Linear PSU failing with overvoltage

According to timestamps on components, I have owned this linear power supply for 30 years (rms K135). It powered my first CB station (Midland Alan 48) and who remembers how many other devices. Lately it has been powering a 12V LED strip in the shack-lab.

One of these days I wanted to take a picture of a multimeter (will be feature on the blog, don't worry) and I needed both something to measure and less light in the shack. So I unplugged the LED strip and wired the PSU to the multimeter.

To my great surprise I read 24V. What?! That would have burned the LED strip and everything else I had connected in the meantime. But, with the LED strip powered, it behaved as expected and outputted 12V or so. Of course I cross-checked the reading with several DVMs and they all agreeed.

A bad capacitor? I opened up the PSU and I was greeted by a large amount of dust. And two 78S12 in parallel, as I remembered. Hard to remove the dust, but once components were in the clear, I still had 24V out without a proper load.

Capacitors looked OK. No leaks, no deformation. Unwillingly I reached for the solder side of the PCB and disconnected one 78S12 at a time (just the output pin, lazy me). One regulated at 12.5V, the second at 24V. Well, it didn't regulate at all then!

I temporarily replaced the bad one with a 7812 and the output is still 12.8V.

So, watch out for this simple technology too! Until today I would test the Amps rating of a PSU. Now I will check the open circuit voltage!