If you're lucky the mobo or psu have polymer fuses that reset after a while.

If you're less, but still lucky, it's a proper fuse:

If you don't have a multimeter, test the psu with... Anything that demonstrates to you that it has power. A fan is good. It'll spin on 3,5,12v

If that's OK look at the motherboard for little orange glass looking things. Kinda like little bugs. Those are. Polymer fuses.
Look for tiny black resistors with 0 or burnt.
Use your nose.

Post pics.

A multimeter will be a big help. Got one?

myne wrote on 2024-09-29, 06:14:

If you're lucky the mobo or psu have polymer fuses that reset after a while. […]
Show full quote

If you're lucky the mobo or psu have polymer fuses that reset after a while.

If you're less, but still lucky, it's a proper fuse:

If you don't have a multimeter, test the psu with... Anything that demonstrates to you that it has power. A fan is good. It'll spin on 3,5,12v

If that's OK look at the motherboard for little orange glass looking things. Kinda like little bugs. Those are. Polymer fuses.
Look for tiny black resistors with 0 or burnt.
Use your nose.

Post pics.

A multimeter will be a big help. Got one?

+1 for the nose knows, burnt electronics are one smell the nose never forgets.

Picking up a board and gently giving it a sniff is oddly one of the first things I do when I get new parts, the burnt smell never leaves the part even if its been sitting in storage for years before being repaired.

Works super well for power supplies and GPUs.

If you smell anything besides musty old electronics then further investigation is warranted, especially if the part may contain RIFFA caps.

I didn't smell that acrid burnt smell. I did smell something subtler though, although the PC had been giving that smell before. Pictures of board attached - I put back in the Matrox Millennium II I had before which worked.

I can see two capacitors with bulged tops, so further attempts at troubleshooting are likely of little use until the bad capacitors are replaced.

It's...back to working fine now. Now idea why, not going to try swapping in that GeForce again though. PSU tester also says is fine.

My Asus A7V133-C had a similar experience two years ago: Asus A7V133-C probably got fried by a faulty GeForce 6200

It managed to start up later on but still behaved with some irregularities.

Polymer fuses.

myne wrote on 2024-09-30, 01:31:

Polymer fuses.

Sure sounds right, thanks for the tip. Inspired me to give it another shot. Glad it's working again and I'm going to be extra careful with it going forward.

Glad it worked.

BTW, those pictures are useless for a visual inspection.
Next time, at least remove all the things in the way.

This ain't going to be no poly fuse problem. At least in my 15 years of doing computer repair, I have not seen poly fuses used on the AGP or PCI-E slots on *any* motherboard. Only USB and PS/2 ports typically have it... and even that's not a given if it's a cheapo board like ECS (though some older ASUS also did this crap.)

The problem was outlined by Repo Man11 above: you have bad capacitors on your board. The board may work "OK" now, but don't expect it to remain this way, as those bad caps will only get worse with time. Given their location, I'd say they are for filtering power either to the Northbridge or the RAM (or both). Good chance if they keep getting worse they could fry the regulators that are feeding to/from them. So you may want to get those replaced sooner rather than later.

As for the GF 4 MX 440 not working: that's either due to the bad caps on the board or it could also be one of those weird incompatibilities between AGP 2x boards and 4x "universal" cards.
IME, most universal /4x cards I tested in various AGP 2x boards worked fine without any issues. As long as the PSU is good and board doesn't have bad caps or other hardware irregularities, things should work.
Occassionally, I've ran into cheap crap cases that either pull or push too much on the AGP card bracket and/or make it sit higher than it should, thus preventing it from making proper contact in the AGP slot. So beware that's another possibility some cards may not appear to work.

myne wrote on 2024-09-29, 06:14:

If that's OK look at the motherboard for little orange glass looking things. Kinda like little bugs. Those are. Polymer fuses.

No, those are mostly regular 1N4148 diodes or similar... or possibly Zener diodes (though I don't ever recall seeing one on a mobo.)
SMD fuses are rectangular and opaque and certainly not glass. Typically they will also have a code letter or number written on them for their current rating.
As for polymer fuses, those are typically green with a symbol in the middle that appears... how to describe it here?... a bit like the Volkswagen "VW" logo. 🤣

I was more suspecting a short.
Which would explain the completely dead behaviour. Which could be explained by the psu or mobo having polymer fuses. The sudden revival over time backs that theory up.

I may be mistaken on what they look like though.
I only have ancient memories of dealing with pc chips boards that exhibited dead, and then revived behaviour and I was fairly sure at the time the little glass looking orange things were responsible.

https://theretroweb.com/motherboard/image/m72 … 30859355411.jpg
See below the com2 header on the left?
Those orange things are what I was thinking of.
Though now, 25 years late, I realise they're marked D for diode.

myne wrote on 2024-10-01, 02:46:

I was more suspecting a short.
Which would explain the completely dead behaviour.

That is plausible indeed.
However, without knowing more details about the system (PSU) used, we can just chuck it up to either a crappy PSU and/or PSU with bad caps, or just the short-circuit protection in the PSU actually working properly and shutting down when it detected the short-circuit. With some PSUs, the short-circuit protection does not reset until the PSU is unplugged from the wall for a certain period of time (until the input caps discharge.)

myne wrote on 2024-10-01, 02:46:

Which could be explained by the psu or mobo having polymer fuses. The sudden revival over time backs that theory up.

PSUs don't have poly fuses, that much I can tell you with 100% confidence, as that's the type of servicing I specialized in.
And motherboards don't have poly fuses on any of the larger power rails. Only things like external ports are usually protected with poly fuses (again, PS/2, USB, Firewire, VGA, and etc.)

myne wrote on 2024-10-01, 02:46:

I may be mistaken on what they look like though. I only have ancient memories of dealing with pc chips boards that exhibited dea […]
Show full quote

I may be mistaken on what they look like though.
I only have ancient memories of dealing with pc chips boards that exhibited dead, and then revived behaviour and I was fairly sure at the time the little glass looking orange things were responsible.

https://theretroweb.com/motherboard/image/m72 … 30859355411.jpg
See below the com2 header on the left?
Those orange things are what I was thinking of.
Though now, 25 years late, I realise they're marked D for diode.

Yup, those orange SMD components are diodes indeed - typically low-power type used for signals.
You can see them on a lot of places on that motherboard's picture.
On that note, the only two fuses I see on that motherboard are as follows:
- one at the very end in the upper left corner of the motherboard (looks like a tan-colored 1/2W through-hole resistor, but only has a single black band on it)
- one at the lower-left corner of the AGP slot next to what I presume is a parallel port header (and again, it looks the same as the fuse described above.)
Both of these are not poly fuses from what I can tell (and I think poly fuses didn't really become a thing until late PII / early P3 era PCs.) So if you blow them, they'll be dead until physically replaced.