(4.70)
E qn. (4 .7 0 ) leads directly to a condition fo r synchronization since th e right side is bounded
b y p lu s and m inus one. T herefore, th e condition for synchronization is given by
f
(4.71)
> 2Q
<On
O f special in terest is th e locking bandw idth (BW l) w hich req u ires finding th e maximum
Aco w h e re synchronization is m aintained. A s long as the change in phase v e rsu s frequency
is g iv e n by th e slo p e n ear ce>o, th e tw o-sided locking bandw idth is given by
5 ^ L= 2 A a m*
mw= Q| ~£
(4.72a)
w ritin g th e eq u atio n s in term s o f injection and oscillator p o w er
A/mtx_
/.
1
1 /f
(4.72b)
2Q \ P
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
73
T ransient analysis o f eqn. (4 .6 5 ) gives insight into th e process o f how an oscillator is
pulled into th e resulting frequency. T w o cases are possible, w hen th e injected signal is the
sam e as th e o scillato r frequency and w hen it is not.
A ssum ing that the injected signal is
precisely a t th e oscillator’s frequency sets A*y0 = 0 in eqn. (4.65).
In such a case, eqn.
(4 .6 5 ) will b e satisfied for any injected signal and Q -facto r o f th e oscillator.
da
—E o)n
-5 r - r i e sin(a>
(473)
Solving th e differential equation
da
s in ( a )
\
s in ( a )
C-E, co.
dt
E 2Q
)
<4 7 4 >
l-cos(g)_ ^ % >
s in ( a )
So th a t any valid solutions a re in th e range - n < a < n . F o r cases w here a is very small,
s in (a )« a
and the solu tio n is simply th e exponential o n the right hand side o f eqn. (4.74).
This sta te s th a t th e oscillator phase approaches th e phase o f the injected signal
exponentially. T he speed o f th e approach depends o n the relative strength o f the injected
to o scillato r signal a n d th e Q -facto r o f the oscillator.
In th e seco n d ca se w hen A (0Q* 0. Eqn. (4.65) becom es
f - =^ s i n ( a )+M
E A<a0
W e first su b stitu te K = 2 0 —
(4.75)
and rew rite the synchronization condition
E> <»o
E A (On
K =2 Q - - f - < l .
Integrating eqn. (4.7 5 ), follow ing th e sam e p rocedure as in eqns. (4.74).
(4.76)
results in
w h ere th e solution, a , is given by
, f i
-J k 2- i 4 A<uo ( / - f 0) J J rT -x
a = 2 arctan — + ------------ tan
K
2
K
K
„
(4.78)
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
74
When the synchronization condition is not met, |£ | > 1, there will be an average angular
beat frequency present which is described by
/^•2
i
Afi> = A<y0— - —
(4.79)
Acoo is the beat frequency present if the oscillator maintains if free-running frequency. The
term on the right hand side drops to zero for K =1 and approaches one for large values of
K. Substituting for K,
E A oj0 IQ
-1
(4.80)
Ato is plotted versus Aeo0 in Figure 4.8. The figure shows the reduction in beat frequency
due to injection-locking or synchronization of the free-running oscillator.
So the
instantaneous beat frequency varies periodically between the free-running and maximum
locking frequency,
± A<omn , for Afflj, > A<o.
/ \ ( j J — A GJ'o
A c jc
Figure 4.8. Change in beat frequency due to injection-locking.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
75
A study on nonlinear synchronized LC oscillators has been carried out by Odyniec.[45 ]
The study uses harmonic balance techniques and averaging to determine characteristics of
these synchronized oscillators.
A useful algorithm for finding stability zones and
oscillations for typical oscillators was presented.
Further studies describing the noise properties of synchronized oscillators have been
presented by Kurokawa [46 ] and Schlosser.[47 ] An extension of these studies was
shown by Schunemann and Behm in 1979. [48 ]
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
76
CH APTER V
GUNN & FET-INTEGRATED INVERTED STRIPLENE ANTENNAS
5.1
Introduction Previous chapters have dealt with the ISA structure primarily as a
passive device.
In chapter II, various probe-fed antennas exhibited very good
performance, smooth radiation patterns and the ability to non-destructively test various
antenna parameters on a single test fixture. Also, directivity and bandwidth were shown to
be a function o f the patch-to-cavity diameter. In Chapter III, this structure was integrated
with diodes to create passive components. The probe-fed varactor- and PIN-integrated
ISAs demonstrated excellent component performances with good antenna characteristics.
Integrated antennas can be easily used for switching [33], amplitude modulation[33],
frequency modulation [26, 33], frequency selectivity [26, 33, 49 ] and RF detection [50 ].
If the devices integrated are active, they can be used for various other types components
such as oscillators [51 ,52 ], amplifiers [53 ], mixers [54 ] and multipliers [55 ]. Complete
doppler sensors [56 ] and transceivers [57 ] can also be realized. In Chapter IV, oscillator
characteristics were discussed in detail as a starting point in the design of active integrated
antennas.
Active integrated antennas are those radiators directly integrated with active devices to
generate or amplify RF power. Unlike conventional antennas, an integrated antenna is
affected by device packages and bias lines.
Although integration reduces the overall
component size, weight and cost, active integrated antennas have shown a deterioration of
both the antenna and component performance.
The combination o f components and
antennas changes operating frequencies, lowers conversion efficiencies and causes higher
cross-polarization levels (CPL). If an active integrated antenna can maintain reasonable
component specifications with little degradation in the radiation characteristics, the
approach would be attractive for many commercial and military systems. Specifically, as
low-cost doppler sensors for perimeter surveillance or low-cost sources for decoys,
communications, beacons and other applications.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
77
Active antennas have been realized using two terminal and three terminal devices (i.e.
diodes, transistors). The choice between transistors and diodes depends on the type of
microwave component required, operating frequency range, RF power output desired,
available DC power input and other considerations. For integrated antennas, the choice of
solid-state device also involves its effect on the radiation characteristics.
Field-effect
transistors (FETs) and other three terminal solid-state devices have shown higher DC-toRF conversion efficiencies over two terminal devices but diodes reach higher operating
frequencies and greater power levels. When used as a radiating oscillator, the antenna
provides a tank circuit, matching transformer or feedback mechanism which sustains
oscillations. Individual active antenna power levels are low due to the limitations o f the
integrated devices but a large number o f these antennas can be synchronized to provide
higher power levels.
Over the past decade, the microstrip patch antenna has been integrated with diodes and
transistors for active, planar, low-cost radiating elements [58 ,59,60 ]. The microstrip
structure provides a resonant patch for oscillations and a ground plane for efficient heat
sinking.
However, the patch has exhibited narrow bias tuning ranges, high cross­
polarization levels, and wide output power deviations.
A tuning diode and bipolar
transistor has recently been integrated within a multi-layer patch to obtain 4.4% tuning of
the operating frequency [61 ]. As an alternative, the endfire notch has many desirable
characteristics including broad impedance matching bandwidth, uniplanar nature and easy
integration o f other planar solid-state devices. Its scalability is attractive for use well into
the sub-mm wave region.
[62 ,63 ,64 ].
Active varactor tunable notch antennas have been reported
However, an array of such endfire antennas would require a brick style
approach for power combining arrays. Although the brick style approach offers large heat
dissipation volume and circuit area with modular replaceability of components, it tends to
be bulkier than an array using a tile style method. For tile type approaches in power
combining, endfire elements may not be as suitable as broadside microstrip patches,
bowties, grids and dipoles. The ISA configuration provides an useful alternative with
various advantages over conventional microstrip patch antennas.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
78
5.2
Single Active Antenna
In chapter IV oscillator conditions and various
characteristics were derived. Similarly, an active antenna is simply a one-port radiating
oscillator. It is primarily made up of an active device and an antenna. The antenna
presents a load to the device which determines whether oscillations can be maintained.
The active device (i.e. Gunn, IMPATT) can be modelled as a nonlinear element
( yn(A ) = - Gd(A) + jB D(A)) whose characteristics depend on voltage amplitude (A).
An analysis for a single active antenna was described by Stephan in 1986 [65 ] using the
canonical oscillator circuit shown in Figure 5.1. The oscillator is described by the parallel
combination o f a capacitor, inductor and resistor. The injected current (/(/)) is then the
sum o f the currents flowing through the individual components
i( 0 = C ^ + G1v + i-Jv£/r + rov
at
(5.1)
L
where the tank circuit consists o f an inductance (L), capacitance (C) and load
conductance^).
— ------------------ o d «
Figure 5.1. Canonical oscillator circuit.
The voltage (v) in equation (5.1) is assumed to be sinusoidal with angular frequency (<y,),
magnitude (A(t)) and instantaneous phase (<fi(t))
v = A(t)cos(a)it + 0(t))
(5.2)
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
79
The magnitude and phase in eqn. (5.2) are assumed to be slowly varying functions in time
with respect to the reference frequency (<u,). This simplification allows us to omit higher
order terms which arise from the insertion of eqn. (5.2) into eqn. (5.1). Integration by
parts yields the following expression for the current
/=
^ sin(tu,/ + 0 ) + ^ - cos(ta/ +
+(G l - Gd)( A cos(<a,f + <(>)- BdA sin g(eo,t + ^))
+X((4—^^)sin(<y-/+
M s& i*+
(5.3)
<*>)
The injected current is then assumed to have an in-phase cosinusoidal component of
magnitude ( l c( t )) and a quadrature sinusoidal component with magnitude(/,(/))• This is
given by
'(0 = /cCOcos^z+^OH^OsinOV + ^O)
(5-4)
Eqn. (5.4) is inserted to the LHS o f eqn. (5.3). The principle o f orthogonality allows us to
reduce the expression into separate components.
First, multiplying both sides by
sin(ffly/ + 0(f)) and integrating removes all cosine terms
<55)
^dt (\ c+±m'L>h Bo+(°>c - eo.L
- 7 = 7A
Similarly, multiplying both sides o f eqn. (5.3) by cos(&/ + 0(/)) and integrating removes
all sine terms
04,
^dt( c -
^
r
Given the oscillator free-running frequency ^<y0 = - ^ =
)
<
5«
and assuming the injected
frequeny (a),) is near coo, a frequency deviation is defined by Aeo = 6), -<na The terms in
equations (5.5) and (5.6) are approximated by the following
<a,C+— *2(<»,-<y0)C=2AaC
ojL
(5.7)
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
80
(5.8)
The simplification shown in equations (5.7) and (5.8) are substituted into equations (5.5)
and (5.6). The differential equations for the phase and amplitude variation o f the oscillator
voltage become
(5.9)
(5.10)
In the absence of injection current (Ic = l s = 0), the steady-state amplitude (A 0) is reached
when (G d ( A ) - G l = 0) making “% = 0. Stephan also noted that the in-phase component
Ic of the injection current has a first order effect on amplitude while the instantaneous
frequency ( d*/dt) is primarily influenced by the quadrature component /,.
These equations assist us in developing a basic understanding for the dynamics o f the
active antenna oscillator. The oscillator can be surrounded by many other oscillators
operating at various operating frequencies and power levels. Some of these effects can be
modelled by the equations above.
5.3 Gunn-Integrated ISAs The cavity-enclosed inverted stripline circular patch antenna
is an improvement over a conventional microstrip patch in that it does not require drilling
through the circuit substrate to accomodate shunt diodes to ground.
Drilling on the
housing is still required but it is not as critical. Accurate height of the patch over the
housing is critical for repeatability. The structure allows the housing to be designed such
that the feed or diode position is easily optimized in a non-destructive fashion.
The
configuration's enclosure chokes out surface waves and introduces an increase in metal
volume for heat dissipation. DC bias is provided to the diode from the ground plane side
with a feed-through capacitor. This scheme does not require direct contact with the patch
radiator and can improve the impedance match to the diode.
