1979-1980 2 Meter Transceiver Refurbishment

(See the new EPROM pcb rationale)

 Now that the equipment seems to have a half decent chance of working now that I've investigated its circuitry etc I'll start work with the most important element. I'd noticed the VXO tuning mechanism was in rather poor condition with lots of rust so I began by detaching enough parts to overhaul it. The tuning assembly was made from two main parts viz. a dual gang tuning capacitor with what looks like a 500pF and a 450pF gang carrying a built-in geared slow motion drive and secondly a plastic scale probably taken from an old portable radio.
 

 

 Above, you can see the VXO assembly detached (with difficulty) from the chassis. I'd intended to remove the flywheel around which wraps a dial cord which passes around the tuning spigot then around a pair of small pulleys which are part of the dial shown on the top right. The flywheel was screwed longitudinally to the end of the tuning capacitor spindle together with a pair of securing screws... belt and braces.. but despite removing the three screws I couldn't remove the flywheel. The outer surface was completely rusted and resulted in too much friction allowing the dial cord to tangle and jam the tuning action and I had to clean off as much as possible using emery cloth. The spigot on which the tuning knob fits is a combination of steel and brass, the former of which was also rusty causing the dial cord to jam. The dial cord is ordinary string and in poor condition so this needs replacing. Also, looking at the position of the spigot, I think operation might be more reliable if this was located slightly further from the diecast box, but this modification was impossible due to the geometry.

 

 

 Above, with the new dial cord in place. A puzzle is the dial pointer position as the total string movement of the horizontal part is about 260mm but the width of the dial aperture is only 80mm. I don't recall a problem when it was all working so the tuning linearity must be limited to only a partial rotation of the tuning condenser. Presumably the settings must be such that +/- 50KHz shift takes place over about one third of the total rotation? However, all did become clear as I progressed so read on... In fact the condition of the tuning arrangement is pretty poor and I may have to design something new to replace it. One interesting method might be to do away with the mechanical setup, measure the frequency of the VXO and use an analogue to digital converter to display say three significant figures to be added to the 100KHz channel. For example 10.240MHz = "000" and 10.241MHz = "250". This might make a challenging addition to the project?

Why not just power the VXO and see if it works? So, I applied 14 volts and progressively increased the current until things started to come alive. With a limit of 600mA and the rig drawing about 495mA, a relay or two could be heard to have clicked and the centre display showed a red 7 with the 144MHz LED lit (=144.7MHz). The lock LED was a solid red and a 10MHz signal appeared on the spectrum analyser. Twiddling the tuning altered the frequency somewhat but it was best to use a shortwave monitor receiver (my Kenwood R2000 in CW mode), to hear the exact frequency change as the VXO knob was turned. The range turned out to be 10.2464 to 10.2318, a shift of 14.6KHz representing +6.4KHz to -8.2KHz with respect to 10.240MHz. Surely not enough for the required swing of +/-50KHz.. or is it.. read on.

 

 I imagine the thing tuned OK back in 1980 so how does the shift of something like +/-6KHz or so at the reference oscillator shift the VHF VCO by at least 50KHz? A spot of electronic magic maybe?

The phase lock loop circuit works by comparing two frequencies and locking one to the other. The basic reference frequency, as far as the loop is concerned, is fixed at 10.24MHz and the VHF VCO is synchronised to this by altering the oscillator tuning via a varactor diode fed from a suitable filter. The process is accomplished by dividing the reference oscillator frequency by exactly 8,192 to produce 1250Hz; the VHF oscillator is then adjusted so that its divided-down frequency is made precisely the same viz. 1250Hz. Take for example a 2m RF output of 145.000MHz which is arrived at by adding 10.7MHz to a VCO frequency of 134.300MHz. This latter is divided by 5 in the PLL module down to 26.86MHz. The divide figure to produce 1250Hz from 26.86MHz is 21,488.

