Restoration of a WW2 Receiver, R1132A Certainly looking it's age!

If all goes well I'll start by making good any deficiencies and replace any bad components, then get it working Step two, I'll dismantle the main mechanical parts and refurbish them, then reassemble it. Finally, I'll test the receiver and see how its performance compares with its specification.

Above is my second R1132A. I bought this one in October 2014 and after removing it's outer case discovered someone had swiped most of the valves. There should be either 11 or 12 according to the circuit diagram, but alas there were now only two. These were hidden away under a metal shield on the right and were two examples of the ubiquitous VR65 or Mazda SP61.

I have a very large collection of valves and I hoped I'd be able to find enough in my various boxes to fill the empty sockets. First I had to check the numbers and see what the commercial equivalents were. Although I'd got a previous model of the R1132A working in the past, it had come with all its valves and I hadn't needed to check what types these were. I was more than a little surprised to discover the audio amplifier was an octode. After some rummaging around I found a suitable valve. This is a 6A8G. I'd initially found a small metal cased 6A8 which was rather rusty and as the top cap lead needed bending a lot I carried on till I found the "G" version. After an hour or so I'd found sufficient valves to get the receiver fitted up. One valve was a problem. This was a P61, a Mazda triode (coded by the Air Ministry as a VR66) from the same stable as the SP61 but having its grid connected to the base rather than the top cap. I did find a triode which I must have had for ages, however, was it a P61? It had no metalling, just plain glass and no number visible and, thinking very hard, I reckon I'd experimented with it back in the 1950s when I'd been interested in receiving VHF transmissions using super-regen receivers. I suspect it was from the first set I'd ever built and to make the valve work at the highest possible frequency I'd removed the base. Later I used a twin-horned triode and then a 955 and the P61 had been relegated to the junk box, miraculously surviving 60 years. I could have borrowed the P61 from my other R1132A or even bought one on Ebay, but it seemed a good idea to restore this valve. The leads were just about intact although pretty heavily tarnished and too short to use without extending them.. First I had to sacrifice an SP61. I found a tatty looking example with a wobbly base taped in place and unsoldered its pins. This wasn't difficult. First I added a blob of solder to the end of each pin to act as a heat reservoir then gently pulled the glass envelope to release pairs of heated pins from their grip on the wires. Little by little the pins freed until the base came off. After cleaning up the pins and checking that there was no solder obstructing the holes I turned my attention to the old P61. It's leads needed extending so I carefully cleaned off tarnishing and soldered short lengths of 24SWG wire to their ends after pushing short pieces of plastic insulation down them to prevent shorting. I checked my valve book to see the connections. MO16 was listed and, after orientating the glass envelope carefully, located the wires connecting the electrodes in their respective pins. The glass fitted nice and snugly and, after soldering the wires to the pin ends, snipped off the excess. A spot of superglue to secure the new base and I was ready to test the valve (and to confirm it was indeed a P61).

See the steps in restoring the P61 below

Not exactly a common base, this is a Mazda Octal and is different to the International Octal, having the heater pins (1 & 8) wider apart than the IO base. This example was taken from an old SP61 and solder inside the pins has been removed ready for the extented wires on the bulb to be inserted.

The base fitted; ensuring the glass bulb is orientated correctly. Later I added more superglue around the base and added aluminium foil. I set up the AVO valve tester dials and checked the restored valve. It worked. It is a P61 and emission is pretty good.

While the valve tester was fired up I checked the EF39s I'd lined up for the receiver. All were fine. Note that the R1132 uses EF39 valves which are variable mu types, somewhat similar to the 6K7 but not identical. They are not interchangeable with non-variable mu types, although one could use the latter in positions not connected to the AGC line (such as the BFO).