The inverted stripline active patch antenna can also be integrated with other solid-state
devices for increased power, frequency switching and/or tuning. The axial symmetry of
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
81
the antenna can be used for circuit level power combining with more than one active
device placed under the patch radiator [66 ].
Monolithic integration of this type of
antenna in a tile type planar array could be accomplished using monolithic chip inserts at
each antenna location. Each insert would contain the patch and active devices. For active
integrated devices, excess heat can be dissipated efficiently by carefully matching the
thermal expansion coefficients of substrate materials and housing alloys. This, in turn, can
be used for hermetic sealing which improves system durability and reliability[67 ]. DC
biasing of devices can be achieved on the substrate or through the ground plane
underneath.
Gunn-integrated active inverted stripline antennas have been demonstrated for beam
steering and spatial power combiners. [3,1] These active antennas exhibit good radiation
patterns, low cross-polarization levels, easy device integration and good heat sinking
capacity.
Unlike PIN- and varactor-integrated antennas, Gunn integrated antennas
generate RF power output. The antenna serves as the resonator as well as the radiator for
the oscillator.
Oscillations occur at the frequency where the diode and the antenna
reactances cancel out.
As discussed in chapter IV, oscillation startup occurs if the
placement o f the diode is such that the magnitude of the diode negative resistance is
greater than the circuit resistance presented to its terminals.
These conditions are
summarized below
|Re[^diod,J - R-e[2c,reui»]
(5 .1 1 )
= - Im[ZdraJ
One could supply DC bias to the patch conductor and then make contact with the diode.
Our design biases the diode directly from under the ground plane using a feed through
capacitor. This biasing scheme allows proximity coupling to the patch antenna and a
method to tune out the diode's series inductive component. One can use the method to
experimentally tune over a small range of frequencies for a given mode. An electronic
method of varying the frequency of operation involves the bias voltage. Varying the bias
voltage of the Gunn diode causes a change in the diode junction reactance and a
corresponding shift in operating frequency. The overall tuning range of the active antenna
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
82
is 140 MHz. The 3 dB bias tuning range is 107 MHz at 9.09 GHz or 1.2 %. The active
antenna pattern remains well-behaved with a cross-polarization level of at least 10 dB
below the maximum.
S u b s t r a t e is p r e s s
f i t t e d to Cavity —
Gunn Cap is n o t
s o l d e r e d to
patch
A d j u s t ab l e
Top p la t e
C= 12.7 m m
D= 10.4 m m
P = 2 mm
T
M ec ha n i c a l
tuning
Bias
Figure 5.2. Gunn-integrated ISA.
Figure 5.2 shows the Gunn integrated ISA using the adjustable test fixture. The package
connection to the heat sink and the added metal volume of the metallic walls provide
efficient heat sinking for the active device.
Screw-type packaged Gunn diodes were
integrated with inverted patch antennas of diameters ranging from 8 to 11 mm. These
diameters and dimensions create active antennas at X-band frequencies. In chapter III, a
larger cavity diameter was used with a wide range of patch diameters for probe-fed ISAs.
These passive antennas operated from S- to X-band frequencies. The differences between
the structures is in the intended applications.
Gunn-integrated antennas are primarily
intended for spatial power combining. In this application, the ability to put a large number
o f sources in a small area is important. To realize this, the cavity diameter is made very
close to the antenna diameter. This increases the source packing density which affects the
overall power output.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
83
By placing metallic walls very near the edges of the antenna, a small and compact active
antenna is created. The enclosure proximity to the radiating edges does, however, affect
the active antenna operating frequency and the radiation performance. The large diode
package used (mainly intended for waveguide applications) also presents a significant
perturbation on this X-band radiating structure which adversely affects the radiation
characteristics and disturbs the oscillating frequency. The measured probe-fed antenna
operating frequencies and corresponding Gunn oscillator frequencies and de-embedded
power outputs are listed below in Table 5.1.
Table 5.1 Gunn-Integrated ISA Operating Frequencies, Power vs. Diameter.
Patch Diameter
(mm)
8
9
10
10.4
11
Passive F„
(GHz)
13.300
12.285
11.415
11.101
10.661
Measured F0
(GHz)
9.997
9.637
9.438
9.240
8.908
Oscillator
Power* (mW)
63.99
56.76
57.04
63.98
66.98
* Calculated using a passive antenna gain of 6.65 dBi. Cavity diameter is 12.7 mm.
The large deviation between the passive operating frequency in the second column and the
active antenna oscillating frequency in the third column is due to the Gunn diode
equivalent circuit. The Gunn diode loads the antenna and the combination will oscillate at
the frequency specified by equation (5.11). The antenna load impedance as a function of
frequency and the device impedance as a function of voltage determines the tuning
characteristics of the active antenna just like any other oscillator.
The Gunn integrated ISA operating frequency can be tuned over modest bandwidth o f a
few percent. It depends on the diode’s impedance characteristics as well as the imbedded
circuit characteristics as given in equations (4.1) and (5.11). As shown for this active
antenna, there is approximately 225 MHz of bias tuning bandwidth centered at 9.37 GHz.
The threshold voltage is approximately 4 Volts.
Figure 5.3 shows the typical tuning
characteristics of a Gunn-integrated ISA.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
925
Threshold V oltage
9.2
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Bias Voltage (Volts)
Figure 5.3. Tuning characteristics of a Gunn-integrated ISA. (D = 10.4 mm)
The Gunn oscillator power in Table 5.1 was calculated by assuming that the active antenna
gain isequal
to the gain of a probe-fed antenna withsimilar dimensions.
oscillatorpower
at the active
The Gunn
antenna element was thenfound byusing the
Friis
transmission equation.
where
• P t is the oscillator power transmitted by the active antenna oscillator.
• P r is the received power at the test horn antenna.
• R is the range length.
• A is the operating wavelength.
• Gr is the standard horn gain.
• G0 is the active antenna gain.
Radiation patterns for the Gunn integrated ISA are shown in Figure 5.4.
The cross
polarization level is at least -10 dB for a Gunn diode 2 mm off-center. The HPBWs are
100 degrees in the E-plane and 70 degrees in the H-plane. A similar antenna using a probe
feed exhibited HPBWs o f 105 and 80 degrees in the E- and H-plane, respectively with a
cross-polarization level of -16 dB.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
85
(a) E -plane
(b) H -plane
F igure 5.4. Typical .X-band G unn-integrated ISA radiation patterns.
T h e differences in radiating perform ance are attributed to th e active an te n n as’ diode
p a c k a g e and bias lines. Im proved perform ance can be achieved by reducing th e size o f the
d ev ice packages.
A n o th er im provem ent is to m ove the cavity walls aw a y from the
rad iatin g edges.
T here is, how ever, a tra d e-o ff in packing density v ersu s individual
an te n n a perform ance.
5 .4
F E T - I n te g r a te d IS A s
T he IS A structure has also been integrated w ith an FE T
tran sisto r. [68 ] T he use o f th e FE T tran sisto r im proves conversion efficiency and noise
ch aracteristics w hile it reduces therm al requirem ents.
H ow ever, the a n ten n a m ust be
m odified to acco m o d ate th e FE T as show n in Figure 5.5. F E T integration requires three
D C b lo ck s for th e drain, g a te and so u rce terminals.
DC biasing can be achieved from
b ehind th e g ro u n d plane o r etched to th e non-radiating edges o f the antenna.
A chip
re sisto r co n n ected across th e source to g a te simplifies biasing to th e device. T h e operating
freq u en cy is d eterm ined by th e loads at each transistor port. T he ISA te st fixture can be
u sed w ith a coaxial p robe feed to de-em bed approxim ate im pedance levels at each
tra n sisto r port.
T h e d a ta is then u sed in a lumped m odel to determ ine oscillating
freq u en cies fo r th e active antenna. A nother advantage o f th e transistor is th a t it can be
in d u ced to oscillate a t low er frequencies to reduce effects from th e device pack ag e and
bias lines.
L o w e r frequencies o f operation can be accom plished w ith larg er patch
diam eters.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
86
A 30 m m p atch w as m odified to insert th e FET. A t th e center o f the patch, 0.4 m m gaps
isolate th e so u rc e from th e gate and drain term inals.
A 0.1 mm gap is etched from the
cen ter to th e non-radiating edges o f the patch to provide D C isolation betw een th e gate
and drain term inals. T h e FE T drain lead is soldered at the center o f the patch and th e gate
lead is approxim ately 2 m m off-center. T hree D C lines bias th e source, g a te and drain o f
th e transistor.
Alternatively, the source bias line can be replaced by a resistor.
The
frequency o f oscillation depends on th e im pedance loads at the F E T ports. T he loads are a
function o f th e position along th e inverted patch antenna. H eat generated is dissipated by
th e patch. T h e lack o f a low therm al-im pedance p a th from th e device to the m etal heat
sink m ay cause therm al problem s in higher p ow er devices. This could be alleviated w ith a
shorting pin at th e cen ter o f th e patch to provide a low im pedance path to th e housing.
Exploded View of P a t c h M eta liza tio n
F igure 5.5. FE T -integrated IS A configuration.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
87
S ince th e antenna w a s m odified and the device term inal are soldered, it is not possible to
experim entally o p tim ize the F E T position as in the G unn-integrated design.
H ow ever,
since th e tra n sisto r is n o t directly attached to the base like th e G unn d io d e, the base can be
u sed as a variable g ro u n d plane.
T he ground plane is essentially a sliding short w hich
ch anges th e an ten n a cavity d e p th and alters the oscillation frequency o f th e oscillator. The
cav ity d ep th is also instrum ental in improving spectral and radiation characteristics o f this
ty p e o f active antenna.
T h e F E T provides 57 m W a t 5.69 G H z w hen biased at 3.8 V and 26 m A . T he 3 dB bias
tu n in g ran g e is approxim ately 1 % for a 1 V change in v*.
a llo w s a m echanical tuning ran g e o f nearly 6 %.
T he sliding g ro u n d plane
A t a cavity d ep th o f 4.15 m m , the
m easu red oscillation frequency rem ains stable at 5.695 + 0.002 G H z o v e r the an ten n a test
sw eep. T h e H P B W in the E - and H -plane patterns are 46 and 64 degrees, respectively.
T h e cro ss-p o larizatio n level is -19.3 dB below th e maximum. Figure 5 .6 show s principal
p lan e and cross-po larizatio n p a tte rn s o f the active antenna.
90
90
120,
120,
s0—
0
(a ) E -plane
(b) H -plane
F igure 5.6. FE T -integrated IS A radiation patterns.
T h e sm o o th rad iatio n p attern s and low cross-polarization level c o m p ares favorably with
prev io u sly re p o rte d active antennas. Probe-fed passive antennas w ith cavity depths o f 3
m m exhibited H P B W s o f 58 an d 62 degrees in th e E - and H -plane p attern s, respectively.
T h e cro ss-p o larizatio n level o f th e passive antenna is also -19.3 dB w ith a gain o f 10.2
dBi. B iasing m odifications an d cavity depth differences in th e active an ten n a may account
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
88
for the changes. T h e effective isotropic radiated p ow er o f the active antenna is 594 mW.
A pproxim ating the active antenna gain w ith th e passive an tenna gain o f 10.2 dB results in
an oscillator p ow er o f 57 m W and a D C -to-R F conversion efficiency o f 57 %.
F urther integrations has since been carried o u t by Flynt et. al. [69 ]
A com pletely
integrated IS A transceiver has been dem onstrated. It uses the FE T IS A as th e transm itter
and local oscillator. T he L O pum ps a pill-type m ixer diode. T he transceiver is unique in
th a t it in tegrates oscillator and m ixing functions w ithin th e antenna volum e.
Isolation is
provided th ro u g h polarization. Prelim inary results at 5.8 G H z exhibit a 5.5 dB isotropic
m ixer conversion loss fo r a 200 M H z IF frequency.
