DESIGN OF A LOW-LOSS SERIES-FED MICROSTRIP ARRAY ANTENNA M. R. Mahbub* and C. G.

Christodoulou Electrical and Computer Engineering Department University of Central Florida Orlando. FL 32816 M.C. Bailey
NASA Langley Research Center Hampton. VA 2368 I

Abstract The design and analysis of a series-fed, low-loss, inverted microstrip amay antenna. operating at 1.413 CHz is presented. The array antenna is composed of two sub arrays. Each sub array consists of an equal number of microstrip patches all connected together through a series microstrip line. The first element of each sub array is coaxially fed but 180 degree out of phase. This approach ensures a symmetric radiation pattern. The design approach, is accomplished using the IE3D code that utilizes the method of moments. All experimental and siniulated data are presented and discussed. Introduction. One of the most basic requirements of a radiometer antenna is to have very low radiation losses in order to maintain a desired radiometric resolution. In this particular space application, portability of the antenna is another requirement that must be satisfied, since it must be transported in space. In this paper, it is shown that these requirements and a low sidelobe ratio are satisfied by using a series-fed, foldable and inverted microstrip antenna configuration. This design minimizes the path length between the elements and thus minimizes ohmic losses 11-21, This reduction of losses makes the series configuration very attractive for radiometer applications. The inverted microstrip configuration is chosen because it provides less dispersion and dielectric losses than the conventional microstrip. For the same characteristic impedance, substrate thickness and a comparable air gap, the attenuation due to conductor loss (usually the dominant loss mechanism) IS improved by a factor of typically 2-3 131. The microstrip patches are supported by foam material (Rohacell material) to reduce possible dielectric losses. The total antenna is composed of two sub array antennas, each consisting of an equal number of rectangular microstrip patch elements in series. The sub arrays are 180 degrees out of phase. The ground plane is also inverted to reduce radiation losses. This design yields not only low radiation losses but a radiation pattern with a narrow beainwidth and low sidelobes. A comparison between the measured and simulated data is presented and discussed. Design Procedure. The first step in the design begins with the calculation of the dimensions of a single inverted microstrip patch. Once this is accomplished

~~~0-7803-4478-2/98/$ Q 1998 IEEE I0.M)

I146

additional patches are added to the array, a patch at a time to achieve PI certain desired radiation pattern and input impedance. Here, the inverted microstrip patch is be considered as a microstrip patch covered with dielectric layer and having the Rohacell material as its dielectric substrate.
Calculation o the resonant length f The resonant length of a single patch microstrip antenna can be calculatcd from the empirical formula 161:

L = c x ( l - 2 . V R ) x ( 2 . F , .&)-I

-2,AI...

.................................

(I)

Where, AI = Capacitive cutback factor given by :

AI =0.412H x (E,

+ 0.3)(W / H - 0.2674) x (E,

- 0.258)- (WI H + 0 A - I .........(2)

with c = speed of light, FR = patch resonant frequency, & = effective dielectric constant of the patch without dielectric cover, & = effective dielectric constant of the patch with dielectric cover, L = patch length, W = patch width, H = substrate thickness and VR = variation in resonant frequency. The optimum width of the microstrip antenna can be found as.
W = c x ( 2 . F R ) - x [ ( e , I ) / 2 f 1 . . . .......................................... +
(?I)

Calculation o the effective dielectric constant f The effective dielectric constant of a dielectric covered patch antenna is given by (61:
E ,

=c, /CO..................................................................

(4)

Where:C, =Capacitance I unit length without dielectric layer present, and Cd = Capacitance I unit length with dielectric layer present.

The dielectric-covered microstrip lines capacitance per unit length can be found as,

e,,xlanh( BdlH)+I c,, E,,+tanh(pd/H)

C&(

B)]]

dj3.......................................................................

Where p is the Fourier transform variable, E, is the permittivity of free space and d is the height of the superstrate. The integration in the last equation is performNed twice - once to find Cd, the capacitance per unit length of dielectric covered antemna, and once calculate CO,the capacitance per unit length with no dielectric layer present.
Optimization of the array dimensions

Once the parameters of the single patch have been determined the array is assemhled and optimized for best match in input impedance and side lobe ratio. The patches can be connected with straight, curved, or zigzagged inverted microstrip lines depending on the cross-polarization and sidelobe requirements. This particular array,

1147

designed for uniform excitation. yielded undesirable grating lobes when straight transmission lines were used. To reduce the size of grating lobes a curved microstrip line was used to make the connections between the patches. Thus, the physical distance between the patches was less than &J2 but the actual electrical length was still 442 as shown in Figure I . The results to be presented will show that the cross polarized component is indeed very small (in the order of -30 dB).

Fig. I The two sub-array antennas comprising the total antenna

Results: Using the IE3D simulator the input impedance of each sub array is matched to a SO Ohm impedance at 1.413 GHz. The measured data yielded an input impedance of 5 1.42 R. The simulated return loss was calculated as -60.43 dB and the measured one was -24 dB at 1.413 GHz. The measured bandwidth was 4.75 percent at the operating frequency and the simulated one as only 2.83 percent. The radiation pattern is shown in Figure 2. The difference in the pattern is attributed to the fact that a wide grid was used in simulating the patches. A finer grid in this case becomes very-time consuming to run and more computer memory is required. Results related to radiation losses from this configuration will be presented as well.
Conclusion The design of a series fed microstrip array antenna using the IE3D simulator (51 is The optimization starts with defining the correct presented and discussed. dimensions for one patch and then build the array so that the input impedance is matched to any desired value. The patches vary in dimensions in order to obtain a certain sidelobe ratio and bandwidth. Care must be taken in designing the distance of separation between adjacent patches in order to reduce any grating lobes. In the present design curved microstrip lines are used to feed each adjacent element instead of a straight microstrip. The level of grating lobes diminishes significantly. without getting any sigificant level of cross-polarization. The width of the connecting microstrip line is also carefully chosen to limit the ohmic losses along the antenna and thus the total radiation losses.

1148

- SimulatedData Measured Data

Fig. 2. Comparison of measured and simulated radiation patterns.

References

[ I ] Tom Metzler, Microstrip Series Arrays. Proceedings of The Workshop on Printed Circuit Antenna Technology, October 17-19, 1979. [2] K.-L. Wu, M. Spenuk, J. Litva. D.-G. Fang, Feed network effect!; on the radiation pattern of series-fed microstrip antenna arrays, IEE Proceedings-H, Vol. 138. No. 3, June, 1991. 131 I. M. Schellenberg. CAD Models for Suspended and Inverted Microstrip, IEEE Trans. on Microwave Theory and Techniques, Vol. 43. No. 6. June. 1995. 141 Gildas P. Gauthier, Alan Courtay and Gabriel M. Rebeiz, Microstrip Antennas on Synthesized Low Dielectric-Constant Substrates. IEEE Trans. On Antennas and Propagation, Vol. 45, No. 8. August, 1997. 151 Zeland Software Inc., IE3D simulator, January. 1997. 161 Randy Bancroft, Accurate Design of Dual - Band Patch Anitennas. Microwave & RF, September, 1988.

1149