A 10GHz Gunn Diode controller.

This article first appeared in P5 - March 1995

As many people have found out, its very easy to produce a TV picture from a  Gunn diode oscillator but very difficult to achieve high quality images with sound. This design overcomes many of the problems experienced with simpler  circuits yet is still easy to construct and set up.

All the components, with exception of the etched PCBs are available from  Maplin Electronics. The boards can be home made and the track layout is included  in this article.
Changing the voltage across a Gunn diode causes the frequency and amplitude  of its oscillations to shift. Since most receivers can effectively ignore  amplitude variations at video rate, only the frequency modulation effect will be  used. To see how the design was developed we must first analyse the deficiencies  and merits of each method of driving Gunn effect devices and then utilise the  most suitable method in a way that allows home construction.

The goal is to present the video information, sound subcarrier and adjustable  DC supply voltage to the Gunn device simultaneously. This isn't too difficult to  achieve if the mixing of these signals can be done very close to the Gunn diode  and sufficient video and sound drive can be produced to feed the low impedance  at this point. In the real world, the Gunn diode may be located up a mast or  high on a wall where accessibility is poor and long cable runs would be needed.  Apart from the requirement for separate audio, video and DC feed cables, the  load impedance for each is quite different and matching components would be  necessary with the inherent losses they introduce.

Three options are open:
1. to run three cables to a modulator unit co-located with the Gunn  oscillator,
2. to mix the signals together at source and feed them to the oscillator via  a single cable,
3. to mix audio, video and a tuning signal together, fed them through a  single cable and add the DC supply at the oscillator end.

Option 1 is most expensive in terms of cabling, also losses in the impedance  matching networks require that considerable signal power is generated. Option 2  is least expensive to implement but the complex and varying impedance of the  Gunn diode makes matching to co-ax very difficult. The mismatch would show as  video ringing, ghosting and probably missing frequencies due to phase  cancellation in the standing waves along the cable. The missing frequencies  could well include the sound or colour subcarriers. Option 3 is technically most  complicated but requires only one co-ax cable to carry the signals and a  separate unscreened cable to carry the DC supply. The third option is the one  chosen in this design.
Circuit description:

Overall, the design occupies two small PCBs, one mounted at the "shack  end" of the cable which deals with the signal processing, the other is  located close to the Gunn head and provides the adjustable DC supply. The co-ax  cable is driven and loaded with 75 ohm impedances to correctly match the cable  and prevent signal distortion, even when long lengths are used. Changing the  values of R26 and R101 to 51 ohms will allow 50 ohm impedance co-ax to be used  instead but the supply current will be slightly increased.

Looking at the main PCB (95-0001) first; the circuitry around U1 is a two  stage audio amplifier giving sufficient gain to allow direct connection to a  microphone. It also provides high frequency pre-emphasis necessary to give an  overall flat response when received through a satellite receivers de-emphasis  circuits. The audio signal is then DC blocked by C6 and fed to a varicap diode  which has a steady 4 volt reverse bias to centralise its capacitance swing. As  the varicap changes capacitance it frequency modulates the subcarrier oscillator  formed from components around TR1. The oscillator and varicap supplies are  stabilised by D2 to prevent frequency drift if the power source fluctuates.  Video arriving at J3 is fed through a pre-emphasis network and 6MHz trap. The  trap slightly upsets the video phase response but with the values suggested its  phase shift passes through 0 degrees at almost exactly 4.4 MHz so colour  distortion is minimal. If desired the trap can be omitted by not fitting L2 and  C16.

The video and sound are mixed with the tuning voltage at pin 3 of U3 which is  configured with a gain of 2 to offset the halving of the signal in the co-ax  feed and load resistors R26 and R101. In order to reduce current flow through  the co-ax and hence its load resistor, the tuning is achieved by sitting the  combined sound and vision signals on a DC offset of between +1 and -1 volts. If  adjusted correctly the offset should be 0v and no load current will flow. To  generate the negative voltage a modular DC-DC inverter is used (U2). This gives  +12 and -12 volt outputs which track each other fairly closely. By using the +12  output from U2 instead of the main power rail, any variation is supply is  balanced and has no effect on tuning voltage. Pin 3 of U3 is a high impedance  point and therefore offers little loading to the sound or vision signals and  permits isolating resistors R22 and R23 to be used.

C17 filters noise from the tuning potentiometer wiring and is returned to the  +12 rail so its switch-on charge momentarily raises the Gunn voltage, this has  been found necessary on some diodes to "kick start" them into  oscillation.

On the second PCB, the co-ax cable is matched into load resistor R101, if the  tuning is correctly set there will be no DC voltage across this resistor. D101  and R102 lift the signal from the cable so it sits 8.2 volts above its previous  level. TR101 is used as a current amplifier to drive the Gunn diode itself. As  its emitter voltage will sit about 0.7 volts below that at its base pin, the  Gunn should have 7.5 volts across it, nicely central in its operating range. By  adjusting the tuning control the voltage across R101 should swing approximately  +1 to -1 volts so the Gunn voltage will swing approximately 6.5 to 8.5 volts  which are about its safe limits. D102 and D103 prevent the Gunn voltage dropping  more than 6.8v below supply or 9.1v above ground should the tuning voltage  exceed safe limits. R103 and C104 appear to the Gunn diode as a 470 ohm shunt  which helps reduce their tendency to oscillate in undesirable modes. If the  wires to the Gunn module are longer than about 75cm (3") fit them at the  module instead of on the PCB. Finally, the relay RLY1 will only close and connect  the Gunn voltage if the supply voltage exceeds about 9.5 volts. This is a  protective measure as a low supply voltage can damage the Gunn diode if it allows current to fall below the devices negative resistance range.

Assembly:

All PCB holes are 0.8mm except the fuse clips which are 1.5mm, TR101 and the  L1 can legs which need 1mm holes.
All resistors, diodes and fixed inductors are mounted on 0.4" spaced  holes, 0.1uF capacitors and electrolytics are on 0.3" spacing and ceramic capacitors are on  0.2" spacing. Pre-forming the leads before assembly will greatly speed  construction.

Assembly order isn't critical but I suggest fitting the two fuse clips first  as these need their pins folding together on the track side, a job more easily  done before fragile components are fitted. Fit U2 and L1 last as these are the  tallest parts and are easily damaged during handling.

The only awkward soldering is around TR1, be careful not to short its pins  together.

Parts List:

Tuning control either 10 turn or single turn but MUST be 10K value and  preferably linear track. This part is not mounted on the PCBs. Use a type that  suits your preferred box or enclosure. A heatsink can be fitted to TR101 as it  runs quite warm. Fold a 30mm x 15mm aluminium strip at 90 degrees, half way  along its length. Then drill a 3mm hole in the centre of one of the 'wings', 5  mm from one side. Provision has been made on the PCB for the heatsink to sit  with its folded side flush with the component side of the board.

The assembled indoor unit in a small metal case.

Click HERE for downloadable PCB files.