All the electrolytic capacitors. even the smallest ones.

weedeewee wrote on 2023-06-11, 14:33:

All the electrolytic capacitors. even the smallest ones.

I fail to see how any failing capacitor could be the cause for a dying mains power switch. They don't even fail short.

yes, but you definitely should change all of them to be sure.
as for the wine & pop.... full bridge rectifier, mosfet & double diode package, inrush limiting NTC (could be the green thing next to where the power comes in on the pcb), oh yeah and a capacitor. the whine was likely an attempt at starting up.

"Recap just to be sure" is not something I do often, it is usually a very unnecessary repair step. I would rather diagnose the problem first.

appiah4 wrote on 2023-06-11, 18:20:

"Recap just to be sure" is not something I do often, it is usually a very unnecessary repair step. I would rather diagnose the problem first.

I agree, it's a better idea to diagnose first.

appiah4 wrote on 2023-06-11, 13:52:

2. I think the power switch may be broken? I turn it to ON position, and the two ground terminals have continuity. The two mains terminals however, do not. That does not seem to be how it is intended to work..

Not sure what you mean by this... these switches usually have two pairs of terminals, one pair for live (brown/black wires) and one for neutral (blue/white wires). But normally there is no ground involved. Can you take a picture of the back of the switch indicating which terminals are connected when you turn it on?

appiah4 wrote on 2023-06-11, 13:52:

3. I don't think the dead switch is the main culprit; ie. something else is wron gin the PSU that probably killed the switch, but what? I can replace the switch, if I can find a suitable replacement.. But I think I should check other things as well. I just don't know what.

Most of the components (except transformers/inductors) are easy to test and cheap to replace, but it takes some time and you will have to remove some of them to test them properly. If you are willing to do this I can list a few usual culprits so you can test. First thing I would do is remove the PCB and give it a good clean to spot burn marks or broken components (as well as the back side) and check that all connections are secure.

Another option to consider is to replace the guts of this unit with another AT style PSU... from what I can see in the pictures the PCB size looks like it could be the same as a regular AT PSU?

TheMobRules wrote on 2023-06-11, 19:15:

Not sure what you mean by this... these switches usually have two pairs of terminals, one pair for live (brown/black wires) and one for neutral (blue/white wires). But normally there is no ground involved. Can you take a picture of the back of the switch indicating which terminals are connected when you turn it on?

There are four terminals, two with black wires and two with white wires. There is one black and one white wire coming from the mains connector. With the switch in 0 position there is no continuity between neither the White terminals nor the Black terminals. With the switch in 1 position, the Black terminals are connected, but the White terminals remain unconnected (no continuity). I think they should be connected mechanically if the switch worked right?

I will take a photo later, but for reference:

10 Position:      1 Position:      1 Position (Expected:)
2
3  W   W            W   W            W---W
4
5  B   B            B---B            B---B

appiah4 wrote on 2023-06-11, 19:27:

There are four terminals, two with black wires and two with white wires. There is one black and one white wire coming from the mains connector. With the switch in 0 position there is no continuity between neither the White terminals nor the Black terminals. With the switch in 1 position, the Black terminals are connected, but the White terminals remain unconnected (no continuity). I think they should be connected mechanically if the switch worked right?

Yes. according to what you measured and already concluded, your switch is broken.

appiah4 wrote on 2023-06-11, 19:27:

There are four terminals, two with black wires and two with white wires. There is one black and one white wire coming from the mains connector. With the switch in 0 position there is no continuity between neither the White terminals nor the Black terminals. With the switch in 1 position, the Black terminals are connected, but the White terminals remain unconnected (no continuity). I think they should be connected mechanically if the switch worked right?

OK, I understand now. I assume you tested this at the metal terminals of the switch itself right? Just to be sure that there are no breaks in the wires that are connected to the switch. In this case then yes, it is a problem since the neutral line (white) is not getting connected to the PCB. I'm not sure if the PSU failing would have killed the switch since what goes through it is basically the mains power (and these switches are rated to handle that), unless live and neutral got shorted at some point, but the breaker in your house should have popped.

Maybe the switch was already marginal to begin with, if the contacts inside were already dodgy making a poor connection in the "ON" position, the high resistance could have produced a lot of heat that caused them to completely break down. Could this have been the cause of the PSU kind of "attempting to work" the first time when there was still some continuity in the neutral line?

If you're 100% sure the switch is dead, you can try to disassemble it and check it yourself. Do you have another 4-terminal switch (rated for mains voltage) that you can use to replace this one and wire it to the PSU? It doesn't have to be a rocker switch like this one, a button also works, obviously you can try to buy a proper replacement later on but it will be useful to know if the switch is the only problem here.

I just checked my box of switched and I only have 2 or 3 terminal switches. I ordered a few switches of the exact dimensions to replace the one on the PSU though. This one was a black rocker, the one I found convenient to order is a red rocker with an LED so I guess we are adding some RGB to my 286 build 🤣 I'll let you know how that goes after I replace the faulty switch.

technically not an led in those switches but a little neon bulb.

In the mean time, if you're up to it, you could measure resistance between the white and black wire connector on the pcb.
If you take out the pcb, measure voltage on the two large capacitors. DC, max range. there shouldn't be any, but depending on the circuit, there might be a charge lingering and it's not quite pleasant to the touch.
So if there's voltage, just let it lay for a while, or short it out.
Then you can more safely measure in diode mode, the black box next to the two large capacitors, it's a full bridge rectifier and it's there to turn the full AC into DC...
I think you shouldn't measure any continuity, aside from a diode reading of 400 ~ 600 there's four diodes in the black box. google full bridge rectifier for an easy schematic. black on the line, red on the arrow.

The rocker switches are here.

I will make the measurements prior to installing the switch and trying to turn it on tonight..

I replaced the switch (the soldering was a pain due to how tight it was) and it now works!

Hapoy ending to this story!

Now it is time to clean and reteobriggt the case, then put together a mighty 286 build 😎