K5OE, Build This No-Tune Dual-Band Feed for Mode L/S)

Build This No-Tune Dual-Band Feed for Mode L/S

Excerpted From The AMSAT Journal, Jan/Feb 2002

The AO-40 transponder has two uplink receivers active most of the time for CW/SSB activity. Most operators use the "U-band" at 435 MHz (70 cm). Also available, however, are two L-band (23 cm) receivers: L1 at 1269 MHz and L2 at 1268 MHz (1). The reasons for going to L-band can be varied, but there is no arguing the benefits in reduced antenna size and AGC-suppression. The types of L-band antennas are varied as well. Many use helices. Others use beams and arrays of beams. Still others use dishes, small and large (2, 3).

I recently acquired an antique UHF TV dish measuring 1.2 m in diameter. I wanted to use it for both receive on "S-band" at 2401 MHz (13 cm) and the uplink on L-band. I covered it with aluminum mesh and built a dual helix feed for it, but was unhappy with the L-band performance (4). It seems the concentric helices interacted with each other to a fairly substantive effect. Having had very good success with a couple of patch feeds on S-band, I designed, built, and installed a dual-patch feed on a 1.5 m solid dish for Field Day 2002 (5). This arrangement worked superbly on uplink (with 25 W), but was embarrassingly deaf on receive. This second dual-band feed failure led me to experiment for months with different configurations, leading ultimately to the design presented here. The project goals were 1) good performance on both S-band receive and L-band uplink and 2) an easily reproduce-able model using common hardware and simple hand tools.

Patches are better than helices as dish feeds. This revelation came to me while doing investigation and experimenting with helix antennas (6). Whilst in the middle of this investigative foray, I saw the radiation pattern for the G3RUH patch feed published on James Miller's web site (7). When I modeled that pattern and input it into the W1GHZ feed pattern program, it produced an amazing 72 percent efficiency (8). The best helix I ever modeled has about 60 percent efficiency. I8CVS recently ran his own antenna range tests of a design similar to the G3RUH patch and produced a similarly impressive pattern. Then I came across the "truncated corners" square patch design popularized by K3TZ (9). This specific AO-40 design is attributed to 7N1JVW, JF6BCC, and JG1IIK: I find references in the literature going back over a decade for this now-common commercial design. The first model I built outperformed my best helix-in-cup design by a full S-unit (delta over the noise) on my FT-100 portable setup (10). Compared to a helix, the patch simply has better illumination efficiency with less spillover from side lobes.

Patch theory is beyond the scope of this article, but can be summarized as building a shape resonant at the desired frequency, compensated in size by the capacitive inductance between itself and the reflector. A patch can be practically any shape as it basically acts like a parallel-plate transmission line (11). Current in a patch flows from the feedpoint to the outer edge(s), where all the radiation occurs. The reputed, but often disputed, circularity of the truncated corner patch is accomplished by effectively designing two antennas into the patch element (the two different diagonal lengths) and feeding them 90 degrees out of phase (12, 13).

For my 1.2 m dish, a little parabolic dish math tells me I can expect 21 dBi gain on L-band and almost 27 dBi on S-band with an assumed 50 percent efficiency:

G = 10 * Log10[((4(/(2)*A], where (=efficiency, (=wavelength, A=aperture of the dish (14)

G = 10 * Log10[0.50(4(/86.72)*(3.14*576)], where A= pi*r2
G = 10 * Log10[0.0724*1807]
G = 21.2 dBi at 1269 MHz

G = 10 * Log10[0.50(4(/24.5)*(3.14*576)] , where A= pi*r2
G = 10 * Log10[0.2564*1807]
G = 26.6 dBi at 2401.5 MHz

Where does the feed get mounted? The focus is where the parabolic shape of the dish concentrates the reflected signal. The author placed the antenna flat on the garage floor to measure the depth, and thus determined:

f = D2 / 16d, where D=diameter of the dish and d=depth of the dish (15)
f = 482 / 16*7.25 (inches)
f = 2304/116
f " 19.8 in. (50.5 cm)

Presented here is just one version of the countless combinations of hardware and patch designs. Inherent in this design, though, are five key design and construction features developed from building and empirical testing of a number of patch feeds.

