Overclock.net - An Overclocking Community

Overclock.net - An Overclocking Community (https://www.overclock.net/forum/)
-   NVIDIA Cooling (https://www.overclock.net/forum/73-nvidia-cooling/)
-   -   980ti memory or vrm overheating (https://www.overclock.net/forum/73-nvidia-cooling/1714432-980ti-memory-vrm-overheating.html)

Sugita2Junko 11-27-2018 01:48 AM

980ti memory or vrm overheating
 
EVGA 980ti rebooting from overheated VRM or Memory. Fixable?

I use a Raijintek morpheus cooler on my EVGA 980ti (06G-P4-4995-KR) for the past 3 years. GPU core temp was always 50-55c while gaming, pretty cool, but recently it started hard rebooting suddenly. Discovered after troubleshooting for a long time the VRM or memory is overheating.

The 980ti doesn't have VRM temperature monitor so I placed a thermal probe on top of the GPU plate, right above one of the VRM or memory chips. Idles around 25-30c, but around 50c it would reboot, usually takes 2-3hrs of gaming, pretty sure the chip underneath is actually hotter than that. Read VRM often are rated at 100-120c. Not sure about memory.

If I limit GPU power to 60% using msi afterburner the probe temp will be around 42c, it won't reboot, could play all day. However, my 980ti performance has been reduced to that of a 970 or worse.

Has my GPU degraded beyond hope or repair? What are my options? Will replacing the thermal pad in between the plates help?

mattliston 11-27-2018 02:45 AM

dont use power limit to reduce vrm temps. That simply makes the card struggle to stay stable even more flipping back and forth with voltages and frequencies.


Use Afterburner to reduce clocks and voltage directly.



Id say a good negative 300mhz core offset with a big chunk of voltage reduction would be better than lowering the power limit. If it was a VRM temperature issue, it will "solve" the issue right then and there.


This does not rule out a component failure though, simply a temperature issue.

Desolutional 11-27-2018 03:41 AM

Have you tried running Furmark on it yet? Let that run for a few hours, it will max out the power target, might help you crash earlier than 2-3 hours.

Sugita2Junko 11-27-2018 12:25 PM

Quote:

Originally Posted by mattliston (Post 27733096)
dont use power limit to reduce vrm temps. That simply makes the card struggle to stay stable even more flipping back and forth with voltages and frequencies.


Use Afterburner to reduce clocks and voltage directly.



Id say a good negative 300mhz core offset with a big chunk of voltage reduction would be better than lowering the power limit. If it was a VRM temperature issue, it will "solve" the issue right then and there.


This does not rule out a component failure though, simply a temperature issue.

Max I can lower for core clock is -90 and memory clock -201. Core voltage the slider is at the min, seems I can only add voltage.

What component failure as in PSU or Mobo? I've replaced the PSU thinking it was at fault but turns out didn't fix it.

Sugita2Junko 11-27-2018 12:27 PM

Quote:

Originally Posted by Desolutional (Post 27733122)
Have you tried running Furmark on it yet? Let that run for a few hours, it will max out the power target, might help you crash earlier than 2-3 hours.

Furmark last about 5 seconds if it has already rebooted prior and still hot.

From normal/cold state it would die around 15-20m

The Pook 11-27-2018 12:36 PM

Don't use Furmark and stop suggesting people use it. It serves 0 purpose and most cards will cripple themselves to a lower P state if it detects it's running. Just run a benchmark if you want to test things out.

@Sugita2Junko : Where did you put the temp probe on? You said the "VRM" but the VRM consists of chokes/coils/inductors/capacitors/MOSFETs/voltage regulators and not all of them really put out that much heat or need a heatsink. The only one that should have one are the mosfets.

Sugita2Junko 11-27-2018 12:42 PM

Quote:

Originally Posted by The Pook (Post 27733888)
Don't use Furmark and stop suggesting people use it. It serves 0 purpose and most cards will cripple themselves to a lower P state if it detects it's running. Just run a benchmark if you want to test things out.

@Sugita2Junko : Where did you put the temp probe on? You said the "VRM" but the VRM consists of chokes/coils/inductors/capacitors/MOSFETs/voltage regulators and not all of them really put out that much heat or need a heatsink. The only one that should have one are the mosfets.

http://images.hardwarecanucks.com/im...X-980-TI-1.PNG

It has a back & front plate covering. I just put the probe on the front plate which I felt was hottest, so hot I can't leave my finger for more than a sec. In between the expose mosfet (square things?) & gpu core.

The Pook 11-27-2018 12:50 PM

1 Attachment(s)
The bigger gray square "things" in that picture are chokes and they don't need a heatsink.

The partly melted thing in this picture is a mosfet and what you should be checking the temps of. If it doesn't have a heatsink, stop using the card until you get some on it.

Sugita2Junko 11-27-2018 01:07 PM

Quote:

Originally Posted by The Pook (Post 27733900)
The bigger gray square "things" in that picture are chokes and they don't need a heatsink.

The partly melted thing in this picture is a mosfet and what you should be checking the temps of. If it doesn't have a heatsink, stop using the card until you get some on it.

https://i.imgur.com/GmS9JK7.jpg

These are the areas too hot to touch for me. I circled it green. I only touched the plate. Probe location is in between there in red circle.

https://www.fudzilla.com/images/stor...ti_sc_pcb1.jpg

Those square chips surrounding the GPU core are VRAM?

https://qph.fs.quoracdn.net/main-qim...838524ae8118-c

The VRM when I touched it was scorching hot too. Didn't put a probe on it to know how hot.

8051 11-27-2018 01:39 PM

The green areas look like where the VRAM is located.

What kind of fan do you have blowing down on the VRM side of the card? You might want to consider putting a more powerful shrouded fan over that location.


All times are GMT -7. The time now is 08:16 AM.

Powered by vBulletin® Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.

User Alert System provided by Advanced User Tagging (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.
vBulletin Security provided by vBSecurity (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.