Tuning waveguide to coaxial adapters

There are several different methods of tuning sections of waveguides and waveguide to coaxial adapters. Here are a couple of the more popular methods. Do not take any of these methods as being the optimum, but only as ideas. Consider the tools and test equipment you have available and pick a method that makes sense for your needs.

Signal source, directional coupler, diode detector useful at the frequency of interest, a 50 ohm termination for the frequency of interest, and connecting cables. I would suggest that any directional coupler used to make these measurements have a directivity of at least 20 dB with 30 dB being preferred. Since waveguide is fairly broadband in nature, a stable signal source is not required. If your source drifts around 100 kHz or so it will not make any real difference. For additional related information read the web page of Return Loss by going here. Also, see the chart below for suggestions of tuning screw sizes for various waveguides.

Spectrum analyzer, a second directional coupler that is calibrated to the first one, calibrated waveguide to coaxial transistion at the frequency of interest. The spectrum analyzer allows you to easily measure directivity and coupling levels of the directional coupler.

Waveguide impedance is a nominal 200 ohms but does vary a bit over its frequency range. This makes any waveguide to coaxial adapter not only an adapter but also a 3.9:1 transformer. Open air has an impendance of about 377 ohms and would therefore show a return loss of about 10 dB, or a nominal 1.9:1 VSWR.

I am not sure who first came up with this process but a good guess would be Paul Wade, W1GHZ. All you need is a good magnet (preferably a small diameter one) and a steel ball bearing. Depending on exactly what you are working with, the procedure will vary. For example; if you are working strictly with waveguide you need to do one thing but if you are working with waveguide to coaxial transitions the process will be slightly different.

Let's go over the procedure used to adjust a coaxial to waveguide transition. The easiest way to do this is to connect the transition to a 50 ohm dummy load (** see note at bottom) that is known to cover the frequency of interest. Insert a steel ball bearing into the transition before bolting it to the termination. Next connect a directional coupler to the transition using only a coaxial adapter and not a length of cable. If you have two different calibrated couplers they can be connected in series but backwards from each other so that one measures forward power while the other measures reflected power. Connect the output of each coupler to your measuring device. This can be two diode detectors (calibrated to each other) or maybe equal length coaxial cables through a microwave SMA relay to a spectrum analyzer. I like to use a spectrum analyzer so I can "see" what I am doing. If you only have a single directional coupler then you will first need to make measurements of forward power, then reverse the coupler, and measure reflected power. I would suggest doing this at least twice (but more is better) to average the individual measurements. Each measurement should not really change very much, if at all, but your signal source could change output levels on its own or even by changes in loading as you disconnect it to reverse the coupler. Also, if you turn the source off while reversing the coupler, it might not come back to exactly the same power level. Once you are set up hold the assembly so that the ball bearing is off to one side and not able to short out the coaxial probe to ground. Now apply a low level signal source and with the small magnet move the ball bearing around the walls of the transition. Look for a spot where reflected power is at the lowest level that you can find. Mark that spot as accurately as possible. This is where you will drill and tap a hole for insertion of a tuning screw.

Once you have drilled and tapped the hole for the appropriate size screw, put a nut onto the screw followed by a lock washer and thread the screw into the hole. Now reconnect the transition to the termination and apply the signal source. Snug but do not tighten the nut. Now adjust the screw for minimum reflected power as measured on the directional coupler. You may need to again snug the nut up as you adjust the depth the screw penetrates into the waveguide. Once you have minimized reflected power tighten the nut to hold the screw at that depth. You may have to go back a forth a few times as tightening the nut will slightly change the thread depth and therefore the tuning adjustment. Once it is locked in place, and depending on the match you have achieved, you can go back to the ball bearing and look for a second location to add another tuning screw. In almost all cases, each additional screw will provide considerably less of an impact on the tuning. You will also find that the tuning screws should not protrude very far into the waveguide or the tuning adjustment will be very touchy and narrow band. You might be able to get the reflected power quite low with a single screw threaded in say 1/2" but it will be a critical adjustment and one that may not hold (for example) with temperature changes. A second, or even a third screw only threaded in .1" might well be a better and more broadband option.

### So far I have referred to minimizing reflected power. Another way to say the same thing is to optimize (increase) return loss. The return loss link above will further explain this. What we are looking for is maximum forward power and minimum reflected power. This is also equal to maximum return loss. ###

There are "standard" locations to place tuning screws on the widest side of the waveguide. The following chart explains this process:

Trial and error (well slightly better then that) can be used with reasonable success. By placing tuning screws a set distance from the coaxial probe towards the open end flange most loads can be properly adjusted. Start with the first screw being placed at 1/8 wave from the probe and each additional probe 1/8 wavelength from the previous one. No screw should protrude very far into the waveguide. A set of 3 tuning screws should be able to tune nearly anything within reason.

** It should be pointed out that in reality there is no such thing as a waveguide 50 ohm load. A termination is the correct term. What these do is absorb power with high losses. The return loss of a waveguide termination is really determined by how much power is actually adsorbed by the ferrite material. There is not a 50 ohm resistor as you would find for a HF or even UHF "dummy load."

Some of the satellite LNAs have WR-75 waveguide inputs. They have been optimized for low noise figures (generally around 1-1.5 dB) at the satellite frequency, typically 11-12 GHz. That does not provide nearly as good of a noise figure at 10 GHz, which can be 3 dB or more.

In order to bring the noise figure down for 10 GHz operation some small tuning screws may do the job. You can carefully drill and tap the center line of the broad side of the waveguide with two or three holes for small screws. If there is not enough waveguide to do that, you can add an extension piece to it. You might start with two holes for screws. You can always add another one if it is needed but once you have a whole drilled you can't remove it. Start with the first hole about 1/4" from the end and space subsequent holes at 1/4". Use 0-80, 2-56 or even 4-40 screws for this purpose. If the waveguide is brass, I would suggest using brass tuning screws. If you can, silver plate them. If it is aluminum, I would suggest aluminum screws. Put a locking nut and lock washer on each screw to hold them in place. Depending on the application, you may also be able to solder them in place once the nut is tight. However, be aware that this could potentially move them slightly as you heat them up. You can also hold them inplace with glue or something else that will harden in place.