A Murphy 62B Overhaul

 Now I've started to work on my military radio collection (and any others sets delivered to my workshop), starting with the R1132, it may be useful to record the odd note on the various sets as I tackle them.
Published military documentation is jolly good and very comprehensive, if available, but doesn't take the place of first hand experience.
For a start, many of these early sets were incredibly heavy, and manipulating them on the workbench is not always easy.
In the case of the Murphy 62B, the weight of over 100 pounds is particularly tricky. It may not be quite as heavy as the AR88 but the shape is awkward and there are lots of bits poking out once you get started.

The set looked as if it had been stored at the back of a barn, giving it a generally scruffy appearance, not helped by the whisps of straw sticking out most of its orifices. The interior metalwork plating was covered with a light powdering of oxide but fortunately the set was not beyond redemption. In fact I vaguely recall buying this set on Ebay donkeys years ago and collecting it from a barn on a farm near Salisbury.

Powering it up using a variac indicated that all was relatively well with the power supply. I say relatively as the dial lamp kept going off and on. This turned out to be a loose fuse-holder in the 13-amp plug. Once it had been tightened the dial lamp stayed on but there was no sound from the internal speaker.

To find out what was amiss I plugged in a pair of headphones. I couldn't hear much, beyond a bit of a hum and the odd crackle when fiddling with switches. Connecting a signal generator, screwed full up, to the aerial input showed that the set had a 500KHz IF amplifier that seemed to be working tolerably well.
Why the set was not able to tune anything became clear when a scope was connected to the anode of the local oscillator: in this model an EF91.
Just in case the EF91 was very low in emission, I swapped it with another of the same type used as an RF amplifier but with no change…. the local oscillator wasn't oscillating on any band.
In the 62 set there's provision for crystal control. Oddly, although the marked positions were On and OFF, the switch had about 8 positions. At first I couldn't see why this was so, but later I worked out the switch had probably been replaced. In the OFF position the local oscillator should run but it wasn't.

I checked the resistors in the front-end and found several that were miles away from their marked values. Resistors aren't anything like as critical in valve circuits as their transistor counterparts, but after changing several of the worst offenders, I confidently switched on and was again rewarded with only sensitivity to a powerful 500KHz signal.

Capacitors can be notoriously bad in these old receivers as they often leak, and in parallel with impedances of meg-ohms can play havoc with valve biasing. I checked the capacitors around the oscillator valve but all read, as they should.

Suddenly, out of the corner of my eye, I spotted a signal appear on the oscilloscope. It was a big signal with tens of volts and twiddling the tuning knob showed the local oscillator at last to be oscillating. I removed the scope probe and connected the signal generator, ready to check the front-end alignment. No signals! I scratched my head before looking again with the scope. Nothing on the anode of the local oscillator. I prodded around without much inspiration and suddenly I heard a signal from the loudspeaker. I removed the probe and the signal went away. I checked the grid of the oscillator valve and there was the very large signal again. I looked on the anode and the signal was gone. There must be a short? I tried a different EF91 but to no avail. I disconnected the coupling capacitor from the oscillator anode to the frequency changer but still no signal on the anode. I checked the grid again and the signal was still there.

Puzzled, I connected two scope probes. One to the grid and one to the anode. Both had good solid signals on them but when I removed the probe from the grid the signal on the anode went away.
There was the answer, the few pF of loading from the probe was sufficient to start the oscillator.
I added a handy small 27pF capacitor and the local oscillator burst into life. Very odd, because the fixed capacitors in the circuit all measured within a few pF of what they should be.

Next I tried to figure out why I could hear only weak, very distorted sound in the headphones.

The 62B has a separate chassis at the bottom of the main casting, carrying the power supply and the audio amplifier. It's detached by removing the gain control knob, unplugging a couple of multi-way connectors then loosening a pair of large screws at the rear of the set. The chassis can then be slid backwards and removed.

I started by fitting a new coupling capacitor between the audio amplifier and the output valve. This is a must in all old valve sets as even the tiniest leak here will severely reduce the life of the output valve. Then I measured the various resistors… a tricky task in this receiver as nearly all the resistors had lost their markings and were just off-white cylinders. Using a circuit diagram and a component layout drawing I immediately found an open circuit 82kohm resistor feeding HT to the anode of the AF amplifier.
Once I'd cut this off and fitted a new resistor I confidently slid the chassis back into the set and plugged it back in.
Powering up the set then gave me lots of volume in the headphones and the speaker worked.

Next I decided to align the IF strip and the front end. Surprisingly, once the high frequency trimming with the trimmers was completed and the low frequency end with the cores, the dial markings were surprisingly accurate across each range.

Quite a few odds and ends now needed to be tidied up. The BFO wasn't working and the crystal switch is too confusing to leave as it is.
A thorough clean and some lubrication and adjustment of the mechanical components wouldn't come amiss.

