Performed a repasting/thermal pad application to fix a defect on my card's VRMs, how do I test that it's working?

Cyber Akuma

Cyber Akuma

Gawd
Joined
Jan 3, 2009
Messages
645
I have a EVGA GTX 1070 FTW ACX 3.0 (Model Number: 08G-P4-6276-KR)

These cards apparently had a defect that could short out the VRMs and basically destroy the card. To "fix" it... sorta... EVGA issued a firmware update that just simply runs the fans at a higher RPM and a set of thermal pads/paste to help cool the card better.

I finally got around to installing them, and I am not sure if I did it right. I ran into several setbacks, had to un-do and re-do everything due to reassembling it incorrectly in a way that was bending the backplate at an angle (I really hope I didn't damage anything from that), got one of the original thermal pads that was supposed to stay a little dirty and there was a dirty oily substance around some parts of the card (I am told this is apparently normal and a result of the thermal pads? It was mostly centered around these four existing pads around the back of the card). Also since I had a lot of my Noctua NT-H2 lying around I used that instead of the plastic packet of non-descript thermal paste that they gave me.

I tried running a stress test in Furmark for about an hour and the highest GPU temperature was 74C, which seems within range from what I understand. But the main issue that this mess was supposed to mitigate was the VRMs overheating and shorting out. Is there any way I can test if these are operating correctly or cooler now? I am not aware of any way to measure VRMs on a video card, I only see ways to measure GPU temp and the speeds/voltages of the GPU and RAM.
 
there likely won't be temperature probes for the VRM on the video card. If there are, you should be able to see them with software like hwinfo64 or hwmonitor. Does EVGA not have any documentation on how to test for the defect?
 
elite.mafia said:
Does EVGA not have any documentation on how to test for the defect?
Click to expand...

Not really, IIRC they tried to deflect/downplay it back when it was new. I only found out about it recently because apparently cards were exploding again a few months ago for some reason for the same defect and I had to argue with tech support to get them to send me one of those old repair kits for mine.

The defect caused the VRMs to overload and blow, destroying the card, not sure how you would test for that.
 
Cyber Akuma said:
Not really, IIRC they tried to deflect/downplay it back when it was new. I only found out about it recently because apparently cards were exploding again a few months ago for some reason for the same defect and I had to argue with tech support to get them to send me one of those old repair kits for mine.

The defect caused the VRMs to overload and blow, destroying the card, not sure how you would test for that.
Click to expand...
Well if you installed the kit, and it seems like it was installed correctly, I'd just try not to think about it and assume its fine.
 
Basically trust or remove the cooler to verify it made contact, depending on the pad material you can squish it slightly for remounting.
 
If there's no sensor(likely there isn't) then the only way to check is by taking it apart again and seeing if the thermal pads have indentations on them.
 
There was a later revision of the EVGA cards from that generation that added sensors for monitoring this (ICX, I think it's called), but if your card is one of the ones where they sent out the retrofit kit, you probably don't have it.

If you were able to reassemble it so that it works, it's probably fine, but what I would do is use Precision XOC to set a custom fan curve that keeps the fans running at all times. They don't have to run very fast - just enough to keep the air moving. You'll be amazed at how much lower the idle temperatures are with the fans at 30% vs 0.
 
You must log in or register to reply here.
Back
Top