1. The specified dimensions are critical for no-tune operation. Figure 2 depicts the dimensions necessary to build the dual-feed patch. I recommend you reproduce this sketch accurately on graph paper. When you cut your patches you can lay them on the paper template for checking. I repeat: the dimensions are critical. Even a 0.5 mm error will throw your resonance off considerably--patches are not broadbanded.
2. Mechanical rigidity is required of the reflector. Spacing between the driven element (patch) and the reflector affects the resonant frequency. I found 0.025" aluminum sheet and 26 gauge copper sheet acceptable for a single S-band patch feed, but too flimsy for an L-band reflector. Use more rigid material or provide additional stiffening for the L-band reflector, such as shown in Figures 4, 5 and 6.
3. The patches must be electrically isolated from each other. A metallic center support works for a single patch but creates harmonic coupling problems when patches are stacked for mult-band use. I found using nylon machine screws and nuts solved the vexing problem of the S-band patch coupling to the L-band patch.
4. The "straight corners" of the truncated corner patch must be kept clear of any proximate metal. This includes the edges of the feed support or "cup," if used. See Figure 5.
5. Feeding the patches at 90 degrees to each other minimizes the electromagnetic interaction between the two antenna fields.

Figure 2: Dual-Band Patch Feed Dimensions (mm)

One final design issue deals with the first harmonic of the L-band antenna: you must select a method to reduce (considerably) the potentially destructive effect from the 1269 MHz signal's first harmonic. Severe de-sense of your receive signal can occur and potentially even overload and damage the first active device in your system. Sensitive preamps and downconverters without a pre-RF-amplifier filter will need an external filter installed. I have used a G3WDG stub filter rated at 100 dB rejection to good success ahead of my preamp (16). My current setup, however, uses a TSI 3731AA downconverter with it's internal combline filter providing adequate filtering (17). Using the downconverter directly at the feed has a noise figure (NF) of 1.0, compared to the cumulative NF of 1.6 using a filter and a preamp.

Construction of the feed begins with selection of material for both the electrical parts (the antennae) and the mechanical parts (the support structure). The L-band antenna is constructed using a 6" x 6" double-sided circuit board for the reflector and a piece of 26 gauge copper sheet for the driven element (patch). A flanged female N-connector is used for the feed connection. The S-band antenna is constructed of two pieces of 26 gauge copper sheet and the feed connection is made with a short piece of UT-141 (0.141" copper clad semi-rigid conduit) terminated in a male SMA fitting. Figure 3 illustrates the assembly of the L-band reflector with the nylon center support bolt, the L-band N-connector, and the S-band semi-rigid coax terminated onto an SMA-to-N adapter through the circuit board.

Figure 3: Assembly of the L-band Reflector

The support structure began life as a "virgin" paint can, measuring 155 mm in diameter and cut down to 15 mm depth. The middle of the bottom of the can was cut out and the PC board is trimmed to fit inside the can bottom. Stainless # 4 x 3/8" bolts, washers, and nuts were used to secure the PC board to the can bottom. A # 6 x 1-1/2" nylon bolt is secured through the center of the PC board with two nylon nuts to provide the requisite 6 mm spacing for the L-band patch. Figure 4 shows the L-band patch in position and ready to be soldered to the N-connector. Note the hole through the L-band patch for the S-band UT-141 coax to pass (without making contact).

> Figure 4: The Support and L-band Reflector and Patch

The remainder of the antenna is then assembled in order: first the L-band patch is secured with two nylon nuts and soldered to the N-connector; then the S-band reflector is secured with one nylon nut, to provide 3 mm spacing, and the UT-141 coax shield is soldered to the S-band reflector; and, finally, the S-band patch is secured with a single nylon nut (3 mm spacing) and soldered to the center conductor of the UT-141 coax. To repeat the order: L-band reflector, two nylon nuts, L-band patch, 2 nylon nuts, S-band reflector, 1 nylon nut, S-band patch, and 1 nylon nut.

> Figure 5: The Completed Dual-Band Patch Feed

An electrical check with an Ohmmeter of the completed feed should show the two reflectors connected, but the patches are isolated from the reflectors and from each other. Figure 5 depicts the completed feed. Note how the "sides" of the support are cut out to avoid proximity to the L-band patch and how the L-band and S-band patches are at 90 degrees to each other. Figure 6 shows the back of the feed, complete with an angle support for the downconverter. The flanged N-connector is for the L-band coax and the male-N adapter is secured from the other side of the feed with the SMA fitting on the UT-141 coax.

For those tempted to "tune" the patch, I recommend you do it with the feed installed on the antenna--as the dish surface affects the feedpoint impedance slightly. You can adjust the feedpoint impedance, thus the resonant frequency, quite a bit by adjustment of the spacing of just the straight corners. You do not need to change the spacing at the center or the feed-just slight bending of the straight corners up or down will change the tuning. Careful: a little bit goes a long way. I found this patch design very repeatable and it will work adequately (an SWR below 1.5:1) with no adjustments.

