A few years ago I was given a PDP 11/24, followed a few months later by the cabinet and two RL02 drives. Not long after, once I had first tried the PSU with a dummy load, I attempted to power it up with just the CPU (M7133) and Unibus Map Module (M7134) installed, no memory. It worked right away and I got a console prompt. Sadly though, after entering a couple of commands on the console, there was a slight click and the machine powered itself off.

The H7140 PSU had failed, and a long saga began. This is a very complex PSU (for me at least). I had a lot of help from the ClassicCmp community. It turned out that the Bias and Interface board was not supplying 5V to the control logic. One of the first things I did was to replace the 555 timer, which had failed, when I powered it on again there was a bang and more stuff broke. I made various attempts to fix it over the years, but then put it to one side.

Recently I was told of a company, not far from me, that repairs power supplies, including vintage stuff. So, having realised that this was a repair that was probably beyond me, I decided to take the plunge. The PSU finally came back to me a couple of weeks ago, unfortunately having suffered some damage during shipping, partly because I had not put back the covers before sending it off.

During previous attempts at repairing the PSU I have had it partially working and I noted a smell from somewhere in the CPU enclosure when it was powered up. Inspecting the boards I noticed a rather mottled looking capacitor on the M8722 memory board:

I replaced this and the one next to it with 40V rated ones as I couldn’t find axial ones rated at 20V or 30V (they seemed to be marked with both voltages). I also replaced the air filter at the front of the CPU enclosure, as it was crumbling.

I reassembled the machine with the M7133, M7134 and M8722 modules and with the newly repaired PSU. When I powered it up the fans turned but no LEDs lit up on any of the modules. The DC ON light was flashing, suggesting the voltages weren’t right. Furthermore a red LED on the M8722 was on and the burning smell came back. I powered it all off and traced the smell to the M8722, it seems to be coming from some other capacitors rated 8uF, they all measured about 5.5 on the ESR meter except one that measured 0.8 (the middle one in the picture below). My guess is that the odd one out has failed short and is pulling down the DC voltage, causing DC ON to flash and the module to show the red LED.

With the M8722 removed the smell goes away and the DC ON light is steady. So the M8722 is faulty. However, none of the LEDs on the M7133 light up either, which suggests the M7133 is not happy too.

A few days later, I decided to connect a terminal to the Serial Line Unit (SLU) to see if there is any life in the CPU still. While doing so I fixed a few problems with incorrect screws, so that it all re-assembled properly with the proper covers fixed correctly. However, the power supply then failed to switch on at all. I didn’t hear the relay click when I closed the circuit breaker. This also happened when I was first trying to repair it a few years ago. I think there may be a dry joint in the input assembly, and perhaps the movement from when I tried to sort out all the fixings affected the bad joint.

So I have now sent the power supply off once more to the company that repaired it, to see if they can find the problem in the input assembly.