Hello.
I've managed to get the board out of the rack so that I can work on it more easily on the bench. As I've stated, for future reference, in my version of the rack there are more parts to remove and screws than stated in the service manual that's available online: take your time and study the situation.
I've been reading about switching PSU topologies and considering the high-current output (35A at 5V) I think I am trying to repair a flyback. Besides the custom inductors, the circuit has something that looks like a custom-made IC as Tek used to do (in my experience with them) even if the power supply came from a different supplier. That component is the controller IC as there isn't a chip on the AC side. Under the larger transformer there is a smaller one that is probably powering the control circuitry.
I have located an interesting document from Texas Instruments "Common Mistakes in Flyback Power Supplies and How to Fix Them" that guides through troubleshooting a unit and includes how's, why's and waveforms. As long as I'm learning new stuff, I'm fine. Here's a slideshow for those of us that don't need full-text explaination:
https://www.ti.com/seclit/ml/slup398/slup398.pdfIn order to troubleshoot the AC-side circuit I'll need to "lift from earth" either the PSU or the oscilloscope, keep one hand in the pocket and adopt all high-voltage safety procedures. OTOH for the output side I can operate normally without the insulating transformer.
Update. Doubtful that the PSU would need to see a load to regulate (fans), I added a resistor on the +12V and an LED on the +5V output. No changes, but while poking around with the DVM I noticed that 10 minutes after power up the LED started flashing. That could mean only one thing: heat makes it work. I confirmed my theory beaming a hairdyer to the circuit and sure enough, after a bit of "coughing", I got a steady output that vanished when temperature returned towards 20°C ambient. I'm not sure how warm the board got, probably in the range of 40°C.
I think I am after either a cold/unstable solder joint (unlikely as it would react to physical "stress" of the board), or a component that fails at room and lower temperatures. I'm not sure tantalum capacitors do this, but electrolytics for sure.
Paolo