According to the serial number key on the Eddystone Classic British Receivers pages by Alan Clayton, it must have been built in 1960. It raises some eyebrows to see that Stratton & Co were using Rimlock valves as late as 1960, when they had long been obsoleted by Noval and 7-pin miniature valves. According to Alan's model directory this model has been brought to market in 1954 so it was a bit outdated by 1960.
The 840A comes in a steel cabinet with a die-cast aluminium front. Cabinet and front have a grey hammered metal finish. This gives it a nice 1950's professional look. Housing an AC/DC radio in a metal cabinet is a bit dangerous, though. One of the many isolating washers or mounting strips may become conductive by deterioration or filth, or a piece of metal may get stuck in the wrong place and cause a short between chassis and case. When I first tried the radio in early 2004, I thought I noticed a slight tingling sensation when I touched the front, even though I powered it through an isolation transformer out of precaution. I didn't like it and decided I had to fit a mains cord with a safety ground connection before I would start to use the radio. And I suspected there was a lot more work on this radio. Its sensitivity wasn't quite what you might expect from a communication receiver. And when I checked the inside of the tuning assembly (the subframe holding the band switch, the tuning coils and trimming capacitors), I discovered a previous owner had sprayed a kind of oil inside, maybe to clean the waveband switch. Because I had other priorities than this radio, I left it on a shelf for over 2 years.
In April 2006, I rearranged my hobby shack and wanted to give the 840A a place where I could listen to it. So I had to get it in working order. I exchanged the power cord for a decent (but ugly modern) power cord with ground lead. I connected the ground lead to the case and the other two to the mains connector. When I turned the radio on, the house went dark as the ground fault circuit interruptor in the switch box tripped. This confirmed the wisdom of grounding the metal case. I went downstairs to turn the lights back on and set all the digital clocks in the house. Then I checked again. At first, my ohmmeter did not show a short circuit between case an chassis, but when I did a continuity test with a 500V transformer, I was able to confirm the short. The 500V test voltage seemed to have broken some weak isolation, because now the ohmmeter could also show the short.
This left me in a puzzle. How I could find the location where the case and the chassis were connected? The short could be in any of about 20 places and I didn't like the idea of finding it by trial and error. I decided to try to trace the current path by detecting the magnetic field caused by it. I took an old high impedance headphone, a "telegraphist's" headphone. I made a test circuit consisting of an AF generator, a 100 Ω resistor and some test wires. I set the generator to 12 V square wave at 1000Hz. Then I experimented a bit with some coils from my junk box, and found one suitable to detect the magnetic field around the wires. Of course, I could have done this with a scope instead of a headphone, and I would have gotten more sensitivity, but this way I felt I would be able to better distinguish real from false signals. Now I connected the two wires to the frame and case, respectively. I changed the connections a few times and decides the short should be somewhere near the volume pot. When I released the nut holding the volume pot, the short disappeared. So my trick worked and I had found the short.
So what was wrong? The potmeters are mounted on the front with a number of insulating and metal washers. On the volume pot, the washers were in the wrong order, resulting in a short between the pot and the metal of the front plate. It looked like an error made after a repair. I corrected this and the short was gone.
The radio wasn't very sensitive, and the stations were on the wrong position on the dial. So I decided to realign it. I suddenly realised that all the trimming capacitors were set to minimum capacity. This was a clue that somebody had been fiddling with them. I had just acquired a GM2883 RF generator, and welcomed the occasion to use it. First I aligned the IF stages. The GM2883 has a special 400-500 kHz range, allowing for a precise setting of the IF frequency. I found two different resonance frequencies with different amplitude, showing the IF transformers were a bit out of alignment. After peaking them to 450 kHz the sensitivity was slightly improved. Then I turned to the oscillator and tuning circuits in the tuning assembly. For all the 4 ranges I first aligned the coils and then the trimming capacitors as indicated by the service manual. For every range, I had to repeat the aligment twice. As you know, the coil and trimmer settings influence each other, and it is necessary to repeat the procedure to get a precise alignment. One has to remove the bottom plate of the tuning assembly to get to the cores of the coils. The trimming capacitors can be set through holes in the bottom plate. I made a final correction of the trimming capacitors to compensate for small changes of the wiring capacitances when the bottom plate was put back in place.
Rereading the service manual, I found that until now I hadn't made the right aerial connection. One of the aerial receptacles had been replaced by a previous owner. There should be a jumper connecting that receptacle and a ground receptacle to connect the input coil to ground, The jumper was missing, though. When I made the connection using a short wire with two plugs, the sensitivity immediately improved and the background noise was reduced. Fantastic!
Overall, my 840A is in good condition. The valves look nearly new and the case hardly has any scratches or rust. Some previous owner has damaged the speaker membrane a bit. But I'm very happy to have it, now it's receiving well again.