Putting together a station for working satellites, even the "easy sats," can be a bewildering exercise. The current crop of amateur satellites present many options for mode of communication, hardware requirements, and antenna selection. The ultimate antenna system consists of high-gain beam antennae with remote polarity switching and fully automated computer-controlled tracking. Many amateurs work satellites with a simple, omnidirectional antenna—but not nearly as well. What about something in between?

This article presents an easy-to-replicate, medium-gain, circularly polarized antenna that can be easily built for both 2 meters and 70 centimeters. The only test equipment required is an SWR meter used for calibration. All materials are available at your local hardware store for under $10.00. I have successfully worked ALL the current Low Earth Orbit (LEO) analog satellites with prototype versions of these antennae, and can copy the 2 meter downlink of AO-10.¹ The final, installed prototypes shown above are set at a 30 degree elevation.

Background:
My quest for a "better" antenna for LEO satellites began with dissatisfaction with available antennae, both omnidirectional and high-gain beams.

I built several of the popular "eggbeaters" and found them to be an irritating compromise. Without radials, they performed better than turnstiles, and much better than a 1/4 wavelength ground plane for low elevation passes, but had severe overhead nulls. With radials they performed exceptionally well for high elevation passes, but were dead below 20 degrees elevation. I believe the enhanced high-elevation performance is due as much to the excellent circularity of the basic design above 45 degrees as it is to the added gain from the "reflector" radials.² Another significant benefit I found to the basic eggbeater design is the higher input impedance greatly eases the matching burden for multiple driven elements when compared to a dipole-turnstile arrangement.

While my experience in building high-gain yagi antennae has met with mixed results, the successful antennae were also disappointing for LEO communications. I found manually tracking the LEOs to be too much work while I compensate for Doppler shift and try to make log notes. The narrow beamwidth of a high gain antenna is an unfortunate tradeoff. What I really wanted was a lower gain, thus a wider beamwidth, antenna with good circularity. For simplicity I choose to stick with right-hand-circular polarization (RHCP) rather than build in a switching scheme.

Design:
I thought if I could improve the overhead lobe of the basic eggbeater design, I could mount the antenna on a small rotator and tilt it 45 degrees and have adequate performance, while requiring only one or two azimuth adjustments per pass. A 90 degree half-power beamwidth in both azimuth and elevation, as opposed to the eggbeater’s omnidirectional pattern, is the goal.

The prototype design is formed around a pair of full-wave loops. A full wave loop, while having a radiation pattern similar to a dipole, has about 1 dB gain over a dipole (in the "normal" plane off the sides) and a radiation resistance of approximately 100 Ohms.³

Unlike the eggbeater however, the "driven" loop in this design is formed as a square, similar to a quad or quagi beam. This is done to move the high current portion of the loop parallel to a parasitic reflector element positioned "below" the loop. Note this arrangement provides gain in the axis perpendicular to the boom (in the "coaxial" plane off the top) of a common quad or quagi. The reflector is placed 0.1 wavelength from the driven element to provide maximum forward gain.⁴ Empirical results indicate this lowers the input impedance to almost 75 Ohms and introduces some negative reactance. (More on this later). Testing with the field strength meter proved to be quite revealing, and provides the elevation and azimuth patterns shown in Figure 3. Theoretical gain is 6 dBi, with an estimated front-to-back ratio of 11 dB.

Construction:
Materials are not critical. 10 gauge wire stripped from common 10/2 house wiring (leaving on the PVC jacket) and 1" PVC pipe and fittings are used in the prototype models. The 2 meter model uses 3/16" aluminum rod for the reflectors, but 1/8" bronze welding rod would work equally well. ⁶ Use of 12 gauge wire and 3/4" PVC are perfectly adequate for the 70 cm antenna. All stainless steel hardware, size 6-32, is utilized, but brass would work as well. Crimp-style ring lugs are used to connect the coaxial cables to the hardware.

Cut the first driven element to the specified dimension shown in Table 1. I allowed 3/4" at each end to be stripped and "curled" for connection of the hardware. These dimension are for the design frequencies of 146 MHz and 436 MHz.. Form into a square, as shown in Figure 4. Prepare a PVC coupling by drilling 4 holes, 90 degrees apart. Using 1/2" long 6-32 bolts, washers, and nuts, connect one 50 Ohm test cable (RG-8X used in the prototypes) of either 1/2 wavelength (2 m) or 1 wavelength (70 cm). This test cable will provide accurate SWR readings unaffected by the impedance mismatch of the cable to the antenna.

