I'd remove the power supply, and check it outside the system. see if the voltages are correct. if not, open it and look for things that have let out their magic smoke.

Make sure all cards are sitting in their slot and the riser is correctly in place.

You may want to try it without any add on cards, the riser, and unplug the disk drives to see if something is messing with the power good signal. Make sure the ram and CPU are not damaged and installed in their correct orientations.

Remember, if the power good circuit gets tripped, it will take a moment or two for the caps in the PSU to drain enough so it lets go.

Check for rot caps both on the board and the psu, a lot of us here have had tantalum caps go bad as well.

Is there silkscreening on the board near the power supply that says what voltage is supposed to go to what pin?

i486_inside wrote:

Is there silkscreening on the board near the power supply that says what voltage is supposed to go to what pin?

doesn't matter. The drives will use a standard power connection. That gives you +5 & +12v. testing the rest should reveal a -5v & -12v. anomalous readings is the first thing to look for.

Is the fuse in the PSU actually good? Is the switch actually wired?

Thanks for the suggestions guys, I'll take a look when I go back to my office tomorrow. I have to find a way to add it to the pile at home in a stealth-like manner, so for the last few days I've been tinkering at work. I did have the PSU out of the machine and opened up, didn't see anything obvious but didn't really dig into it yet.

I really don't know what I'll do with this machine but I love how clean and simple it is. By 1991, it must have been one of the later 286s?

luckybob wrote:

i486_inside wrote:

Is there silkscreening on the board near the power supply that says what voltage is supposed to go to what pin?

doesn't matter. The drives will use a standard power connection. That gives you +5 & +12v. testing the rest should reveal a -5v & -12v. anomalous readings is the first thing to look for.

His concern was that the computer uses a proprietary PSU, a lot of older boards have silk screening near the PSU connectors , my suggestion was to look what voltages are listed by the pins and cross check it with the AT pin out just to make sure it is actually an AT power supply. If he finds out it uses a standard AT supply then he could at least test the board with a spare AT supply.

i486_inside wrote:

luckybob wrote:

i486_inside wrote:

Is there silkscreening on the board near the power supply that says what voltage is supposed to go to what pin?

doesn't matter. The drives will use a standard power connection. That gives you +5 & +12v. testing the rest should reveal a -5v & -12v. anomalous readings is the first thing to look for.

His concern was that the computer uses a proprietary PSU, a lot of older boards have silk screening near the PSU connectors , my suggestion was to look what voltages are listed by the pins and cross check it with the AT pin out just to make sure it is actually an AT power supply. If he finds out it uses a standard AT supply then he could at least test the board with a spare AT supply.

The form-factor is proprietary, the output isn't. My Acer board will run off a standard ATX PSU so that would be the first thing to check (if you've got one). But yeah, check any silkscreening just in case.

Actually that's a good point! It's AT, it works if connected to the motherboard or not!
You can test the PSU outside the case. Just attach a fan, HDD, something for a bit of load and hit the PSU's power switch. You'll know right then if its the PSU or not.

If it is the PSU, things are looking good as the 2 motherboard power cables are following the standard black to black wire arrangement, not that is a 100% accurate test (dell!) but still better then something totally different

If the PSU is dead, since the form factor of that PSU doesn't seem to be standard, you could still take one of these

with an ATX to AT adapter

Hey guys,

Thanks for all the tips. Work is crazy the next few days so I've only had time to snap some better pictures and take apart the PS. I did remove the PS from the computer and it wouldn't power on even by itself. Of course I can't load-test it like that, because there are no rails/cables coming out of it. All the power cables for drives/fan etc. come from the MB itself. That must be uncommon?

Anyway, here are a few pictures. The fuse in the PS looks unbroken to my eyes, but we'll have to see.

Deksor wrote:

If the PSU is dead, since the form factor of that PSU doesn't seem to be standard, you could still take one of these https://ima […]
Show full quote

If the PSU is dead, since the form factor of that PSU doesn't seem to be standard, you could still take one of these

with an ATX to AT adapter

A little bit off-topic, do you have experience with that? Are all power rails needed by an old/AT system supplied? Could be a nice way to silence some old PC's here 😜

You would need an ATX to AT adapter with -5V.

I've never used one of these, all I know is that they're tiny, they are liked by tiny pc makers and that they are powerful enough to power any retro PC made before the year 2000 unless you're using a lot of hdds maybe

tizzdizz wrote:

All the power cables for drives/fan etc. come from the MB itself. That must be uncommon?

Indeed it is uncommon - my Acer 386 is the only PC I've ever seen that has the same arrangement - it was a way of keeping things compact, I guess. I recently used a splitter to power an additional hard drive and a 5.25" floppy drive from the PSU. Probably shouldn't have done this as it's rated at a pathetic 43 Watts. It got really hot, as it's passively cooled (as is yours). Have you managed to find any documentation for your system online? I have found a grand total of fuck-all. Definitely try to get this system working as they are super-super-rare. If you have a multimeter, you could check the continuity of the fuse or, if you don't, you could make a simple circuit from a battery pack and an LED using the fuse to complete the circuit (or not, as the case may be). Unless anyone with more electronics experience than me can explain why this could be a really bad idea?

It's difficult to say what's broken.
The transformers don't look that healthy, but maybe that's just because of heat/grease.

I would check :
- diodes (little red and big black ones)
- power transistors
- regulators (green circle)
- huge ceramic resitors (blue circle)

- caps

Also have a look at the thing I circled (pinkish). That's either a power mosfet, triac, thyristor or some sort of regulator.
If it is defective, the power-good circuit wont't allow the PSU to power up.

Also make sure someone else is in the room when you tinker whith this stuff.
Experimenting with these things is dangerous! Seriously. So at least if you are going to
to turn it on after repair, make sure someone else is there.

I got zapped a few times already. For example, one time I touched the insides of a portable PC.
Even though, it was disconnected from power for hours, the built-in CRT was still very, um, energetic. 🤣

I'd be also good if you have safety switches (residual-current devices , RCDs) in the house.
In Europe, these are usually installed in the fuse box. One for each room, floor..
But you can also get them as "in-between cables" (I'm pretty sure that's no valid English term, haha 😊).

The pink/salmon circle is around the line input bridge rectifier. Watch out, as the output legs on that thing will be energized with quasi-DC at ~340V when the supply is on and until discharged. At that voltage, the bulk caps discharging in your body have enough energy for a good chance of stopping your heart.

If nothing else looks broken, one common point of failure is a high-valued resistor (often largish in size too) that feeds the switching controller from the main bulk cap before the switching starts and can provide power to it.