rIL / CO_ POST-TUNED SINGLE-FEED CIRCULARLY POLARIZED PATCH ANTENNAS Guey-Liou Lan* and Dipak L. Sengupta Radiation Laboratory Department of Electrical Engineering and Computer Science The University of Michigan Ann Arbor, Michigan 48109 SUMMARY It is generally known [1] that circular polarization can be obtained from an ordinary patch antenna by using two orthogonally phased feeds located geometrically 90 degrees apart. In situations where space is at a premium both on the antenna circuit board and at the rear of the board,'the incorporation of an extra feeding network may be disadvantageous. A number of approaches have therefore been proposed to achieve circular polarization using a single feed [2-4]. Although they are different from each other, they all operate on the principle of slightly detuning the degenerate modes of the symmetrical structure, i.e., they rely on certain asymmetry in the radiator. The present paper describes the design conditions for obtaining circular polarization from a single-feed circular patch antenna tuned by passive metallic or shorting post(s) appropriately placed in the input region of the antenna. A similarly tuned rectangular patch antenna has been experimentally developed in [5]. Consider a post located along the p = 45 degree line close to the center of a circular patch antenna fed from the back by a coaxial probe as shown in Fig. 1. It can be shown [6-7] that in the presence of the post, the two degenerate orthogonal modes, named a and b, respectively, will produce approximately equal input impedance while their resonant frequencies split with two distinct angular frequencies, say wa and wb' respectively. The resonant frequency Wa of mode a remains the same as that of the untuned antenna (i.e., o ) and that of mode b (i.e., wb) shifts upward from w0 by an amount depending on the radius and location of the tuning post. Near resonance the impedance variation due to each mode can be obtained from an equivalent parallel R-L-C circuit and is given by R a,b in (1) in 6 1 + jQ Wa,b Now with RCA Laboratories, Princeton, NJ 08540. RL-782 = RL-782

-2 - where R. is the resonant input impedance due to each mode, Q is the quality factor, and 6w is the frequency deviation from resonance. Figure 2 shows the input impedance and corresponding phase variation vs normalized frequency for these two modes. If the antenna is operated at a frequency intermediate between wa and wb, it can be seen from Fig. 2 that the impedance due to the a and b modes will be capacitive and inductive, respectively. When the operating frequency p is chosen properly, the total impedance of the antenna will be real at p, i.e., the resonance condition will be obtained. If the degree of mode splitting is chosen such that the phase angles of the modal impedances are +45 degrees apart at the operating frequency, then the conditions for circular polarization-are achieved. Under these conditions as shown in Fig. 2, the frequency shift satisfies the following condition Wb - Wa 1 Q * (2) Thus, from a knowledge wo (= wa) and Q of the untuned antenna, wb is obtained from (2). The dimension and location of the post can be obtained [6,7] from a knowledge of wb. Left and right circular polarizations are obtained for a post located at $ = 45 degree and $ = 135 degree radials, respectively. Circular polarization is obtained only over a very narrow bandwidth centered at w. For an axial ratio within 3.0 dB which would produce a polarization mismatch loss less than 0.25 dB, the bandwidth will be limited to a fraction of the impedance bandwidth. The useful circular polarization bandwidth can also be obtained from Fig. 2. If the frequency deviates from wo by 6w, the axial ratio is approximately given by 2 2QAw 1 + j2Q - o0 1Aw f 2 (3) 1 + j2Q Aw + 6w 1 + 1 + 0 where Aw = wo - w - b For a 3.0 dB axial ratio, eq. (3) p a b p yields 6w/Aw = 0.35 which indicates that the useful circular polarization bandwidth is only 35 percent of the bandwidth of the original linearly polarized antenna. A plot of axial ratio vs normalized frequency for a typical antenna is shown in Fig. 3 where some available experimental [3] results are given for comparison.

References [1] I. J. Bahl and P. Bartia, Microstrip Antennas, Artech House, 1980. [2] J. L. Kerr, "Microstrip antenna developments", Proc. Workshop on Printed Circuit Antenna Technology, New Mexico State University, pp. 3-1 to 3-20, 1979. [3] S. A. Long and L. C. Shen, "The circularly polarized elliptical printed-circuit antenna," IEEE AP-S Intl. Symp. Digest, pp. 731-734, 1980. [4] W. F. Richards, Y. T. Lo and P. Simon, "A perturbational app'roach to the design of circularly polarized microstrip antennas", IEEE AP-S Intl. Symp. Digest, pp. 339-342, 1981. [5] D. H. Schaubert, F. G. Farrar, A. Sindoris and S. T. Hays, "Microstrip antenna with frequency agility and polarization diversity," IEEE Trans., Vol. AP-29, No. 1, pp. 118-123, January 1981. [6] G-L Lan and D. L. Sengupta, "Frequency agile circular patch antennas", to be published. [7] G-L Lan, "Investigation of a class of tunable circular patch antennas", Ph.D. dissertation, Department of Electrical Engineering, University of Michigan, 1984. SHORITING POSTS FE ED Fig. 1: Single feed circularly polarized patch antenna using shorting post.

-4 - vA, TDT Tr T'T'M CIRCULAR POLARIZATI o,/ t^ \J p^. ^ \1 < ---- -900o - -... 2 A a / Lo Fig. 2: Condition for generating circular polarization. dB iL.J- L J1. J UUj/L 0. 8 + 0.6 + Measured by Long and Shen Theory 0.5 U 0.5 p w 0-.35 --- Fig. 3: Axial ratio vs normalized frequency.