Above you can see the IF amplifier and audio stages. The four IF transformers have chunky tuning slugs best twiddled (carefully) with a metal screwdriver having a blade width of around 3mm. There wasn't much evidence of hand capacity effects. The primary aim is to see if the receiver is restorable without too much effort, so I soldered three wires to the Jones Plug. HT, LT and ground and connected these to a power supply. I set the HT to the recommended 210 volts and checked the current consumption. It was not far from the 50mA suggested in the manual. I plugged in a pair of headphones to the monitor socket. Nothing, completely dead so back to the circuit diagram. I located the output transformer and found a length of TV coax cable was soldered to its output. Connecting the phones again produced nothing so I checked the HT connection to the transformer primary. This read 199 volts but the 6J5 anode read only 0.6 volts... odd. I turned off the power and checked the transformer primary. Open circuit: explaining no doubt why the set had been ditched and its valves removed. Why had this happened? Often a decoupling capacitor at the output valve anode is responsible, but in this set the anode connects directly to the transformer and nothing else so it'll have to remain a mystery. Turning to my transformer junk box I selected a small low voltage mains transformer (with three secondary windings) and wired it in place of the old output transformer. It'll do to prove the receiver works at least. I connected a small loudspeaker to one of the low voltage windings and turned on the power. I heard a faint hum from the speaker... so far good. I connected a coax cable from my signal generator to the new PL259 aerial socket (the previous owner had removed the wartime coax socket so I found a panel-mounting "UHF" socket which fitted perfectly in the original hole). There was a healthy crackle and I hadn't even turned on the generator output... very promising... I tuned the R1132A dial to mid-way, turned on the generator RF output set to 100mV having selected 110MHz then stepped up the frequency little by little. Suddenly the loudspeaker burst into life.. The set was actually receiving a 120MHz AM signal modulated by a 1KHz tone. I twiddled the dial and discovered the set tuned 106 to 130MHz. The coils had seemed intact but were pulled apart explaining the shift of 6MHz from the original spec of 100 to 124MHz. Reducing the generator output revealed the set could just about hear 20uVolts, fading out completely at 10uVolts. What about the image? The IF is said to be 12MHz so tuning the set to 120MHz -24MHz =96MHz should bring in the image. With the dial set to receive 120MHz I set the generator to 96MHz and could only hear this when the level was 10mVolts. So, the set is generally working. Sensitivity needs to be improved and my guess is this can be put right by looking at the AGC and gain control circuits. Currently the pre-set gain control has been disconnected and who knows how the grids and cathodes are wired. There is only a tiny change in the S-meter reading if the input is cranked up to hundreds of milli-volts. The AGC circuit is one of most complicated I've ever seen so probably there's a problem somewhere, inadvertently put in by the last experimenter... Before I started I decided to regularise the output transformer modification and removed the original with the open circuit primary and substituted the replacement which was roughly the same size and shape. I did a quick check with a wattmeter and this seemed to indicate that the best of the three output winding measured around 100 ohms. My junk box speaker was marked 25 ohms, however I need to carry out further tests once other problems are sorted out. I started to investigate poor gain by checking the IF stages and quickly found a screen resistor marked 68Kohm that was difficult to measure, but I finally decided it was more than a megohm. I fitted a new 68Kohm resistor and the screen voltage increased to about a hundred volts or so and the overall gain stepped up a notch. After another half an hour I'd found that adding a couple of decoupling capacitors added a further notch of gain. At this point I must admit to have mis-remembered the IF value and set the signal generator from its setting of 120HMz to 6MHz. I noticed that the best response from the loudspeaker was about 5.9MHz so started setting the slugs in the transformers to set this response to 6MHz. Gain started to improve quite dramatically and eventually I found the IF strip would burst into oscillation as slugs were peaked. This receiver uses lots of metal plates and special decoupling arrangements to minimise feedback, but in my example is let down by valves with poor metalizing. In fact the last IF stage uses a red EF39 with a loose base and clearly a break between its metal coating and pin 1 of its base, so I unplugged the valve and temporarily wrapped stranded wire in the gap which had developed between the glass and bakelite base, connecting the bare wire to pin 1. This immediately reduced the unstable gain and killed the oscillation. I then decided to move back to tests using VHF input frequencies. The set was noticeably much more sensititive and was responsive to inputs of only 5uVolts. Previously an input of 20uVolts produced nothing and around 50uVolts could be heard just comfortably. I was now able to check the action of the beehive trimmers and discovered they were able to be peaked on incoming test signals. A short length of wire enabled me to hear very weak broadcasts from Bournemouth Airport, but overall the set is far too unresponsive and audio is too low.

Above is the VHF front end as it was before I started restoration. Somebody has tweaked the tuning condensed vanes to produce less swing than the designers' intended and pulled apart the coils to increase the maximum received frequency. Below shows the new aerial socket and P61 plus the oscillator voltage stabiliser.