F igure 5.7 show s the top and side
view o f a com plete IS A transceiver front-end. T he to p view differs from Figure 5.5 only
in th e addition o f a m ixer diode used for receiving. T he sid e view sh o w s the bias lines for
th e g ate and drain o f the F E T as well as th e lo w pass filter u sed to rem o v e th e IF. P a rt o f
th e F E T p o w e r is u sed to pum p th e m ixer diode. T he transm itt and receive polarizations
are orthogonal to each o th er to provide som e isolation. T he principal radiation patterns
are show n in Figure 5.8.
T he transceiver radiation perform ance is extrem ely good
considering all the m odifications endured by this antenna a s w ell as th e integration o f an
F E T , mixer, chip capacitors, chip resistors and bias lines.
T h e addition o f an o th er device affects th e transm it and receiv e radiation perform ance o f
the active integrated antenna. T h e E - and H -plane H P B W s a re 49 and 67 degrees which
differs fro m th e 46 an d 64 d egrees m easured w ithout th e m ixer diode. Differences may be
du e to differences in th e antenna dim ensions, feedlines, cavity depth an d diode package.
Figure 5.9 show s h o w to u se th ese transceivers for typical com m unication link. Transm it
and receive functions are orthogonal to each other.
U n it 1 transm its vertical polarized
w aves w hich are received by unit 2. Similarly, unit 2 transm its horizontal polarized w aves
w hich a re received b y unit 1.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
89
Transmit
Polarization
re sisto r
Gate,Drain \
Bias lin e s
Patch
A n te n n a
, .
R e c e iv e
\ Polarization '
FET
Mixer
D io d e
Figure 5.7. IS A transceiver front-end configuration.
ODeg.
ODag.
r45
JOdB
-45
dP
■20
90
135
■10
-135
-90 90
-90
135
-135
180
Figure 5.8. IS A tran sceiver front-end radiation perform ance.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
90
'U N I T
Mixer
UNIT 2
1
FET
T ransm it f 2
FET
f,
'Patch
(a)
T ransm it f 2
T ra n sm it f t
Bias and
FET
Osc.
Patch
an ten n a
x
X
Receive f 2
/
/
Patch
a n ten n a
Bias and
M odulation
Receive f t
Mixer
Mixer
LPF
(b)
Figure 5.9. Two-way communication link using ISA transceiver front-end.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
91
CHAPTER VI
S P A T IA L P O W E R C O M B IN IN G A N D B E A M S T E E R IN G
6.1 I n tr o d u c tio n M any system s req u ire high o u tp u t p o w ers and large antennas with very
directive radiation patterns.
In applications such as radar, com m unication links and
astronom y, th e antennas m ay function as transm itters, receivers o r both simultaneously.
T o produce th ese pencil beam s, larg e reflector antennas have been u sed extensively.
T h ese beam s a re o ften required to be pointed in m any directions.
V arious m echanical
drive schem es a re com m only used to point the antenna beam o v e r several axes.
For a
rad ar, th e m o to r p o sitio n s a re co rrelated to retu rn signals fo r track in g while a
com m unication link obtains a peak in the received signal strength.
pointing
M echanical beam
m ethods a re relatively inexpensive and th e antenna perform ance is not
com prom ised th ro u g h o u t th e scan b u t th ey tend to be large, bulky and slow . Instead o f
reflecto rs, large p lanar arrays w hich consist o f several hundred to several thousand
antennas can also provide narrow , high-gain pencil beam s. T hey m ay be arranged in one,
tw o o r three-dim ensions and pointed mechanically. Since o n e has access to each radiator,
electronic m ethods can also be em ployed for beam scanning.
Load
C o r p o r a te
P arallel
S arlaa
Spacs
F igure 6.1. V ario u s feed structures used fo r antenna arrays.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
92
A ntenna arrays a re not as bulky as com parable reflector antennas, but they o ften require
com plicated feeding netw orks.
Feed netw orks w hich d istrib u te energy to each rad iator
include co rp o rate-, parallel-, series- and space-fed arrays as show n F igure 6.1.
T hese
n etw o rk s increase com plexity, cost and losses in an array ap ertu re b u t th ey can be
in teg rated w ith com ponents to provide both beam w idth flexibility and electronic steering
capability.
A lth o u g h it is m o re com plex and expensive, electronic beam steering is m any o rd e rs o f
m agnitudes fa ste r than m echanical beam pointing.
E lectro n ic beam steering can be
achieved using frequency scanning, beam switching, tim e delay o r phase delay. E lectronic
steering allow s beam m anipulations in tens o f nanoseconds u sin g com pact, m echanicallyrigid structures. D u e to th e increase in com plexity th ere is so m e loss in gain and p attern
d eterio ratio n a t large scan angles but, in many applications (i.e. military radar) beam
steering agility and speed m ay be the m ost critical param eters.
T h e single active antennas described in C hapter V have lim ited applications d u e to low
o u tp u t pow ers. H ow ever, a group o f active antennas can b e synchronized to w o rk as a
single h igh-pow er source. A ctive antennas w ith spatial p o w e r com bining techniques have
b een devised to overcom e low p o w er levels and fabrication c o sts o f mm and sub-m m w ave
system s.
By designing th e antenna and oscillator on a single substrate, one avoids the
design com plexity o f a p o w er distribution netw ork and th e transition/transm ission line
losses associated w ith it.
T h e integrated antenna is also a necessary step to w ard s co m p lete m onolithic sub-system
integration. A t m m and sub-m m w avelengths, m onolithic c o n stru ctio n becom es necessary
to o v ercom e m any fabrication difficulties.
H ow ever, so lid -state devices p ro d u ce very
small p o w e r levels at th ese frequencies. The use o f active anten n as and spatial o r quasioptical p o w er com bining techniques can overcom e m any o f th e se lim itatio n s[7 0 , 71 ].
G eneral p o w er com bining techniques are reviewed in [70] w hile th e current state o f quasioptical p o w e r com bining technology is described in [72 ].
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
93
P o w e r com bining m ethods enable arrays o f active antennas to o p e ra te as o n e coherent
tran sm itter w ith predictable effective radiated pow er, beam w idth and tuning bandw idth.
P ow er-com bining techniques include chip-level, circuit-level and spatial com biners.
S patial an d open reso n ato r com biners involve a large num ber o f tran sisto rs o r diodes. The
individual free-running oscillators m ust b e injection-locked to p ro d u c e a co herent higher
p o w e r R F source.
Injection locking m ay be obtained via m utual coupling, external
feedback, o r an external source.
T he result o f using spatial o r quasi-optical p o w er
com bining techniques is to c re a te a single, coherent and h ig her-pow er signal from many
lo w -p o w e r radiating sources.
Spatial o r quasi-optical p o w er com bining is n o t as lim ited
by size, ohm ic o r dielectric losses o r m oding problem s and allow s th e com bination o f a
g re a te r num ber o f active devices.
T he p atch and g rid have been used fo r m ost active
an tennas and in p o w er com bining [73 , 7 4 , 75 , 76 , 77 ].
6.2
Injection-Locked Oscillators
F o r a single active antenna, an injection-locking
experim ent can be carried o u t to determ ine locking gain and locking bandw idth o f the
radiating oscillator.
T h e locking gain (G l) determ ines th e relative p o w er required to
externally injection lock th e active antenna elem ent. T h e locking b an d w id th is th e range o f
frequencies ov er w hich th e active antenna stays synchronized to th e external so u rce (for a
given injection p o w er level).
T hese tw o characteristics a re contributing facto rs to the
success o f a spatial p o w e r com biner. G£ is defined in d B as
Gl = 10 log
( 6 . 1)
w h ere P0 is th e free-running oscillator p o w er and P, is th e injected signal p o w er level.
A sim ple experim ent can be carried out to determ ine these characteristics. A high-quality,
stable so u rce such a s an H P -83622A synthesized sw eep er can be used alo n g w ith a
standard gain horn. T h e h o rn is placed a know n distance aw ay fro m the ac tiv e antenna.
W hen th e synthesizer freq u en cy is near th e free-running frequency o f th e active antenna, it
w ill attem p t to pull th e free-running oscillator. Its success will d epend o n th e injection
p o w er, active antenna p ow er an d th e spectral quality o f the active antenna.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
94
E xternal injection-locking experim ents w ere perform ed to determ ine the locking-gain and
locking-bandw idth o f th e active antenna.
T he signal quality show s a noticeable
im provem ent in noise level and spectral purity after injection-lock. A s th e injection lock
signal p o w e r decreases, th e re is a corresponding decrease in obtainable locking bandw idth.
A useful fig u re o f m erit w h ich relates these quantities is the external Q uality fa cto r o f the
active an ten n a circuit w h ich is found by using
Q.= — f e
‘ 2 A F \P 0
(6.2)
w h ere Qe is th e external Q -fa c to r , F0 is the operating frequency and AF is th e injectionlocking bandw idth.
T h e locking-gain vs. lo ck in g bandw idth(A F) results fo r a 10.4 mm patch antenna are
show n in F ig u re 6.2. F o r a locking gain o f 30 dB, a locking bandw idth o f 16 M H z w as
obtained at 9.23 G H z. T h e external Q -factor o f the active cavity-enclosed inverted
stripline an ten n a w as c o m p u te d at 36 w hich nearly doubles previous active antenna
external Q -facto rs rep o rted .
40
-50
-45
-40
-35
-30
-25
-20
-15
-10
■5
0
Relative Power (DB)
F ig u re 6.2. L o c k in g bandw idth results for a typical X -band G unn ISA.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
95
6 .3 S p a tia l P o w e r C o m b in in g A s early as 1968, an idea w as in troduced w h ic h u sed a
single lo w pow er solid-state so u rce w hich could b e am plified at each an ten n a in p u t for
spatial p o w er com bining.
Staim an et. al. introduced th e new p o w er com bining schem e
w hich u sed a co rp o rate n etw o rk to feed solid-state amplifiers and an array o f dipole
an ten n as [78 ]. Unlike o th er com biners, p o w e r com bining occurs in ftee-sp ace instead o f
a w av eg u id e cavity. T h e R F losses incurred in the feed netw orks are at low p o w e r levels
w hich m inim izes th eir effect on th e overall losses.
p a ra m e te rs o f th e am plifiers and fe e d netw ork.
Synchronization d e p en d s o n the
O ne hundred tra n sisto r am plifiers w ere
u sed to provide a net gain o f 4.75 dB at 4 1 0 M H z with 100 W atts o f o u tp u t p o w er. This
id ea can b e expanded to include o scillators o r active antennas.
In 1986, Jam es M ink introduced th e use o f an array o f radiating so u rces in a la rg e open
re so n a to r fo r p o w er com bining [71], T he open reso n ato r serves as a qu asi-o p tical cavity
fo r synchronization and p o w er com bining.
T he array spacing, num ber o f o scillato rs and
ta p e r distribution acro ss th e a p e rtu re determ ines the efficiency w ith w hich th e energy
co u p les to the m odes o f th e o p en reso nator.
T he open reso n ato r has tw o b a sic form s:
o n e fo rm uses tw o m etallic m irrors a n d th e o th er uses o n e m etallic and o n e dielectric (o r
partially reflecting) m irror. T he first approach is, essentially, a high p o w er so u rc e w hile
th e seco n d approach also provides a fixed, directive antenna pattern.
A spatial o r quasi-optical p o w e r com b iner can be easily used w ith previously discussed
m echanical beam pointing m ethods.
H ow ever, if th e m irrors o f th e o p en re so n a to r are
rem oved, th e large array o f active antennas can b e synchronized via m utual coupling,
circuit interconnections, a partially reflecting surface o r an external space-feed.
T he
p o w er, frequency and phase o f each so u rce can th en be m anipulated to alter th e radiation
characteristics o f the entire array.
