50MHz Moxon Rectangle




[Tech Topics]

The FT-857 is the first radio that I've owned with 6m capability, and so I needed a 50MHz antenna. The HF wire works in a pinch, but what I really wanted was a beam. The balcony mast can only hold very small and light antennas, so I thought I'd try a Moxon rectangle. I first discovered this antenna on W4RNL's site. I originally said on this page that the mechanical design was insufficient for permanent home station use, but I've changed my tune. The Moxon has been on my chimney for nearly nine months as of April 2007 through up to 50 MPH winds and some light icing with no problems.

The Moxon above the trees. You can see the HF magnet wire in the foreground.

A design goal of this particular antenna was to use only parts I could find at Home Depot, to make it a particularly suitable antenna for the weekend builder. I made a trip and pretty quickly found what I needed. The antenna can be made fairly cheaply, especially if you can use some items on hand rather than buying every piece of hardware.

What you need, at least:

  • 3x 8' lg. 1/2" square aluminum U-channel (C-channel?), 1/16" thick
  • 1x 4' lg. 1" square aluminum tubing, 1/16" wall
  • 2x fiberglass driveway marker stakes
  • 1x 3' lg. 1" equal leg aluminum angle, 1/8" thick
  • Large quantity of black cable ties (the bigger ones)
  • Insulating cable staples, see pictures
  • Assembly hardware - various 8-32 machine screws and nuts, maybe rivets

A shot of the installed antenna with the now rotating mast

The antenna dimensions are calculated using Moxgen by AC6LA for a frequency of 50.200 MHz. Moxgen takes input of the wire diameter, and, well, I didn't use wire, but I have it on good authority that you can model a square element by using a round wire of diameter 1.18 times the square's side length. I input a diameter of 0.59 inches for the 0.5 inch square element and the antenna was spot on. The Moxgen program gives you the side-to-side and front-to-back measurements of the rectangle as well as the gap between the element tips, which is very, very critical. The gap measurement cannot be sloppy. I measured to the nearest millimeter. This is probably excessive precision, but I've got a millimeter graduated tape measure, so I did it anyway.

A shot of the driven element with driveway marker insulator.

As it turns out, a piece of fiberglass driveway marker with a cable tie wrapped around it gives a nice snug fit to the inside of the U-channel material. A couple of cable ties can be used to hold the resulting insulating bar tightly into the element. This is what was done at the feedpoint where the driven element must be split. I originally used a double bar for strength at the feedpoint and then had to cut much of it out to allow the boom-to-element clamps the access they needed, but there's still about 2 inches of doubled material right at the break.

A closeup of the feedpoint split.

The element gaps also need mechanical support. They get the same treatment, but with a slight modification. A piece of fiberglass is cut to EXACTLY the gap spacing. It's cable-tied to a piece several inches longer. This allows you to easily set the gap spacing very accurately, because the element tips actually butt up against a physical object. I did this so that I could break down the antenna for portable use and repeat the gap setting effortlessly.

The element gap with the insulator and spacer.

The U-channel is especially suited to Moxon construction because it makes the square corners so easy to fabricate. You just overlap the walls of the two pieces and drill a hole all the way through. Since this antenna is intended to be broken down and toted around in the car, I decided that I wanted the element tips to FOLD. This requires that one corner on the element and one corner on the tip be rounded. I took a #8 washer (I'm using 8-32 hardware) and placed it as far in the corner I was trying to round as I could. I then traced both the outside and the inside of the washer. The inside trace told me where to drill the bolt hole. I used a Dremel and cutoff wheel to remove the corner outside the outer trace. This gave me a nice rounded corner concentric with the pivot hole. You can sort of see one of the necessary roundings in the corner photo below. It's inside the channel. I used an 8-32 bolt and nylon insert locknut to secure the joint.

Corner junction detail.

The elements are secured to the boom using some modified cable staples made of some tough grey plastic. They were meant to be nailed into a surface. I pulled the nails, enlarged the holes and cut the legs down a bit so that I could use them as insulating clamps to isolate the elements from the boom. They bolt into aluminum angle pieces that are riveted to the boom, but there are other approaches that would work just as well here.

Element clamp detail with riveted bracket.

The reflector is mounted exactly the same way as the driven element, and it's even easier because it's not split. I found, though, that an additional short piece of U-channel overlapped with the center of the reflector element gave much greater stiffness and strength than the unreinforced element. The short piece is riveted to the element to make a short section of essentially square tubing, and this also helps to keep from crushing the element with the clamps. If you wanted to build a more rugged version of this antenna, you could double the reflector all the way out to the ends in this fashion.

Reflector-boom junction detail.

The coax is attached using a BNC connector in this design, but you could just run the center conductor and shield of the coax to either side of the driven element. I like to use connectors when I have them. A ferrite bead choke balun is used at the feedpoint for lots of good reasons. It keeps the coax from radiating, and probably more importantly in an apartment application, it keeps CONDUCTED noise from the shack off of the antenna. The outer shield of the coax can conduct currents from switching power supplies, computers, and other sources of RF noise. This common-mode current can't find its way into the radio if it cannot reach the antenna. Why? The radio only responds to differential voltages between the shield and the center conductor of the coax at the antenna feedpoint. The noise can be conducted into the "ground" side of the driven element if there's no balun. It's not conducted onto the "hot" side, though, so you have a differential between the half of the antenna connected to the shield and the half of the antenna connected to the center conductor and the radio can hear it. The choke balun breaks the path that conducted noise would take onto the shield half of the antenna, and the antenna will be quieter for it.

The BNC is connected to the element with solder lugs. I used some Coax-Seal for weatherproofing.

Finally, the antenna needs to be attached to a mast. I'm using this for SSB/CW work so horizontal mounting is appropriate. I just used a U-bolt with a saddle and some wingnuts for easy on, easy off mounting. You can see the bead balun in the picture below. It's just 50 or so FT-43 material ferrite toroids over a piece of RG-58. If you can't find these you can use a coil of coax. I had the beads, and I know, you can't get them at Home Depot but I decided that I would define the antenna as ending at the coax connector....

Boom to mast clamp and choke balun.

The antenna is a pretty good beam. The front-to-back is quite respectable and the gain is quite a bit higher than my HF wire antenna. I haven't done any non-qualitative pattern measurements yet, but as a measure of front-to-back I can take an S7 signal and make it S0 and in the noise by pointing the antenna away. The gain is modest, and I've not measured it, but I'll include some EZNEC model results to give a rough idea. This is the antenna designed for 50.200 MHz using Moxgen (which will spit out an EZNEC model file!), over real, high accuracy ground (0.002S/m, Dielectric Constant 13) at a height of 28 feet, plotted for an elevation angle of 10 degrees. This is an inaccurate representation of my antenna's living conditions in that I've got plenty of nearby metallic objects to introduce pattern distortion, but it's better than nothing.

Azimuth pattern 50.125 in the clear at 28 feet. Click for one you can actually read.

I did try to make an estimate of the effect that nearby objects had on the antenna pattern. I came up with the following results, which are probably not especially accurate given that I'm not sure what the overall distribution of metal in my apartment roof is, but this does give me some idea of the real pattern in various directions. The antenna patterns show up more clearly in PDF, so here's the report: moxon_realitycheck.pdf ~390kB

The formula for equivalent diameter for a square conductor, 1.18*L where L is the side length of the square is awfully close to what I got when I took the geometric mean of the diameter of the circle that will just fit inside the square and the circle that will just fit around the square. I doubt this is a coincidence. I got the 1.18*L formula from W7EL.

over here?
© 2005-2014 Daniel S. Zimmerman, N3OX