The next step was to readjust the IF transformers using the correct input of 12MHz. I noticed that the set responded to 10uVolts injected at the top cap of the SP61 mixer whereas previously it needed 10mvolts to get the same response, this being an indication of the level of the 2nd harmonic from my Wavetek signal generator. This time I tuned the first IF transformer which I'd initially omitted to do as it was awkward to reach. Both windings were off tune and resetting them resulted in the set responding to an input of only 2uVolts at 125MHz. As the audio level from my speaker is very low unless I crank up the input to 500uVolts I need to determine what's wrong. Also the signal strength meter is pretty unresponsive until very large inputs are applied. This points to the AGC line and possibly a poor choice of audio output transformer. As a start I measured the resistors around the EB34 valve and found very large discrepancies between the marked and actual values viz. 6A8 anode load 100kohm read 169kohm and following 6J5 grid leak 330kohm read 550kohm with 6A8 anode at 185volts. 6A8 screen resistors 68kohm and 100kohm read 79kohm and 1.2Mohm with the screen voltage at 31 volts. At the EB34 various 330kohm read 460kohm, 800kohm, 379kohm and 1.05Mohm and a 220kohm read 1.5Mohm. I haven't yet checked all the various decoupling capacitors, many of which may be leaky. Some are waxed types and others are mounted in metal cans fixed to the chassis.

Here's a collection of components removed from the set. Often you'll find one specific value resistor is bad; in this case 330kohm and a couple of 68kohm, a 100kohm and a 220kohm. Old WW2 decoupling condensers are always leaky although it isn't necessary to swap them all unless they're very bad or in a critical circuit position, such as in an AGC line where impedances are vey high and leaky condensers will severely degrade the AGC performance.

Long ago the control adjacent to the RF gain knob has been disconnected and various wires left floating. I've also discovered that twiddling the BFO tuning knob dramatically affects overall sensitivity which is a bit odd, and I haven't found the tiny diode valve in the AGC circuit. This is supposed to be an EA52. Was it removed or perhaps a later design change? After another session with the old receiver, I changed the audio output transformer from the 12 volt mains type to one from my junk box which looked better as it had a primary resistance of something like 1000 ohms. It gave me more output and I'm now using a standard speaker. To correct the deafness I changed lots of resistors and capacitors. Each change improved the audio output by just a little bit but all told the audio is now pretty good and loads of air traffic signals are coming in with a length of wire as an aerial. Below you can see the replacement audio output transformer next to the audio filter choke. It's carrying temporary wires to a loudspeaker and wattmeter (the latter is what I use for peaking the output). New parts are fitted temporarily.

Because this set has loads of gain the designers' have taken a lot of trouble to provide common earthing points at each valve to prevent unwanted feedback, hence the strange 5-pin solder tag assemblies located above the valves.

The next step will be to check the IF response as now that signals are reasonable I can detect some humpiness when tuning across them. At the low end of the band I can hear extremely strong (and very distorted) FM broadcast stations and there seems to be a lot of spurious signals across the whole band including what sounds like an analogue video signal somewhere around 120MHz. Sometime I need to add some aluminium foil around the repaired P61 because it has no metalising. As far as sensitivity is concerned I can now comfortably tune a 1.5microvolt signal. The S-meter is a 5mA milliammeter monitoring the anode current of V5, a variable-mu EF39. The AGC applies a negative bias to its grid which reduces the anode current. The design caters for really strong signals viz. No signal =4.7mA; 20uV =4.5mA; 100uV =3.4mA; 1mV =2.9mA; 10mV = 2mA and 100mV = 1.4mA. I'll check the actual readings next time I turn on the receiver. As the output circuitry has all been disconnected I'd like to re-establish this but, because the replacement audio output transformer has only a single output winding I'll either need to add a virtual earth centre-tap or leave the output floating. I'll also check the BFO which I think is working OK. I re-established the output circuit but the existing low frequency choke attenuated the audio so much I shorted it out. Now the attenuator works OK and the audio can reach 200mW. The disconnected preset gain control was really redundant so I connected it up as a tone control. After looking fruitlessly for V12 I suspect this example of receiver doesn't use such a valve and never has. I suspect the diodes in the EB34 are used an AM detector and AGC rectifier and V12 maybe was a modification wired as a noise limiter? Basically, I'm now happy with the performance of the R1132A. It's not rebuilt as original but it now makes a useful air-band receiver. I guess the next step is to tidy up the wiring and then detach the front panel, repaint it and add new markings for the controls.