E a c h activ e antenna o r oscillator will have its ow n frequency, m agnitude a n d phase
co m ponents. Synchronization o f each oscillator will determ ine the overall p o w e r output.
I f N oscillators are arranged in a linear array as show n in F igure 6.3, then cou p lin g w ithin
th e inner elem ents can be assum ed to be prim arily due to its nearest neighbors.
This
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
96
coupling occurs th ro u g h the adm ittances (Tc = GC+jB c). This coupling is assum ed to be
con stan t and identical throughout th e array. F o r a m ore general case, this coupling could
also b e frequency dependent.
ooo
Osc. 2
O sc. 1
O se. 3
O sc. N
ooo
Figure 6.3. Linear netw ork o f oscillators.
T h e first source is th e reference (injection source) w ith a current ( / , ) and phase (0=0)
while th e last so u rc e has current (IN) and phase (6!/). T he resulting c u rren ts (/,, i2, • • •, iN)
and p hase co m p o n en ts can be calculated w ith respect to each oscillator. T hese constants
and variables a rran g ed in the linear cascade show n in Figure 6.3 can b e used to calculate
th e instantaneous frequency and oscillator am plitudes over tim e
dt
.= -A a > ,
'
_ l_
2C
2C
2BC+
(Gc sin ($ - <jtiA) - Bc c o s (# -
))
(6.3)
A
+^s\n(<t>i - 0 ,) + ^ f- ( G cs i n ( - <f>M) - Bc cos(<f>t - <f>t+x))
A
A
^ = ± [ a (g Di -O ', - 2 G e) + / , « * * - * , ) ]
+ ^ [ 4 - i ( GcM A -
) + Be sin (4 -
))]
cos( ^ -
)+■Bc M k -
))]
(6.4)
E quations (6.3) a n d (6.4) describe the change in m agnitude and frequency fo r each
oscillator in the array . They can b e integrated to obtain a tim e-dom ain solution for the
entire system.
T h e accuracy o f th e analysis depends on the characterization o f the
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
97
individual oscillato r adm ittan ce (YD(A) = -G D(A) + jB D(A)) as w ell as th e coupling
n etw o rk .
T h e spatial p o w e r com bining concept u ses many low pow er sources to p rovide a single
co h eren t, h igher p o w er o u tp u t.
T o dem onstrate spatial p ow er com bining via m utual
coupling, tw o fo u r elem en t arrays w e re constructed.
E ight active inverted stripline
an ten n as w ere built and te s te d fo r the square and diam ond array. Since m utual coupling is
u sed fo r injection-locking, th e locking gain level betw een tw o o f these active
antennas
w o u ld b e high w ith a co rresponding n arro w locking bandw idth. M echanical tu n in g o f the
d iodes ensured th a t several antennas w o u ld operate w ithin the narrow locking bandw idth.
T his allow s th e use o f a single p ow er supply at a fixed bias voltage for all d io d es and
successful p o w e r com bining. A ntennas 1,2 ,3 and 4 w ere used in a four elem ent square
a rray w hile antennas 5, 6 , 7 and 8 w ere used in a fo u r elem ent diam ond array. T able 6.1
lists th e o p eratin g frequency, oscillator p ow er and efficiency o f each antenna and array as
w ell as the EERP. T he sq u a re array uses 17 mm spacing betw een elements. T h e diam ond
a rray spacing b etw een elem ents is 17 m m along the diagonal. F igure 6.4 show s the array
configurations, spacing an d d io d e position.
17
Patch
#5
17
2
#7
17
i
#3 /
i
#4
/
Figure 6 .4 . 2 x 2 X -band active array configurations.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
98
T h e p o w e r com bining efficiency^ 77) is defined by
7] = ^
“ner x l0 0 %
(6 .5 )
Z -l n
N
Pn is th e p o w e r o f n-th activ e antenna (U sin g G0 as th e gain o f a sim ilar passive antenna)
an d Pcombiner *s t^ie p o w e r o f in jection-locked, pow er-com bined signal (U se th e gain as
NG0 fo r an N elem ent array).
E and H -p lan e p o w e r com bining arrays sh o w high com bining efficiencies. A fo u r elem ent
sq u are a rray w ith 89% com bining efficiency show ed an injection-locked bias tuning
ban d w id th o f 60 M Hz.
Similarly, a fo u r elem ent diam ond array show ed over
com bining efficiency and 50 M H z b ias tu n e d bandw idth.
re su lts o f individual active an tennas as w ell as a
T able
6
2
x
2
Table
6
86%
.1 show s experim ental
square and diam ond arrays.
.1. O perating Frequency, O scillato r P o w e r and Efficiency o f each A ntenna in A rray.
Antenna
Oscillator
Frequency
(GHz)
Oscillator
Power
(mW)
EIRP
(mW)
[dBm]
Combining
Efficiency
(% )
DC to RF
Efficiency
(% )
1
2
3
4
Square
Array*
5
6
7
8
Diamond
Array**
9.498
9.499
9.498
9.497
9.511
9.330
9.340
9.330
9.340
9.380
52.63
56.80
59.32
60.14
204
50.55
49.51
56.72
55.55
184
243.4
262.6
274.3
278.1
3773
233.7
228.9
262.3
256.8
3403
r23.91
[24.2]
T24.41
[24.41
T35.8]
[23.7]
[23.61
[24.2]
[24.11
[35.31
89.03
1.75
1.89
1.98
2.00
1.63
86.57
1.69
1.65
1.89
1.85
1.47
* Antennas 1, 2,3 and 4 are used in the Square Array with 17 mm Spacing.
** Antennas 5, 6, 7 and 8 are used in the Diamond Array with 17 mm Spacing along the diagonal.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
99
T he sq uare a rray w as used to te st tw o elem ent injection-locking and pow er com bining via
m utual coupling fo r E and H -plane elem ents.
T he tw o elem ent H -plane array show s a
p o w e r com bining efficiency o f 86.4% w hile th e E -plane array's efficiency w a s 99.8% . The
fo u r elem ent sq u a re array exhibited p o w e r com bining efficiency o f 89%. T he overall DC
to R F efficiency o f th e fo u r elem ent array is 1.63%. The tw o element p a ttern s for E and
H -plane p o w e r com bining show nearly 2 to 1 beam w idth sh arp en in g o f o v e r a single
elem ent.
T he antenna p attern s for th e square p o w e r com biner array is show n in Figure
6.5. T h e array can b e bias tu n e d from 10.7 to 14 V olts w ithout losing injection lock over
a 60 M H z b an d w id th from 9 .467 to 9.527 GHz. T he output p o w er level varied less than
0.8 dB o v er th e bias tu n ed range.
s, *10
$ -15
•20
•25
-30
•40
•90
•30
Angle (Degrees)
F igure 6.5. R adiation pattern s o f 2 x 2 square array.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
100
The four elem ent d iam o n d array w as also tested for injection-locked p o w e r combining.
The diam ond array p o w e r com bining efficiency is 86.57% w ith an overall D C to RF
conversion efficiency o f 1.47% . T h e array can be bias tim ed from 9.5 to 6.2 V olts w ithout
losing injection lock o v er a 50 M H z bandw idth.
T he radiation p a tte rn s o f the diamond
array are show n in F ig u re 6.6. T he 3 dB beam w idth is similar to th at o f square array. The
cross-polarization lev el o f th e square array is 3 dB low er than th e diam ond array.
-30
-35
-40
-90
Angle (Degrees)
F igure 6.6. R adiation patterns o f 2 x 2 diam ond array.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
101
6.4
Injection Locked Phased Arrays
Phased arrays provide an interesting and
challenging application for active antennas and spatial power combiners. A spatial power
combining phased-array using many iow-power solid-state oscillators to feed antennas was
shown in 1974 by Al-Ani et. al.[79 ] All of the sources are externally synchronized to a
single master oscillator to ensure high power output. These investigators also unveiled a
novel method for beam steering in a spatial power combiner. Several 1 GHz transistor
oscillators were injection-locked to a master source operating at the 4 GHz. Steering of
18 degrees was achieved using DC pulses applied to the bias of each transistor oscillator.
Although novel and potentially useful, this approach stirred little interest. The state of
solid-state device technology in the 1970s and the sparse use o f millimeter wavelengths
would not push for the use o f spatial power combining for over a decade.
In 1986, Stephan showed that spatial power combiners can serve to, not only solve the
need for power at millimeter wavelengths, but also provide a method for beam steering
[80 ]. He used distributed oscillators interinjection locked together in a linear cascade for
coherent power combining. Initially, three VHF transistors oscillators were studied and
simulated [81 ].
State variable analysis of the linearized equations led to closed-form solutions for one and
two-dimensional phased arrays [82 ].
The feasibility and behavior o f a planar spatial
pow er combining phased-arrays was studied at 220 MHz.
A 4x4 planar array was
simulated at broadside and 5.5 degrees o f beam steering. In this investigation, effects of
device failures on the radiation patterns was also simulated.
The results are directly
related to claims o f graceful degradation in distributed systems which may not hold in
beam steerable spatial combiners.
A model was then developed to understand the interaction o f two radiation coupled
oscillators [83 ]. Measurements of the phase and frequency dependence of two coupled
radiating oscillators versus separation allows a more accurate determination of the
oscillator coupling characteristics [84 ]. During the investigation, X-band Gunn-integrated
antennas oscillators were also developed.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
102
X-band microstrip Gunn diode oscillators were used in conjunction with a 1x4 array of
linear tapered slot antennas to demonstrate 7.4 Watts of effective isotropic radiated
power(EIRP) at 10 GHz with + 5 degrees of beam steering [85 ]. Several oscillators are
interinjection-locked to an external source and the single phase shifter was varied over 300
degrees to induce the 10 degrees o f total beam steering [86 ].
6.5 Beam Steering in Active Antenna Arrays
Stephan developed a description and
analysis for a set o f distributed oscillators such that a single, stable power combining
frequency would result and he showed how a progressive phase-shift could be introduced
in a linear cascade o f oscillators to provide beam steering.
For a linear cascade of
oscillators spaced <p electrical degrees apart, the progressive phase-shift (A<f>) which will
steer the array beam <f>a degrees away from boresight is given by
-i
= sin 1
(6 6 )
Where d is the physical separation between sources at the wavelength operation (/I). For a
9
given spacing, the maximum one-sided beam angle allowed
maximum phase progression per element (zl^ax).
depends on the
The progressive phase-shift can be
controlled by the free-running frequencies and thus the bias voltages of each active
antenna element.
As the phase-progression between elements approaches 90 degrees, coupling between
adjacent oscillators approaches zero. This results in a loss of synchronization which may
be due to oscillators with too high of a Q-factor or low mutual coupling. For whatever
reasons, these parameters reduce phase progression between sources and limit the
maximum scan angle. As an example, if one assumes a maximum phase progression of 60
degrees, the maximum attainable scan angle for a typical half-wave array is ±19.5 degrees.
Wider beam steering can be accomplished using more densely packed sources. For the
same phase-progression, a quarter-wavelength spacing has a scan limit o f ±42 degrees.
Re-arranging equation (6.6), the required phase progression
between sources needed
to steer a beam ($,) degrees away from boresight is given by
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
103
A^ = ~
sin (& )
(6 -7)
Following Stephan’s work, other investigators have used various methods to induce wider
scan angles and improved performance. Table 6.2 shows several other investigators which
have used active antennas in phased array demonstrations.
Table 6.2. Beam Steering Demonstrations.
Reference
No.
