Introduction to Microstrip Antennas

Introduction to Microstrip Antennas
David R. Jackson
Dept. of ECE
University of Houston
Contact Information
David R. Jackson
Dept. of ECE
N308 Engineering Building 1
University of Houston
Houston, TX 77204-4005
Phone: 713-743-4426
Fax: 713-743-4444
Email: djackson@uh.edu
Purpose of Short Course
Provide an introduction to microstrip antennas.

Provide a physical and mathematical basis for understanding how
microstrip antennas work.
Provide a physical understanding of the basic physical properties
of microstrip antennas.
Provide an overview of some of the recent advances and trends in
the area (but not an exhaustive survey directed towards
understanding the fundamental principles).
Additional Resources
Some basic references are provided at the end of these viewgraphs.
You are welcome to visit a website that goes along with a course at the
University of Houston on microstrip antennas (PowerPoint viewgraphs
from the course may be found there, along with the viewgraphs from this
short course).
ECE 6345: Microstrip Antennas

http://www.egr.uh.edu/courses/ece/ece6345/web/welcome.html
Note: You are welcome to use anything that you find on this
website,
as long as you please acknowledge the source.
4
Outline
Overview of microstrip antennas
Feeding methods
Basic principles of operation
General characteristics
CAD Formulas
Radiation pattern
Input Impedance
Circular polarization
Circular patch
Improving bandwidth
Miniaturization
Reducing surface waves and lateral radiation
5
Notation
c speed of light in free space
c 2.99792458 108 m/s
0 wavelength of free space
0 c / f
k0 wavenumber of free space
k0 0 0 2 / 0
k1 k0 r
k1 wavenumber of substrate
0 intrinsic impedance of free space
1 intrinsic impedance of substrate

r relative permtitivity (dielectric constant) of substrate
reff effective relative permtitivity
(accouting for fringing of flux lines at edges)
eff
rc
complex effective relative permtitivity
(used in the cavity model to account for all losses)
0
376.7303
0
1 0 / r
1
0 0
0 4 107 H/m
0
1
8.854188 1012 F/m
2
0c
6
Outline
Feeding methods
CAD Formulas
Radiation pattern
Input Impedance
Circular patch
Improving bandwidth
Miniaturization
Overview of Microstrip Antennas

Also called patch antennas
One of the most useful antennas at microwave frequencies (f > 1 GHz).
It usually consists of a metal patch on top of a grounded dielectric substrate.
The patch may be in a variety of shapes, but rectangular and circular are the
most common.
Microstrip line feed
Coax feed
8

Common Shapes
Rectangular
Square
Elliptical
Circular
Annular ring
Triangular

History
Invented by Bob Munson in 1972 (but earlier work by Dechamps
goes back to1953).
Became popular starting in the 1970s.
G. Deschamps and W. Sichak, Microstrip Microwave Antennas, Proc. of

Third Symp. on USAF Antenna Research and Development Program,
October 1822, 1953.
R. E. Munson, Microstrip Phased Array Antennas, Proc. of TwentySecond Symp. on USAF Antenna Research and Development Program,
October 1972.
R. E. Munson, Conformal Microstrip Antennas and Microstrip Phased
Arrays, IEEE Trans. Antennas Propagat., vol. AP-22, no. 1 (January
1974): 7478.
10

Advantages of Microstrip Antennas
Low profile (can even be conformal, i.e. flexible to conform to a

surface).
Easy to fabricate (use etching and photolithography).
Easy to feed (coaxial cable, microstrip line, etc.).
Easy to use in an array or incorporate with other microstrip circuit
elements.
Patterns are somewhat hemispherical, with a moderate directivity
(about 6-8 dB is typical).
11

Disadvantages of Microstrip Antennas
Low bandwidth (but can be improved by a variety of techniques). Bandwidths of
a few percent are typical. Bandwidth is roughly proportional to the substrate
thickness and inversely proportional to the substrate permittivity.
Efficiency may be lower than with other antennas. Efficiency is limited by
conductor and dielectric losses*, and by surface-wave loss**.
Only used at microwave frequencies and above (the substrate becomes too
large at lower frequencies).
Cannot handle extremely large amounts of power (dielectric breakdown).
* Conductor and dielectric losses become more severe for thinner
substrates.
** Surface-wave losses become more severe for thicker substrates
(unless air or foam is used).
12

Applications
Applications include:
Satellite communications
Microwave communications
Cell phone antennas
GPS antennas
13

Microstrip
antenna
Filter
DC supply Micro-D
connector
K-connector
LNA
PD
Fiber input with
collimating lens
Diplexer
Microstrip Antenna Integrated into a System: HIC Antenna Base-Station for 28-43 GHz
(Photo courtesy of Dr. Rodney B. Waterhouse)
14

Arrays
Linear array (1-D corporate feed)

22 array
2-D 8X8 corporate-fed array
4 8 corporate-fed / series-fed array
15

Wraparound Array (conformal)
The substrate is so thin that it can be bent to conform to the surface.

16

Rectangular patch
y
Js
W
h
Note: L is the resonant dimension (direction of current flow).

The width W is usually chosen to be larger than L (to get higher
bandwidth). However, usually W < 2L (to avoid problems with the
(0,2) mode).
W = 1.5L is typical.
17

Circular Patch
y
a
h
The location of the feed determines the direction of current flow and hence
the polarization of the radiated field.
18
Outline
Feeding methods
CAD Formulas
Radiation pattern
Input Impedance
Circular patch
Improving bandwidth
Miniaturization
19
Feeding Methods
Some of the more common methods for

feeding microstrip antennas are shown.
The feeding methods are illustrated for a rectangular patch,

but the principles apply for circular and other shapes as well.
20
Feeding Methods
z
Coaxial Feed
y
Surface current
A feed along the centerline

is the most common
(minimizes higher-order
modes and cross-pol).
W
Feed at (x0, y0)
x
21
Feeding Methods
Coaxial Feed
x0
R Redge cos
(The resistance varies as the

square of the modal field
shape.)
Advantages:
Simple
Directly compatible with coaxial cables
Easy to obtain input match by adjusting feed position
W
Disadvantages:
Significant probe (feed) radiation for thicker substrates
Significant probe inductance for thicker substrates (limits
bandwidth)
Not easily compatible with arrays
x0 , y0
22
Feeding Methods
Inset Feed
Advantages:
Simple
Allows for planar feeding
Easy to use with arrays
Easy to obtain input match
Microstrip line
Disadvantages:
Significant line radiation for thicker substrates
For deep notches, patch current and radiation pattern may show distortion
23
Feeding Methods
Inset Feed
Recent work has shown
that the resonant input
resistance varies as
Wf
2 x0
Rin A cos 2
B

2 L
x0
S
L
The coefficients A and B depend on the notch width S but (to a good
approximation) not on the line width Wf .
Y. Hu, D. R. Jackson, J. T. Williams, and S. A. Long, Characterization of the Input Impedance of
the Inset-Fed Rectangular Microstrip Antenna, IEEE Trans. Antennas and Propagation, Vol. 56,
No. 10, pp. 3314-3318, Oct. 2008.
24
Feeding Methods
Proximity-coupled Feed
(Electromagnetically-coupled Feed)
Advantages:
Less line radiation compared to microstrip feed
Can allow for higher bandwidth (no probe inductance, so
substrate can be thicker)
Patch
Microstrip line
Top view
Microstrip
line
Disadvantages:
Requires multilayer fabrication
Alignment is important for input match
25
Feeding Methods
Gap-coupled Feed
Advantages:
Can allow for a match even with high edge impedances, where a notch
might be too large (e.g., when using high permittivity)
Gap
Patch
Microstrip line
Patch
Top view
Microstrip
line
Disadvantages:
Requires accurate gap fabrication
Requires full-wave design
26
Feeding Methods
Aperture-coupled Patch (ACP)
Advantages:
Slot
Feed-line radiation is isolated from patch radiation

Higher bandwidth is possible since probe inductance is
eliminated (allowing for a thick substrate), and also a
double-resonance can be created
Allows for use of different substrates to optimize
antenna and feed-circuit performance
Disadvantages:
Requires multilayer fabrication
Alignment is important for input match
Top view
Microstrip
line
Patch
Slot
Microstrip line
27
Outline
Feeding methods
CAD Formulas
Radiation pattern
Input Impedance
Circular patch
Improving bandwidth
Miniaturization
28
Basic Principles of Operation

The basic principles are illustrated here for a rectangular patch, but the
principles apply similarly for other patch shapes.
We use the cavity model to explain the operation of the patch antenna.
z
h
PMC
Y. T. Lo, D. Solomon, and W. F. Richards, Theory and Experiment on Microstrip Antennas,

IEEE Trans. Antennas Propagat., vol. AP-27, no. 3 (March 1979): 137145.
29

Main Ideas:
The patch acts approximately as a resonant cavity (with short-circuit (PEC)
walls on top and bottom, open-circuit (PMC) walls on the edges).
In a cavity, only certain modes are allowed to exist, at different resonance
frequencies.
If the antenna is excited at a resonance frequency, a strong field is set up inside
the cavity, and a strong current on the (bottom) surface of the patch. This
produces significant radiation (a good antenna).
Note: As the substrate thickness gets smaller the patch current radiates less, due to
image cancellation. However, the Q of the resonant mode also increases, making the
patch currents stronger at resonance. These two effects cancel, allowing the patch to
radiate well even for small substrate thicknesses.
(See next slide.)

30

A microstrip antenna can radiate well, even with a thin substrate.
As the substrate gets thinner the patch current radiates less, due to image
cancellation (current and image are separated by 2h).
However, the Q of the resonant cavity mode also increases, making the
patch currents stronger at resonance.
These two effects cancel, allowing the patch to radiate well even for thin
substrates (though the bandwidth decreases).
1
Js
h
Js
31

Thin Substrate Approximation
On patch and ground plane:
E z E z
Et 0
Inside the patch cavity, because of the thin substrate,

the electric field vector is approximately independent of z.
E x, y, z z Ez x, y
Hence
E z x, y
32

Magnetic field inside patch cavity:
1
H
E
j
1
z x, y
zE
j
1
z Ez x, y
33

1
H x, y
z Ez x, y
j
Note: The magnetic field is purely horizontal.
(The mode is TMz.)
E z x, y
H x, y
34

Magnetic-wall Approximation
On the edges of the patch:
J s n 0
(Js is the sum of the top and bottom surface currents.)
On the bottom surface of the patch

conductor, at the edge of the patch,
we have:
bot
s
n 0
assuming J
bot
s
t
J stop
Also,
J sbot z H
Js
H tbot 0
H tbot 0
35

Since the magnetic field is approximately
independent of z, we have an approximate PMC
condition on the entire vertical edge.
y
L
H t 0 (PMC)
W
or
Js
n H x, y 0
H
h
Actual patch
edge
t
PMC
h
PMC Model
36

y
n H x, y 0
1
H x, y
z E z x, y
j
W
Hence,
n z Ez x, y 0
z n Ez x, y 0
Ez
0
n
Js
n z Ez x, y z n E z x, y Ez x, y n z
(Neumann B.C.)
h
PMC
37

Resonance Frequencies
Ez k Ez 0
2
k k1 k0 r
E z x, y
From separation of variables:

m x
n y
cos
W
L
Ez cos
PMC
(TMmn mode)
L
m n
We then have

L
W
Hence m n
L W
k Ez 0
2
1
k12 0
38

We thus have
y
2
m n
k

L
W
E z x, y
2
1
PMC
Recall that
k1 k0 r 0 0 r
L
2 f
Hence
c
2 r
m n
L W
c 1/ 0 0
39

y
Hence
Ez x, y
f f mn
(resonance frequency of (m,n) mode)
PMC
f mn
c
2 r
m n
L
W
40

Dominant (1,0) mode
y
This structure operates as a fat planar dipole.
Current
This mode is usually used because the

radiation pattern has a broadside beam.
c 1
f10

2 r L
x
Ez cos
H x, y y
J s x
x
sin
L
L

W
x
L
The resonant length L is about 0.5 guided
wavelengths in the x direction.
x
sin
L
L
41

Resonance Frequency of Dominant Mode
The resonance frequency is mainly controlled by the
patch length L and the substrate permittivity.
Approximately, (assuming PMC walls)
2
1
m n

L W
(1,0) mode:
k1 L
This is equivalent to saying that

the length L is one-half of a
wavelength in the dielectric.
0 / 2
L d / 2
r
Note: A higher substrate permittivity allows for a smaller antenna

(miniaturization) but with a lower bandwidth.
42

The resonance frequency calculation can be improved by adding a
fringing length extension L to each edge of the patch to get an
effective length Le .
Le L 2L
c 1
f10

2 r Le
L
L
x
Le
Note: Some authors use effective permittivity in this equation.

(This would change the value of Le.)
43

Hammerstad formula:
W

eff
0.3
0.264
h
r

L / h 0.412
eff 0.258 W 0.8
h
r
reff
r 1
2
r 1
h
1
12

Note: Even though the Hammerstad formula

involves an effective permittivity, we still use
the actual substrate permittivity in the
resonance frequency formula.
1/ 2
c
1
f10
L
2 r
44

Note:
L 0.5 h
This is a good rule of thumb to give a quick estimate.
45

Results: Resonance Frequency
Hammerstad
Measured
0.95
0.9
0.85
0.8
0.75
0.01
0.02
0.03
0.04
0.05
0.06
0.07
h
0
h //
0
r = 2.2
W/ L = 1.5
The resonance frequency has been normalized by the

zero-order value (without fringing):
fN = f / f0
46
Outline
Feeding methods
CAD Formulas
Radiation pattern
Input Impedance
Circular patch
Improving bandwidth
Miniaturization
47
General Characteristics
Bandwidth
The bandwidth is directly proportional to substrate thickness h.
However, if h is greater than about 0.05 0 , the probe inductance (for a
coaxial feed) becomes large enough so that matching is difficult the
bandwidth will decrease.
The bandwidth is inversely proportional to r (a foam substrate gives a high
bandwidth).
The bandwidth of a rectangular patch is proportional to the patch width W (but
we need to keep W < 2L ; see the next slide).
48
Width Restriction for a Rectangular Patch
c 1
f 01

2 r W
W 2L
f mn
c
2 r
m n
L W
f 02 f 01
c 1
1
r W 2 L
f10
c 1

2 r L
f 02
c 2

2 r W
fc
f01
f10
f02
W = 1.5 L is typical.
L
49
Some Bandwidth Observations
For a typical substrate thickness (h /0 = 0.02), and a typical substrate
permittivity (r = 2.2) the bandwidth is about 3%.
By using a thick foam substrate, bandwidth of about 10% can be achieved.
By using special feeding techniques (aperture coupling) and stacked
patches, bandwidths of 100% have been achieved.
50
Results: Bandwidth
30
25
r r=10.8
10.8
20
15
10
2.2
2.2
5
0
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
0.1
hh/ /
0
The discrete data points are measured values.
The solid curves are from a CAD formula (given later).
W/ L = 1.5
r = 2.2 or 10.8
51
Resonant Input Resistance
The resonant input resistance is fairly independent of the substrate
thickness h unless h gets small (the variation is then mainly due to
dielectric and conductor loss).
The resonant input resistance is proportional to r.
The resonant input resistance is directly controlled by the location of the
feed point (maximum at edges x = 0 or x = L, zero at center of patch).
(x0, y0)
W
L
L
x
52
Resonant Input Resistance (cont.)
The patch is usually fed along the centerline (y0 = W / 2)
to maintain symmetry and thus minimize excitation of undesirable modes
(which cause cross-pol).
Desired mode: (1,0)
Feed: (x0, y0)
W
x
L
53
For a given mode, it can be shown that the resonant input resistance is
proportional to the square of the cavity-mode field at the feed point.
This is seen from the cavity-model eigenfunction analysis
(please see the Appendix).
Rin E
2
z
x0 , y0
For (1,0) mode:
x0
Rin cos
(x0, y0)
W
54
y
Hence, for (1,0) mode:
x0
Rin Redge cos
(x0, y0)
W
L
The value of Redge depends strongly on the substrate permittivity

(it is proportional to the permittivity).
For a typical patch, it is often in the range of 100-200 Ohms.
55
Results: Resonant Input Resistance
200
The discrete data points are from a CAD
formula (given later.)
150
r r=10.8
10.8
y0 = W/2
x0 = L/4
Region where loss is important
100
2.2
2.2
(x0, y0)
50
W
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
hh //
0
r = 2.2 or 10.8
0.08
W/L = 1.5
56
Radiation Efficiency
Radiation efficiency is the ratio of power radiated into
space, to the total input power.
Pr
er
Ptot
The radiation efficiency is less than 100% due to
Conductor loss
Dielectric loss
Surface-wave excitation
57
Radiation Efficiency (cont.)
y
TM0
surface wave
Js
x
cos () pattern
58
Hence,
Pr
Pr
er
Ptot Pr Pc Pd Psw
Pr = radiated power
Pc = power dissipated by conductors
Ptot = total input power
Pd = power dissipated by dielectric

Psw = power launched into surface wave
59
Some observations:
Conductor and dielectric loss is more important for thinner substrates (the
Q of the cavity is higher, and thus more seriously affected by loss).
Conductor loss increases with frequency (proportional to f 1/2) due to the
skin effect. It can be very serious at millimeter-wave frequencies.
Conductor loss is usually more important than dielectric loss for typical
substrate thicknesses and loss tangents.
Rs
0
Rs
Rs is the surface resistance of the metal.

The skin depth of the metal is .
f
60
Surface-wave power is more important for thicker substrates or for
higher-substrate permittivities. (The surface-wave power can be
minimized by using a thin substrate or a foam substrate.)
For a foam substrate, a high radiation efficiency is obtained by making the
substrate thicker (minimizing the conductor and dielectric losses). There is no
surface-wave power to worry about.
For a typical substrate such as r = 2.2, the radiation efficiency is maximum for
h / 0 0.02.
61
Results: Efficiency (Conductor and dielectric losses are neglected.)
100
2.2
2.2
80
60
r10.8
10.8
40
exact
CAD
20
0
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
0.1
h
h // 0
r = 2.2 or 10.8
W/L = 1.5 Note: CAD plot uses Pozar formula (given later).
62
Results: Efficiency (All losses are accounted for.)
100
2.2
2.2
80
exact
60
CAD
rr = 10.8
10.8
40
tan 0.001
20
0
3.0 107 [S/m]
r = 2.2 or 10.8
0.02
0.04
0.06
hh // 00
0.08
0.1
W/L = 1.5 Note: CAD plot uses Pozar formula (given later).
63
Radiation Pattern
y
E-plane: co-pol is E
H-plane: co-pol is E
Js
E plane
Probe
H plane
Note: For radiation patterns, it is usually more

convenient to place the origin at the middle of the
patch
(this keeps the formulas as simple as possible).
64
Radiation Patterns (cont.)
Comments on radiation patterns:
The E-plane pattern is typically broader than the H-plane pattern.
The truncation of the ground plane will cause edge diffraction, which
tends to degrade the pattern by introducing:
Rippling in the forward direction
Back-radiation
Pattern distortion is more severe in the E-plane, due to the angle dependence
of the vertical polarization E on the ground plane.
(It varies as cos ()).
65
Radiation Patterns
Edge diffraction is the most serious in the E plane.
E varies as cos
y
Space wave
L
Js
E plane
H plane
66
Radiation Patterns
E-plane pattern
Red: infinite substrate and ground plane
Blue: 1 meter ground plane
0
30
Note: The E-plane pattern

tucks in and tends to
zero at the horizon due to
the presence of the infinite
substrate.
-30
-10
60
-60
-20
-30
-40
90
-30
-20
120
-10
-90
240
150
210
180
67
Radiation Patterns
H-plane pattern
Red: infinite substrate and ground plane
Blue: 1 meter ground plane
0
45
-10
-45
-20
-30
-40
90
135
-30
-20
-10
-90
225
180
68
Directivity
The directivity is fairly insensitive to the substrate thickness.
The directivity is higher for lower permittivity, because the patch is
larger.
69
Results: Directivity (relative to isotropic)
10
r r=2.2
2.2
10.8
10.8
6
4
exact
CAD
2
0
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
0.1
hh//0
r = 2.2 or 10.8
W/ L = 1.5
70
Outline
Feeding methods
CAD Formulas
Radiation pattern
Input Impedance
Circular patch
Improving bandwidth
Miniaturization
71
CAD Formulas
CAD formulas for the important properties of the
rectangular microstrip antenna will be shown.
Radiation efficiency
Bandwidth (Q)
Resonant input resistance
Directivity
D. R. Jackson, Microstrip Antennas, Chapter 7 of Antenna Engineering Handbook, J. L.
Volakis, Editor, McGraw Hill, 2007.
D. R. Jackson, S. A. Long, J. T. Williams, and V. B. Davis, Computer-Aided Design of
Rectangular Microstrip Antennas, Ch. 5 of Advances in Microstrip and Printed Antennas,
K. F. Lee and W. Chen, Eds., John Wiley, 1997.
D. R. Jackson and N. G. Alexopoulos, Simple Approximate Formulas for Input
Resistance, Bandwidth, and Efficiency of a Resonant Rectangular Patch, IEEE
Trans. Antennas and Propagation, Vol. 39, pp. 407-410, March 1991.
72
CAD Formulas
Radiation Efficiency
er
1 e
hed
r
where
l d
erhed
Rsave
1 3 r
0 h / 0 16 p c1
h / 0
Comment: The efficiency becomes small as the substrate gets thin.
l d tan loss tangent of substrate
Rs surface resistance of metal
Rsave Rspatch Rsground / 2
Note: hed refers to a unit-amplitude horizontal electric dipole.

73
CAD Formulas
erhed
Psphed
Psphed Pswhed
where
hed
sp
P
hed
sw
1
Pswhed
1 hed
Psp
1
2
2 k0 h 80 2c1
0
1
1
3
2 k0 h 60 1
0
r
Note: hed refers to a unit-amplitude horizontal electric dipole.

Note: When we say unit amplitude here, we assume peak (not RMS) values.
74
CAD Formulas
Hence, we have
hed
r
1
1
1
3
1 k0 h 1
4
c1
r
Physically, this term is the radiation efficiency of a

horizontal electric dipole (hed) on top of the substrate.
75
CAD Formulas
The constants are defined as
1
2/5
c1 1
2
r
r
a2
2
4
2
3
1
p 1 k0 W a22 2a4
k
W
c
k
L
2
0
0
10
560
5
2
2
1
a2 c2 k0 W k0 L
70
c2 0.0914153
a2 0.16605
a4 0.00761
76
CAD Formulas
Improved formula for HED surface-wave power (due to Pozar)
hed
sw
r x 1
k
8 r 1 x1 ( k0 h) x02 1 1 r2 x1
2
0
2
0 0
x02 1
x1
r x02
0 s tan k0 h s
x0 1
3/2
Note: x0 in this formula is

not the feed location!
r2 01 r r2 2 01 02
r2 12
k0 h s
1
1 tan k0 h s
2
s
cos k0 h s
s r 1
D. M. Pozar, Rigorous Closed-Form Expressions for the Surface-Wave Loss of Printed
Antennas, Electronics Letters, vol. 26, pp. 954-956, June 1990.
Note: The above formula for the surface-wave power is different from that given in Pozars paper by
a factor of 2, since Pozar used RMS instead of peak values.
77
CAD Formulas
Bandwidth
1
BW
2
Rsave
1
1
16 p c1 h W

l d

hed
0 h / 0 3 r 0 L er
1
Q
2 BW
Comments: For a lossless patch, the bandwidth is

approximately proportional to the patch width and
to the substrate thickness. It is inversely
proportional to the substrate permittivity.
For very thin substrates the bandwidth will
increase, but as the expense of efficiency.
BW is defined from the frequency limits f1 and f2 at which SWR = 2.0.
f 2 f1
BW
f0
(multiply by 100 if you want to get %)

78
CAD Formulas
Quality Factor Q
Us
Q 0
P
U s energy stored in patch cavity

P power that is radiated and dissipated by patch
1
P
Q 0U s
P Pd Pc Psp Psw
1
1
1
1
1
Q Qd Qc Qsp Qsw
79
CAD Formulas
Q Components
Qd 1 / tan

Qc 0
2
(k0 h)
ave
Rs
3 r L 1
Qsp

16 pc1 W h / 0
erhed
Qsw Qsp
hed
1
The constants p and c1 were defined previously.
erhed
1
1
1
3
1 k0 h 1
4
c1
r
80
CAD Formulas
Resonant Input Resistance
Probe-feed Patch
R R
max
in
Redge
x0
Redge cos
40
Rs
1
l d
h /
16
h

0
p c1
W
1
h
hed
L
0 er
Comments: For a lossless patch, the resonant resistance is approximately

independent of the substrate thickness. For a lossy patch it tends to zero as
the substrate gets very thin. For a lossless patch it is inversely proportional to
the square of the patch width and it is proportional to the substrate permittivity.
81
CAD Formulas
Directivity
3
D
pc1
2
r tan k1h
2
tanc
k1h
k1 k0 r
where
tanc x tan x / x
The constants p and c1 were defined previously.

82
CAD Formulas
Directivity (cont.)
For thin substrates:
3
D
p c1
(The directivity is essentially independent of the substrate thickness.)
83
Outline
Feeding methods
CAD Formulas
Radiation pattern
Input Impedance
Circular patch
Improving bandwidth
Miniaturization
84
Radiation Pattern
There are two models often used for calculating the radiation pattern:
Electric current model
Magnetic current model
Note: The origin is placed at the center of the patch,

at the top of the substrate, for the pattern calculations.
z
Patch
Probe
Coax feed
85
Radiation Pattern
Electric current model:
We keep the physical currents flowing on the patch (and feed).
Patch
Probe
Coax feed
bot
J spatch J top
J
s
s
J spatch
J sprobe
86
Radiation Pattern
Magnetic current model:
We apply the equivalence principle and invoke the (approximate) PMC condition
at the edges.
Equivalence surface
J se
n H
Patch
Probe
M se n E
Coax feed
The equivalent
surface current is
approximately zero
on the top surface
(weak fields) and
the sides (PMC).
We can ignore it on
the ground plane (it
does not radiate).
M se
M se
M se n E
87
Radiation Pattern
Theorem
The electric and magnetic models yield identical patterns
at the resonance frequency of the cavity mode.
Assumption:
The electric and magnetic current models are based on the fields of a single cavity mode, corresponding to an ideal cavity with PMC walls.
D. R. Jackson and J. T. Williams, A Comparison of CAD Models for Radiation from Rectangular
Microstrip Patches, Intl. Journal of Microwave and Millimeter-Wave Computer Aided Design, vol.
1, no. 2, pp. 236-248, April 1991.
88
Radiation Pattern
Comments on the Substrate Effects
The substrate can be neglected to simplify the far-field calculation.
When considering the substrate, it is most convenient to assume an infinite
substrate (in order to obtain a closed-form solution).
Reciprocity can be used to calculate the far-field pattern of electric or
magnetic current sources inside of an infinite layered structure.
When an infinite substrate is assumed, the far-field pattern always goes to
zero at the horizon.
Microstrip Patches, Intl. Journal of Microwave and Millimeter-Wave Computer Aided Design, vol. 1,
no. 2, pp. 236-248, April 1991.
89
Radiation Pattern
Comments on the Two Models
For the rectangular patch, the electric current model is the simplest since
there is only one electric surface current (as opposed to four edges).
For the rectangular patch, the magnetic current model allows us to classify
the radiating and nonradiating edges.
For the circular patch, the magnetic current model is the simplest since there
is only one edge (but the electric surface current is described by Bessel
functions).
x
J s x A10 cos
y
x
Radiating edges
Js
Ez sin
On the nonradiating edges, the

magnetic currents are in opposite
directions across the centerline (x = 0).
M se n E
M se
Nonradiating edges
90
Radiation Pattern
Rectangular Patch Pattern Formula
(The formula is based on the electric current model.)
Infinite ground plane and substrate

H-plane
The origin is at the

center of the patch.
(1,0) mode
J s = x cos
y
x
E-plane
The probe is on the x axis.
L
91
Radiation Pattern
The far-field pattern can be determined by reciprocity.
Ei (r , , ) E
hex
i
WL
r , ,
k y W
sin
2
ky W
k x L
cos

2
2
2

k x L

2
2
i or
k x k0 sin cos
k y k0 sin sin
The hex pattern is for a
horizontal electric dipole in the x direction,
sitting on top of the substrate.
Js
x
L
Microstrip Patches, Intl. Journal of Microwave and Millimeter-Wave Computer Aided Design, vol. 1,
no. 2, pp. 236-248, April 1991.
92
Radiation Pattern
Ehex r , , E0 sin F
Ehex r , , E0 cos G
where
j 0 jk0 r
e
4 r
E0
F 1
TE
2 tan k0 h N
tan k0 h N j N sec
G cos 1 TM
N r sin 2
2 tan k0 h N cos
r
tan k0 h N j
cos
N
Note: To account for lossy substrate, use
r rc r 1 j tan
93
Outline
Feeding methods
CAD Formulas
Radiation pattern
Input Impedance
Circular patch
Improving bandwidth
Miniaturization
94
Input Impedance
Various models have been proposed over the years for calculating the
input impedance of a microstrip patch antenna.
Transmission line model
The first model introduced
Very simple
Cavity model (eigenfunction expansion)

Simple yet accurate for thin substrates
Gives physical insight into operation
CAD circuit model

Extremely simple and almost as accurate as the cavity model
Spectral-domain method
More challenging to implement
Accounts rigorously for both radiation and surface-wave excitation
Commercial software
Very accurate
Can be time consuming
95
Input Impedance
Comparison of the Three Simplest Models
Circuit model of patch
Transmission line model of patch
Cavity model (eigenfunction expansion) of patch
r 2.2
tan 0.001
h 1.524 mm
L 6.255 cm
x0 6.255 cm
W / L 1.5
y0 0
3.0 107 S/m
a 0.635mm
Results for a typical patch show that the first

three methods agree very well, provided the
correct Q is used and the probe inductance is
accounted for.
96
Input Impedance
CAD Circuit Model for Input Impedance
The circuit model discussed assumes a probe feed.
Other circuit models exist for other types of feeds.
Note: The mathematical justification of the CAD circuit model comes
from the cavity-model eigenfunction analysis (see Appendix).
The transmission-line model and the cavity model are

discussed in the Appendix.
97
Input Impedance
Probe-fed Patch
Near the resonance frequency, the patch cavity can be approximately modeled
as a resonant RLC circuit.
The resistance R accounts for radiation and losses.
A probe inductance Lp is added in series, to account for the probe inductance
of a probe feed.
Patch cavity
Probe
Zin
Lp
L
R
98
Input Impedance
Z in j Lp
1 jQ
R
Q
0 L
1
BW
2Q
R
f
f 0

f0 f
BW is defined here by SWR < 2.0 when the

RLC circuit is fed by a matched line (Z0 = R).
0 2 f 0
Zin
Lp
1
LC
L
R
Z in Rin jX in
99
Input Impedance
Rin
f
f 0
1 Q

f0 f
Rinmax Rin
f f0
R is the input resistance at the resonance of the patch cavity

(the frequency that maximizes Rin).
f f0
Lp
max
in
(resonance of RLC circuit)
L
R
f f0
100
Input Impedance
Z in j Lp
1 jQ
R
f
f 0

f0 f
The input resistance is determined once we know four parameters:

f : the resonance frequency of the patch cavity
0
CAD formulas
for the first three R: the input resistance at the cavity resonance frequency f
0
parameters
have been given Q: the quality factor of the patch cavity
earlier.
Lp: the probe inductance
Lp
(R, f0, Q)
L
Zin
101
Input Impedance
Results: Input Resistance vs. Frequency
Note: exact means the cavity model will all infinite modes.
80
Rectangular patch
70
60
CAD
exact
50
40
Frequency where
the input resistance
is maximum (f0):
Rin = R
30
20
10
0
r = 2.2
4.5
5
FREQUENCY (GHz)
W/L = 1.5
5.5
L = 3.0 cm
102
Input Impedance
Results: Input Reactance vs. Frequency
Frequency80
where the input
Note: exact means the cavity model will all infinite modes.
resistance is maximum (f0)
Rectangular patch
60
CAD
exact
40
20
Xp
Shift due to probe reactance
-20
-40
r = 2.2
4.5
5
FREQUENCY (GHz)
W/L = 1.5
5.5
Frequency where the

input impedance is real
L = 3.0 cm
103
Input Impedance
Approximate CAD formula for probe (feed) reactance (in Ohms)
a = probe radius
h = probe height
0
Xp
k0 h ln
2
r k0 a
This is based on an infinite parallel-plate model.
X p Lp
B 0.577216
2a
(Eulers constant)
0 0 / 0 376.7303
104
Input Impedance
Feed (probe) reactance increases proportionally with substrate
thickness h.
Feed reactance increases for smaller probe radius.
0
Xp
k0 h ln
2
r k0 a
Important point:
If the substrate gets too thick, the probe reactance will make it difficult
to get an input match, and the bandwidth will suffer.
(Compensating techniques will be discussed later.)
105
Input Impedance
Results: Probe Reactance (Xf =Xp= Lp)
40
35
CAD
exact
30
Rectangular patch
r = 2.2
25
20
W/L = 1.5
15
(x0, y0)
Note: exact means the cavity model with all infinite modes.
10
5
0
L
0
0.1
0.2
Center
xr = 2 ( x0 / L) - 1
0.3
0.4
0.5
Xxr
r
0.6
0.7
L
0.8
h = 0.0254 0
0.9
a = 0.5 mm
1
Edge
The normalized feed location ratio xr is zero at the center

of the patch (x = L/2), and is 1.0 at the patch edge (x = L).
106
Design Example
Design a probe-fed rectangular patch antenna on a substrate having a relative permittivity of 2.33
and a thickness of 62 mils (0.1575 cm). (Rogers RT Duroid 5870). Choose an aspect ratio of W / L =
1.5. The patch should resonate at the operating frequency of 1.575 GHz (the GPS L1 frequency).
Ignore the probe inductance in your design, but account for fringing at the patch edges when you
determine the dimensions. At the operating frequency the input impedance should be 50 (ignoring
the probe inductance). Assume an SMA connector is used to feed the patch along the centerline (at
y = W / 2), and that the inner conductor of the SMA connector has a radius of 0.635 mm. The copper
patch and ground plane have a conductivity of = 3.0 107 S/m and the dielectric substrate has a
loss tangent of tan = 0.001.
1) Calculate the following:
The final patch dimensions L and W (in cm)
The feed location x0 (distance of the feed from the closest patch edge, in cm)
The bandwidth of the antenna (SWR < 2 definition, expressed in percent)
The radiation efficiency of the antenna (accounting for conductor, dielectric, and surfacewave loss, and expressed in percent)
The probe reactance Xp at the operating frequency (in )
The expected complex input impedance (in ) at the operating frequency, accounting for
the probe inductance
Directivity
Gain
2) Plot the input impedance vs. frequency.
107
Design Example
Results from CAD formulas
Feed at (x0, y0)
1) L = 6.07 cm, W = 9.11 cm

2) x0 = 1.82 cm
3) BW = 1.24%
4) er = 81.9%
y0 = W/2
5) Xp = 11.1
6) Zin = 50.0 + j(11.1)
7) D = 5.85 (7.67 dB)
8) G = (D)(er) = 4.80 (6.81 dB)
108
Design Example
Z in jX p
f0 = 1.575 109 Hz
R = 50
Q = 56.8
Xp = 11.1
1 jQ
Results from CAD formulas
R
f
f
0
f0 f
60
50.255
50
Rin
40
30
Rin( fghz)
Xin( fghz)
20
Xin
10
0
10
13.937
20
1.5
1.5
1.525
1.55
1.575
fghz
f (GHz)
1.6
1.625
1.65
1.65
109
Outline
Feeding methods
CAD Formulas
Radiation pattern
Input Impedance
Circular patch
Improving bandwidth
Miniaturization
110
Circular Polarization
Three main techniques:
1) Single feed with nearly degenerate eigenmodes (compact but small
CP bandwidth).
2) Dual feed with delay line or 90o hybrid phase shifter (broader CP
bandwidth but uses more space).
3) Synchronous subarray technique (produces high-quality CP due to
cancellation effect, but requires even more space).
The techniques will be illustrated with a rectangular patch.
111
Single Feed Method
The feed is on the diagonal.

The patch is nearly
(but not exactly) square.
(0,1)
(1,0)
L W
L
Basic principle: The two dominant modes (1,0) and (0,1) are
excited with equal amplitude, but with a 45o phase.
112
Design equations:
f CP
fx f y
2
The optimum CP frequency is the

average of the x and y resonance
frequencies.
BW
1
2Q
(SWR < 2 )
1
f x f CP 1 m
2Q
1
f y f CP 1
2Q
Top sign for LHCP,

bottom sign for RHCP.
The frequency f0 is also the resonance frequency:
(0,1)
W
(1,0)
x0 y0
Z in Rin Rx Ry
The resonant input resistance of the CP patch is the same as what a

linearly-polarized patch fed at the same position would be.
113
Other Variations
Note: Diagonal modes are used as degenerate modes
L
x
L
Patch with slot
x
L
Patch with truncated corners
114
Here we compare bandwidths (impedance and axial-ratio):
Linearly-polarized (LP) patch:
LP
SWR
BW
2Q
(SWR 2)
Circularly-polarized (CP) single-feed patch:

CP
SWR
BW
(SWR 2)
CP
BWAR
0.348
Q
AR
2 (3dB)
The axial-ratio bandwidth is small when using the single-feed method.

W. L. Langston and D. R. Jackson, Impedance, Axial-Ratio, and Receive-Power Bandwidths of
Microstrip Antennas, IEEE Trans. Antennas and Propagation, vol. 52, pp. 2769-2773, Oct. 2004.
115
Dual-Feed Method
y
Phase shift realized with delay line:
RHCP
L
P
L
P+g/4
116
Phase shift realized with 90o quadrature hybrid (branchline coupler)
RHCP
Z0
Z0 / 2
Z0
Feed
g/4
50 Ohm load
Z0
g/4
This gives us a higher bandwidth than the simple power

divider,
but requires a load resistor.
x
117
Synchronous Rotation
Multiple elements are rotated in space and fed with phase shifts.
-180o
-90o
-270o
0o
Because of symmetry, radiation from higher-order modes (or probes)
tends to be reduced, resulting in good cross-pol.
118
Outline
Feeding methods
CAD Formulas
Radiation pattern
Input Impedance
Circular patch
Improving bandwidth
Miniaturization
119
Circular Patch
y
a
h
120
Circular Patch
Resonance Frequency
From separation of variables:
Ez cos m J m k1
PMC
k1 k0 r
Jm = Bessel function of first kind, order m.
Ez
0
a
J m k1a 0
121
Circular Patch
Resonance Frequency
J m k1a 0
PMC
This gives us
k1a xmn
(nth root of Jm Bessel function)
f mn
c
2 r
xmn
1
0 0
122
Circular Patch
Resonance Frequency
Table of values for xmn

n /m
3.832
1.841
3.054
4.201
5.317
5.416
7.016
5.331
6.706
8.015
9.282
10.520
10.173
8.536
9.969
11.346 12.682
13.987
Dominant mode: TM11
c
f11
x11
2 a r
x11 1.841
123
Circular Patch
Dominant mode: TM11
y
y
a
Circular patch
Square patch
W=L
The circular patch is somewhat similar to a square patch.
124
Circular Patch
Fringing extension
ae a a
c
f11
x11
2 ae r
a + a
Long/Shen Formula:
2h
ae a 1
a r
a
ln
1.7726
2h
PMC
or
h
a
r
a
ln 2h 1.7726
L. C. Shen, S. A. Long, M. Allerding, and M. Walton, "Resonant Frequency of a Circular Disk PrintedCircuit Antenna," IEEE Trans. Antennas and Propagation, vol. 25, pp. 595-596, July 1977.
125
Circular Patch
Patterns
(The patterns are based on the magnetic current model.)
2a
h
H-plane
The origin is at the

center of the patch.
Infinite GP and substrate
y
x
E-plane
a
In patch cavity:
J1 k1
Ez , cos
J
k
a
1
1
The probe is on the x
1

h
(The edge voltage has a maximum of one volt.)
axis.
k1 k0 r
126
Circular Patch
Patterns
E0
E r , , 2 a
tanc k z1h cos J1 k0 a sin Q
0
R
J1 k0 a sin
E0
E r , , 2 a
tanc k z1h sin
P
0
k0 a sin
R
where
P cos 1
TE
tanc ( x ) = tan ( x) / x
2 jN
cos
tan k0 hN jN sec
r
2 j
cos
N
TM
Q 1
r
tan k0 h N j
cos
N
Note: To account for lossy substrate, use r r 1 j tan
j 0 jk0 r
e
4 r
E0
N r sin 2
127
Circular Patch
Input Resistance
J12 k1 0
Rin Redge 2
J1 k1a
k1 k0 r
128
Circular Patch
Input Resistance (cont.)
Redge
where
1
er
2 Psp
er = radiation efficiency
/2
2
Psp
k0 a tanc2 k0 hN
80
0
Q 2 J 2 k a sin P 2 J 2 k a sin sin d

1
0
inc
0
J inc x J1 x / x
Psp = power radiated into space by circular patch with maximum
edge voltage of one volt.
129
Circular Patch
Input Resistance (cont.)
CAD Formula:
Psp
( k0 a ) 2 I c
80
e0 1
e2 0.400000
4
I c pc
3
pc k0 a
k 0
e4 0.0785710
2k
e2 k
e6 7.27509 103
e8 3.81786 104
e10 1.09839 105
e12 1.47731 107
130
Outline
Feeding methods
CAD Formulas
Radiation pattern
Input Impedance
Circular patch
Improving bandwidth
Miniaturization
131
Improving Bandwidth
Some of the techniques that have been successfully

developed are illustrated here.
The literature may be consulted for additional designs and variations.
132
Improving Bandwidth
Probe Compensation
L-shaped probe:
As the substrate
thickness increases the
probe inductance limits
the bandwidth so we
compensate for it.
Top view
Capacitive top hat on probe:
133
Improving Bandwidth
SSFIP: Strip Slot Foam Inverted Patch (a version of the ACP).
Bandwidths greater than 25% have been achieved.
Increased bandwidth is due to the thick foam substrate and
also a dual-tuned resonance (patch+slot).
Note: There is no probe inductance to worry about here.
Patch substrate
Patch
Foam
Microstrip
substrate
Microstrip line
Slot
J.-F. Zrcher and F. E. Gardiol, Broadband Patch Antennas, Artech House, Norwood, MA, 1995.
134
Improving Bandwidth
Stacked Patches
Bandwidth increase is due to thick low-permittivity antenna
substrates and a dual or triple-tuned resonance.
Bandwidths of 25% have been achieved using a probe feed.
Bandwidths of 100% have been achieved using an ACP feed.
Patch substrates
Coupling patch
Top Patch
Microstrip
substrate
Microstrip line
Slot
135
Improving Bandwidth
Stacked Patches
0
-5
-10
-15
Measured
Computed
-20
-25
-30
-35
-40
Stacked patch with ACP feed
10
11
12
Frequency (GHz)
Bandwidth (S11 = -10 dB) is about 100%

136
Improving Bandwidth
Stacked Patches
10
0.5
0.2
180
90
13 GHz
4 GHz
-90
Stacked patch with ACP feed
Two extra loops are observed on the Smith chart.

137
Improving Bandwidth
Parasitic Patches
Radiating Edges Gap Coupled
Microstrip Antennas
(REGCOMA).
Mush of this work

was pioneered by
K. C. Gupta.
Non-Radiating Edges Gap

Coupled Microstrip Antennas
(NEGCOMA)
Four-Edges Gap Coupled

Microstrip Antennas
(FEGCOMA)
Bandwidth improvement factor:

REGCOMA: 3.0, NEGCOMA: 3.0, FEGCOMA: 5.0?
138
Improving Bandwidth
Direct-Coupled Patches
Radiating Edges Direct
(REDCOMA).
Non-Radiating Edges Direct
(NEDCOMA)
Four-Edges Direct Coupled

Microstrip Antennas
(FEDCOMA)
Bandwidth improvement factor:

REDCOMA: 5.0, NEDCOMA: 5.0, FEDCOMA: 7.0
139
Improving Bandwidth
U-Shaped Slot
The introduction of a U-shaped slot can give a

significant bandwidth (10%-40%).
(This is due to a double resonance effect, with two different modes.)
Single Layer Single Patch Wideband Microstrip Antenna, T. Huynh and K. F. Lee,
Electronics Letters, Vol. 31, No. 16, pp. 1310-1312, 1986.
140
Improving Bandwidth
Double U-Slot
A 44% bandwidth was achieved.
Y. X. Guo, K. M. Luk, and Y. L. Chow, Double U-Slot Rectangular Patch Antenna,

Electronics Letters, Vol. 34, No. 19, pp. 1805-1806, 1998.
141
Improving Bandwidth
E Patch
A modification of the U-slot patch.

A bandwidth of 34% was achieved (40% using a capacitive washer to
compensate for the probe inductance).
B. L. Ooi and Q. Shen, A Novel E-shaped Broadband Microstrip Patch Antenna,
Microwave and Optical Technology Letters, vol. 27, No. 5, pp. 348-352, 2000.
142
Multi-Band Antennas
A multi-band antenna is sometimes more desirable than a broadband
antenna, if multiple narrow-band channels are to be covered.
General Principle:
Introduce multiple resonance paths into the antenna.
143
Multi-Band Antennas
Low-band
Low-band
Feed
High-band
Feed
Low-band
High-band
Dual-band E patch
Dual-band patch with parasitic strip
144
Outline
Feeding methods
CAD Formulas
Radiation pattern
Input Impedance
Circular patch
Improving bandwidth
Miniaturization
145
Miniaturization
High Permittivity
Quarter-Wave Patch
PIFA
Capacitive Loading
Slots
Meandering
Note: Miniaturization usually comes at a price of reduced bandwidth!

Usually, bandwidth is proportional to the volume of the patch cavity,
as we will see in the examples.
146
Miniaturization
High Permittivity
r 1
r 4
Size reduction
W W / 2
(Same aspect ratio)
L L / 2
L
The smaller patch has about one-fourth the bandwidth of the original patch.
(Bandwidth is inversely proportional to the permittivity.)
147
Miniaturization
Quarter-Wave patch
Ez = 0
Short-circuit vias
L L / 2
The new patch has about one-half the bandwidth of the original patch.
Neglecting losses:
Us
Q 0
Pr
U s U s / 2
Pr Pr / 4
Q 2Q
Note: 1/2 of the radiating magnetic current

148
Miniaturization
Smaller Quarter-Wave patch
A quarter-wave patch with the same aspect ratio W/L as the original patch
Ez = 0
W W / 2
Short-circuit vias
Width reduction
L L / 2
L L / 2
The new patch has about one-half the bandwidth of the original quarter-wave
patch, and hence one-fourth the bandwidth of the regular patch.
(Bandwidth is proportional to the patch width.)
149
Miniaturization
Quarter-Wave Patch with Fewer Vias
Use fewer vias
L L / 2
L L
Fewer vias actually gives more miniaturization!
(The edge has a larger inductive impedance: explained on the next slide.)
150
Miniaturization
Quarter-Wave Patch with Fewer Vias
Inductance
Short
Open
The Smith chart provides a simple explanation for the length reduction.
151
Miniaturization
Planar Inverted F (PIFA)
Feed
Shorting strip or via
Top view
A single shorting strip or via is used.

This antenna can be viewed as a limiting case of the via-loaded patch,
or as an LC resonator.
152
Miniaturization
PIFA with Capacitive Loading
Feed
Shorting plate
Top view
The capacitive loading allows for the length of the PIFA to be reduced.
1
LC
153
Miniaturization
Circular Patch Loaded with Vias
Feed
2a
b
Patch
Metal vias
The patch has a monopole-like pattern

The patch operates in the (0,0) mode, as an LC resonator
(Hao Xu Ph.D. dissertation, University of Houston, 2006)
154
Miniaturization
Circular Patch Loaded with Vias
Example: Circular Patch Loaded with 2 Vias
315
-20
45
-10
270
-40
-30
-20
90
-10
-10
-30
-5
-15
1.5
2.5
Frequency [GHz]
180
135
-20
0.5
225
E-theta
E-phi
Unloaded: resonance frequency = 5.32 GHz.

(Miniaturization factor = 4.8)
155
Miniaturization
Slotted Patch
Top view
90o
0o
Linear
CP
The slot forces the current to flow through a longer path,

increasing the effective dimensions of the patch.
156
Miniaturization
Meandering
Via
Feed
Feed
Meandered quarter-wave patch
Via
Meandered PIFA
Meandering forces the current to flow through a longer path,

increasing the effective dimensions of the patch.
Meandering also increases the capacitance of the PIFA line.
157
Outline
Feeding methods
CAD Formulas
Radiation pattern
Input Impedance
Circular patch
Improving bandwidth
Miniaturization
158
Reducing Surface and Lateral Waves

Reduced Surface Wave (RSW) Antenna
z
Feed
b
Shorted annular ring
Feed
Ground plane
SIDE VIEW
TOP VIEW
D. R. Jackson, J. T. Williams, A. K. Bhattacharyya, R. Smith, S. J. Buchheit, and S. A. Long,

Microstrip Patch Designs that do Not Excite Surface Waves, IEEE Trans. Antennas Propagat.,
vol. 41, No 8, pp. 1026-1037, August 1993.
159

Reducing surface-wave excitation and lateral radiation reduces
edge diffraction.
Space-wave radiation (desired)

Lateral radiation (undesired)
Diffracted field at edge
Surface waves (undesired)
160

Principle of Operation
TM11 mode:
1
Ez , V0
cos J1 k1
hJ1 k1a
M s
k k1 k0 r
At edge:
V0
Ez cos
h
z
M s n E zE
M s Ez a,
M s
V0
cos
h
161

y
Substrate
M s
M s
V0
cos
h
Surface-Wave Excitation:
EzTM 0 ATM 0 cos H1 2 TM 0 e jk z 0 z
ATM 0 AJ1 TM 0 a
Set
(z > h)
J1 TM 0 a 0
162

y
Substrate
M s
Hence
Patch resonance:
Note:
TM 0 k1
TM 0 a x1n
For TM11 mode:
x11 1.841
TM 0 a 1.841
k1a 1.841
The RSW patch is too big to be resonant.
163

z
Feed
b
Shorted annular ring
Feed
Ground plane
TOP VIEW
SIDE VIEW
TM 0 b 1.841
The radius a is chosen to make the patch resonant:
J 1 k1a
Y1 k1a
k1 x11
J 1
TM
k1 x11
Y1
TM
164

Reducing the Lateral Radiation
y Ground plane
M s
a
Assume no substrate outside of patch

(or very thin substrate):
M s
Space-Wave Field:
V0
cos
h
SP
z
1 jk0
ASP cos e

(z = h)
ASP CJ1 k0 a
Set
J1 k0 a 0
k0a 1.841
165

y
Substrate
For a thin substrate:
TM 0 k0
M s
The same design reduces
both surface-wave fields and
lateral-radiation fields.
Note: The diameter of the RSW antenna is found from
k0 a 1.841
2a
0.586
0
Note: The size is approximately independent of the

permittivity (the patch cannot be miniaturized by
choosing a higher permittivity!).
166

E-plane Radiation Patterns
Measurements were taken on a 1 m diameter circular ground plane at 1.575 GHz.
Measurement
Theory (infinite GP)
0
30
-30
30
60
-60
-20
-40
-60
-20
-30
-30
90
-30
-10
-10
60
-30
-20
-10
-90
240
120
150
210
180
Conventional
Conventional
-40
90
-30
-20
120
-10
-90
240
150
210
180
RSW
RSW
167

Reducing surface-wave excitation and lateral
radiation reduces mutual coupling.
Space-wave radiation
Lateral radiation
Surface waves
168

Reducing surface-wave excitation and lateral radiation reduces mutual coupling.
0
S12 [dB]
RSW - Measured
E-plane
-10
RSW - Theory
-20
Conv - Measured
-30
Conv - Theory
-40
1/r
-50
-60
-70
1/r3
-80
-90
-100
0
10
Separation [Wavelengths]
Mutual Coupling Between Reduced Surface-Wave Microstrip Antennas, M. A. Khayat, J. T. Williams, D.

R. Jackson, and S. A. Long, IEEE Trans. Antennas and Propagation, Vol. 48, pp. 1581-1593, Oct. 2000.
169
References
General references about microstrip antennas:
Microstrip Patch Antennas, K. F. Fong Lee and K. M. Luk, Imperial
College Press, 2011.
Microstrip and Patch Antennas Design, 2nd Ed., R. Bancroft, Scitech
Publishing, 2009.
Microstrip Patch Antennas: A Designers Guide, R. B. Waterhouse,
Kluwer Academic Publishers, 2003.
Microstrip Antenna Design Handbook, R. Garg, P. Bhartia, I. J. Bahl,
and A. Ittipiboon, Editors, Artech House, 2001.
Advances in Microstrip and Printed Antennas, K. F. Lee, Editor, John
Wiley, 1997.
170
References (cont.)
General references about microstrip antennas (cont.):
CAD of Microstrip Antennas for Wireless Applications, R. A. Sainati,
Artech House, 1996.
Microstrip Antennas: The Analysis and Design of Microstrip Antennas
and Arrays, D. M. Pozar and D. H. Schaubert, Editors, Wiley/IEEE
Press, 1995.
Millimeter-Wave Microstrip and Printed Circuit Antennas, P. Bhartia,
Artech House, 1991.
The Handbook of Microstrip Antennas (two volume set), J. R.
James and P. S. Hall, INSPEC, 1989.
Microstrip Antenna Theory and Design, J. R. James, P. S. Hall, and
C. Wood, INSPEC/IEE, 1981.
171
References (cont.)
More information about the CAD formulas presented here
for the rectangular patch may be found in:
Microstrip Antennas, D. R. Jackson, Ch. 7 of Antenna Engineering

Handbook, J. L. Volakis, Editor, McGraw Hill, 2007.
Computer-Aided Design of Rectangular Microstrip Antennas, D. R.
Jackson, S. A. Long, J. T. Williams, and V. B. Davis, Ch. 5 of Advances
in Microstrip and Printed Antennas, K. F. Lee, Editor, John Wiley, 1997.
172
References (cont.)
References devoted to broadband microstrip antennas:
Compact and Broadband Microstrip Antennas, K.-L. Wong,
John Wiley, 2003.
Broadband Microstrip Antennas, G. Kumar and K. P. Ray,
Artech House, 2002.
Broadband Patch Antennas, J.-F. Zrcher and F. E. Gardiol,
Artech House, 1995.
173
The End
174
Appendix
Transmission line model
Cavity model (eigenfunction expansion)
175
Input Impedance
Transmission Line Model for Input Impedance
The model accounts for the probe feed to improve accuracy.

The model assumes a rectangular patch
(it is difficult to extend to other shapes).
We think of the patch as a wide transmission line resonator (length L).
176
Input Impedance
y
Physical patch dimensions (W L)
x0e x0 L
Js
We
PMC
y0e y0 W
( x0e , y0e )
x
Denote
Le
ke k0 rceff
rceff r 1 j leff
leff
1
1
1
1
1

Q Qd Qc Qsp Qsw
rceff
Note:
L is from Hammerstads formula
W is from Wheelers formula
A CAD formula for Q has been given earlier.

177
Input Impedance
Commonly used fringing formulas
ln 4
W h
reff 0.300
L 0.412h eff
0.258
r
r 1
reff
r 1
(Wheeler formula)
0.262
h
W
(Hammerstad formula)
0.813
h
1
h
1 12
178
Input Impedance
x0
x Le
x x0e
Z 0effc , kceff
Lp
X p Lp
Z in Z inTL jX p
Zin
where
0
Xp
k0 h ln
2
Y0effc 1 / Z 0effc
h
1 h
Z 0effc ceff
0
We
rceff We
YinTL 1 / Z inTL jY0effc tan kceff x0e
jY0effc tan kceff Le x0e
r k0 a
(from a parallel-plate model of probe inductance)
B 0.57722
(Euler's constant )
179
Input Impedance
Cavity Model
Here we use the cavity model to solve for the
input impedance of the rectangular patch antenna.
It is a very efficient method for calculating the input impedance.

It gives a lot of physical insight into the operation of the patch.
The method is extendable to other patch shapes.
Y. T. Lo, D. Solomon, and W. F. Richards, Theory and Experiment on Microstrip Antennas,

IEEE Trans. Antennas Propagat., vol. AP-27, no. 3, pp. 137-145, March 1979.
180
Input Impedance
Physical patch dimensions (W L)
( x0e , y0e )
We
x0e x0 L
PMC
y0e y0 W
x
Le
Denote
rceff
ke k0 rceff
leff
eff
rc
r 1 j leff
Note:
1
1
1
1
Qd Qc Qsp Qsw
L is from Hammerstads formula

W is from Wheelers formula
A CAD formula for Q has been given earlier.

181
Input Impedance
Commonly used fringing formulas
ln 4
W h
reff 0.300
L 0.412h eff
0.258
r
r 1
reff
r 1
(Wheeler formula)
0.262
h
W
(Hammerstad formula)
0.813
h
1
h
1 12
182
Input Impedance
Next, we derive the Helmholtz equation for Ez.
H J i j ceff E
E j H
(Amperes law)
(Faradays law)
Substituting Faradays law into Amperes law, we have:
E J i j ceff E
j
E j J i ke2 E
E 2 E j J i ke2 E
2 E ke2 E j J i
183
Input Impedance
Hence
2 Ez ke2 Ez j J zi
Denote
( x, y ) E z ( x, y )
Then
2 ke2 f ( x, y )
(scalar Helmholtz equation)
where
f ( x, y ) j J zi ( x, y )
184
Input Impedance
Introduce eigenfunctions mn (x,y):
mn ( x, y ) mn ( x, y )
2
2
mn
mn
0C
n
For rectangular patch we have, from separation of variables,
m x
n y
mn ( x, y ) cos
cos
L
W
e
e
m 2 n 2
2
mn

Le W
e
185
Input Impedance
Assume an eigenfunction expansion:
( x, y ) Amn mn ( x, y )
m,n
This must satisfy

Hence
2 ke2 f ( x, y )
2
2
A
k
mn mn e Amn mn f ( x, y)
m,n
m ,n
Using the properties of the eigenfunctions, we have

