NEED HELP! MB seemed to have killed GPU.

[Sabotaged_Enigma]

After entering OS for about two minutes, it suddenly shut down. Then there was no response when I pressed power-on button.

At first I thought it was because of CPU overheat since I noticed it was approaching 90 degrees C, but that was not the case. It stayed unable to boot until I pulled the GPU from the PCIe slot and every other part started to function properly.

Then I realised there could be something wrong with my GPU. Capacitors could be to blame. But I wasn't really sure, so I took out another board. And history happened again.

And here happened the weirdest thing: I pulled out all the power connectors and started to try one by one. 6-pin was already connected to GPU and when I inserted the 24-pin connector, board booted automatically, but since GPU was not installed so I had no way of knowing whether it booted properly. And with this weird thing, I cut off the power and installed GPU, but it wouldn't boot again.

So I took out my multimeter and found there was breakover of many capacitors and even resistors.
And then I booted the board and measured the input voltage of 6-pin from the power supply, which was so normal at 12.10 V.

Then power supply was not to blame.
And I measured the capacitor beside PCIe slot on the board, and found it was a breakover, too!

So could it be that it was the mainboard that killed my GPUs? (Two GPUs, they died the same way. Well, they are not totally dead, their fans could still spin. And one needs 6-pin and the other doesn't, but it behaves the same when either of them is installed.)

I am now upset and confused. I never saw this kind of thing and I can't do anything for I can't fix it...

Update: Damn. The capacitor of the left DRAM slot is also breakover.

 

[ThaiTaffy]

Considering the board is about $20 wouldn't it be better to bin it and buy another rather than wasting time and money trying to fix it?

 

[Sabotaged_Enigma]

Considering the board is about $20 wouldn't it be better to bin it and buy another rather than wasting time and money trying to fix it?

Yes that's true. But it has made my GPUs fail. I want to know why. Even if I can't bring them back to life, I'll avoid this kind of thing if I learnt why.

 

[Bill_Bright]

You cannot properly or conclusively test a capacitor, especially with a multimeter, while that capacitor is "in-circuit".

 

[Ferd]

So I took out my multimeter and found there was breakover of many capacitors and even resistors.

Can you explain what does this mean? What is breakover ?

 

[ThaiTaffy]

I'm guessing he means continuity but testing like that with the capacitor still attached to the board and possibly still charged isn't a great idea.

 

[cst1992]

True. Capacitor should be discharged before testing, or it could lead to accidents.

 

[Bill_Bright]

It is not about the capacitor being charged. It would get discharged as soon as the meter leads touched both capacitor leads. With such small caps and charges, the "spark" or "arc" would be tiny and most any decent multimeter would be unaffected. It could be dangerous if testing a large cap - as might be found inside the power supply. But not a motherboard cap.

The problem is, when still in circuit, other devices in the circuit would affect any reading. So the capacitor needs to be removed, then tested.

 

[Sabotaged_Enigma]

Can you explain what does this mean? What is breakover ?

Maybe that's not the word? I mean normally a capacitor doesn't allow DC current to go through, but multimeter detected that DC current is going through so it must be broken.

The problem is, when still in circuit, other devices in the circuit would affect any reading. So the capacitor needs to be removed, then tested.

Oh dear. Then there's literally nothing more I can do... I can't analyse PCB design anyway... Now I just want to know whether anyone has encountered this kind of thing.

 

[ThaiTaffy]

Maybe that's not the word? I mean normally a capacitor doesn't allow DC current to go through, but multimeter detected that DC current is going through so it must be broken.

Do you know specifically if the cap is showing continuity or could it be passing through elsewhere on the board.

It is not about the capacitor being charged. It would get discharged as soon as the meter leads touched both capacitor leads. With such small caps and charges, the "spark" or "arc" would be tiny and most any decent multimeter would be unaffected.

Op is testing caps in a circuit and calling continuity something random I don't think he's using a true RMS Fluke.

True. Capacitor should be discharged before testing, or it could lead to accidents.

Yes any cap should really be discharged before testing, little caps like the ones he's testing wouldn't kill anyone. Maybe blunt the tip of your probe at worse damage the ADC circuit.

 

[Ferd]

Maybe that's not the word? I mean normally a capacitor doesn't allow DC current to go through, but multimeter detected that DC current is going through so it must be broken.

I don’t know what the original word you intended to use so idk , maybe you mean shorted? As in short circuit?
Anyway you can forget about that , pointing multimeter probes to a motherboard is only useful if you want to read voltages , or current , all the other readings will mostly be wrong, you need to completely remove the specific components you want to test from the motherboard and test it separately...

capacitors do allow current through, I think you’re using continuity mode ? So the reason why you hear it beeping is not because it’s short , it actually start beeping when the capacitor is charging from the probes of your multimeter.... that doesn’t mean the capacitor is bad , and doesn’t mean it’s good either ... maybe your multimeter has a capability to measure capacitance , what model is it ?