Array
Class
Antenna Array Device Frequency
Type
Size
Type
(GHz)
[79]
Distributed Sources Conventional
[80,81]
[87]
Distributed Sources Conventional
1x4
JFET
Synchronization
Method
Steering
(angle)
1 GHz
External Circuit: 4th Harmonic
18 degrees
1x4
JFET
10 GHz
External Circuit
10 degrees
Parasitic
leaky-wave
Single Source
Parasitic
Leaky-wave
Distributed Sources
Active
lxll
Gunn
11.5 GHz
N. A.
8 degrees
1x12
Gunn
8.1G H z
N. A.
4 degrees
1x4
FET
2.3 GHz
External Probe-fed
40 degrees
[105]
Distributed Sources
Active
2x2
Gunn
9.5 GHz
M utual Coupling
15 degrees
[98,99]
[100]
[102]
[104]
Distributed Sources
Active
1x4
MESFET
10 GHz
M utual Coupling
27.5 degrees
Distributed Sources Conventional
1x6
MESFET
4.2 GHz
External Circuit
70 degrees
Distributed Sources Conventional
1x3
FET
6 GHz
External Circuit (unilateral)
27 degrees
Distributed Sources
1x2
MESFET
10 GHz
External Circuit (Phase-Locked
Loop)
33 degrees
[88]
[89]
Single Source
Active
In 1992, Fralich and Litva used a 1 x 11 active leaky wave antenna array to obtain beam
steering as a function of oscillator frequency [87 ]. A total o f 8 degrees of steering was
shown for a oscillator tuning range o f 380 MHz at 11.5 GHz. Kirk and Chang [88 ]
simultaneously developed the same active image-line leaky wave antenna in 1992. The
resulting theoretical beam angle for this periodic grating antenna depends on the operating
wavelength(/l0), guided wavelength o f the structure^) and the spacing of the grating (d)
f
<P0 = sin
f i
1
- i
—
K
V
n
Y\
(6 .8)
i—
*
JJ
where n = 0,±1,±2,--- is the n-th space harmonic o f the Floquet modes for the periodic
structure. The dominant space harmonic is known to be n= -1.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
104
Hall and Haskins used two transistors and two external injection locking inputs to
demonstrate some of the most impressive beam steering demonstrations o f active antennas
[89 ]. An edge-fed dual FET integrated microstrip patch antenna oscillator operating at a
frequency o f 2.28 GHz was integrated with two external injection-locking signals to
provide phase shifting capability. The probes at each patch location for the injectionlocking signal provides accurate and stable injection-locking signals for dependable beam
steering. Array patterns at 2.3 GHz are smooth with beam steering of over 40 degrees.
Hall and Haskins also demonstrated a unique polarization agile active antenna using a
similar injection locking scheme [90 ]. Several factors were investigated including radiated
power and phase versus locking frequency for a single active antenna as well as effects of
mutual coupling on the relative phase difference between two oscillators and the radiated
power. Similar investigations have been carried out by Drew, Fusco and McDowall [91,
92],
The use o f interconnecting transmission lines as shown by Stephan in [93 , 94 ] increases
the coupling level and provides a more reliable control mechanism for beam steering. The
ISA array surface shown in Figure 6.4 can be easily modified to include such
interconnections. However, the most reliable and effective method to increase scanning
angle and active array performance was shown by Hall and Haskins in [89], The method
uses an external signal which is fed to each active antenna and precisely locks each source
to the desired frequencies.
The method ensures good repeatable performance and
provides the most control over the array. However, it is more intricate and complex than
other approaches.
Some of the most thorough analysis and experimental demonstrations in active antenna
arrays was carried out by York and Compton. [95 ] In their investigations, they used both
diode and transistor integrated microstrip patch antennas in linear and planar arrays. For
their coupled-oscillator theory, the individual oscillators were modelled by simple , single
tuned resonant circuits with a negative lumped resistance for the active device.
coupling coefficient between oscillators i and j is assumed to be
The
. For N oscillators,
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
105
assuming only nearest neighbor interactions, a method of averages approach yields the
following equations for the individual oscillator amplitudes and phase as a function of time
[96]
(6.9)
( 6 . 10)
where A„ co, and at are the instantaneous amplitude, free-running frequency and freerunning oscillator amplitude at the i-th element. The instantaneous phase of each oscillator
is ,9i = £»,/ + <j>r Q is the quality factor o f the oscillator embedding circuit. Equations (6.9)
and (6.10) describe a set o f coupled Van der Pol oscillators.
When the coupling between oscillators is weak, the individual amplitudes A t remain very
close to their free-running values and the dynamics o f the system are primarily governed
by the phase description o f equation (6.10). If these free-running frequencies are within
the locking bandwidths o f the oscillators and the coupling levels are within the locking
gain, the set o f sources will injection lock to a single frequency
= oij for all i. This is
the basis for spatial power combining where the combiner frequency is given by
(6.11)
where (/ = 1,2, •••,# ) and ay is the steady state frequency. Given that one phase variable
can be set to zero as a phase reference, eqn. (6.11) is a set of N equations and N
unknowns.
It can be solved for the unknown phase distribution and the steady-state
synchronized frequency.
Although 2n_1 phase distributions satisfy eqn. (6.11), not all are stable solutions.
A
perturbation analysis can be used to investigate mode stability. [97 ] The equations are
linearized around the various solutions creating a stability matrix. The various solutions
which make up the solution vector is perturbed by a small amount. The perturbed solution
is stable only if it decays with time.
This, for a linear array, occurs when all of the
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
106
eigenvalues of the stability matrix have negative real parts. This requirement will eliminate
most o f the solutions found in the solution vector.
A linear array o f loosely coupled oscillators was then used by Liao and York to
demonstrate beam steering of linear array of active antennas in 1993. [98 ] From the
analysis, the oscillator frequencies were distributed to develop a phase progression across
the array aperture.
For N loosely coupled oscillator with a coupling phase (<!>), the
following individual free-running frequencies (<a,) achieves a constant phase progression
required in phased arrays:
(l+-£sin(<D+A0)) '
if
(l+^sin(<I>)cos(A0)) * i f
Where A d = $t
>=l
\< i < N
(6.12)
is the progressive phase-shift between elements, ev = e and <J>V= C>.
In this fashion, a 1 x 4 linear array was steered total o f 27.5 degrees was shown using
mutually coupled active antennas. [99 ] The spacing which, as shown earlier, limits the
total scan available with this method is used to achieve a the mutual coupling angle(tf>) of
zero degrees.
Setting 0 = 0 in equation (6.12) simplifies the notation and reduces the distribution
frequencies across the array aperture to
1
if
(l - 2c sin(A6>))_1 i f
1< / < N
(6.13)
i=N
For the X-band demonstration, four FET integrated patch antennas were used. Table 6.3
lists the frequencies and maximum scan angles realized.
The limitations o f mutual coupling schemes led York to other synchronization methods
which would allow more flexibility, tighter control and improved performance. Following
the interinjection locked approach of Stephan, York developed a six-element beam
steerable array which scanned over 70 degrees. [100 ]
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
107
Table 6.3. Frequency Distribution in Scanning Demonstration.
Linear
1x4
Arrays
H-PIane (1x4)
Active
Antenna
Active
Antenna
Active
Antenna
Active
Antenna
MaximumScan
Angle
#1
#2
#3
#4
(GHz)
(GHz)
(GHz)
(GHz)
10.0075
10.00
10.00
9.9925
+12.5 degrees
9.9850
10.00
10.00
10.015
-15 degrees
Several modifications to the original active antenna configuration have been performed in
these beam steerable power combiners.
The FET is no longer integrated with the
microstrip patch antenna. The source is realized in a well-defined microstrip oscillator
which feeds a conventional patch at the nearest radiating edge. The radiating edge on the
far side is used to interinjection lock the sources together. The approach provides much
stronger coupling between antennas and the oscillators can be realized more consistently at
the intended designed frequencies. The element separation at the operating frequency of 4
GHz is about 0.3 /1which would limit the scan angle to ±59 degrees. The array exhibited
continuous H-plane beam scanning
from -30 to 40 degrees for this remarkable
demonstration.
Other investigations have yielded some very good results. An eight element interinjection
locked linear array o f Gunn diode oscillators was demonstrated by Nogi et. al. [101 ] Lin
et. al. introduced a novel unilateral injection locking approach to bring about 33 degrees of
beam scanning in a 2 element array and 27 degrees in a 3 element linear array. [102 ] The
coupling is enhanced with an amplifier which ensures strong coupling between sources.
Kurokawa’s injection-locking theory [103 ] was used to describe the induced phase-shift
and phase progression for beam steering.
r (Qfrce-o),nj^
A^ = sin"1
(6.14)
where the maximum locking range is given by
(L , GP VPou< sm(a)
(6.15)
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
108
where (omj is the injection locking frequency, Qa t is the external quality factor of the
imbedding resonant circuit, P inj is the injection locking power, Pmt is the oscillator output
power, a is the angle between the impedance locus and device line, Gs is the maximum
stable gain of the two-port oscillator and Gp is the square root of the output power ratio of
the two ports. The radiated beam angle induced by the phase progression is given in
equation (6.6). Up to 33 degrees of total beam scanning was shown for a 2 element array
and 27 degrees for a 3 element linear array.
One of the latest demonstrations for active antenna beam scanning was shown by Martinez
and Compton in 1994. [104 ] The configuration uses an FET and an antenna for the free
running source which is coupled to a balanced mixer. A reference source is used to drive
a second microstrip patch antenna which is electromagnetically coupled to pump the
balanced mixer. An offset voltage is used to vary the phase difference between the two
sources. Fifteen degrees of received beam steering was demonstrated.
Navarro and Chang introduced a 2 x 2 beam steering array in 1993 using active inverted
stripline antennas. [105 , 106 ] The array operated at 9.5 GHz with an EIRP o f 3.8 Watts
and an RF combining efficiency o f 89 %. The beam steering performance for the 2x2
square array is shown in Figure 6.7.
V1.V2
Cras*-pcl level
-25
-35
-
40-
Figure 6.7. Beam steering patterns of the 2 x 2 square array.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
109
This beam steering occurs because o f the phase induced on each oscillator as it is pulled
from its free-running to the combiner operating frequency.
The active antenna free-
running frequency depends on the imbedded circuit and applied bias voltage.
Single­
output DC power supplies are usually used to operate an entire array to reduce the overall
cost.
However, a variable voltage can be used to change individual active antenna
operating frequencies.
The difference between self-oscillating frequencies o f active
antennas produces a phase-shift in the injection-locked power combiner. The phase-shift
can be used to electronically steer the beam o f the active antenna power combiner.
Two identical active antennas with different bias voltages oscillate at different frequencies.
When the difference between the self oscillating frequencies is within the lockingbandwidth for a given locking-gain, the antennas will injection-lock at a single oscillating
frequency for power combining. A row o f active antennas can be made to operate at
various self-oscillationg frequencies using a ramped DC voltage.
The elements will
injection lock and the power combiner antenna beam will steer off-broadside. The beam
steering angle is directly proportional to the difference in self-oscillating frequencies which
depends on the applied DC bias voltage.
The square 2 x 2 power combiner exhibited 15 degrees H-plane beam-steering. For Hplane beam steering, one bias voltage(V]) is applied to antennas 1 and 3, and another bias
voltage(V2 ) is applied to antennas 2 and 4.
The self-oscillating frequencies of each
antenna were adjusted to within 0.02% at 12 V bias. When Vj=V 2 the resultant power
combining beam is broadside at 9.511 GHz with 89% combining efficiency. The bias V \
was then lowered to 10.5 V while maintaining V2 at 12 V. Similarly, the bias V2 was
adjusted to 11 V while maintaining V \ at 12 V. The two extremes give a 15 degree Hplane beam steering. All four elements remain injection-locked throughout the tuning
range. As shown, the overall beam remains within 1.2 dB of the maximum at 0 degrees.
