Those voltage swings on +5V and +12V rails are huge, if your voltimeter doesn't detect such swings, then it's MB related - a health monitoring chip or BIOS related issue. Does ATX connector get warm?

I had to tear down the set up so I'm not able to test if the power connector gets hot and it might be a while before I can set it back up.

However, I tore it down directly after turning it off and I don't recollect anything out of the ordinary in terms of heat. I don't recall the power connector feeling warm. What stood out though was that I was expecting the CPU heat sink to have some temp. to it but it actually felt cold, just like the CPU itself. The banshee heat sink was warm during operation (but not "hot"). When the system was running I did do a "smell" test a couple of times, over the years I have noticed that when something get excessively hot there is usually a typical "smell", but none of that either.

I was just looking at mine a couple of hours ago and noticed the System Management Bus is brought out to a header, I wonder if that is collecting noise. I'd hang a multimeter on 5V and 12V on a spare drive connector and if there wasn't anything happening there, then would feel content ignoring it, unless other symptoms began to show. I don't know though if the PSU itself if ATX 2.0 would be regulating on the 12V, which probably has only 10W of the hdd spinning until you put a CD in, which might not be enough for it. Try plugging any spare HDDs into any spare drive connectors, see if it calms down with 4 or 5 on there.

What OEM is the power supply? My guess is you are using a really old design power supply that isn't regulating properly.

I generally won't use any power supply for a build that is not at least 80+ bronze and it must also have Active PFC.

I've had a motherboard that reported voltages with a fixed offset, but not seen one with swings like that. Do you have a second PSU you can compare with? It might be difficult to check for a noisy supply with a multimeter. I know that my cheap meter is quite slow, so it averages the voltage and will miss any fast (<0.5s) noise. I think that the motherboard is more taking a series of point measurements, so would be more likely to catch any transient spikes.

It's an old thread.

But, the motherboard just takes a sample every second or so and displays it. If there is ripple on the power supply, it is taking a sample at a point on the sin wave somewhere, so it will jump around when sampled once every second or so.

A cheap multimeter or poorly designed multimeter will not show the fluctuations either. You pretty much need something like a Fluke TrueRMS model to get a high sample rate. The best way to see if there's ripple like that is to use a good multimeter and set it to AC voltage measurement and read what the 5v, 12v, 3.3v power supply outputs show. Anything above 120mVAC needs a recap job. Other than that, an oscilloscope will be able to do the same.

Dang, sorry, I necroed it by accident. Not sure how I thought it was current.

Edit: Heh, mystery solved Re: Bought these (retro) hardware today
Edit2: annnnd it's gone.. 🤣 I'm not losing it honest, pan dropped a mention.

Hey guys. Yes, it's an old thread. Still love to have peoples input on it re. troubleshooting. I wouldn't have linked to this thread if I wasn't interested 😀

Unfortunately, I haven't been able to do anything with the board since I first encountered the problem. One day I will! (I hope 😀)

Okay... I managed to recreate this set up. After linking to this thread from various other threads re. this board [1] over the past year or so I thought lets spend some time to see what's going on.

After I recreated the set up, same hard drive, graphics (Voodoo banshee AGP) and CDROM drive, it immediately showed the same behavior. I thought, maybe swapping the CPU would make a difference. So I swapped the K6-2 450 for an MMX 233, but no difference.

Then I thought, maybe add some more load, so I added an extra CDROM drive. No difference.

Then I thought, maybe more peripherals. So I added a sound card (cheapo ESS ISA) and a 3com PCI network card. Hey... noticeable difference. The +5 is now slowly fluctuating between 5.0 and 5.1 and the +12 is now mostly settled on 12.6, sometimes dropping to 12.5 and sometimes spiking to 12.7, but not higher. Still a little bit high but not 13!

Then I removed the second CDROM drive. Similar conditions.

Then I removed the 3com PCI network card. Similar conditions.

Then I removed the ESS sound card. Back to heavily fluctuating.

Adding the ESS sound card back. Again, +12 hovering around 12.6 with a bit of swing and the +5 slowly bopping between 5.0 and 5.1.

So, the ISA sound card did the trick. Still a little worried about that +12 rail but definitely better than what it was.

Other than that, this board seems to run okay, just running some games etc. Maybe if I do some actual work on it I might find stability problems. 😀

[1] ASUS P5A and P5AB problems

Change the PSU ... you are almost certainly seeing the effects of either a poorly designed, or a very tired PSU.

For a K6 or P-MMX, you could go with nearly any modern PSU for testing.

Okay, I had some time to play around with this some more. I changed PSU. The PSU I was using was a Startech ATX2POWER350 and I switched to a Seasonic SS-600ES model (It's very difficult to find anything below 600W nowadays other than SFX). Both PSU's were purchased new by myself. Result? Pretty much the same. With no load other than the bare essentials, the Seasonic doesn't error on the +5V though. With load added the behavior with both PSU's seems identical. Unfortunately I don't have the equipment to do proper testing but it might very well be that there is a problem with the board it self.

Some pictures of the PSU's, someone told me you nerds like this kinda stuff... 😀

The Startech (It uses Teapo caps which I believe are shit?):

The Seasonic (It uses KZE caps which I believe are of a better quality?):

I have more photos but I can only attach 5 per post. Happy to post more...

From the previous posts it seems you haven't hooked up a multimeter to check the voltages directly from the power supply. You really want to do that, even with a cheap meter.

The BIOS voltage readings from a board this old are just not reliable. If you're not encountering problems when using the board then I'm 99.9% sure that both power supplies are just fine, and you are being misled by the BIOS.

TheMobRules wrote on 2021-10-16, 09:30:

The BIOS voltage readings from a board this old are just not reliable. If you're not encountering problems when using the board then I'm 99.9% sure that both power supplies are just fine, and you are being misled by the BIOS.

Even with bad PSUs, voltages normally do not fluctuate that much when there are no fluctuating loads.
So I think @TheMobRules is right, as the PSUs you used do not look like "gutless wonders".