What power supply? How old is it?

Could also be capacitors on th motherboard or even voltage circuitry

Are the other processors you are trying known to be good?

I would memtest the hell out of that ram. I have tons of old "pc133" modules that fail memtest86 at 133 but are fine at 66 or 100.

pete8475 wrote on 2020-11-18, 02:37:

I would memtest the hell out of that ram. I have tons of old "pc133" modules that fail memtest86 at 133 but are fine at 66 or 100.

Could also just be an incompatibility issue with the RAM. Some boards were super picky about what RAM they worked with back then.

pete8475 wrote on 2020-11-18, 02:37:

I would memtest the hell out of that ram. I have tons of old "pc133" modules that fail memtest86 at 133 but are fine at 66 or 100.

I have tried other modules too, but with the Celeron these modules have always worked well even in this board. And when I tried the 866mhz PIII and it booted, I got into the BIOS and put the RAM at 100, but it always froze into Windows or in it.

To sum up, with the Celeron cpu this board runs fine, RAM and everything.

cyclone3d wrote on 2020-11-18, 02:30:

What power supply? How old is it?

Could also be capacitors on th motherboard or even voltage circuitry

Are the other processors you are trying known to be good?

It's a Pentium IV supply, so I imagine 17 years maybe? Yes they are, that I can confirm

My guess is the power supply or the motherboard. Sounds to me like voltage regulation is too far off for it to work properly with the faster CPUs.

Could be too much ripple (powwer supply or bad caps on the motherboard) or the voltage is going out of range... probably too low if it is.

I would try with a new power supply first. The Pentium III setup should work fine with a new 350w power supply. I always make sure to get power supplies that are from a good brand (Seasonic or something like that) and are at least 80 Plus Bronze rated. Active PFC is a given.

As others have suggested, first thing to do would be remove any things in excess, so that means extra cards and memory sticks.
I had a socket 754 driving me crazy last week with freezes and resets and I was afraid the caps were long gone, but after testing every stick individually, I found a lot of them unable to keep up. Once I had a couple of sticks that I had tested thoroughly and knew were good, the freezes went away.

As such, you should install the CPU you want to run with the board and memtest86+ the RAM until you find one stick that passes the test at least twice. Once that is done, try again.
If the problems persist, I'd try a newer, decent PSU. Tualatins never exceeded 30-32W of TDP and as such, don't require exotic PSUs by today's standards such as those 60-70W Athlons and Athlon XPs.

If that fails too, then you might need a recap on the board. The reason this would be the case is that the processors you tried draw more power and the board can't provide it due to caps going bad. However, the Celeron 600 and Pentium 3 650 you tried are 2-3W apart, which is why I think you might be having a combination of RAM and PSU issues here. Give it a try and don't give up. If you feel frustrated, leave it for a couple of days and come back to it with a fresh set of ideas.

I'd first remove all the stuff you don't need to power on the board.
Remove the AGP card and test using a PCI card of onboard if your board supports it. Also remove NIC, USB (except perhaps for USB keyboard/mouse but with s370 it's usually better to use PS/2 anyway) and remove any sound card.
Only use keyboard (mouse is not really needed for troubleshooting), CPU with HSF, ONE stick of SDRAM, onboard graphics or use a known good PCI card (like a Trio or Virge 2MB or something) and remove any AGP card unless this is like the only graphics card you have.
Removing the AGP card is not as necessary btw but when I troubeshoot I really want to rule any weird stuff out. GF4MX doesn't usually have a large power draw, so it shouldn't matter much with regards to your PSU.

I'd also vote to have switching with a different PSU to test to be somewhere at the top of your priority list. Switch to known working PC-133 if the board gives the same issues with the test-PSU (and your PC-133 thus far is not known to work at 133MHz afaict).

Use only a single stick of RAM when testing. Add more stuff once you get something stable working together.

Btw, also make sure your CPU HSF is up for the tast of cooling your CPUs sufficiently.

As soon as I get a chance I will follow your instructions. After what I tried and all the possibilities you indicate, one of those you propose will surely be the culprit. Thank you for all your help and replies

andre_6 wrote on 2020-11-18, 15:46:

As soon as I get a chance I will follow your instructions. After what I tried and all the possibilities you indicate, one of those you propose will surely be the culprit. Thank you for all your help and replies

Good luck with your bughunt =)

After trying everything I found the real issue, I changed the remaining working CD-ROM drive (I had a dead one too) as an absolute last resort after running everything you told me and that I could think of. It worked. I found it had a not exactly bent pin, but more like a sunken one on the drive's IDE connector. It's amazing the amount of issues such a thing could cause, I mean the CD-ROM drive?...

Again, thanks a lot for all your help and suggestions, Vogons never fails to deliver!

andre_6 wrote on 2020-11-21, 15:26:

After trying everything I found the real issue, I changed the remaining working CD-ROM drive (I had a dead one too) as an absolute last resort after running everything you told me and that I could think of. It worked. I found it had a not exactly bent pin, but more like a sunken one on the drive's IDE connector. It's amazing the amount of issues such a thing could cause, I mean the CD-ROM drive?...

Again, thanks a lot for all your help and suggestions, Vogons never fails to deliver!

Nice! Another bug squished! 😁

These unexpected (and often rather unique or very unlikely) scenarios are a main reason why one will want to reduce all potential unnecessary factors and start with a base working setup, then add items one by one. The messier the troubleshooting is handled, the messier the outcomes may become (and as a bonus doing messy troubleshooting may result in more broken parts or worse if you're not careful).