Tuesday, October 30, 2012

24GHz the old way

In the 80's wide band FM was common in the 10 GHz band. When 24 GHz devices were available at surplus, wideband FM was overridden by narrow band modes (SSB), so there was very little wide band FM activity at 24 GHz. Even today there is almost no wideband activity in any microwave bands (Except for some ATV signals). Why?

Of course SSB signals have clear advantages over any FM mode. But FM signals have also interesting advantages: Wideband microwave transceivers are very easy to construct for the beginner and also very cheap. They provide a easy path to the microwaves world and can provide very interesting contacts. Also nowadays is very easy to construct such wideband equipment because of the availability of general coverage receivers.

Looking for microwave devices in eBay, I found some interesting (and cheap) 24 GHz gunn modules (Doppler modules a.k.a. door openers) so I decided to build a pair of wide band FM transceivers, using the old techniques but with modern approaches.

The modules I acquired have a gunn transceiver with waveguide output and a small horn antenna. They are identical to the 10GHz units used in the 80's, but working at 24 GHz. The only difference is the gunn diode voltage: 24GHz units work at 5 volts and they produce around 5mW of RF power.

The RF section

The RF section is very simple. The gunn diode voltage is protected by a 6.2 volts zener just in case of overvoltages or spikes. The 33 ohms resistor and the 100nF capacitor just add some AC load to the line to ensure the modulator works nicely.

The signal from the mixer diode is feed to an ERA-3 mmic amplifier. I chose ERA-3 just because I have some of them lying around but almost any MMIC will work here as long as you tweak the polarization resistor accordingly (The 270 ohms one).

The circuit is so simple that it can be mounted on a piece of unused PCB just making some pads with the help of a drill machine. It permits to use IF frequencies from about 10 MHz up to around 100-200 MHz. The lower frequency is determined by the 10 nF capacitors and the upper frequency is determined by the mixer diode feed through capacitor in the gunn diode cavity.

The Low Frequency section

The low frequency section is divided in two parts: The gunn diode modulator and the tone / beacon circuit.

The gunn diode modulator is based on the LM317 voltage regulator. The 330 / 1k ohms resistors set the output to 5 volts. The signal from the microphone is limited by the pair of 1N4148 diodes and applied to the regulator pin of LM317. There is no need to amplify the signal. Most electret microphones give enough signal to fully modulate the gunn diode. Applying this signal to LM317's regulator pin produces a nicely modulated voltage.

The tone / beacon modulator produces an alternating pair of tones. It is very useful when aiming antennas or looking for signals in band. It's based on the CD4093 and there are a pair of spare gates just in case you need them for any other function. The 10k variable resistors adjust modulation level for the microphone and the tone / beacon circuit.

The whole circuit is so simple it was constructed in a small breadboard and mounted in a small box not much larger than the gunn cavity:

And of course, unless you live in an area with lot of 24 GHz wideband activity, you must construct a pair of transceivers.

This is a view of the box back panel:

The DC jack is used to apply 12 volts DC from a battery. The BNC outputs the IF signal to an external receiver, like you general coverage scanner or walkie talkie. The 3.5mm jack is used to connect the microphone and the small switch turns on or off the tone / beacon generator. In fact the box is so small than you can use it hand-held.

Maybe the first 24GHz handheld in the world.

It should be obvious, but if you use a walkie talkie as WFM receiver, make sure you never transmit into the unit or you will burn the MMIC and maybe the mixer diode.


The last step is the tuning of the cavities. In Spain the 24 GHz band is limited to only the primary segment (24.00 - 24.05 GHz) so precise tuning using a low IF is a must. Tuning was easy: I used a 1.2GHz walkie, a 10dB attenuator, and a diode. Even a cheap 1N4148 worked fine:

A 24 GHz signal generator

The diode was placed in front of the horn at few centimeters:

Tuning the 24GHz cavities

And then, connect a frequency counter, or even better, an spectrum analyzer at the IF output. Of course there are better diodes to create signals in the 24 GHz range, but the 1N4148 worked just fine, it is cheap, and we all have some dozens of them in the junk box.

The tuning consists in calculate the right harmonic being received:

First you must calculate which harmonic is the one you are receiving. The tuning is so broad that you can receive easily several GHz around 24 GHz. Place a carrier at 23cm and measure the frequency output from the mixer diode. Now move the carrier at 23 cm for example 100 kHz up. Measure the drift at the mixer diode. if you measure 1800 kHz drift you are receiving around 22 GHz. If you measure a 2000 kHz drift your tune is around 25 GHz. As you would guess, the right drift is 1900 kHz (x19). Now you are on the 24 GHz band.

Lets do some numbers. If your 23cm walkie is on 1265.525 MHz, your harmonic is on 24.045 GHz. Lets assume you are measuring a 20 MHz signal at the IF output. This means your Gunn diode is oscillating either at 24.025 GHz, or 24.065 MHz. These both signals mixed with 24.025 will give a 20 MHz IF output. How to determine which one is the real oscillating frequency?

It's very easy. Move your walkie's carrier 25 kHz up while watching the IF output. If IF output goes UP, this means the LO is UNDER the RX frequency (24.025 GHz). If the IF goes DOWN, this means the LO is ABOVE the RX frequency (24.065 GHz).

With this method is easy to tune a pair of cavities for any given IF.

In fewer words, the IF drift divided by the walkie TX drift gives you the harmonic number you are receiving. 19 is the right one. And the direction of the drift tells you if the LO is above or under the RX frequency.

I admit is easier to do it than to explain it :-)

Does it work?

Of course it works! The two units can be used fixed with the help of a small tripod or even handheld as seen in the above picture. They produce easy contacts up to 5 km (the longest path I tested by now) with strong signals and IF values around 20-30 MHz having visual line.

It is quite interesting using 24GHz while walking or moving in a car. Vegetation completely blocks 24 GHz signals. A tree or even the smallest bush blocks the signal. It is also very interesting to see how the 24GHz signal is reflected by walls or other objects. Sometimes the contact can be made by looking at these reflections when direct path is obstructed. It is also interesting to see how a simple wall attenuates the signal almost to DX levels when the signal goes through it. When using a window from home or from a car, remember to open it: Glass is completely opaque at 24 GHz!

The small horn antennas produce about 30 degrees beamwidths, so they are easy to aim even using them handheld.



  1. Some folks are building EmDrives based on these, so thanks for the very useful tips!

  2. Excellent project. I am curious about the rf modules you got. I have not been able to find anything similar.
    73 de Gerry

  3. Interesting post. I think these would work great for ATV
    Like you say you need to make two or make it a club project providing you can get enough of the units.
    Thanks for sharing your experiments.
    73 Mark G0NMY

  4. Thank you Miguel for the detailed post! I am playing on 24 GHz with modern radar modules (CDM-324, or IPM-165 if you buy the original), as we've done on 10 GHz with HB-100.
    My updates at for those interested.

    73! Paolo IK1ZYW

    1. Gunn diode based 24 GHz modules are now quite rare so your experiments are quite interesting. Good work!

  5. Have a search for HONEYWELL INTELLISENSE DT7550A ADT SENSORS. They have PIR and a 24GHz cavity.
    Could be quite useful source.