A SHORT DESCRIPTION OF GUNNPLEXOR CONVERSION FOR 10 GHZ AMATEUR RADIO USE
This is taken from several email messages that Larry Filby, K1LPS had sent in response to an open question on how to put salvaged Gunnplexer asemblies to use for ham radio 10 GHz WBFM.
The lack of electrical varactor tuning in most surplus units, limits their use for ham work. The frequency can also be tuned by varying the Gunn voltage... but this is often not a very satisfactory approach for a number of reasons. However, if you have ham microwave people in your area that do have tunable units... the surplus types can be set up, fixed tuned, on one of the standard frequencies... (10.250/10.280 for a 30 Mhz duplex setup.) and the tunable units can find them. Trying to communicate with two "fixed" tuned units as a pair on the duplex frequencies is an exercise in frustration... especially on any distance over a few miles, or less.
As for voltages of Gunn diodes: 10V is a common, but not universal Gunn voltage. Some of the surplus units, such as Alpha or NEC will peak at the 8V area. Determining their optimum voltage requires the right setup and some experience and vital information.
Testing a diode for it's characteristics, frequency, power, optimum voltage and current... requires that the unit be installed in a working cavity and test equipment such as a microwave power meter, variable current limited supply, and absorption cavity wavemeter. A lot of experimenters lack the right equipment, so you might have to find someone who does.
Not to make this too long... just a couple of observations about Gunn diodes. Once they are out of the cavity... determining polarity for supply purposes becomes difficult. (Physically, one end looks like the other, and too small for any markings.) The diode shows no front to back resistance. Usually around 1 ohm either way. Applying the wrong polarity will probably blow the diode. That's where a working cavity, power meter and current limited supply come in... but that procedure is beyond what I'll get into here. I'll be glad to provide the information to any who are really interested.
In summary... I would not salvage a good speed radar as there are far too many really cheap (Often free) alternatives. For loose Gunn diodes... you can test them in a cavity with the proper equipment... or take/send them to someone who has the equipment and the experience.
73 DE Larry (K1LPS/Vermont)
O.K... here is the additional info for testing Gunn diodes.
You need to put them in a known working cavity. You'll need to couple energy out of the cavity and into some sort of measuring device. A ferromagnetic isolator between the output of the Gunn oscillator and the measuring equipment will make things somewhat more stable. A microwave power meter is suitable. Apply a voltage of arbitrary polarity via a current limited supply with a current meter in series. Current limiting can be as simple as small pilot bulb of appropriate voltage and current rating. A 10 MW diode will likely draw around 140 MA at rated output. 35 MW diodes could draw as much as 350-400 MA. Start the voltage at about 4-5 volts. If you see output from the start, you have the polarity correct. If you see no output, start raising the voltage VERY slowly while observing current draw and any sign of output. If you note a very rapid increase in current for a rising voltage, and still no sign of any output... you likely have the wrong polarity applied to the diode. In this situation, I wouldn't advance the voltage beyond 8 volts or so. If it does the same thing with polarity reversed... you probably have a bad diode. (Presuming, of course, you are working with a known working cavity.) If you do have output, you can keep increasing the voltage... (And increase current limiting if appropriate.) until the RF output flattens out. This should be no more than 10-12 volts maximum. Once you've noted the slightest drop in output as you increase voltage... back off half a volt, and that will be your optimum supply voltage for the diode in question. 1 or 2 volts above this point where the output starts dropping, could blow the diode. Most diodes I've tested optimized in the 8-11 volt range. We've generally run X band M/A-COM diodes with 10V three terminal regulators, even though some diodes test out at slightly higher voltages.
Note that some 24 Ghz diodes are operated at much lower voltages... in the vicinity of 5V. I have not tested any of these diodes... but I see no reason why the same procedure wouldn't work.
For frequency testing... I've found good, calibrated absorption wavemeters to be adequate for the intended applications of the Gunn diodes. (Ie; WBFM mode.) Suitable spectrum analyzers and microwave frequency counters are not often in the inventory of most home experimenters.
As for tuning the output of a Gunn system by varying the bias voltage... you must limit the voltage excursions, and in no case, exceed the maximum rating of the diode at any time. At lower bias voltages... output can drop off rapidly and sometimes the cavity will break out into spurious oscillations. I would limit excursions between optimum rating down to no more than 2V less than optimum, and set the mechanical cavity screw to center the frequency at the midpoint of desired operation. Considerable "fiddling" of voltage and mechanical tuning may be required to achieve useful operation using this method of tuning. At that... you can still expect considerable difficulty in making two "fixed tuned" Gunn systems "talk" to each other over any respectable distance.
Good luck and happy experimenting.
73 DE Larry (K1LPS/Vermont)