The cross-polarization level increases slightly over the broadside combiner but remains at
least 15 dB below the maximum. E-plane beam-steering can be accomplished by pairing
elements 1 with 2 and 3 with 4.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
110
Experiments on linear 1 x 4 arrays were carried out to further demonstrate this beam
steering concept on a more realistic and useful size array. Four element E- and H-plane
arrays o f active Gunn-integrated inverted stripline antennas were used with identical
dimensions to the 2 x 2 arrays. The 1 x 4 E- and H-plane configurations are shown in
Figure 6.8 along with their respective radiation pattern measurements. Maximum E-plane
beam scanning demonstrated was 36 degrees while maximum H-plane beam scanning was
34 degrees. Table 6.4 shows the frequency distribution of the two linear arrays along
with the resulting scan angle.
Table 6.4. Linear Array Frequency Distributions for Beam Steering.
Linear
1x4
Arrays
E-Plane (1x4)
H-plane (4x1)
Active
Antenna
Active
Antenna
#1
#2
(GHz)
(GHz)
Active
Antenna
#3
(GHz)
Active
Antenna
Maximum
#4
Scan Angle
(GHz)
9.3307
9-3157
93157
93001
+20 degrees
93097
93244
93247
93397
-16 degrees
93639
93422
93395
93410
93400
93179
93553
+ 21 degrees
93273
-1 3 degrees
The beam scanning shown above used only mutual coupling to maintain injection locking
between the four sources. There was no interconnections between active antennas or any
external locking inputs. Although simple, the scan angle is limited by the low coupling
levels. When using this method, lower quality oscillators and increased mutual coupling
can increase the scan angle. Mutual coupling due to space waves is increased by the use
of a partially reflecting surface such as a low dielectric substrate. The increase in mutual
coupling between active antennas helps to maintain injection lock over a wider frequency
range which results in wider scanning angles.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
Ill
(a) E -plane
(b) H -plane
F igure 6.8. P a tte rn s o f 1x4 E - and H -plane beam steering arrays.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
112
CHAPTER VII
CONCLUSIONS AND RECOMMENDATIONS
T his d issertation presents som e innovative w o rk in the area o f integrated antennas and
active a rra y beam scanning. P IN and v aracto r integrated ISA s d em onstrated very good
co m p o n en t as well as radiation perform ance.
G unn and FET in teg rated antennas also
d em o n strated som e exceptional com ponent as well as antenna perform ance.
These
antennas m ay have possible u ses in the security m arket w here a low c o st D o p p ler sensor is
needed to provide intruder detection.
m ilitary applications.
T hese antennas also m ake ideal decoys fo r som e
W ith enough pow er, these sources could b e u sed as feeds for
reflecto rs fo r increased directivity. Given th e right stabilization schem es and m odulation
circuitry, in tegrated transceivers m ay be used for w ireless com m unication links.
T he electronic beam steering dem onstrations represent th e latest, leading edge approaches
fo r achieving beam scanning. H ow ever, the use o f a large num ber o f sources is currently
not a cost-effective approach fo r m ost m illim eter-w ave applications.
are th e p rim ary cost-drivers in most millim eter w ave systems.
distrib u ted sources increases b o th the com plexity and cost.
O scillator sources
T herefore, th e use o f
A s m onolithic fabrication
tech n iq u es m ature, th e perform ance o f vital com ponents at m illim eter and sub-m illim eter
w av elen g th s will continue to im prove. V ery large scale production o f th ese com ponents
will lo w e r p er unit costs and allow the use o f distributed system approaches. F o r th e tim e
being, com m ercial applications which have th e prom ise o f mass p ro d u ctio n a re primarily
driven b y cost.
C urrent, state o f the art techniques solve m ost com m ercial applications
w ith conventional system approaches.
This investigation has dem onstrated beam steering as w ell as introduced several novel
integ rated antennas.
Som e k ey accom plishm ents o f this investigation are sum m arized
below :
1. Probe-fed inverted antenna configuration.
=> Efficient radiator w ith higher directivity o v er a conventional patch.
=> N o drilling o r soldering required fo r probe o r shunt diode insertion.
=> B eam w idth flexibility w ith various param eters.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
113
2. P ro b e-fed d io de-integrated inverted antenna configuration.
=> P IN integration for switching o r m odulation applications.
■=> V a ra c to r integration for frequency tuning/m odulation applications.
3. G unn-integrated activ e antenna.
=> Inexpensive, relatively stable radiating oscillators.
=> B ias tunability fo r frequency m odulation.
=> S m o o th radiation patterns and relatively low cross-polarization.
4. F E T -integrated active antenna.
=> Inexpensive, relatively stable radiating oscillators.
=> B ias tunability and low pow er consum ption.
=> V ery sm ooth radiation patterns and very low cross-poiarization.
=> H ig h D C -to -R F conversion efficiencies o f 59 %.
5. Efficient G unn-integrated spatial p ow er combiners.
=> 2 x 2 arrays w ith nearly 90 % R F conversion efficiencies.
=> Stable p o w er an d frequency over tuning bandw idth.
6. B eam -steerable spatial pow er combiners.
=> 2 x 2 phased a rra y w ith nearly 15 degrees o f beam steering.
=> 1x4 E and H -plane arrays w ith nearly 40 degrees o f beam steering.
7. F E T -in teg rated IS A Transceiver.
=> F E T -in teg rated active antenna transm itter.
=> O rthogonally polarized receiver using schottky m ixer diode.
In the fu tu re, there are several m icrow ave com ponents w hich w ould be a direct extension
o f the resu lts presented in this dissertation.
A G unn o r FE T -integrated antenna could be
u sed alo n g w ith a P IN or v aracto r diode fo r active antenna switching, tuning and
m o d u latio n functions.
O th e r passive and active integrations should be investigated fo r
single an ten n a beam con tro l and polarization agility.
could
be
investigated
for
sensitivity
and
A ctive antenna D o p p ler m odules
manufacturability.
C om plete
w ireless
com m unication links should b e co nstructed along with m odulation schem es fo r intelligence
transfer.
W ith re sp e c t to spatial p o w e r com bining and phased-arrays, larger and m ore efficient
com biners should be investigated.
A com bination o f m utual coupling, space w aves,
external circu its and external so urces should b e sought for im proved injection-locking
schem es.
M o re cost-effective solutions are needed to m ake an im m ediate im pact in th e
cu rren t com m ercial m arketplace.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
114
REFERENCES
[1]
J. A. N avarro, Lu Fan and K. Chang, "A ctive Inverted Stripline C ircular P atch
A ntennas fo r Spatial P o w er Com bining," IEEE Transactions on Microwave Theory and
Techniques, Vol. 41, N o. 10, pp. 1856-1863, O c to b er 1993.
[2] J. A. N av arro an d K. Chang, “ Inverted Stripline A ntennas In teg rated w ith P assiv e and
A ctive Solid-State D evices,” IEEE Transactions on Microwave Theory and Techniques,
Vol. 43, N o . 9, pp. 2059 - 2065, Septem ber 1995.
[3] J. A. N av arro and K. Chang, "E lectronic B eam Steering o f A ctive A ntenna A rrays,"
Electronics Letters, Vol. 29, N o. 3, pp. 302 - 304, 4 th February 1993.
[4] K. Chang, Microwave Solid-State Circuits and Applications, N ew Y ork, N Y :
John
W iley & Sons, 1994.
[5] R. M. B arrett, “M icrow ave Printed -C ircuit-A H istorical S urvey,” IRE Transactions
Microwave Theory and Techniques, Vol. 3, N o . 2 , pp. 1 - 9, M arch 1955.
[6]
H . A. W heeler, “Transm ission Line P ro p erties o f Parallel Strips S eparated by a
D ielectric S heet,” IEEE Transactions Microwave Theory and Techniques, Vol. 13, N o . 2,
pp. 172-185, M arch 1965.
[7]
T. C. E dw ards, Foundations fo r Microstrip Circuit Design, N ew Y ork, N Y :
John
W iley & Sons, 1981.
[8]
M . V. Schneider, “M icrostrip Lines fo r M icrow ave In teg rated C ircuits,” The Bell
System Technical Journal, Vol. 48, pp. 1703-1726, July-A ugust 1969.
[9] M. V. Schneider, B. Glance, an d W . F. B odtm ann, “M icrow ave and M illim eter W ave
In teg rated C ircuits fo r R adio System s,” The B ell System Technical Journal, pp. 14211444, M ay - June 1969.
[10]
A . K . G anguly and B. E. Spielman, “D ispersion C haracteristics o f A rbitrarily
C onfigured
T ransm ission
M edia,”
IEEE Transactions Microwave
Theory and
Techniques, Vol. 25, N o. 12, pp. 1138-1141, D ecem ber 1977.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
115
[11]
J. P. V illotte, M. A ubourg an d Y. G arault, “M odified Striplines fo r M icrow ave
In te g ra te d C ircuits,” Electronics Letters, Vol. 14, N o. 18, pp. 6 0 2 -6 0 3 , 31 A ugust 1978.
[12] C. P. W en, “C oplanar W aveguide: A Surface S trip T ransm ission Line Suitable for
N o n -R e c ip ro c a l G yrom agnetic D evices,”
IEEE Transactions Microwave Theory and
Techniques, V ol. 17, N o. 12, pp. 1087-1090, D ecem ber 1969.
[13]
V . F. H anna, “Finite B o undary C orrections to C o planar Strip-L ine Analysis,”
Electronics Letters, Vol. 16, N o. 14, pp. 604-606, July 1980.
[14]
S. B . C ohn, “ Slot L ine on a D ielectric S ubstrate,” IEEE Transactions Microwave
Theory and Techniques, Vol. 17, N o. 10, pp. 768-778, O ctober 1969.
[15]
P. J. M eier, “Integrated Finline M illim eter C o m ponents,”
IEEE Transactions
Microwave Theory and Techniques, V ol. 22, No. 12, pp. 1209-1216, D ecem ber 1974.
[16]
P. J. M eier, “ M illim eter W ave Integrated C ircuits S u spended in th e E-plane o f
R e c ta n g u la r W aveguide” IEEE Transactions Microwave Theory and Techniques, Vol.
26, N o . 10, pp. 726-733, O cto b er 1978.
[17]
E. A. J. M arcatili, “D ielectric R ectangular W aveguide and D irectional C oupler for
In te g ra te d O p tics,” Bell System Technical Journal, V ol. 48, pp. 2133 -2 1 6 0 , Septem ber
1969.
[18] R. M . K nox, “D ielectric W aveguide M icrow ave In teg rated C ircuits-A n O verview ,”
IEEE Transactions Microwave Theory and Techniques, Vol. 2 4 , N o . 11, pp. 806-814,
D e c e m b e r 1976.
[19]
R. M . K n o x and P. P . T oulios, “Integrated C ircuits fo r th e M illim eter Through
O ptical F req u en cy R ange,”
Proceedings o f the Symposium Submillimeter Waves,
P o ly tech n ic P re ss o f th e Polytechnic In stitute o f B rooklyn, N ew Y o rk , pp.497-516, 1970.
[20]
I. J. B ahl a n d K. Ely, “M o d e m M icrow ave S ubstrate M aterials,”
Microwave
Journal, pp. 131-146, Septem ber 1990.
[21] K. W akino, T. N ishikaw a, H . T am ura and T. Sudo, “D ielectric R e so n a to r M aterials
an d T h eir A pplications,” Microwave Journal, pp. 133 - 148, Ju n e 1987.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
116
[22] D. K . Ferry, Gallium Arsenide Technology, Indianapolis, IN: H o w ard W. Sam s &
Co. Inc., 1985.
[23] K. F . L ee and J. S. D ahele, “C haracteristics o f M icrostrip Patch A ntennas an d Som e
M eth o d s o f Im proving F requency Agility and B andw idth,” in C hapter 3 o f M icrostrip
Antennas Handbook, ed ited by J. James, Vol. 1
[24]
I. J. Bahl an d P. B hartia, Microstrip Antennas.
N o rw o o d , M A :
A rtech H ouse,
1980.
