I previously posted about the failure of the H7842 power supply in my DEC Rainbow. I reported that after replacing some parts I was getting an overcurrent detection in the -12V output. This was under bench test conditions using a bench PSU to power the startup circuit.

I did a lot of tests, comparing a working spare with the failed PSU, but I just could not find a reason for the overcurrent detection to be wrong. Strangely, that problem disappeared after a time. It has been suggested to me that the test conditions were not really representative enough of real conditions to be reliable, and that even an increase in temperature with the arrival of spring and summer might have been a factor. I decided to ignore the problem for now as it has gone away.

At this point I could try once more to see if the PSU would now work. As a precaution I used two 100W filament bulbs in series with the mains input as a current limiter, to make sure that things wouldn’t go too badly wrong if it didn’t work. I attached a dummy load to the 5V and 12V outputs. In a working PSU this causes the 100W bulbs to pulsate and never to go very bright. However in my bad PSU they shone much more brightly, there was no pulsing and the DC outputs stayed at zero.

I checked the supply to the startup 7812 regulator. It was only about 3.5V, so there was no way that the PWM was going to work or control the chopper transistor. In the same conditions this voltage was much higher on the working spare. I wondered if perhaps the startup transformer had some shorts. This seemed to be borne out after I lifted the diodes on the outputs of the startup transformer, because they tested fine.

However, on receiving further advice, I looked at the mains bridge rectifier. On removing it I found that it had two pins shorted together. When I tested the startup transformer with the rectifier removed and also all the diodes on the startup transformer output removed, the startup transformer output was 24VAC. So the startup transformer was OK. I think the rectifier must have been shorted when the main chopper transistor blew up.

I replaced the mains bridge rectifier with a new 600V part and the power supply started to work. I tested it first with the 100W bulbs to be sure, but when that went OK I took them out. With a dummy load the PSU had reasonable ripple and the voltages were correct.

I reassembled the machine and it ran perfectly. The picture at the top is of the Rainbow running once more on its original power supply.

I would like to thank the people who offered my so much advice, without whom I would never have got this fixed.