In a situation like this where you have a configuration where it comes on, and the normal way where it does not, the best thing you can do is just do one very tiny assembly step at a time and then immediately try a power on and see if it works. So first step you can try is very carefully lay the board in the case with the motherboard side laying flat without any screws so it is loose but normally positioned. Hook up power carefully, keeping the board in the right spot. If that works, then you have one step eliminated. Then insert the I/O shield, try again. The each screw one by one trying each time. Then each front panel connection one by one. Then the video card, etc. And finally the hard drive. You will eventually hit a step that does not work, assuming there is a fault somewhere.

Edit: Per the question on the CPU temperature and system stability, I do recommend measuring the output of those voltage regulators and power supply as you move through the steps above.

Vagabund2k wrote on 2023-07-31, 09:25:

The CPU temperature steadily climbed to 60°C, and then a blue screen appeared – this time without a warning signal from the PC speaker. At that time, the system had no case fans.

Try applying a bit of pressure (with your finger) to the mobo chipset (both north and southbridge). You have BGA chips and some of these can, due to manufacturing defects or improper soldering, develop cracks in the solder balls. In its early stages, when very few balls are affected, the mobo might appear to work but be unstable when flexed or with temperature changes.

60C on the CPU is not great but also not terrible. That being said the CPU socket iteself might also have issues related to heat cycling - if it's not cracked solder then working the lever a few times might help seat the pins better. Try it.

-

the3dfxdude wrote on 2023-07-31, 13:31:

In a situation like this where you have a configuration where it comes on, and the normal way where it does not, the best thing you can do is just do one very tiny assembly step at a time and then immediately try a power on and see if it works. So first step you can try is very carefully lay the board in the case with the motherboard side laying flat without any screws so it is loose but normally positioned. Hook up power carefully, keeping the board in the right spot. If that works, then you have one step eliminated. Then insert the I/O shield, try again. The each screw one by one trying each time. Then each front panel connection one by one. Then the video card, etc. And finally the hard drive. You will eventually hit a step that does not work, assuming there is a fault somewhere.

Edit: Per the question on the CPU temperature and system stability, I do recommend measuring the output of those voltage regulators and power supply as you move through the steps above.

This did the trick. Put it back together screw by screw, cable by cable. Thank god, it was just the Cable to the resetbutton. Would have never expected that.
Thanks to you and everybody else who gave tips.

I will now stress test the system and check if I have managed to address the temperature issue, then I will report back.

But out of curiosity, are there any current power supplies available on the market that are compatible with ATX Socket 7 / Super Socket 7? (Excluding the -5V rail)

Okay, that was a ride. The system booted cleanly with the existing installation (Windows ME) from the previous owner. I spent 30 minutes exploring everything, lots of beautiful old software – like a time capsule from the year 2000. That was already an improvement compared to before. Installed Orpheus 2 drivers... installed GUS drivers. 2-3 restarts. The CPU temperature was consistently around 56 to 58°C. Then I started the game "Raptor," which was among many others on the hard drive. I went through the menu, started the game, played for 5 seconds... and the PC speaker alarm went off again (temperature warning). The monitoring tool (MBProbe) came to the foreground, CPU temperature blinking red (62°C), and the gunshot sound from Raptor kept looping because the game froze. What a cacophony – I felt like I was on a crashing spaceship 😁 I took a photo, attached it below. In my opinion, the voltage looks okay (K6-2-500), what do you think? Strangely, the chassis fans are not recognized, neither by the BIOS nor by MBProbe, but they are running.

BTW: I I followed Deunan's tips.

Vagabund2k wrote on 2023-07-31, 18:56:

The monitoring tool (MBProbe) came to the foreground, CPU temperature blinking red (62°C)

Hm, that fan you have on photos, is it working? Not going very slow or something? I have a mobo with K6-2 400MHz and the fan is 12V but I run it at 5V - and it's enough too cool the CPU. Plus it's a slow, quiet fan to begin with. I do have a bigger heatsink though.

Things to try/check:
- remove the heatsink, make sure somebody didn't forget to peel off the plastic protector from thermal pad
- if thermal pad is badly damaged or gone (or was never there) then clean the CPU and bottom of the heatsink and reapply some thermal grease
- try a bigger of faster fan, see if it improves anything (but only after you do above steps), if it does but the new fan is too loud then you need bigger heatsink for this CPU
- lastly, depening on how your case is filled with cards and cables, you might need additional fan(s) to cool the interior

If the heatsink seems to be stuck for good then don't force it, let the CPU heat up to 50C or more first and then power the PC off and try to twist the heatsink off the CPU.

BTW the alarm temperature can usually be set in BIOS, I think 60C is within reason (although try improving it) so perhaps you can just bump this setting to 70C and maybe that stop the alarms. If there is no BIOS setting for this then your monitoring software is doing it - configure it there.

Deunan wrote on 2023-07-31, 19:36:

Hm, that fan you have on photos, is it working? Not going very slow or something? I have a mobo with K6-2 400MHz and the fan is […]
Show full quote

Vagabund2k wrote on 2023-07-31, 18:56:

The monitoring tool (MBProbe) came to the foreground, CPU temperature blinking red (62°C)

Hm, that fan you have on photos, is it working? Not going very slow or something? I have a mobo with K6-2 400MHz and the fan is 12V but I run it at 5V - and it's enough too cool the CPU. Plus it's a slow, quiet fan to begin with. I do have a bigger heatsink though.

