J.S. Mandeep, N. Misran, H. Abdullah, and Tan Chiy How

Satellite communication (satcom) systems employ one or more fixed earth stations and a constellation of earth-orbiting satellites to provide transmission/reception of voice, video, and data as well as position information.1 A satcom system is typically configured with a directive antenna, which will follow the motion of an object relative to a given satellite. When the position of the object changes, the azimuth angle from the object relative to the satellite will change. What follows are details on the design of a microstrip-fed patch array antenna that can meet the requirements of these systems.

For such satcom applications, the antenna beamwidth should be relatively narrow in the azimuth plane, with wide beamwidth in the elevation plane. A number of different antennas can meet these requirements, but typically in too bulky or complex a configuration for integration into a satcom earth terminal.

An optimal antenna for this application should be small and simple. For cost-effectiveness and ease of integration, an ideal antenna would be based on a planar printed structure that can be realized with microstrip.2 A proposed solution is a corner reflector patch array antenna that is driven by a microstrip line feed. The antenna is comprised of metallic reflectors arranged for a desired beamwidth direction. The antenna features eight axial dipole elements, with each element formed with half of the dipole on the front-side substrate surface and half on the back side. An SMA end-launch connector carries signals to and from the dipole elements.

The corner reflector patch array antenna includes a feed network printed on the same dielectric substrate as the radiating elements, and a corner reflector consisting of two metal plates forming an 180-deg. angle (Fig. 1).3,4 A suitable distance must be maintained between the dipoles. Too little spacing between adjacent elements can lead to adverse effects of mutual coupling. But too large a spacing between adjacent elements will cause grating lobes in the radiation pattern.

As a suitable compromise, a typical spacing in the range of 0.6 to 0.8 of the free-space wavelength was chosen for optimum performance. In this particular design, the distance between dipoles is three-quarters the wavelength distance (0.75λ) at the center frequency to obtain maximum array gain. The distance of the corner edge from the apex of the radiating elements is 0.4λ. The antenna incorporates a balun with continual taper to make the transition between symmetrical (balanced) microstrip and conventional microstrip. The power divider's linear taper contributes to maximum power transfer from the source (feed) to the load (the radiating elements).

The layout for the array antenna was created with the help of simulations performed on the Momentum electromagnetic (EM) simulation software tool within the Advanced Design System (ADS) electronic-design-automation (EDA) suite of software design tools from Agilent Technologies.5 The steps used to design the antenna were as follows:
1. Create the layout.
2. Choose the Momentum operating mode.
3. Define the proposed array antenna substrate material.
4. Solve for the substrate parameters.
5. Define the antenna ports.
6. Generate the circuit mesh, and simulate array antenna performance.
7. View the simulated results.

Figure 2 shows the fabricated patch array antenna. It was constructed on RO4003C laminate substrate material from Rogers Corporation. The material has a thickness o 0.813 mm and a nominal dielectric constant of 3.38 in the zaxis at 10 GHz. An SMA end-launch connector was soldered at the end of the microstrip line of the proposed antenna for test purposes.

A simple setup (Fig. 3) based on a calibrated microwave vector network analyzer (VNA) was used to measure the S-parameters (S11) of the patch array antenna. During the measurements, the patch array antenna's 50- connector was connected to the VNA's port 1 by means of a low-loss 50- coaxial test cable. The VNA was used to evaluate the antenna's performance from 10 to 14 GHz. Measured S11 log magnitude results from the VNA were compared with the simulated results from Momentum. From the VNA, a resonant frequency of 13.3 GHz was selected for the radiation pattern measurements.

For transmitter measurements, a microwave signal generator is connected to the patch array antenna through the end-launch SMA connector, using a low-loss 50- coaxial connector. The signal generator is set to the antenna's resonant frequency of 13.3 GHz, and its output power level is set to +10 dBm. For the receiver setup, a microwave spectrum analyzer with frequency range of 10 to 14 GHz is connected to a second patch array antenna by means of a low-loss 50-Ω coaxial cable. The distance between the receiver and transmitter is set to 1 m or more so that the measurements are made of the farfield radiation pattern. During testing, the receiver antenna is rotated in various angles while the transmitter antenna is at a fixed angle. The receiver antenna is rotated from 0 to 360 deg. in 10-deg. steps, with the spectrum analyzer used to measure power levels in dBm. An alternate approach to the radiation pattern measurement setup is to replace the transmit antenna with a horn and parabolic antenna.

The gain measurement setup is similar to the setup used to measure the radiation pattern of the patch array antenna. From the Friss transmission formula, the transmit and receive antennas, which are identical, will have same gain based on calculations made by means of Eq. 1:

G0t = G0r = 0.510}(4πr/λ) + 10log10(Prec/Pt)> (1)

where
G0t = transmit antenna gain (in dB);
G0r = receive antenna gain (in dB);
Pt = the transmitted power (in W);
Prec = the received power (in W);
λ = operating wavelength (in m); and
r = the separation distance between antennas (in m). Figure 4, Figure 5, and Figure 6 and the table show the simulated and measured results.

Figure 4 shows return loss. Differences in the two plots could be due to mismatches between the transmission line and the coaxial connector since this was not taken into account during the simulations. The input port is at the edge of the substrate during simulation; in the actual measurement, the SMA connector is disposed inside the antenna substrate, making the actual transmission line shorter than the transmission-line model used in the simulation. Figure 5 and Figure 6 show antenna beamwidth results. The beamwidth in the elevation plane (Fig. 6) is relatively wide compared with the beamwidth in the azimuth plane (Fig. 5). This wide antenna beamwidth in the elevation plane allows the antenna to receive signal sources at wide elevation angles without rotating the antenna in the elevation plane. The differences between the simulated and measure radiation patterns are due to some interference and obstructions between the proposed antenna and the transmitter during the radiationpattern measurements, since the measurements were not made in a room with adequate free space.

REFERENCES
1. Klein S. Gilhousen, Franklin P. Antonio, Irwin M. Jacobs, and Lindsay A. Weaver, Jr., "Alternating Sequential Half Duplex Communication System," United States Patent No. 4979170, issued December 18, 1990.
2. David M. Pozar and G. I Costache, Microstrip Antennas, IEEE Press, New York, 1995.
3. A. Nesic, I. Radnovic, and Z. Micic, "Printed Antenna Arrays with High Side Lobe Suppression," Hindawi Publishing Corporation, Active and Passive Electronic Components, Vol. 2008, Article ID 542929.
4. D. G. Laramie, "Printed Circuit Antenna Array Using Corner Reflector," United States Patent No. 5708446, issued January 13, 1998.
5. Momentum planar EM simulator, ADS 2005, Agilent Technologies.