If we wish to fudge the VCO to 134.350MHz (ie. shift it up by 50KHz so that it results in 145.050MHz or half way to the next higher channel frequency) the frequency in the PLL would need to be 26.87MHz and, with the multiplier unchanged at 21,488, the reference will need to be about 1250.465 making the reference oscillator 8,192 times this at 10.243812MHz which is 3.8KHz higher than nominal.

That means the VXO tuning dial markings of +/-50KHz actually represent a shift of a mere +/- 3.8KHz, so the recorded measurement of +6KHz to -8KHz is fine and is probably only slightly different from the setting back in 1980. The requirement of a total of 7.6KHz explains the size of the front panel tuning aperture of less than half of the tuning cord movement. The frequency coverage through the dial aperture was arranged to conform with a fairly linear section of the VXO tuning, probably by careful selection of fixed capacitors and coils.

I found the circuit diagram and pcb layouts below which seem to line up with the printed circuit board adjacent to the dual gang tuning condenser.

 

 

 The pcb layout below has only the FET section of this circuit so I guess the transistor amplifier was an early afterthought because it's fitted to the final pcb. Looking at the final design below I must have made a breadboard layout before etching the pcb. Possibly the FET output needed buffering to improve oscillator stability as the extra stage is an emitter follower not a voltage amplifier?

Note the wiring of the dual gang tuning condenser. The larger 500pF section must be "large C" and the smaller 450pF section "small C". Presumably I found this circuit in a 1979 magazine?

  This seems to be the only surviving pcb layout in my records. Not quite up to formal Plessey Drawing Office standards (the company for whom I worked at the time).Top right (pcb upper surface) is the 10.245MHz crystal.

The power supply connection was measured at 11 volts (so the rather odd choice of regulators are working as they should) and the RF output goes off under the chassis to the PLL module (the next picture- below).

 

 
 

 

 Above, marked "G3PIY Dec1979" is the phase lock loop pcb. The wiring gets in the way and I should really tidy this up as it makes access awkward. Of interest is the transistor with a heatsink (top right) which seems to be fed by two external DC or LF circuits. In fact I later found several pcb's with similar TO5 devices which are 6 volt regulators. The loop filter (bottom right) is interesting. There are four pots and a couple of germanium diodes. I recall I had trouble locking the circuit and I'm pretty sure I needed twin bridge T filters to eliminate 1250Hz modulation from both transmit and receive. I traced the coax lead (lower edge-right of centre) to a front panel pot and so that collection of resistors, capacitors and diodes seems to be for narrow band FM. I must have decided that generating FM in the PLL was the best solution. The small pot must provide a limit to the modulation and the front panel control the amount of deviation. It'll be interesting to hear the effect of this technique as the PLL will be fighting to control the VCO shift and remain within lock range..

 

 2m RF Output PA

 

 This module fits into an aluminium case and seems to be the 2m linear output amplifier. Interestingly it includes a notch filter marked 135MHz (the coil and tubular trimmer top, left of centre) so my design must have had too much VCO content. As that frequency is in the aircraft band it's definitely a bad idea to allow this to leak through the system. The mult-turn pot must be for adjusting the transistor bias to optimise linearity?

 

 

 

 Having got the VXO working I decided to tackle the 10.7MHz crystal oscillator (extract from the one of the original circuits shown here), then the PLL with the VCO next. That TO5 device just visible below is a 6 volt regulator.

 

  The 10.7MHz signal popped up readily and twiddling the beehive trimmer next to the crystal proved it could be set to precisely 10.7MHz. Ignore the RF level because I was using an uncalibrated lossy probe.

As there are lots of loose leads visible maybe I should trace their ends, label them and solder them back in place before checking the PLL. Also I should consider rewiring some of the areas as existing wiring is a tangle and obscuring access to the various circuit boards.

Another option is maybe to remove each module and test it independently?

 
 

 The first check on the VCO showed it wasn't working as all I could see around 135MHz was a low-level signal at around 120MHz, probably the 11th harmonic of the VXO as the SL680 10.7MHz crystal oscillator is reckoned to have minimal harmonic content (better than -30dB at 20MHz). If I'm successful in this venture, perhaps I should add a low pass filter at the output of the VXO in case the air band is at risk (we're very close to Bournemouth Airport)?