A brief description of the set is worthwhile at this stage as it's pretty unusual. Being originally designed for the Royal Navy to replace the old CR100, the set is built like a battleship. I imagine that only a near hit by a 14" shell would shift its frequency. The front end is remarkably similar to that of my old R206, using a large drum, into which are slotted four wedge-shaped coils for each of its six ranges. Cogs that wouldn't look out of place in a Centurian tank gearbox turn the drum.
The weight of the receiver is about 100pounds, and unlike earlier versions, the B40 and B41, the 62B uses "modern" B7G and B9A valves.

The next day I decided to investigate the reason for the lack of a BFO.
The switch that presumably sets the BFO in action is marked "High", "Low", "Tune", "R/T" and "Cal".
The "CAL" setting seemed to mute the set but selecting the other options made hardly any difference.

I identified the BFO as being an EF93 located at the end of the IF strip. It was lit and there was evidence of volts at some of its pins, but alas no sign of oscillation.

Close by, a capacitor had been disconnected and a replacement fitted. Maybe this was evidence of previous fault-finding around the BFO?

I spotted a crystal marked "500KHz" close by. Maybe this was the BFO oscillator crystal? Then again the "Cal" setting might refer to a calibration oscillator?

I prodded around the EF93 with a voltmeter and noticed that when I switched to "Cal", the screen grid voltage disappeared and was replaced with some sort of pulsating signal. I switched the multi-meter to "Hz" and it read almost 2,500KHz. Odd, why 2,488KHz when there was a 500KHz crystal? Maybe it's running in 5th overtone mode? That would account for the discrepancy of 12KHz but surely the designers would have selected a crystal that worked precisely at 2.5MHz?

The amplitude of the 2,488KHz signal measured many tens of volts on the scope and it was a near perfect sine wave with a period that looked like 0.4uSecs.

I tuned the receiver and sure enough I heard strong signals every 2.5MHz. At least not exactly 2.5MHz, more like every 2.488MHz.

I looked at the circuit diagram, which was in pretty poor shape, being quite faded and hard to read… probably an early Xerox copy that had suffered from being left in the light?

After prodding around for a few minutes I decided to adjust the frequency to exactly 2.5MHz. On the adjacent tag-strip I found a 0.02uF capacitor to which the oscillation seemed a little sensitive and replaced it with 390pF. The oscillator frequency now read 2.520MHz and after adding in parallel a couple of handy 220pF capacitors wired in series the Racal counter read almost exactly 2.5MHz.
At least the calibration oscillator was now working OK I thought.

Next I decided to investigate the reason for the lack of a BFO. There were several coils in the vicinity and I twiddled each in turn but with no success as no signs of oscillation appeared in any of the BFO settings. Wait a moment though. I'd left the switch in the "Cal" position and as I twiddled the core in the transformer can next to the end of the strip the receiver heterodyne wailed. What's this? Maybe the fifth overtone coil is in the can that I thought was the final IF amplifier coil?

I judiciously twiddled and realised that I could set the calibration oscillator to exactly 2.5MHz with the original 0.02uF capacitor in place. A bigger capacity means a more stable oscillator and I had noticed that the fifth overtone seemed to take ages to settle down.

I set the core so that the oscillator worked on precisely 2,500KHz. Good, I thought, I'm pleased that the calibration circuit is OK.

Now the BFO. Why isn't it oscillating?

I measured all the resistors in the vicinity but, although some where miles out I couldn't really see why the EF93 wasn't oscillating, after all, the anode and screen voltages seemed to be reasonable. The cathode moved from 3 volts in the "Cal" setting to 0.3volts in the BFO settings. Clearly no oscillation. I tried a different valve but that was the same.

What about the various capacitors in the BFO circuit?

Some were inside the oscillator coil can. That would be a pig to remove. I started at the end of the tag strip. First a 100ohm resistor. I hadn't spotted that before.. but it measured OK. Next a couple of capacitors. They were OK as well. Wait a minute though. What's this capacitor sticking up in the air? I could see a new capacitor soldered to a brown wire, but when I looked under a metal can carrying two more coils I spotted a second unsoldered capacitor. I checked it with my meter. It looked OK and no leakage so I removed the replacement and soldered the brown wire to it. Now what about the first disconnected capacitor? The one with an end not soldered to anything?
I examined the faded circuit diagram and worked out it was probably one that seemed to connect between two of the switch wafers. I looked further and found it was actually a cathode decoupling capacitor, drawn in a confusing way.

After a few moments I found the point to where it probably had been originally soldered and soldered it back in position.

I noticed out of the corner of my eye the Racal counter was showing 500.001KHz. Odd, why was that I wondered?

Then the penny dropped. The calibration oscillator was indeed 500KHz not 2.5MHz. Now the slug in the penultimate can had absolutely no effect on the frequency of the crystal oscillator. It was rock steady, as one might have supposed a crystal oscillator should be.

I switched from the "Cal" to the BFO settings and was rewarded by a tone in each setting. The "Tune", "High" and "Low" settings corresponded to about zero beat, plus or minus about 1KHz from 500KHz, the nominal BFO frequency.

I reset the penultimate coil-tuning slug to its proper setting and was rewarded by a sharp increase in signals.

Of the set's three oscillators none had been working.

This exercise was carried out in 2015 and continued in 2020 here (Click)

return to entrance