[25]
W. C. C h ew and J. A . K ong, "Effect o f Fringing Fields on th e C apacitan ce o f a
C ircular M icrostrip D isk,"
IEEE Transactions on Microwave Theory and Techniques,
Vol. M T T -28, N o. 2, pp. 98-104, February 1980.
[26]
P. B hartia and I. J. Bahl, “Frequency Agile M icrostrip A ntennas,”
Microwave
Journal, pp. 6 7-70, O c to b e r 1982.
[27] A. G . D em eryd, “ A T heoretical Investigation o f the R ectagular M icrostrip A ntenna
E lem ent,” IEEE Transactions on Antennas and Propagation, Vol. 26, N o . 7, pp. 532535, July 1978.
[28] R. B . W aterhouse, “M odelling o f Schottky-B arrier D iode L oaded M icrostrip A rray
E lem en ts,” Electronics Letters, V ol. 28, N o. 19, pp. 1799 - 1 8 0 1 ,1 0 S eptem ber 1993.
[29]
R. E . M unson, “ C onform al M icrostrip Antennas and M icrostrip P hased A rray s,”
IEEE Transactions on Antennas and Propagation, Vol. 22, N o. 1, pp. 74-78, January
1974.
[30]
D. L . Sengupta, “R esonant Frequency o f a Tunable R ectangular P a tc h A n ten n a,”
Electronics Letters, V ol. 20, N o. 15, pp. 614 - 6 1 5 , 19th July 1984.
[31]
G. L. L an an d D . L. Sengupta, “Tunable Circular P atch A ntennas,”
Electronics
Letters, V ol. 21, N o . 22, pp. 1022 - 1023, 2 4 th O ctober 1985.
[32] D. J. R oscoe, L. Shafai, A. Ittipiboon, M . Cuhaci a n d R. Douville, “T unable D ipole
A ntennas,” IEEE AP-S Symposium, Ann A rbor, M I, pp. 672-675, 1992.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
117
[33] J. A. N av arro an d K. C hang, "L ow -C ost Integrated Inverted Stripline A ntennas w ith
S olid-S tate D evices for C om m ercial A pplications,"
IEEE Microwave Theory and
Techniques International Symposium, San D iego, California, pp. 1771-1774, M ay 1994.
[34] A. S. D aryoush, K. B o n tz o s and P. R. H erczfeld,”Optically T uned P atch A ntenna for
P h ased A rray A pplications,” IEEE AP Symposium Digest, Philadelphia, PA , pp. 361-364,
1986.
[35]
K . K urokaw a, “ Som e B asic C haracteristics o f B roadband N egative R esistance
O scillator C ircuits,” Bell System Technical Journal, Vol. 48, pp. 1937-1955, July 1969.
[36]
K. M . Johnson, “L a rg e
Signal G aA s M E S F E T O scillator D esign,” IEEE
Transactions on Microwave Theory and Techniques, Vol. 27, N o. 3 , pp. 2 1 7 -2 2 7 , M arch
1979.
[37] G. R. B asaw ap atn a and R . B. StanclifF, “A Unified A pproach to the D esign o f W ideB and M icrow ave Solid-State O scillators,” IEEE Transactions on Microwave Theory and
Techniques, V ol. 27, N o. 5, p p .379-385, M ay 1979.
[38]
A. P. S. K hanna and J. O bregon, “M icrow ave O scillator Analysis,”
IEEE
Transactions on Microwave Theory and Techniques, Vol. 29, N o. 6 , pp. 6 0 6 -6 0 7 , June
1981.
[39] D . J. E sdale and M. J. H o w ies, “A R eflection Coefficient A pproach to th e D esign o f
O ne P o rt N egative Im pedance O scillators,”
IEEE Transactions on Microwave Theory
and Techniques, Vol. 29, N o. 8, pp. 770-776, A ugust 1981.
[40] S. Y ngvesson, Microwave Semiconductor Devices, B oston, M A: K luw er A cadem ic
Publishers, 1991.
[41]
B. D. B ates and P. J. K han, “ Stability o f M ultifrequency N egative-R esistance
O scillators,” IEEE Transactions on Microwave Theory and Techniques, Vol. 32, N o. 10,
pp. 1 3 1 0 -1 3 1 8 , O cto b er 1984.
[42]
E. L. H olzm an, Solid-State Microwave Power Oscillator Design, B o sto n , MA:
A rtech H ouse, 1992.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
118
[43]
J. O b reg o n and A. P. S. K hanna, “E xact D erivation o f th e N onlinear N egative-
R esistan ce O scillator Pulling F ig u re,”
IEEE Transactions on Microwave Theory and
Techniques, Vol. 3 0 , N o. 7, pp. 1109-1 111, July 1982.
[44] R . A dler, “ A Study o f L ock in g Phenom ena in O scillators,” Proceedings o f the IRE.
and Waves and Electrons, pp. 351-357, June, 1946.
[45]
M . O dyniec, “N onlinear Synchronized L C O scillators: T h eo ry and Sim ulation,”
IEEE Transactions on Microwave Theory and Techniques, Vol. 41, N o. 5, pp. 774-780,
M ay 1993.
[46]
K
K uro k aw a, “N oise in Synchronized O scillators,”
IEEE Transactions on
Microwave Theory and Techniques, Vol. 16, N o. 4, pp. 234-240, A pril 1968.
[47] W . O. Schlosser, “N oise in M utually Synchronized O scillators,” IEEE Transactions
on Microwave Theory and Techniques, V ol. 16, N o. 9, pp. 732 - 737, S eptem ber 1968.
[48]
K . F. S chunem ann and K. Behm , “N onlinear N o ise T h e o ry for Synchronized
O scillato rs,” IEEE Transactions on Microwave Theory and Techniques, Vol. 27, No. 5,
p p .4 5 2 -4 5 8 , O c to b e r 1979.
[49]
R . B. W aterh o u se and N . V . Shuley, "Dual Frequency M icrostrip R ectangular
P atch es," Electronics Letters, V ol. 28, N o. 7, pp. 606-607, 26th M arch 1992.
[50] D .B . R utledge, D .P . N eikirk an d D .P. Kasilingam, “Integrated C ircuit Antennas," in
C h a p te r 1 o f Infrared and Millimeter-Waves, edited by K. J. B u tto n , Vol. 10, pp. 1 - 90,
A cadem ic P ress In c., 1983.
[51]
H . J. T hom as, D . L. F u d g e and G. M orris, "G unn S o u rce Integrated w ith a
M icro strip P atch ," Microwaves andRF, pp. 87-89, February, 1985.
[52] T. O. P erkins, "Active M icro strip C ircular Patch A ntenna," Microwave Journal, pp.
110-117, M arch, 1987.
[53] B . R o b ert, T. R azban and A. Papiem ik, “C om pact Amplifier Integration in Square
P a tc h A ntenna,”
Electronics Letters, Vol. 28, N o. 14, pp. 1808-1810, 10th Septem ber
1992.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
119
[54] K. D . Stephan and T. Ito h , “A P lanar Q uasi-O ptical Subharm onically P u m p e d M ixer
C haracterized by Iso tro p ic C onversion L oss,” IEEE Microwave Theory and Techniques,
Vol. 32, N o . 1, pp. 97-102, Ja n u a ry 1984.
[55] A. M o rtazaw i and T. Ito h , "A Periodic S econd H arm onic Spatial P o w e r C om bining
O scillator," IEEE M TTM icrowave Symposium Digest, Dallas^ T X , pp. 1213-1216, M ay
1990.
[56]
V. F. Fusco, "S elf-D etection Perform ance o f A ctive M icro strip A ntennas,"
Electronics Letters, V ol. 28, N o . 14, pp. 1362-1363, July 1992.
[57] U. G uttich , "Planar In te g ra te d 20 G H z R eceiver in Slotline and C o p lan ar W aveguide
T echnique," Microwave and Optical Technology Letters, Vol. 2, N o . 11, pp. 404-406,
N o v em b er 1989.
[58] K. C hang, K. A. H um m er, an d G. G opalakrishnan, "A ctive R adiating E lem e n t U sing
F E T S o u rce Integrated w ith M icro strip Patch A ntenna," Electronic Letters, V ol. 24, N o.
21, pp. 1347-1348, 13th O c to b e r 1988.
[59] V. F. Fusco, "S eries F eed b ack In tegrated A ctive M icrostrip A nten n a Synthesis and
C haracterization," Electronics Letters, V ol. 28, N o. 1, pp. 89-91, 2nd Ja n u ary 1992.
[60]
R. A. Y ork, R. M . M artin ez and R. C. C om pton, "Hybrid T ra n sisto r and P atch
A ntenna E lem en t fo r A rray A pplications," Electronics Letters, V ol. 2 6 , pp. 494-495,
M arch 1990.
[61] P. M . H askins, P. S. H all and J. S. D ahele, "A ctive P atch A n ten n a E lem e n t w ith
D io d e T uning," Electronics Letters, Vol. 27, N o. 20, pp. 1846-1847, 2 6 th S eptem ber
1991.
[62] J. A. N avarro, Y. Shu a n d K. C hang, "A ctive E ndfire A ntenna E lem en ts an d P o w e r
C om biners U sing N o tc h A ntennas," IEEE M TT Microwave Symposium D igest, D allas,
T X , pp. 7 9 3 -7 9 6 , M ay 1990.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
120
[63]
J. A . N avarro, Y . Shu and K. Chang, "B roadband Electronically Tunable Planar
A ctiv e R adiating E lem ents and Spatial P o w e r C om biners U sing N o tch A ntennas," IEEE
Transactions on Microwave Theory and Techniques. V ol. M T T -40, N o. 2, pp. 323-328,
February, 1992.
[64]
J. A. N av arro an d K. Chang, "B roadband E lectronically Tunable Integrated C ircuit
A ctiv e R adiating E lem ents," Microwave Journal, V ol. 35, pp. 87 - 101, O ctober, 1992.
[65]
K. D. Stephan, “In ter-Injection-L ocked O scillators w ith A pplications to Spatial
P o w e r C om bining an d Phased A rrays,” IEEE M TT Microwave Symposium Digest,
B altim ore, M D , pp. 159 - 162, June 1986.
[66] R. A. Y o rk and R. C. C om pton, "D ual-D evice A ctiv e Patch A ntenna w ith Im proved
R adiation C haracteristics," Electronics Letters, V ol. 28, N o . 11, pp. 1019-1021, 21st M ay
1992.
[67] J. A. N av arro , K. Chang, J. T olleson, S. S anzgiri and R. Q. Lee, "A 29.3 G H z C avity
E n clo sed A p ertu re-C o u p led C ircular P atch A ntenna for M icrow ave C ircuit Integration,"
IEEE Microwave and Guided Wave Letters, Vol. 1, N o. 7, pp. 170-171, July 1991.
[68]
J. A. N avarro, L. F an and K. C hang, "N ovel Q uasi-O ptical A ctive A ntenna Using
In teg rated F E T In v erted Stripline P atch ,"
Electronics Letters, N o. 8, pp. 655-657, 14
A pril 1994.
[69]
R. Flynt, L. Fan, J. A. N avarro and K. C hang, “L ow -C ost and Com pact A ctive
In te g ra te d A ntenna T ransceiver fo r System A pplications,” IEEE M TT Microwave
Symposium Digest, O rlando, FL , pp. 956 - 959, M ay 1995.
[70]
K. C hang and C. Sun, "M illim eter-W ave P ow er-C om bining Techniques," IEEE
Transactions on Microwave Theory and Techniques, V ol. M T T -31, N o. 2, pp. 91-107,
February, 1983.
[71]
J. W . M ink, "Q uasi-optical P o w e r C om bining o f Solid-State M illim eter-wave
S ources," IEEE Transactions on Microwave Theory and Techniques, Vol. M T T -34, pp.