 

[ThaiTaffy]

Check to see if you meter has the μF symbol.

 

[Sabotaged_Enigma]

capacitors do allow current through, I think you’re using continuity mode ? So the reason why you hear it beeping is not because it’s short , it actually start beeping when the capacitor is charging from the probes

Yeah that's it! Continuity mode it is. It turns out I shouldn't be so upset because what I concluded is wrong lol

 

[pavle]

You would have to de-solder the condensers to measure resistance, but that's really not the way to go; you have to have a good cap-tester with ESR-meter.
You can't just say a cap is bad if they're in-circuit, as this simple image shows:

 

[Sabotaged_Enigma]

Check to see if you meter has the μF symbol.

Yes it does. But unfortunately the range is too low so I can't get results...

 

[Bill_Bright]

Op is testing caps in a circuit and calling continuity something random I don't think he's using a true RMS Fluke.

Doesn't matter. Even a true RMS Fluke does not measure capacitance. And again, even if his model does, the cap must be isolated (taken out of circuit) and measured all by itself.

pavle illustration shows one reason why.

 

[Sabotaged_Enigma]

You would have to de-solder the condensers to measure resistance, but that's really not the way to go; you have to have a good cap-tester with ESR-meter.

Thanks! It seems I'll have to hand it over to a professional. Oh heck, I'd better forget it since it's not worth a lot. But I still would like to find out why.

 

[pavle]

Sorry - I messed up the text by the second circuit before - had to correct it to say no continuity instead of no resistance.

 

[eidairaman1]

After entering OS for about two minutes, it suddenly shut down. Then there was no response when I pressed power-on button.

At first I thought it was because of CPU overheat since I noticed it was approaching 90 degrees C, but that was not the case. It stayed unable to boot until I pulled the GPU from the PCIe slot and every other part started to function properly.

Then I realised there could be something wrong with my GPU. Capacitors could be to blame. But I wasn't really sure, so I took out another board. And history happened again.

And here happened the weirdest thing: I pulled out all the power connectors and started to try one by one. 6-pin was already connected to GPU and when I inserted the 24-pin connector, board booted automatically, but since GPU was not installed so I had no way of knowing whether it booted properly. And with this weird thing, I cut off the power and installed GPU, but it wouldn't boot again.

So I took out my multimeter and found there was breakover of many capacitors and even resistors.
And then I booted the board and measured the input voltage of 6-pin from the power supply, which was so normal at 12.10 V.

Then power supply was not to blame.
And I measured the capacitor beside PCIe slot on the board, and found it was a breakover, too!

View attachment 219050

So could it be that it was the mainboard that killed my GPUs? (Two GPUs, they died the same way. Well, they are not totally dead, their fans could still spin. And one needs 6-pin and the other doesn't, but it behaves the same when either of them is installed.)

I am now upset and confused. I never saw this kind of thing and I can't do anything for I can't fix it...

Update: Damn. The capacitor of the left DRAM slot is also breakover.

Have you tried the gpus in a totally different system. That mobo looks oaf.

 

[ThaiTaffy]

Of your interested here is a decent guide on testing e-caps.

https://www.element14.com/community/community/element14-presents/workbenchwednesdays/blog/2019/06/19/measuring-an-aluminum-electrolytic-capacitors-esr

 

[Ferd]

Have you tried the gpus in a totally different system. That mobo looks oaf.

He did and the second mobo blew up too , or shut-itself down at least lol

 

[Sabotaged_Enigma]

Have you tried the gpus in a totally different system. That mobo looks oaf.

Yes I have. I tried on an AMD board and it triggered self protection as well. So I was thinking if GPU PCB has got a short circuit or something. If so, what caused GPU to be like that...

Of your interested here is a decent guide on testing e-caps.

Thanks. That'll be a lot for me to digest lol

 

[Ferd]

Yeah that's it! Continuity mode it is. It turns out I shouldn't be so upset because what I concluded is wrong lol

It’s okay , don’t get too distracted by this , because if it’s a bad cap it’s not a big deal , they cost less than a dollar
Seeing that the problem happened again I would blame the gpu before the motherboard capacitors....
just curious what gpu isit ?

 

[Sabotaged_Enigma]

Seeing that the problem happened again I would blame the gpu before the motherboard capacitors....
just curious what gpu isit ?

One is 7300 GT which I replaced lots of exploded capacitors myself and successfully brought back to life. The other is Sapphire HD 4850 with GDDR4 memory. 7300 GT is a lost cause and I don't really care about it, but I don't want to lose 4850. Today I used 4850 and it was only 3 minutes or so and it almost died.

 

[OneMoar]

Old hardware is old old hardware has died of old age
./thread