I apologizes for all those typos above, but I can't edit my own topic for some reason (probably because I'm Newbie still)

I don't have a clue what could cause this problem, but i will write anything so i can see what the real masters here has to say about this weird thing 😀 Now it will leave me this topic in my posts 😀 I'm curious to see what happened here?

I once recapped an asrock ? ecs? doesn't matter socket A board and ended up with the same problem. all voltages measured ok, yet two cpus fried on it.

I had to recap it since the old caps were bulging and causing instability. unfortunately after the recap the board fried two cpus.

Never dug deeper into the reason why.

Hi, I'm very pleased that you read what I wrote, just to clarify, I have several boards under repair, unfortunately I can't fix them all, both because I have limitations due to the equipment, and because I don't have knowledge of the circuits and components, I manage to do what I can, but I don't do it as a profession, just as a hobby.
Looking at the photo of the CPU, it is definitely overheated (and therefore no longer functional, almost 100%).
But from memory I remember a CPU that was very hot, trying it in one MB, but in another MB it works and doesn't heat up.
There will probably be a shorted PIN, there are not only the I/O and VCore voltages, I advise you to measure the voltages on all the mosfets (or regulators), they usually have three pins, if you measure those of the VRM, one of the three The PIN should be around 1.5-1.8V, if it is more than 2V, the problem may be there.
But there could be something wrong, for example if it has jumpers, check that they are all set correctly, or that they are all there, finally a careful visual check, make sure there are no tracks or broken SMDs.
If you want you can put the voltage of the three PINs of each mosfet (or regulator) here, maybe there is some anomalous voltage.

Thanks a lot for answering, guys.

PC@LIVE wrote on 2024-11-11, 20:06:

...

PC@LIVE, yeah I've been planning to remeasure everything again. I suspect that if the issue is with either of the bridges then I simply won't be able to confirm that without replacing those. I wish there was a schematic for AU31 but it's too old and people didn't bother to digitize schematics back then 🙁

It occurs to me that those CPUs are quite easy to fry, certainly the first one was fried due to the lack of the heatsink, but the second one could have fried because you have to put a bit of thermal paste on the die (the little square or rectangle in the center of the CPU ), without thermal paste it is possible that the heat will not be dissipated, the consequence being burning.
I saw the photo and in fact it has very few jumpers, so I think that if it does not have values ​​stored in the BIOS with overvolts, the only problem could be a missing or faulty SMD, which causes an overvoltage in one or more pins of the socket, at least I think that is the most likely thing

Makes sense , yeah , But I had a new thermo-paste applied before turning everything on again. Also I forgot to mention I bought a programmer and read the motherboard's bios to see if it's ok. It was ok but still I flashed the new one from theretroweb site.

I will keep digging. how many more CPUs are gonna fry before I fix this mobo or give up...

Ujeen wrote on 2024-11-11, 20:41:

I will keep digging. how many more CPUs are gonna fry before I fix this mobo or give up...

How many more cpus do you have ?

Did you make the markings on the capacitors ?

Ujeen wrote on 2024-11-11, 20:41:

I will keep digging. how many more CPUs are gonna fry before I fix this mobo or give up...

Put only Athlon XP on this board (they should turn off when the temperature goes to 85 or 90)....

weedeewee wrote on 2024-11-11, 20:43:

How many more cpus do you have ?

only two fried ones 😀 but ordered two more... but I guess this is a bad idea...

weedeewee wrote on 2024-11-11, 20:43:

Did you make the markings on the capacitors ?

Yeah my markings 😀 those were looking suspicious 😀

butjer1010 wrote on 2024-11-11, 21:03:

Put only Athlon XP on this board (they should turn off when the temperature goes to 85 or 90)....

Noted! But I think both of them were XP , this didn't save them ...

Ujeen wrote on 2024-11-11, 21:13:

butjer1010 wrote on 2024-11-11, 21:03:

Put only Athlon XP on this board (they should turn off when the temperature goes to 85 or 90)....

Noted! But I think both of them were XP , this didn't save them ...

Then is not overheating the problem i guess....
I have never fried Athlon XP, even without cooler... It should turn off after 90 degrees

There's a thermal diode on later Socket A but no idiotproofing, nevereverevereverever power them without a heatsink on, this is what happens.

butjer1010 wrote on 2024-11-11, 21:18:

Then is not overheating the problem i guess....
I have never fried Athlon XP, even without cooler... It should turn off after 90 degrees

then only short can lead to this .. which I can't find still 🙁

BitWrangler wrote on 2024-11-11, 21:22:

There's a thermal diode on later Socket A but no idiotproofing, nevereverevereverever power them without a heatsink on, this is what happens.

yeah I agree, but the second time it had heatsink+ fan on , working. Got fried same fast: couple seconds and done !

Socket A builds are probably the most tricky of all, especially the part with the cooling/heatsink.

Are you 100% positive that when you installed the 2nd CPU with its heatsink that it was making proper contact with the CPU core/die? If the heatsink was used, there's always a chance the clip could be bent and/or not clamping the heatsink all the way. I just ran into a situation like that last week while testing an SV266A mobo. Luckily, I caught this way way before attempting to power up the board. In my case, the clip for the heatsink was original, but it just did not provide any tension on the heatsink at all. Had I powered it on like that, my CPU would have been fried. Luckily, I had some spare retention clips that fit and put proper pressure on the heatsink. With that, it was fine (despite me actually always being a cheapskate and re-using old... but not dry.. thermal compound from another CPU.)

butjer1010 wrote on 2024-11-11, 21:03:

Put only Athlon XP on this board (they should turn off when the temperature goes to 85 or 90)....

Keyword here: *should*
Reality: they usually don't and burn right away.

BitWrangler wrote on 2024-11-11, 21:22:

There's a thermal diode on later Socket A but no idiotproofing, nevereverevereverever power them without a heatsink on, this is what happens.

It's been proven though: that thermal diode is way too slow to act when there is no heatsink installed on the CPU. Now if a heatsink is installed, but say the CPU fan is stuck, disconnected, or the heatsink is clogged with dust... that thermal diode *might* actually work to save the CPU. But don't count on it.

Ujeen wrote on 2024-11-11, 20:41:

I will keep digging. how many more CPUs are gonna fry before I fix this mobo or give up...

As many as it takes? For science. You monster! 😉

At least try going with some less-rare CPUs, so you don't piss off the collectors. 🤣
If the board supports them, maybe try a low-end ceramic chip Duron - these are, as the name might imply, a little more Durable. They are also lower TDP, so less chance they will fry instantly if the heatsink is not 100% right on the die.

BTW, I looked at the pictures on TRW, and this motherboard appears to use a synchronous buck VRM for the CPU, meaning it *should* be pretty hard for the board to fry the CPU, even if there was a shorted MOSFET (which I don't think is the case here... yet?) In any case, just to (roughly) rule out any high-side MOSFETs being bad, check resistance to ground on all of the pins in the 4-pin 12V CPU power connector. Two of those pins will be ground and should show a short circuit, but the other two are for the 12V rail and should show a relatively high resistance (KOhms range, at least.) Do this test with the PSU disconnected.

Speaking of the PSU...
Have you opened it up to check for bad caps there? I had someone send me a socket A board before that was completely fried - both the RAM and CPU were fried. Turns out, the PSU was a decent build but had a ton of bad caps, outputting ripple so high that it was killing stuff left and right. How the chipset managed to survive, IDK. But the RAM stick and CPU were gonners (probably the died before the chipset, saving it from the PSU.)

butjer1010 wrote on 2024-11-11, 21:03:

Put only Athlon XP on this board (they should turn off when the temperature goes to 85 or 90)....

Keyword here: *should*
Reality: they usually don't and burn right away.

Of course he will put the heatsink on it, he's not an idiot!!! Who mentioned that it will not "burn" without cooler?

BitWrangler wrote on 2024-11-11, 21:22:

There's a thermal diode on later Socket A but no idiotproofing, nevereverevereverever power them without a heatsink on, this is what happens.

It's been proven though: that thermal diode is way too slow to act when there is no heatsink installed on the CPU. Now if a heatsink is installed, but say the CPU fan is stuck, disconnected, or the heatsink is clogged with dust... that thermal diode *might* actually work to save the CPU. But don't count on it.

Same answer....

Ujeen wrote on 2024-11-11, 20:41:

I will keep digging. how many more CPUs are gonna fry before I fix this mobo or give up...

As many as it takes? For science. You monster! 😉

At least try going with some less-rare CPUs, so you don't piss off the collectors. 🤣
If the board supports them, maybe try a low-end ceramic chip Duron - these are, as the name might imply, a little more Durable. They are also lower TDP, so less chance they will fry instantly if the heatsink is not 100% right on the die.

[/quote]

Durons are the socket A CPUs that burns right in seconds when the fan stops spinning!!! What gives You this funny idea that he needs to use Durons???? That is the only reason i suggested him XP Athlons, because they have this thermal diode that Durons don't have! In nineties Intel had the comercial with Durons (if You remember those times, don't know when were You born), where they made fun out of AMD because of that reason. They used Duron in comercial, and they stopped the fan! After few seconds, the smoke came under the cooler, and they done that with Intel pentium also, but the pentium kept working fine! This was their way to people buy intel instead of AMD. So AMD made XP (extra protection) Athlons after that "fiasco"!!! They will not burn in second if the fan is problem....
After all he wrote up there, after all the things he has done, i think he is capable to put the cooler right, don't You think? Don't beleive the problem is in cooler, but that's my opinion....

butjer1010 wrote on 2024-11-11, 21:18:

I have never fried Athlon XP, even without cooler... It should turn off after 90 degrees

butjer1010 wrote on 2024-11-14, 21:54:

Of course he will put the heatsink on it, he's not an idiot!!! Who mentioned that it will not "burn" without cooler?

Uhm... you did. Or is that not your quote above? 😁

butjer1010 wrote on 2024-11-14, 21:54:

Durons are the socket A CPUs that burns right in seconds when the fan stops spinning!!! What gives You this funny idea that he needs to use Durons????

Sure.
Care to show proof?

Laws of physics (specifically, thermal transfer and thermal mass) simply don't work that way.
Stopping the fan for a few seconds won't do shit to the CPU, so long as the heatsink is still properly attached to it. It will take a good minute or more before the temperatures finally rise high enough to kill the CPU. That's because the heatsink has thermal mass, and it takes time to heat that up.
So NO, there's NO WAY the CPU will burn instantly IF the CPU heatsink is installed properly, regardless if the fan is running or not.

butjer1010 wrote on 2024-11-14, 21:54:

That is the only reason i suggested him XP Athlons, because they have this thermal diode that Durons don't have!

And so what?
As I mentioned, these thermal diodes aren't that good at reacting to over-temperature. If you're so trusting in them, why don't you show how that works with one of your own Athlon XP builds? I know I'm not willing to do that, because I've worked in the computer repair business long enough to have seen more than enough examples of how those thermal diodes specifically in socket A CPUs exactly don't really help more than anything else.

As for the reason I suggested low-end Durons (600-900 MHz models): they have a much lower TDP compared to Athlon XP. For example, the 700 MHz Duron Spitfire has a typical power consumption of ~29 Watts and a maximum consumption of a little over 30 Watts. That means this CPU will be spitting out about 30 Watts of heat, tops. For the 900 MHz Duron Spitfire, that's about 40 Watts max. Now take O/P's Athlon XP 2100+: assuming it's the Palomino model, it has a typical power draw of about 65 Watts and a maximum power draw of 72 Watts. This is more than twice the heat dissipation compared to the 700 MHz Duron. So if using the same heatsink (with a stopped fan... or maybe just a badly clogged heatsink with dust), the 2100+ will be much quicker to overheat than both the 700 MHz or 900 MHz Durons. It's simple physics, and that's all there is to it.

butjer1010 wrote on 2024-11-14, 21:54:

In nineties Intel had the comercial with Durons (if You remember those times, don't know when were You born), where they made fun out of AMD because of that reason. They used Duron in comercial, and they stopped the fan! After few seconds, the smoke came under the cooler, and they done that with Intel pentium also, but the pentium kept working fine! This was their way to people buy intel instead of AMD. So AMD made XP (extra protection) Athlons after that "fiasco"!!! They will not burn in second if the fan is problem....

No I don't remember that commercial (was there even one?), but I do remember the Tom's Hardware video from back in the day (mid 2000's) where they did this experiment with a Pentium 3, a Pentium 4, and two socket A CPUs (IIRC, one was a ceramic chip Athlon and the other a newer FCPGA Athlon/XP.)
In any case, stopping the fan for a few seconds will NOT damage the CPU - that's just myth / bullshit if such thing was ever shown in a commercial. Now removing the heatsink entirely from a socket A CPU or just not seating it properly on it, that surely will burn out the CPU in seconds.

butjer1010 wrote on 2024-11-14, 21:54:

After all he wrote up there, after all the things he has done, i think he is capable to put the cooler right, don't You think?

Not saying O/P is not capable of putting on the cooler right. But I have worked long enough in the electronics repair business to know that shit happens all the time and can happen to anyone, regardless of experience level. In my case as mentioned with the SV266A socket A motherboard that I was testing last week, the cooler retention simply was faulty and did not clamp onto the CPU properly anymore... though from a distance, it looked totally fine. It wasn't until I grabbed the whole motherboard by the cooler and noticed that it's actually wobbly and not tight / firm against the CPU. Not exactly sure how that happened, but I verified it several times. The retention spring was somehow damaged / bent out of shape. If I wasn't always double-checking everything like I always do, that CPU on the SV266A motherboard would surely have burned out.

butjer1010 wrote on 2024-11-14, 21:54:

Don't beleive the problem is in cooler, but that's my opinion....

Having an opinion is fine. But when it comes to troubleshooting, I like to work with facts only. So I don't take anything for granted until categorically verified.