2 7 3 -2 7 9 , February, 1986.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
121
[72]
J. C. W iltse an d J. W. M ink, "Q uasi-optical P o w e r C om bining o f Solid-State
S ources," Microwave Journal, pp. 144-156, February 1992.
[73]
S. Y o u n g and K. D . Stephan, "Stabilization and P o w e r C om bining o f Planar
M icro w av e O scillators w ith an O pen R esonator," IEEE M T T Microwave Symposium
Digest, L a s V egas, N V , pp. 185-188, M ay 1987.
[74] K . C hang, K. A . H u m m er and J. L. Klein, "Experim ents on Injection Locking o f
A ctive A n te n n a E lem ents fo r A ctive Phased A rrays and S patial P o w er C om biners," IEEE
Transactions on Microwave Theory and Techniques, V ol. M T T -3 7 , N o . 7, pp. 10781084, Ju ly 1989.
[75] J. B irkeland and T. Ito h , "Planar FE T O scillators U sing P eriodic M icrostrip P atch
A ntennas," IEEE Transactions on Microwave Theory and Techniques,
Vol. M T T -37,
N o. 8, pp. 1232-1236, A ug u st 1989.
[76] Z. B . P opovic, R . M . W eikle II, M . Kim, K. A. P o tte r a n d D . B. R utledge, "B ar-G rid
O scillators," IEEE Transactions on Microwave Theory and Techniques, Vol. M T T -38,
N o. 3, pp. 225-2 3 0 , M arch 1990.
[77] R. A . Y o rk and R. C. C om pton, "Q uasi-O ptical P o w er C om bining U sing M utually
S ynchronized O scillator A rrays,"
IEEE Transactions on Microwave Theory and
Techniques, V ol. M T T -39, N o . 6, pp. 1000 - 1009, June 1991.
[78] D. Staim an, M. E . B reese and W. T. P atto n , “N ew T ech n iq u e fo r Com bining SolidS ta te S o u rces,” IEEE Journal o f Solid-State Circuits, Vol. S C -3, N o. 3, pp. 238 - 243,
S eptem ber 1968.
[79] A. H . Al-Ani, A. L. C ullen and J. R. F orrest, "A P hase-L ocking M ethod for B eam
Steering in A ctive A rray A ntennas," IEEE Transactions on Microwave Theory and
Techniques, V ol. M T T -22, N o . 6, pp. 698-703, Ju n e 1974.
[80] K. D . S tephan, "Inter-Injection L ocked O scillators for P o w e r C om bining and Phased
A rrays," IEEE T ransactions Microwave Theory and Techniques, Vol. 34, N o. 10, pp.
1017-1025, O c to b e r 1986.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
122
[81] K. D. S te p h a n and W. A. M organ, “ Analysis o f Inter-Injection-L ocked O scillators
fo r Integrated P h a se d A rrays,” IEEE Transactions on Antennas and Propagation, Vol.
35, N o. 7, pp. 771 - 7 8 1 , July, 1987
[82] W . A. M o rg a n Jr. and K. D . Stephan, "R adiation C oupling o f Inter-In jectio n L ocked
O scillators," SPIE M illimeter Wave Technology IV and Radio Frequency Power Sources,
V ol. 791, pp, 6 9 - 76, 1987.
[83]
K. D.
S tep h an and
S. L. Y oung, "M ode
Stability o f R adiation-C oupled
In terin jectio n -L o ck ed O scillators for Integrated Phased A rrays," IEEE Transactions on
Microwave Theory and Techniques, Vol. 36, N o. 5, pp. 921-924, M ay 1988.
[84]
W . A. M o rg a n Jr. and K . D. Stephan, "Inter-Injection L ocking- A N o v el Phase
C o n tro l T ech n iq u e fo r M onolithic Phased A rrays,"
12th Conference on Infrared and
M illimeter Waves, pp. 81-82, D ecem ber, 1987.
[85] W . A. M o rg a n an d K. D . Stephan, “A n X -B and Experim ental M o d el o f a M illim eterW av e Inter-in jectio n -L o ck ed P hased Array System ,”
IEEE Transactions on Antennas
and Propagation, V ol. 36, N o. 11, pp. 1641 - 1645, N ovem ber 1988.
[86]
R. Fralich an d J. Litva, "B eam -Steerable A ctive A rray A ntenna,"
Electronics
Letters, Vol. 28, N o . 2, pp. 184-185, 16th January 1992.
[87]
A. M. K irk a n d K. C hang, "Integrated Im age Line Steerable A ctive A ntennas,"
International Journal o f Infrared and Millimeter Waves, Vol. 13, N o. 6, p p .8 4 I-8 5 1 ,
Ju n e 1992.
[88] P. S. H all a n d P. M. H askins, "M icrostrip A ctive P atch array w ith B eam Scanning,"
Electronics Letters, V ol. 28, N o. 22, pp. 2056-2057, 22nd O ctober 1992.
[89]
P. S. H all, I. L . M o rro w , P. M. H askins and J. S. D ahele, “P hase C o n tro l in
In jection-L ocked M icro strip A ntennas,” IEEE M TT Microwave Symposium Digest, San
D iego, C A , pp. 12 2 7 - 1230, M ay 1994.
[90] S. D rew a n d V . F. Fusco, "Phase M odulated A ctive A ntenna," Electronics Letters,
V ol. 29, N o. 10, pp. 835-836, 13th M ay 1993.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
123
[91]
V. F. F usco, S. D rew and D . S. M cD ow all, "Injection L o cking P henom ena in an
A ctive
M icro strip
A ntenna,"
IEE International Conference on Antennas and
Propagation, E dingburgh, U K , pp. 295-298, A pril, 1993.
[92]
P. M. H askins, P . S. Hall an d J. S. D ahele, "A ctive P a tch A ntenna E lem ent W ith
D iode T uning," Electronics Letters, Vol. 27, N o. 20, pp. 1846-1847, 2 6 th S eptem ber,
1991.
[93]
R. A. Y o rk an d R. C . C om pton, "C oupled-O scillator A rrays for M illim eter-W ave
P ow er-C om bining a n d M ode-L ocking," IEEE M TT Microwave Symposium Digest,
A lbuquerque, N M , pp. 429-432, June, 1992.
[94] R. A. Y ork, "N on-L inear Analysis o f P hase R elationships in Q uasi-O ptical O scillato r
A rrays," IEEE Transactions on Microwave Theory and Techniques, Vol. 41, N o. 10, pp.
1799-1809, O ctober, 1993.
[95]
P. L iao an d R . A . Y ork, "Phase-Shifterless B eam -Scanning U sing C o u p led -
O scillators: T h eo ry a n d E xperim ent," IEEE AP Symposium Digest, Ann A rbor, M I, pp.
668 - 671, 1993.
[96] P. L iao and R . A . Y ork, "A N e w Phase-Shifterless B eam -Scanning T echnique U sing
A rrays o f C o upled O scillators,"
IEEE Transactions on Microwave Theory and
Techniques, Vol. 4 1 , N o . 10, pp. 1810-1815, O c to b er 1993.
[97] P. L iao and R . A . Y ork, "A Six-Elem ent Beam -Scanning A rray," IEEE Microwave
and Guided Wave Letters, V ol. 4, N o . 1, pp. 20-22, January 1994.
[98] J. L in a n d T. Ito h , “A M o d e A nalysis and Stabilization o f a Spatial P o w e r C om bining
A rray w ith S trongly C o u p led O scillators,” IEEE Transactions on Microwave Theory and
Techniques, Vol. 4 1 , N o . 10, pp. 1827 - 1837. O c to b er 1993.
[99] J. Lin, S. T. C h e w and T. Ito h , “A U nilateral Injection-L ocking T ype A ctive P hased
A rray fo r B eam Scanning,” IEEE M TT Symposium Digest, San D iego, C A PP- 1231 1234, June, 1994.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
124
[100]
K.
K urokaw a,
“Injection
L ocking o f M icrow ave
S olid-S tate
O scillators,”
Proceedings o f the IEEE, V ol. 61, pp. 1 3 8 6 - 1410, O ctober 1973.
[101] R . D. M artinez and R . C. C om pton, “E lectronic Beam S teering o f A ctive A rrays
w ith P h ase-L o ck ed L o o p s,” IEEE Microwave and Guided Wave Letters, V ol. 4, N o. 6,
pp. 166 - 168, June, 1994.
[102]
R . A. Y ork and R. C. C om pton, "Experim ental O bservation and Sim ulation o f
M o d e-L o ck in g P henom ena in C oupled O scillator A rrays," Journal o f Applied Physics,
pp. 2 9 5 9 -2 9 6 5 , 15th M arch 1992.
[103] J. J. L ynch and R. A . Y ork, "A M o d e L ocked A rray o f C oupled P hase-L ocked
L o o p s,"
IEEE Microwave and Guided Wave Letters, Vol. 5, N o . 7, pp. 21 3 -2 1 5 , July
1995.
Reproduced with perm ission of the copyright owner. Further reproduction prohibited without permission.
125
VITA
Ju lio A ngel N a v a rro w as b o m in C o rd o b a, A rgentina o n N ovem ber 1 3 ,1 9 6 4 . H e received
th e B .S ., M .S . and Ph.D . d eg rees in electrical engineering from T exas A & M U niversity in
1988, 1990 a n d 1995, respectively.
H e w as in G eneral D y nam ics-F ort W o rth ’s co-operative ed ucation program from M ay,
1985 to February, 1991. A t G en eral D ynam ics (now Lockheed-Fort Worth Company), he
w a s a part o f various g ro u p s including: A vionics System s D esign, A dvanced Technology
& System s E ngineering, E m itte rs & Intelligence, A ntenna System s and R a d a r CrossS ectio n R esearch.
H e has desig n ed w ideband, low -observable antennas as w ell as many
o th e r printed circuit c o m p o n en ts and antennas.
A t T exas A & M U niversity, he has
receiv ed a G ra d u a te E n gineering Fellow ship in 1989, a N ational Science Foundation
Fellow ship in 1991, th e E b en sb arg er E lectrical Engineering aw ard in 1992 a n d a N A SA L e w is T raining G rant in 1993.
H e w as a research and teaching assistant fro m January
1989 to A u g u st 1994 w here h e in tro d u ced m any integrated and active integrated antennas.
T h e se include G unn-integrated v aractor-tunable no tch antennas used in spatial pow er
com bining arrays.
H e also in teg rated P IN , varacto r, m ixer and G unn d io d e s as well as
tra n sisto rs w ith inverted stripline p atch antennas u sed for spatial p ow er com bining and
b e a m steering applications. D u rin g this tim e period, he developed novel G unn V C O s, P IN
sw itchable & v aracto r-tu n ab le unip lan ar filters and ring resonators. In 1990, h e designed
th e K a-band ap ertu re-co u p led circu lar patch antennas used in th e N A S A -L ew is KaMIST
su b -array project.
H e has b e e n a t E psilon-L am bda E lectronics C o rp o ratio n since
S ep tem b er 1994 w here he d esig n s and develops m icrostrip and w aveguide an te n n a arrays
fro m X - to W -band.
Specifically, m icrostrip p atch antenna arrays w hich operate at
2 4 .1 2 5 , 35, 5 9 .5 , 76.5 and 94 G H z. H e is currently w orking o n cost-effective approaches
fo r large p h ase scanned arrays a n d low -loss feed netw orks fo r efficient apertures.
H e has published ov er 20 technical papers and h olds several patents in th e m icrow ave
field.
H e m ay be co ntacted at th e D epartm ent o f Electrical Engineering, T e x as A&M
U niversity, C ollege Station, T e x a s 77843-3128.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Документ
Категория
Без категории
Просмотров
0
Размер файла
5 101 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа