Microwave Horn Antenna Design and Test

System

EE198B: Senior Design Project II

San Jose State University
Fall 2003

Presented by:
Vishal Ohri
Ozair Amin
Hiruy Gebremariam
Benjamin Dubois
 Table of Contents
Abstract ………………………………………………………………… 2

Introduction ………………………………………………………………… 3

Background and Theory ………………………………………………………… 4

Basics of Antenna Patterns Measurements ………………………………………… 5

Results and Analysis ………………………………………………………… 8

I Waveguide design and tuning ………………………………………… 8

II Horn Antenna design ………………………………………… 10
III Test Setup and Measurements ………………………………………… 11

Discussion …………………………………………………………………. 14

Concluding Remarks …………………………………………………………………. 14

Reference ………………………………………………………………….. 15

Appendix ………………………………………………………………….. 16

1
Abstract
This paper is our report for our senior design project on Microwave Antenna Design and
Test System. This project requires two primary areas of concern; a pyramidal horn
antenna design and a test system that will determine the performance of our antenna. A
brief theory on microwave horn antennas will be discussed along with the results of our
design. Our results and analysis show that the project was within the scope of our ability
to design and test the horn antennas.

2
Introduction
In today’s technological society, wireless communication has become an increasingly
important part of daily life. We have come to depend on our pagers, cellular phones,
satellite dishes, radios, etc., usually without understanding how they work. The common
element to all of these wireless systems, whether they transmit or receive, is the antenna.
The antenna is responsible for coupling the RF energy from the transmission-line feed
(guided) to free space (unguided), and vice versa. Antennas are characterized using
several parameters, such as geometry, gain, beamwidth, side-lobe level, frequency of
operation, efficiency, and polarization. Keeping this in mind for this senior design
project we designed two microwave horn antennas and implement a test system that will
test the performance of our antenna and the efficiency of our test system. This paper will
address the theoretical and practical construction of a 2.4GHz horn antenna and
methodology used in testing the antennas. The pyramidal horn antenna is part of the
aperture antennas family that has a conical radiation pattern, linearly polarized and is
ideal in high gain transmission and receiving, peer to peer communications, and as a dish
feed.

3
Background and theory
Currently there are many companies developing microwave antennas and highly
sophisticated test systems that range in the millions of dollars. Our aim is to build an
affordable horn antenna, less than $20, and an inexpensive antenna test system setup.
Horn antennas are extremely popular in the microwave region (above 1 GHz). Horns
provide high gain, low VSWR (with waveguide feeds), relatively wide bandwidth, and
they are not difficult to make. There are three basic types of rectangular horns:

Figure 1: Basic types of horn antennas

We are concerned with the pyramidal horn antenna shown in Figure 1(c). The horns can
be flared exponentially, too. This provides better matching in a broad frequency band, but
is technologically more difficult and expensive. The rectangular horns are ideally suited
for rectangular waveguide feeders. The horn acts as a gradual transition from a
waveguide mode to a free-space mode of the EM wave. The open-ended waveguide will
radiate, but not as effectively as the waveguide terminated by the horn antenna. The wave
impedance inside the waveguide does not match that of the surrounding medium creating
a mismatch at the open end of the waveguide. Thus, a portion of the outgoing wave is
reflected back into the waveguide. The horn antenna acts as a matching network, with a
gradual transition in the wave impedance from that of the waveguide to that of the
surrounding medium. With a matched termination, the reflected wave is minimized and
the radiated field is maximized. Designing the horn antenna is easy once we determine
the dimensions of our horn antenna. There are many software programs available for
download that can calculate the E-field and H-field dimensions of our horn for a given
frequency (2.4GHz) and gain of about 19dB.

4
The horn antenna we design fit the following specification:
• 2.4 GHz (S-band)
• Beam width ~17o
• ~ 19 dB gain
• Linearly polarized
• Return loss of > -10dB (SWR < 1.2:1)
• WR430 Waveguide standard (4.30”X2.15”)
• Waveguide to Coax adapter N-type connectors

Basics of Antenna Pattern Measurements

A general system designed for antenna measurements uses the following algorithm for
performing a far-field antenna pattern measurement. An antenna under test (AUT) goes
through all the desired angular configurations, while the AUT’s response to RF stimulus
(illuminated by a still source antenna) is being recorded. The plot of magnitude of the
received signal versus angle displays the pattern directivity. As the process of
measurement (rotating the AUT and recording the pattern) is usually done by the system,
an operator has to set up (install/mount) the AUT and source antenna correctly. There are
two major requirements to be satisfied: Realizing which plane (E or H) is to be used. This
defines the placement of the AUT on the rotating table; Matching the polarization of
AUT with the polarization of source antenna (if not measuring cross-polarization).The
radiation pattern of an antenna describes its far field directional characteristics. When the
antenna is transmitting the pattern indicates the relative power density radiated in
different directions in the plane relative to the antenna principal direction of radiation
(“bore-sight”). When receiving the pattern indicates the variation in the received signal
level relative to bore-sight signal level as the antenna orientation is changed. Figure 2
shows a typical antenna radiation pattern.

P(θ)/P(0) dB

Antenna

θ
Figure 2: typical antenna radiation pattern

The radiation pattern can be measured in different spatial planes, principally the E and H
planes as shown in figure 3 on the next page.

5
 E
H

Figure 3: The E-field and H-field orientation of a horn antenna

Example (measuring pattern of a horn in the E-plane)

As the definition says, the E-plane is determined by the direction of the electrical field. In
Figure 4, a horn antenna is fed by a rectangular waveguide, where the electrical field has
only components parallel to the narrower sides. Thus, the field distribution in the horn
antenna should be similar – parallel to the Y-axis. Under normal circumstances, the horn
antenna has its maximum radiation in the Z-axis. These two directions (the direction of
the electrical field and the direction of the maximum radiation) define the E-plane by the
Y and Z axes (or the plane x=0). Thus, if the antenna must, can rotate an AUT in
horizontal plane, then for measuring in the E-plane, the above-mentioned horn antenna
must be mounted with its E-plane parallel to the plane of rotation (horizontal plane). This
antenna mast (antenna tower) is placed inside an anechoic chamber. Through a window
in one of the walls, a source antenna illuminates the AUT. The E-plane is parallel to the
floor (horizontal plane), the source antenna must be polarized horizontally. For example:
if a similar horn antenna is used for the source antenna, it must be placed the same way as
the AUT (here – with the wider sides vertically).

Figure 4: The E-field orientation

6
Antenna Chart
Name Shape Gain (over Beamwidth -3 Radiation Pattern
isotropic) dB

Isotropic 0 dB 360

Horn
15 dB 15

Source: http://www.tmeg.com/tutorials/antennas/antennas.htm

Table 1: Gain, Beamwidth, and Radiation Pattern of horn relative to isotropic antenna

Propagation Link
Transmitter Receiver

d
Figure 5: The antenna test system block diagram.

Ideally the antennas in the test system are put in a chamber (Faraday Cage) to eliminate
any unwanted interference introduced by the inherent existence of EM waves in space.
Since a perfect conductor is an idealization unavailable in nature, perfect Faraday cages
do not exist. However, extremely good Faraday cages are constructed for electromagnetic
experiments; they are commonly called "screen rooms." Early (and still common) screen
rooms were made of copper mesh screen with a somewhat tighter mesh than typical
aluminum window screen. Screen rooms are also made of welded sheet aluminum or
sheet steel. For all screen rooms, special attention is given to electromagnetic sealing of
closed doors and to metallic penetrations for power or communication. Screen rooms are
typically designed to shield the enclosed volume from low-energy, high-frequency
transmitted electromagnetic waves (e. g., AM/FM radio waves). Figure 6 shows our
expected E-Field radiation pattern simulated using MatLab.

7
 Figure 6: MatLab simulation of the E-field of a typical horn using the equations and
code on Appendix.

Results and Analysis

I. Waveguide Design and tuning:

(a) (b)

Picture 1: (a) top view of the WR430 waveguides attached together, (b) tuning screw
under the waveguide.

We chose the standard waveguide (WR430), which operates in the frequency of our
interest (2.4GHz). Once the two WR430 waveguides were built, we attached them facing
each other with the slit open to allow us the freedom to move the N-connectors to the best
location that have optimum S11 and S21 before attaching them to the horns. Picture 1
shows the setup for tuning the waveguides. After tuning the waveguides, we found out

8
that they had (1.70 -2.60) GHz, range and optimum location for the N-connectors was
1.23 inches from the back of the waveguide.

Picture 2: N-connector with a 1.18 inches of copper wire attached to it

Picture 2 shows the N-connector we used in our project. The length of the copper in the N
connector was adjusted by trimming it little by little until the bandwidth of S11 was wide
enough with an ideal gain of -20dB, and the transmission coefficient S21 was smooth
between the 2.4GHz and 4.5GHz. Tuning the waveguide was the most important step in
our testing procedure. Once we matched the two waveguides the horns was matched
equivalently. A great deal of our time was spent trying to match the waveguides because
we had to drill and open up more slit toward the back. Once we located the optimum
point for the N-connector, we screwed five number 4-6 screws for tuning the waveguide
even more. Picture 3 shows the S11 and S21 graphs displayed in the network analyzer
while tuning the waveguide.

Picture 3: S11 and S21 measurements while tuning the waveguide

S11 measurement on picture 3 was improved to be -10.2dB at 2.4GHz. Ideally S11 is

supposed to be -20dB. We spent days trying to tune the waveguides and obtain an S11 of –
20dB, but it was impossible to reach that goal. The major reason was that the waveguides
9
were built in a machine shop where we did not have access to the equipment used to
build the waveguides; hence the waveguides had manufacturing imperfections that made
the tuning difficult. While doing the project, we found out that an ideal waveguide with
an S11 of –20dB costs a little over $1000. That is why for the price we paid for our
waveguides, we were satisfied with what we got.

II. Horn antenna design:

We used a computer program that generated the E-plane [A] and H-plane [B] dimensions
given a cutoff frequency (2.4GHz), waveguide standard (WR430), and desired gain
(19dB). Once we had the value for A and B, then we used the following equations to
designing the horn antenna.

lH lE
R R
x y
R1 R2
a 2α H z 2α E z
A b B

RH RE

(a) (b)
Figure 7: (a) The H-plane and (b) E-plane of the horn antennas.

Horn antenna design was relatively simple. Once you meet the physical constraints of the
E and H plane of the aperture then you can use plane geometry and symmetry to get the
rest of the dimensions.

10
 R1 / RH = A /( A − a)
Physical constraints:
R2 / RE = B /( B − b)

A− a
1
B = (b + b 2 + 8λRE ) RE = A
3λ
For optimum horn, 2

4π 4π 1 8 A( A − a )
G= ε ap AB = 2 ε ap A (b + b 2 + )
λ2 λ 2 3

3bGλ2 3G 2 λ4
A4 − aA3 + A=
8πε ap 32π 2ε ap
2

Once each parameter was solved, we fabricated the antennas and the waveguides in the
San Jose State University’s machine shops in the engineering building. The parameters
are shown in table 2.

A [in.] B [in.] RE [in.] R2 [in.] R1[in.] lH [in.] lE [in.]

20.50 15.18 22.47 26.19 28.45 30.24 27.27

Table 2: Calculated horn antenna dimensions of each parameter.

III. Test setup and measurements:

The horn antennas were 8 feet apart as shown in picture 4 below. We followed the test
methodology we mentioned earlier, and attained the data shown in table 3. Then,
radiation pattern shown on figure 8 were plotted.

Picture 4: the antenna system setup

The far back antenna is the receiver

antenna where we measured S21 from
and the front antenna is the rotating
antenna. The corresponding S21
magnetite was changing as we rotated
the front antenna. The network
analyzer display of S12 is shown in
picture 5.
11
 Left and Right Rotation Vertical Rotation
Rotations S21 [dB] S21 [dB] S21[dB]
[Degrees] E plane H plane E plane
0 0 -26.6 -35.4
10 -3 -22.1 -37.6
20 -8.85 -19.15 -39.9
30 -17.5 -11.6 -45
40 -23.6 -7.7 -48
50 -35 -0.5 -46.8
60 -37.7 -9.8 -45
70 -39.7 -4.3 -53
80 -51.5 7.8 -63.5
90 -46.5 -1 -75
100 -54 5
110 -44.6 -4.4
120 -44.5 -2.2
130 -42 -7
140 -43.5 -2
150 -35.5 -4.3
160 -31.7 -11.3
170 -29.3 -15.2
180 -28 -13.5
-170 -30.1 -16.3
-160 -32.5 -12.5
-150 -36.3 -4.5
-140 -44.2 -2.3
-130 -43.1 -8.1
-120 -44.8 -2.9
-110 -45 -4.7
-100 -54 6
-90 -45 -2
-80 -52 8
-70 -41.2 -4.9
-60 -38.1 -10.1
-50 -35.3 -0.6
-40 -23.9 -8.1
-30 -18.5 -11.7
-20 -9.1 -20.1
-10 -3 -23.2

Table 3: Measured data obtained using a network analyzer.

12
Picture 5: S12 of magnetite as one antenna is rotated while the other is held in place.

E-field and H-field Radiation Pattrens

0
-10 10 10
-20 20
-30 30
-40 40
-10
-50 50

-60 -30 60

-70 70
-50
-80 80

-90 -70 90
E-field
H-field
-100 100

-110 110

-120 120

-130 130

-140 140
-150 150
-160 160
-170 170
180

Figure 8: Actual measured E and H field radiation patterns obtained from S21 at the
receiver end antenna.

13
Discussion
When comparing the simulated radiation patter with the actual measured one, one can see
significant differences. There are two major reasons for that. The first one is the way the
antennas were fabricated. Although we provided the values for each parameter listed in
table 2 using the equations mentioned above, the values for lH and lE were changed
slightly to make the horns fit.

The other problem that contributed to the errors shown in figure 8, where we get radiation
when the rotating antenna is more than 1100 away from the receiving antenna, is because
the test environment was not ideal. The maximum distance (between the two antennas)
we had to work with was about 8 feet. The room was full of reflective objects such as
metal cabinets, tables, chairs, and the wall.

Concluding remarks
The construction of the horn antennas was simple in terms of paper and pencil. However,
fabrication was far more difficult than anticipated before we started the project, but we
managed to construct the horn antennas and reach our goal. The resulted measured
radiation pattern of the E and H field supports our expected radiation pattern and the
calculation of the dimensions of the antennas. Our data could have been improved if we
had an adequate antenna testing facility accommodated by a chamber. We tested our
antenna in a room full of many objects that were reflecting the radiated signal
significantly.

14
References
1. http://www.tmeg.com/tutorials/antennas/antennas.htm
2. https://ewhdbks.mugu.navy.mil/ANTENNAS.HTM
3. http://www.tele.ntnu.no/radio/fag/SIE2080/AntennaLab/AntPatternMeas.pdf
4. http://www.eecs.umich.edu/emag/labmanual/EECS330_LE9.pdf
5. Tatsuo Itoh, George Haddad, James Harvey, “RF Technologies for Low Power
Wireless Communications” IEEE, pp.305-345, John Wiley & Sons, Inc, (2001).
6. Warren L. Stutzman, “Antenna Theory and Design”, John Wiley & Sons, (1981)
7. Carr, Joseph J, “Practical Antenna Handbook”, TAB BOOKS, (1989).

15
Appendix
Equations used in simulating the E-Field radiation pattern.

1 + cos θ I (θ , φ = 0 0 )
FH (θ ) = (1)
2 I (θ = 0 0 , φ = 0 0 )
where
 π  A  12
j      sin θ + 
I (θ , φ = 0 0 ) = e  8t  λ  2
[C ( s 2 ) − jS ( s 2 ) − C ( s1 ) + jS ( s1 ) ] +
 π   A  12
j      sin θ − 
e  8t   λ  2
[C (t 2 ) − jS (t 2 ) − C (t1 ) + jS (t1 ) ] ( 2)
where
 1 A  1
s1 = 2 t  − 1 −  sin θ  − 
 4t  λ  8t 
 1 A  1
s 2 = 2 t 1 −  sin θ  −  (3)
 4t  λ  8t 
 1A  1
t1 = 2 t  − 1 −  sin θ  + 
 4t  λ  8t 
 1A  1
t 2 = 2 t 1 −  sin θ  + 
 4t  λ  8t 

1 + cos θ  [C ( r4 ) − C ( r3 ) ]2 + [S ( r4 ) − S ( r3 )]2  2
FE (θ ) =
2

 [ ( )
4C2 2 s + S2 2 s ( )] 

( 4)

where
2 2
B2 1 B 1 A2 1 A 1
s= =   , t= =   (5)
8 λR 2 8  λ  R 2 8λR1 8  λ  R1
λ λ
 1 B   1 B 
r3 = 2 s  − 1 −  sin θ   , r4 = 2 s 1 −  sin θ   (6)
 4s  λ   4s  λ 

16
MatLab Code for simulation
lambda = 0.125;
A = 0.5205;
B = 0.3855;
R2 = 0.6652;
a = 0.10922;
b = 0.0546;
b = 0.05461;
R1 = 0.7228;
Rh = 0.571;
t : = R1 * Rh;
t = R1 * Rh;
t = A^2;
t = (A^2)/(8*lambda*R1);
s = (B^2)/(8*lambda*R2);
theta = 0:pi/100:2*pi;
s1 = 2*t^(1/2)*(-1-1/(4*t)*((A/lambda)*sin(theta)-1/(8*t)));
s2 = 2*t^(1/2)*(1-1/(4*t)*((A/lambda)*sin(theta)-1/(8*t)));
t1 = 2*t^(1/2)*(-1-1/(4*t)*((A/lambda)*sin(theta)+1/(8*t)));
t2 = 2*t^(1/2)*(1-1/(4*t)*((A/lambda)*sin(theta)+1/(8*t)));
Cs2 = quad ('cos(pi/2.*(x.^2))', s2, 0);
Cs1 = quad ('cos(pi/2.*(x.^2))', s1, 0);
Ct2 = quad ('cos(pi/2.*(x.^2))', t2, 0);
Ct1 = quad ('cos(pi/2.*(x.^2))', t1, 0);
Ss2 = quad ('sin(pi/2.*(x.^2))', s2, 0);
Ss1 = quad ('sin(pi/2.*(x.^2))', s1, 0);
St1 = quad ('sin(pi/2.*(x.^2))', t1, 0);
St2 = quad ('sin(pi/2.*(x.^2))', t2, 0);
r3 = 2*s^(1/2)*(-1-1/(4*s)*((B/lambda)*sin(theta)));
r4 = 2*s^(1/2)*(1-1/(4*s)*((B/lambda)*sin(theta)));
Cr4 = quad ('cos(pi/2.*(x.^2))', r4, 0);
Cr3 = quad ('cos(pi/2.*(x.^2))', r3, 0);
C2sqrs = quad ('cos(pi/2.*(x.^2))', 2*s^(1/2), 0);
Sr4 = quad ('sin(pi/2.*(x.^2))', r4, 0);
Sr3 = quad ('sin(pi/2.*(x.^2))', r3, 0);
S2sqrs = quad ('sin(pi/2.*(x.^2))', 2*s^(1/2), 0);
Iphi0 =
exp(j*(pi/8*t)*((A/lambda)*sin(theta)+1/2).^2).*(Cs2-j*Ss2-Cs1+j*Ss1)+
exp(j*(pi/8*t)*((A/lambda)*sin(theta)-1/2).^2).*(Ct2-j*St2-Ct1+j*St1);

Iphi0zero =
exp(j*(pi/8*t)*((A/lambda)*sin(0)+1/2).^2).*(Cs2-j*Ss2-Cs1+j*Ss1)+
exp(j*(pi/8*t)*((A/lambda)*sin(0)-1/2).^2).*(Ct2-j*St2-Ct1+j*St1);

Fhtheta = (1+cos(theta))/2 * (Iphi0/Iphi0zero);

Fetheta = (1+cos(theta))/2 .* (((Cr4-Cr3).^2 +

(Sr4-Sr3).^2)./(4*((C2sqrs.^2)+(S2sqrs.^2)))).^(1/2)
polar(theta, Fhtheta, '--r');
hold on;
polar(theta, Fetheta, '--o');

17