Thanks for your lightning fast response! Let me share some more findings as of now.
Some things I forgot to mention in my first post:
- This machine indeed has hundreds of tantalum caps inside. To give you an idea, apart from the PSU, I found only two electrolytics on the other boards. (One in front-panel, rated 350V, that seem to be part of a dc-dc converter for powering VFD anode, and another one on a side board)
- As of now, I have visually inspected every board, looking for physical damage on components. Inspected every single cap (looking for burn marks or any leakage) but found nothing as of now.
- No AC voltage is leaving the PSU. I didnt had the chance to measure current on PSU output lines (will do so), but comparing between load attached and without load, there is very little variation on voltage (e.g. 5.16V goes to 5.08V with load, etc.). This also made me believe that there is not any short on the lines.
Bellow is a photo of the PSU board (please ignore the sharpie marks that I made on the output header), and what I measured on the output lines with load disconnected (Pins 1-2-3 are connected into the same line, so does 4-5-6)
Output (bottom to top)
Pin 1 GND
Pin 2 GND
Pin 3 GND
Pin 4 5.16V
Pin 5 5.16V
Pin 6 5.16V
Pin 7 GND
Pin 8 5.12V
Pin 9 -5.42V
Pin 10 GND
Pin 11 GND
Pin 12 14.99V
Pin 13 -14.84V
Pin 14 GND
Pin 15 3.8V
Now, 3.8V is really obscure, so I tried to follow the path of this signal. Starting from the PSU, this signal goes into the main board assembly. From there I found only one connection of this signal that goes into the main logic-board (which is the fourth plugin card; photo bellow)
Into the main logic-board, the first branch of this signal
goes to the ~CS input of an MC68230P8 parallel interface driver (red arrow) and a second branch goes into a jumper (red arrow). I didnt traced further than that.
As noticed from datasheet 3.8V should be logic HIGH for this IC, setting it into a disabled state. My main suspicion for this signal now, is whether it could be some kind of "PSU ready" signal, preventing output on the bus until it gets low (?). That brings me back into further investigating the PSU...
Will follow up with any more findings, meanwhile any thoughts are more than welcome!