Living in a community that expressly prohibits ham radio towers often means enduring constant frustration trying to work DX with wire antennas or a multi-band vertical with what is always an inadequate radial field (read: less than 200 radials) over the “worse soil in the world”. My mantra of “work ’em on all bands and all modes” that I pursued for many years from my NH QTH has morphed to “hope I can hear them and they can hear me”. My lot can support a 40-foot tower that would be great for a small beam like the Cycle 24 used during the World Radiosport Team Championship a couple of years ago. Unfortunately, I would never get past a permit for the tower base if I ever tried to put one up.

Ah, but there is hope. The frustrations wrought by CC&R limitations often bring out the ingenuity in hams. Thanks to Layne, AE1N, I visited the website of Jeff, AC0C (see AC0C.com). Jeff lived in a condo with the no-tower restrictions. He looked over the attic where he lived and set about figuring out how to build an array of beams using the roof support structure. In a triangular volume measuring 16 feet at the apex, 20 feet long, and 40 feet wide, he managed to cram in a 22-element array that covers 160 – 6 meters! A scan of Jeff’s website is a testament to a ham’s perseverance to build an antenna farm where common sense says you cannot. Jeff used the popular EZNEC modeling program developed and maintained by Roy, W7EL. The overall configuration of the project shown on Jeff’s website looks like the proverbial rat’s nest. Somewhere on his site he mentions using over 1000 feet of coax, hundreds of ferrite cores, dozens of relays, and numerous switching boards to select the desired antenna while holding the RFI beast at bay.

Intrigued by Jeff’s work, I decided to take a look at my garage attic and see what I could do. My attic measures about 17.5 feet long, 19 feet wide, and 5 feet at the center above the floor. There is usable space beyond the garage over the remainder of the house that runs to the back, providing more depth. Unfortunately, my attic runs N-S so that any fixed antenna would have to be situated to favor a pattern to the E/NE direction for DX. I studied Jeff’s design for his 2-element 40 meter Yagi and decided to scale it for 30 meters. Each element resembles an inverted-V with the lower half of the ends bent at right angles to run horizontally along the floor of the attic. The horizontal portions of each element run toward each other so that the antenna resembles a modified “bent” Moxon antenna. (The Moxon antenna owes its design to Les Moxon, G3XN (SK). It is a 2-element beam that resembles a rectangle, is easy to build, and is forgiving in dimensions for a given band.)

I modeled the antenna with a separation of 19 feet between the elements and inserted loading coils in the driven element to shorten them. The apex height of the antenna is only about 14 feet above the ground so I was not expecting spectacular gain in the primary direction (North). The resulting pattern resembled a low dipole with maximum gain of 4.5 dB North and South, dropping to 3.4 dB at 45 degrees. While this may seem respectable, it occurs for an elevation angle of 45 degrees which means it would not work very well for typical DX angles. Indeed, when I looked at the gain performance at 24 degrees elevation, I had less than a dB. (The astute DX’er will notice that good DX performance occurs for elevation angles of 15 – 20 degrees. When limited to a height of 14 feet, the gain profile at 15 degrees is over 8 dB down from its peak.) Figure 1 shows the layout of the antenna. I realized that I’d be better off with a 30-meter inverted-V dipole off the back of the house.

Not to be deterred, I then modeled a 17-meter 2-element beam using relays to disconnect the 30-meter element at its coils and inserting an inverted-V element between the 30-meter elements to act as a director for 17. Figure 2 shows the results where the driven element is on the right and the director is at the center.

The performance for 17 meters was similar to 30 meters with a modified dipole pattern. The gain was a little better at a 24-degree elevation angle but still less than 2 dB. Again, better to stay with my inverted-V that boasts a 3.8 dB gain toward NE.

While I pondered my next move, it occurred to me that what worked for Jeff might not be the best idea for me. All that coax and those ferrite cores pointed to a lot of blood, sweat, and tears to get rid of RFI gremlins that such close quarters tend to foster. I considered using relays to change bands by lengthening or shortening the antenna elements. However, this would mean control cable wires from the shack to the antennas, themselves inadvertent antennas when transmitting. There has to be a better way.

Then, the long-dormant light bulb in my head came on. Why not try a 2-element wire beam? I looked at a model for such a beam for 17 and 15 meters, 2 elements for each band on the same “boom” and each fed separately. The results showed 17 meters behaves as expected with a gain of 6.2 dBi and a F/B of 21.6 dB. However, the 15-meter portion had a peanut-shaped pattern that was reversed from the intended direction with a F/B of less than 5 dB. I adjusted the height of the 15-meter elements within the available limits but to no avail.

Now what? I looked at the Moxon design again for its space-saving feature to see if I could get more isolation between the beams. I pulled up the file for the 2-element 17-meter beam and another file for a 15-meter Moxon beam. I merged the two antennas and juggled the positions so that the Moxon was a foot above the wire beam. Figure 3 shows the EZNEC model.

Figure 4 shows the azimuth gain pattern for the 17-meter beam.

Changing antennas for the 15-meter Moxon antenna produces the pattern shown in Figure 5.

As the model plots suggest, both antennas have nearly identical performance. Success at last! At least on paper minus any interactions and interfering structures such as air conditioning ducts and electrical wiring. Installing two antennas with separate feeds is easy to control with a remote antenna switch. I avoided the need for relays to shorten elements on 17 meters in order to work on 15 meters. I did not want to use relays because of the prospect of RFI causing them to trip, and the need for extra wires from my shack plus ferrite cores for RFI suppression. Now on to the fun part – BUILD IT.