> Figure 6: Rear Of Completed Feed

The antenna performs well to both my expectations and the calculated predictions above. On receive, this antenna performs at 4 S-units better than my 45 cm dish and 3 S-units above my 65 cm dish (both other dishes have similar patch feeds and the exact same downconverter). It also clearly outperforms my previous dual-helix arrangement on the 1.2 m dish, but I was unable to do a side-by-side comparison. On transmit, it does equally well, with a decent signal into the satellite with only 10 W measured at the antenna. The L-band is noticeably improved over the helix predecessor. At low squints I find the L-band uplink to be about 1 S-unit weaker than my U-band uplink.

This design is simple and effective, but is merely one way to construct a dual feed. Cookie tin lids also make excellent supports. I found tin snips to be a good investment and much easier to use than a hack saw. A flat file is then used to de-burr the edges of the patches. Stainless steel hardware was used, most notably #4 x 3/8" machine bolts and nuts for the antenna hardware and #6 x 1/2" for the support structure connections to the support arms (1/2" aluminum tubing). The copper sheet is MUCH easier to solder to than the aluminum. I gave the completed feed a few coats of white enamel to protect the copper and to minimize the visual reflections.

This is not the only dual-band antenna on AO-40. There are many varied, innovative designs available as further food for thought. Others have been found, including G6LVB's simple and effective 1.2 m home-brew stressed "chicken wire" dish with a dual G3RUH helix feed (18). G3WDG has a 3 m dish with L/S-band helices and a K-band (1.3 cm) feed horn, and W0LMD has developed some popular dual- and tri-band "round" patch feeds (19, 20).

References:

1. AO-40 frequencies, "Official Transponder Frequency Bandplan for P-3D"; http://www.amsat-dl.org/p3dqrg.html.
2. Kingery, Mike, KE4AZN, "Setting Up for AO-40 L-Band Uplink," The AMSAT Journal, May/Jun 2002, pp. 14-16, also: http://web.infoave.net/~mkmk518/.
3. Suding, Dr. Robert, W0LMD, "Converting TVRO Dish & Dishes For Amateurs": http://www.ultimatecharger.com/dish.html.
4. Brown, Gerald R., K5OE, "Dual-Band Dish Feeds for 13/23 cm," Proceedings of the 2002AMSAT-NA Symposium, October 2002, pp. 123-131.
5. MacAllister, Andy, W5ACM, "Field Day 2002," 73 Amateur Radio Today, Sept. 2002, pp. 48-52.
6. Brown, Gerald R., K5OE, "A Helix Feed for Surplus MMDS, Antennas" Proceedings of the 2001AMSAT-NA Symposium, October 2001, pp. 89-94, also: http://members.aol.com/k5oe.
7. Miller, James S., G3RUH, "'Patch' Feed For S-Band Dish Antennas; http://www.jrmiller.demon.co.uk/products/patch.html.
8. Wade, Paul, N1BWT (W1GHZ), Online Microwave Antenna Handbook, "Chapter 4, Parabolic Dish Antennas," http://www.w1ghz.cx/antbook/chap4.pdf, Mar. 98.
9. Zibrat ,Timothy S, K3TZ, "2.4 GHz Patch Design": http://www.qsl.net/k3tz/.
10. Brown, Gerald R., K5OE, "Patch Feeds," http://members.aol.com/k5oe.
11. Kraus, John D., Antennas, 1988, "16-12, Patch or Microstrip Antennas," pp. 745-749.
12. Thiel, David & Smith, Staphani, Switched Parasitic Antennas for Cellular Communications, Artech House, 2002, "Chapter 3, Patch Antennas," pp. 79-96.
13. Orban Microwave, "The Basics of Patch Antennas: http://www.orbanmicrowave.com/antenna_application_notes.htm.
14. Krome, Ed., K9EK, Mode S-The Book, p. 96.
15. Ibid, Krome, p. 109.
16. Suckling, Charles, G3WDG "Notch Filters for AO-40 Mode L/S": http://www.g3wdg.free-online.co.uk/notch.htm.
17. Seydler, Robert, K5GNA,"Modifications of the AIDC 3731 Downconverters"; http://members.aol.com/k5gna/AIDC3731modifications.doc.
18. Long, Howard, G6LVB, "My Shack Configuration-Spring 2002": http://www.g6lvb.com/g6lvb_shack_spring_2002.htm.
19. Suckling, Charles, G3WDG, "K-Band Results From AO-40": http://www.g3wdg.free-online.co.uk/kband.htm.
20. Suding, Ibid.