Things to try/check:
- remove the heatsink, make sure somebody didn't forget to peel off the plastic protector from thermal pad
- if thermal pad is badly damaged or gone (or was never there) then clean the CPU and bottom of the heatsink and reapply some thermal grease
- try a bigger of faster fan, see if it improves anything (but only after you do above steps), if it does but the new fan is too loud then you need bigger heatsink for this CPU
- lastly, depening on how your case is filled with cards and cables, you might need additional fan(s) to cool the interior

If the heatsink seems to be stuck for good then don't force it, let the CPU heat up to 50C or more first and then power the PC off and try to twist the heatsink off the CPU.

BTW the alarm temperature can usually be set in BIOS, I think 60C is within reason (although try improving it) so perhaps you can just bump this setting to 70C and maybe that stop the alarms. If there is no BIOS setting for this then your monitoring software is doing it - configure it there.

The CPU fan runs at 5000+ RPM as advertised. There was no thermal pad with the new heatsink/fan and there is also no plastic / foil protector on the heatsink. This may sound stupid but i have never seen anybody using thermal paste / thermal pad with a K6-2. The Heatsink / Fan just gotclipped to the CPU. May this be the Problem?

Vagabund2k wrote on 2023-07-31, 17:19:

But out of curiosity, are there any current power supplies available on the market that are compatible with ATX Socket 7 / Super Socket 7? (Excluding the -5V rail)

All good quality ATX PSUs will work just fine with a SS7 build.

Vagabund2k wrote on 2023-07-31, 19:57:

This may sound stupid but i have never seen anybody using thermal paste / thermal pad with a K6-2. The Heatsink / Fan just gotclipped to the CPU. May this be the Problem?

I would say that this is very likely the problem if the cooler is otherwise OK. I’ve always used thermal paste on most CPUs.
23 years ago, when I upgraded my Pentium MMX 166 to a K6-2 500 I also used a white cheap thermal paste on the latter.

Vagabund2k wrote on 2023-07-31, 19:57:

Deunan wrote on 2023-07-31, 19:36:

Hm, that fan you have on photos, is it working? Not going very slow or something? I have a mobo with K6-2 400MHz and the fan is […]
Show full quote

Vagabund2k wrote on 2023-07-31, 18:56:

The monitoring tool (MBProbe) came to the foreground, CPU temperature blinking red (62°C)

Hm, that fan you have on photos, is it working? Not going very slow or something? I have a mobo with K6-2 400MHz and the fan is 12V but I run it at 5V - and it's enough too cool the CPU. Plus it's a slow, quiet fan to begin with. I do have a bigger heatsink though.

Things to try/check:
- remove the heatsink, make sure somebody didn't forget to peel off the plastic protector from thermal pad
- if thermal pad is badly damaged or gone (or was never there) then clean the CPU and bottom of the heatsink and reapply some thermal grease
- try a bigger of faster fan, see if it improves anything (but only after you do above steps), if it does but the new fan is too loud then you need bigger heatsink for this CPU
- lastly, depening on how your case is filled with cards and cables, you might need additional fan(s) to cool the interior

If the heatsink seems to be stuck for good then don't force it, let the CPU heat up to 50C or more first and then power the PC off and try to twist the heatsink off the CPU.

BTW the alarm temperature can usually be set in BIOS, I think 60C is within reason (although try improving it) so perhaps you can just bump this setting to 70C and maybe that stop the alarms. If there is no BIOS setting for this then your monitoring software is doing it - configure it there.

The CPU fan runs at 5000+ RPM as advertised. There was no thermal pad with the new heatsink/fan and there is also no plastic / foil protector on the heatsink. This may sound stupid but i have never seen anybody using thermal paste / thermal pad with a K6-2. The Heatsink / Fan just gotclipped to the CPU. May this be the Problem?

You need thermal paste between the CPU and heatsink. Every cpu I've ever had from socket 7 to now has some sort of thermal paste or pad.

Vagabund2k wrote on 2023-07-31, 19:57:

This may sound stupid but i have never seen anybody using thermal paste / thermal pad with a K6-2. The Heatsink / Fan just gotclipped to the CPU.

Well, you can get away with that on 486, some Pentiums too, but the problem is the CPU top surface is not perfectly flat, and the the heatsink will not sit on it well. In some cases there might be too little contact surface to pull the heat away. In order to prevent any possible overheating it's always better to apply the thermal paste, no matter what CPU it is. Yes, it's more messy but the results are usually well worth it.

The k6-2 non plus at 500mhz is nearly the hottest for it's family. I would add thermal paste. But I would also check the temperatures directly yourself. I don't think that it would be locking only hitting 62 C. Could there be a thermal monitor halting the system? Could it be wrong? It's happened before. (Yes check the BIOS) There could also be something else wrong. I don't like the leads looking cooked on the vrm. This could be a time to just try another known good CPU to narrow the problem.

The Epox MVP3-Gx mainboards were hit by the capacitor plague big time.
Most instabilities they show are directly or indirectly caused by dried out electrolytics. Sometimes you couldn´t even finish a windows installation. Dead electrolytics also make the chipset and CPU overheat.
It´s highly recommended to replace all the electrolytics in the VRM-circuits before looking for any other reasons for instabilities!