I suspect a missing DC supply connection is preventing the VCO from working? Time to sort out all the loose wires. Sure enough, I checked the power supply connections to all the modules and at the same time tried different settings of the mode switch (I assume to be SSB, AM, CW, FM and FM-Repeater) and found no power was being supplied to the VCO. I found a white wire from a terminal, to which the power lead from the VCO connected, went off to one of two large relays (TX/RX?). Also connected was a long, dangling thin red wire. I suspect I'm following the path I must have taken back in 1980 to find a fault in the rig. For some reason, perhaps pressure of (real) work had halted progress?

I connected the thin red wire to my power supply and carefully cranked up the voltage whilst monitoring the current. I stopped at 11 volts and noticed 22mA drain so I turned on my spectrum analyser, hooked up my probe, and immediately spotted a VHF signal and its smaller second harmonic. Tuning the SA to 135MHz I found the VCO output very gently drifting HF. Of course the PLL is unpowered so I'm looking at only warm-up drift.

Surprisingly the frequency is just about right where it should be for 146MHz RF output (actually 135.313+10.7=146.013MHz). The signal below was measured at the buffered output of the VCO. Sometime later I'll calibrate my probe using a signal generator so I can judge the true RF levels.

 

 So far so good as all the internal frequencies have been checked and seem OK. What remains is to untangle and fix broken connections to a host of things. I've now traced various ends and added fresh labels, in place of the ones that have completely faded, and I know where most should be wired. Next I'll see if the PLL is working. I wasn't too sure how to check the PLL so twiddled the diode selection switches to see what happened. The numeric display hasn't been debugged yet but I did resolder a couple of wires around the rear of the front panel. As I rotated the inner of the two channel switches wired to the diode matrix I noticed the lock LED flashed at a deceasing rate, but at one point it stopped and the LED went out so my guess is this meant the PLL was locked. I connected an unmodulated test signal at 145.1MHz into the receive mixer RF input and the spectrum analyser to the IF output of the module then checked for signals. I could see the 135MHz VCO signal when I monitored the buffered output of the VCO and at the IF output of the Rx mixer I saw the pictures below. The first has a scan width of 50KHz and the second 10KHz.

 

 

 As I've had experience with this transceiver I can interpret what's happening. The peaks are in fact 1250Hz apart and represent modulation from the audio frequency used for comparing the reference crystal and the VCO setting. If the audio section of the receiver was working you'd hear a tone of 1250Hz. Back in 1980 the receiver output was acceptable as was the transmitted signal so the problem is now one of sorting out a fault. The next step will be to check the PLL filter (below).

 

 The varactor diode used for shifting the frequency of the VCO should be driven by what is essentially a DC signal but in this case the filter isn't working. But nevertheless you can see the PLL is working. At rest it's 135.3MHz.

Above (approx):145.1MHz-134.4MHz=10.7MHz.

I varied the signal generator output and the amplitude decreased in line with a reduction but didn't increase much beyond what you see in the scans. Lock (as evidenced by the Lock LED) was tricky so the modulation in the loop was creating a problem.

 

It was at this point I found the display was bad and probably because of its vintage the only replacements were horrendously expensive so because of this and the cumbersome diode matrix arragements I decided to change the design to something better.

Read about swapping the diode matrix for a custom EPROM board.

 

 

 Some progress... the new wiring from the PLL EPROM to the 144/145/146MHz LEDs in place of a mess of transistors and diodes. A new 7-segment display driven from a second EPROM now tested and working.

The new display is a dual-in-line type and I used a 16-pin socket with a plug to aid changing it if necessary.

 

 

 A view of the new EPROM pcb wired to the channel selection switch. The top EPROM handles 20 x 100KHz 2m channels complete with 600KHz repeater swtching with the lower one driving the 7-segment display.

The toggle switch replaces a second 12-way switch by activating an extra address bit in the EPROM.

 pending

 See a description of the 2 Meter Transceiver

 Return to Reception