AT-68/UP Waveguide Horn

I have been interested in microwaves since working with them in labs while attending the University of Waterloo, but until recently had not had any other opportunity to work with them. I had purchased a small amount of waveguide and related equipment over the years, but not enough to make a working system. Then, over a relatively short timeframe, I was given some test equipment by an employer and more equipment and waveguide was generously donated by local amateurs. I was able to buy a few missing pieces at ham radio fleamarkets and local surplus stores. Suddenly, I had enough equipment for a useful test setup.

At the same time, I was also given the M/A-COM White Box 10 GHz equipment, see the White Box page on this site. I could use my waveguide test setup to help with setting up the White Box and related equipment.

One project was tuning up an AT-68/UP horn antenna for use as a test antenna or temporary dish feed. This antenna started life as part of a radar test set and was designed for use at 8.5 to 9.6 GHz, so it requires tuning for best performance in the ham radio band around 10.368 GHz. It is composed of a female N-connector which transitions into a short section of cast x-band waveguide which widens out into a small horn, see Figure 1.

AT-68/UP horn antenna

Figure 1. AT-68/UP Horn Antenna

To start, the return loss of the antenna was measured. The test setup for determining return loss is shown in Figure 2, with the AT-68/UP horn on the right.

AT-68/UP horn antenna under test

Figure 2. AT-68/UP horn antenna under test.

A block diagram of the test setup is shown in Figure 3. The signal source is a Gunn diode transceiver intended for use as a motion detector. It's simple to use, but doesn't have the historic appeal of the small klystrons in the university labs! Prior to any testing, I use my wavemeter to ensure that I am within a few megahertz of the desired frequency.

waveguide test setup block diagram

Figure 3. Block Diagram of the test setup for tuning the antenna.

I found tuning information for the antenna on the web site of Paul Wade, W1GHZ (in the Q & A section at

The solution is to add a tuning screw at an appropriate location in the centre of the broad wall of the waveguide section. To find the appropriate location, a ball bearing is moved around inside the waveguide by a magnet on the outside. The ball bearing adds shunt capacitance in the same way as the tuning screw so that locations where the ball bearing minimizes return loss are good locations for the tuning screw. The location nearest the signal source is used. The tuning procedure is shown in Figure 4.

Tuning the AT-68/UP horn antenna

Figure 4. Tuning the AT-68/UP horn antenna with magnet and ball bearing.

The unmodified antenna had a return loss of 11 dB, which corresponded well with Paul's measured SWR of about 1.8 (10.9 dB R.L.) for his version. I followed Paul's procedure with the ball bearing and magnet and found that the best location for my antenna differed from his by a small distance. I expected that the ball bearing would also lower the return loss at half-wavelength points along the waveguide section of the horn, so I measured the other locations where return loss was at a minimum and found good agreement with the theoretical guide wavelength.

I drilled and tapped the horn to accept a #6 screw at the location that Paul recommended and re-tested it for return loss while adjusting the depth of the screw in the waveguide. Unfortunately, I was only able to improve the return loss by about 1 dB. I removed the #6 screw and tried again with the ball bearing. The result was close to my original location, approximately 15/16" from the closed end of the waveguide, but too close to the existing hole to use a #6 screw, so I used a #4 screw. With the #4 screw in the new location, I was able to tune the return loss to the lowest value my system can read, approximately 27 dB at this power level. There was approximately 3/16" of the screw inside the waveguide.

Because the double-N adaptor used between the transition and the antenna is of unknown quality, I experimented further with a sliding-pin tuner between the directional coupler and the transition. The tuner should tune out any reactance added by the adaptor. I terminated the adaptor and set the tuner for best return loss, then replaced the antenna. With the #6 screw, I could improve the return loss by 3 dB with the tuner in place. With the #4 screw, I could improve the return loss to approximately 19 dB (SWR of 1.25), which is acceptable. The optimum depth for the #4 screw changed by only about 1/2 turn.

Based on these results, it seems likely that I could use this antenna with or without the double-N adaptor and have good return loss. I think my results differed from Paul's for two main reasons (and one minor problem):

Copyright Keith Thomson, 2010.