2
2
A
k
mn e mn mn ( x, y ) f ( x, y )
m,n
186
Input Impedance
Next, we multiply by
mn ( x, yand
) integrate.
Note that the eigenfunctions are orthogonal, so that
mn
( x, y ) mn ( x, y ) dS 0,
(m, n) ( m, n)
Denote
2
mn , mn mn
( x, y ) dS
S
We then have
2
Amn ke2 mn
mn , mn f , mn
187
Input Impedance
Hence, we have
f , mn
1
Amn
2
2
mn , mn ke mn
For the patch problem we then have
J zi , mn
1
Amn j
2
2
mn , mn ke mn
The field inside the patch cavity is then given by
Ez ( x, y ) Amn mn ( x, y )
m ,n
188
Input Impedance
To calculate the input impedance, we first calculate the complex
power going into the patch as
1
Pin E z ( x, y ) J zi*dV
2V
1
h E z ( x, y ) J zi* dS
2 S
1
h Amn mn J zi* dS
2 S m ,n
1
h Amn mn , J zi*
2 m ,n
189
Input Impedance
Hence
1
Pin h Amn mn , J zi*
2 m ,n
mn , J zi
1
1
i*
h j
,
J
2
mn
z
2
2 m ,n
mn , mn ke mn
1
h
2 m ,n
, J i 2
mn
z
j
mn , mn
Also,
1
Pin Z in I in
2
so
Z in
1
2
2
k
mn
e
2 Pin
I in
2
190
Input Impedance
Hence we have
, J i 2
1
mn
z
Z in j h
2
I in m,n mn , mn
where
m,n
m0
1
2
2
ke mn
n0
191
Input Impedance
Rectangular patch:
mn
m x
n y
cos
cos
W e
Le
2
mn
m n
L
W
e
e
ke k0 rceff
where
mn , mn
rceff r 1 j leff
Le
2 m x
cos
dx
Le
0
We
2 n y
0 cos W e dy
192
Input Impedance
so
We Le
1 m0 1 n0
2 2
mn , mn
m0
To calculate
mn , J zi
1, m 0
0, m 0
, assume a strip model as shown below.
ap
( x0e , y0e )
Actual probe
Wp
( x0e , y0e )
Strip model
193
Input Impedance
For a Maxwell strip current assumption, we have:
J sz
I in
2
W p
e 2

y y0
2
W p
y y
,y
2
2
e
0
Wp
Wp
W p 4a p
e
0
( x0e , y0e )
Note: The total probe current is Iin.
194
Input Impedance
For a uniform strip current assumption, we have:
I in
J sz
,
Wp
W p
y y
,y
2
2
e
0
Wp
e
0
3
2
W p a p e B 4.482 a p
Wp
( x0e , y0e )
Note: The total probe current is Iin.
D. M. Pozar, Improved Computational Efficiency for the Moment Method Solution of

Printed Dipoles and Patches, Electromagnetics, Vol. 3, pp. 299-309, July-Dec. 1983.
195
Input Impedance
Assume a uniform strip current model
(the two models gives almost identical results):
y0e
mn , J
i
z
Wp
m x0e
n y 1
I in cos
cos
dy
Le
W e W p
e Wp
y
0
m x0e
I in
cos
Wp
L
e
Wp
n e
W cos We y0 y dy '
p
m x0e
I in
cos
Wp
L
e
y y0e y
Use
Wp
Integrates to zero
n y0e
n y0e
W cos W e cos y ' sin We sin y dy
n W p
m x0e n y0e
I in
cos
cos
W p sinc
Wp
W e
2 We
Le
196
Input Impedance
Hence
mn , J
i
z
n W p
m x0e
n y0e
I in cos
cos
sinc
L
W
2
W
e
e
e
n Wp
Note: It is the sinc
2
W
e
term that causes the series for Zin to converge.
We cannot assume a prove of zero radius, or else the series

will not converge the input reactance will be infinite!
197
Input Impedance
Summary of Cavity Model
, J i 2
1
mn
z
Z in j h
2
I in m ,n mn , mn
mn , J
i
z
1
2
2
k
mn
e
n W p
m x0e
n y0e
I in cos
cos
sinc
L
W
2
W
e
e
e
W L
mn , mn e e 1 m 0 1 n 0
2 2
2
2
mn
m n
L
W
e
e
ke k0 rceff
198
Input Impedance
Summary (cont.)
rceff r 1 j leff
leff
1
1
1
1
1

Q Qd Qc Qsp Qsw
A CAD formula for Q has already been given.
199
Input Impedance
Probe Inductance
, J i 2
1
mn
z
Z in j h
2
I in m ,n mn , mn
1
2
2
ke mn
Note that
(1,0) = term that corresponds to dominant patch mode

(which corresponds to the RLC circuit).
Hence
jX p j h
1
I in
mn , J
( m,n )
(1,0)
i
z
mn , mn
1
2
2
k
mn
e
200
Input Impedance
RLC Model
The input impedance is in the form
Z in Z
m,n
m,n
in
where
m,n
m0
n0
where
m ,n
in
2
i
1 mn , J z
j h
2
I in mn , mn
1
2
2
k
mn
e
After some algebra, we can write this as follows:
201
Input Impedance
Rmn
Z inm ,n
f rmn jQ f rmn
f
rmn
where
Q
f rmn
Rmn
1
leff
and
f
f mn
Pmn
mn
kmn leff
Pmn h
mn , J
1
I in
i
z
mn , mn
202
Input Impedance
2
For f rmn
1 , we have
m ,n
in
Rmn
1 jQ f rmn
f
rmn
This is the mathematical form of the

input impedance of an RLC circuit.
203
Input Impedance
Circuit model
Note: The (0,0) mode has a uniform electric field,
and hence no magnetic field. It also has negligible
radiation (compared to the (1,0) mode).
(1, 0)
(0,1)
(0, 0)
Zin
Note: This circuit model is accurate as long as we are near the
resonance of any particular RLC circuit.
204
Input Impedance
RLC Model
Approximately,
Rmn 0 (m, n) (1, 0)
Hence, an approximate circuit model for operation near the

resonance frequency of the (1,0) mode is
(1, 0)
jX p
Zin
205

Introduction to Microstrip Antennas

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Introduction to Microstrip Antennas

Uploaded by

Copyright:

Available Formats

Introduction to Microstrip Antennas

Purpose of Short Course

Provide an introduction to microstrip antennas.

ECE 6345: Microstrip Antennas

c 2.99792458 108 m/s

0 wavelength of free space

k0 wavenumber of free space

0 intrinsic impedance of free space

1 intrinsic impedance of substrate

reff effective relative permtitivity

(accouting for fringing of flux lines at edges)

complex effective relative permtitivity

(used in the cavity model to account for all losses)

Overview of Microstrip Antennas

Microstrip line feed

Overview of Microstrip Antennas

Overview of Microstrip Antennas

G. Deschamps and W. Sichak, Microstrip Microwave Antennas, Proc. of

Overview of Microstrip Antennas

Low profile (can even be conformal, i.e. flexible to conform to a

Overview of Microstrip Antennas

Overview of Microstrip Antennas

Overview of Microstrip Antennas

Overview of Microstrip Antennas

Linear array (1-D corporate feed)

2-D 8X8 corporate-fed array

4 8 corporate-fed / series-fed array

Overview of Microstrip Antennas

The substrate is so thin that it can be bent to conform to the surface.

Overview of Microstrip Antennas

Note: L is the resonant dimension (direction of current flow).

Overview of Microstrip Antennas

Some of the more common methods for

The feeding methods are illustrated for a rectangular patch,

A feed along the centerline

(The resistance varies as the

Feed-line radiation is isolated from patch radiation

Basic Principles of Operation

Y. T. Lo, D. Solomon, and W. F. Richards, Theory and Experiment on Microstrip Antennas,

Basic Principles of Operation

(See next slide.)

Basic Principles of Operation

Basic Principles of Operation

Inside the patch cavity, because of the thin substrate,

Basic Principles of Operation

Magnetic field inside patch cavity:

Basic Principles of Operation

Basic Principles of Operation

(Js is the sum of the top and bottom surface currents.)

On the bottom surface of the patch

Basic Principles of Operation

Basic Principles of Operation

Basic Principles of Operation

From separation of variables:

Basic Principles of Operation

Basic Principles of Operation

(resonance frequency of (m,n) mode)

Basic Principles of Operation

This mode is usually used because the

Basic Principles of Operation

Approximately, (assuming PMC walls)

This is equivalent to saying that