Minimum SWR should be about 2:1, indicating 100 Ohms impedance. The resonant point should occur around 145.5 MHz and 434 MHz. The lower "test" frequency is designed to offset the negative reactance introduced with the later addition of the reflector element. Next, while reading SWR at the "design" frequency, move the reflector up and down below the square driven element until minimum SWR is achieved. This should be about 1.5:1, indicating approximately 75 Ohms input impedance. This occurred exactly as expected in the prototypes, at 0.1 wavelengths below the full wave "driven" element. Finally, prepare a 1/4 wavelength delay line of 92 Ohm coaxial cable (RG-62) and connect as shown in Figure 5 (view from bottom) with the final 50 Ohm cable such as flexible 9913. The resultant SWR should be below 1.2:1 at the design frequency (calculated impedance of 45 Ohms at resonance)

Performance:
This design in inherently broadbanded due to the full-wave loop driven element. The antenna easily handles all of the current satellite frequencies without retuning. The coaxial performance of this single reflector design provides approximately 90 degrees of azimuth (off the top) and elevation (also measured off the top) beamwidth. My test equipment is not sensitive enough to accurately determine F/B ratio, but my on-the-air observations substantiate it must be at least the anticipated 11 dB, and might even exceed my expectations.

I do note a very noticeably higher, about 3 dB, relative strength reading when the receiving antenna is in the same plane as the 0 degree driven element. I believe this is due to an intrinsic design flaw. The 90 degree phasing line causes unequal currents to flow in each driven element and there is no balun to choke off current flowing on the coax. The use of a balanced matching/phasing section (like that employed in the commercial M² design) would likely fix this mismatch, but adds considerable complexity.⁷ I mounted the 70 cm antenna with the 0 degree driven element horizontal, maximizing horizontal (linear) polarization at the horizon, and mounted the 2 meter antenna at 45 degrees to minimize the height profile. An Alliance U-110 (TV type) rotator, commonly used as an elevation rotator, is employed for azimuth control.⁸ See the photo above.

In actual practice, I find the antenna to be excellent for mode JA birds, especially FO-20 and FO-29, which can be worked from horizon to horizon. For AO-27 I found the antenna adequate, but the horizontal 0 degree driven element actually penalizes me on this bird. I find it effective to "off-point" by about 45 degrees in azimuth.. The 2 meter version has become my standard Mode A uplink antenna for RS-12 and RS-15, with a discernible uplink improvement of at least 6 dB over my 1/4 wavelength ground plane antenna. Later experimentation indicates an elevation angle of 30 degrees is about optimum for maximizing gain at the horizon and still avoiding an overhead null. In summary, these circularly polarized, coaxial-axis parasitic, full-wave turnstile antennae are easy to build and replicate and they work well. Build them yourself and see! "See" you on the birds.

	436 MHz	146 MHz
Loop Length (*)	29"	80"
Reflector Length	13-1/4"	39-5/8"
Reflector Spacing	2-1/2"	7-5/8"
Phasing Line Length (**)	5-5/8"	16-3/8"
Test Cable Length (***)	20-3/8"	30-3/8"

* 3/4" at each end included for connecting to hardware
** based on nominal RG-62A. Verify with coax tee and dummy load procedure.
*** based on RG-8X with VF=.79. Verify with coax tee and dummy load procedure.

Notes:
1. For a more detailed description of these amateur satellites, visit: www.amsat.com.
2. Commercial versions of the eggbeater antennae are available from M², Fresno, Ca.
3. The ARRL Antenna Book, 16th Ed., 1992, p. 5-2.
4. Ibid., p. 3-11.
5. Ibid., pp. 11-2, 11-3.
6. 3/16" solid aluminum rod available from Texas Towers, Plano, Tx.
7. Martin Davidoff, The Radio Amateur’s Satellite Handbook, 1998, pp. 10-16, 10-17.
8. Alliance U-110 rotators available from Norm’s Rotor Service, Frederick, Md.

(C) 2000, Gerald R. Brown, K5OE