MSI 1080 ti possibly dying?

[Lorec]

I have a used Founders Edition card with a EK full cover water block and EK backplate on it.
Card is on same loop as my CPU (360 Rad EK performance edition).
Recently I run my CPU on base clock to lower temps and since then I noticed occasional black outs.
PC would shutdown with power left in mobo (rgb on mobo and ram on).

I noticed it would happen when gaming, so I cranked up settings in GTAV to the max and run the game.
In game I was monitoring GPU core temp which was at ~63C when my PC shutdown.
I opened front cover and touched back of my card, it was searing hot.

I think it could be vram temps as it happens specifically when I fill vram in settings.
In my GTAV test, I set it to use 6Gb vram (of 11Gb available).

I have this waterblock since October last year.
Other info in my system specs.

Is my 1080 ti dying?

Ideas? I think I will just disassemble the loop, check the contact of T pads with vram.
Anything in particular to look for when I do so?

EDIT: GPU-Z Screenshot

 

[Vya Domus]

I can't possible see how the VRAM would overheat with a full block, GDDR5X dissipates very little heat.

 

[jsfitz54]

I have a used Founders Edition card

Does used mean a mining card?

Is it the stock Bios for that card?

Re-paste chip?

 

[xtreemchaos]

my pc was doing the same a while ago but with a 1080, id be playing a game and it would hard reboot, it turned out it was the psu. new psu and its fixed.i also had problem with my waterpump speeding up and down for no reason and the new psu fixed that too maybe it was a early sign that there was something up with my psu voltage.

 

[Lorec]

Does used mean a mining card?
Is it the stock Bios for that card?
Re-paste chip?

Added GPUZ screenshot.
Bios is stock,
I bought the card in Oct 2018 from a used PC parts shop.
Its impossible to tell for sure if it was mined. When buying it was pristine.
I will try repasting!

my pc was doing the same a while ago but with a 1080, id be playing a game and it would hard reboot, it turned out it was the psu. new psu and its fixed.i also had problem with my waterpump speeding up and down for no reason and the new psu fixed that too maybe it was a early sign that there was something up with my psu voltage.

The PSU is new though! Released Q4 2018, I bought it unopened from used PC parts store last year!

 

[xtreemchaos]

might be worth it to try a diff psu if you have one handy just to rule it out, i took the thermtake 600w out my intel build just to try and it didnt miss a beat. my psu was 4 years old so getting on a bit.

 

[Vya Domus]

63C seems unusually high for a 1080ti under water with a 360 rad though.

 

[jsfitz54]

I will try repasting!

Make sure that it's a non electrically conductive paste. (clean the chip area well to make sure you didn't cause a problem when you put the waterblock on)

Don't mix paste with thermal tape, the grease eats the tape.

 

[lsevald]

I rarely see above 35°C on mine at 2100MHz/1.200Vgpu

 

[Lorec]

might be worth it to try a diff psu if you have one handy just to rule it out, i took the thermtake 600w out my intel build just to try and it didnt miss a beat. my psu was 4 years old so getting on a bit.

I do have, will try it out first.

63C seems unusually high for a 1080ti under water with a 360 rad though.

Its was under full load though! 99% usage!
Also its sharing the coolant with a cpu so I assumed its just what it is.

Make sure that it's a non electrically conductive paste. (clean the chip area well to make sure you didn't cause a problem when you put the waterblock on)
Don't mix paste with thermal tape, the grease eats the tape.

I use Arctic MX-4!

I rarely see above 35°C on mine at 2100MHz/1.200Vgpu

View attachment 156870

Jesus christ... but is it because You have two 360 rads?

I just checked unboxings on YT, waterblock doesn't have plastic sticker(happens every now and then)
I might have overtightened the screws or something...

 

[Vayra86]

Case airflow, perhaps? Reduction on CPU load = rad is not pushing or pulling as fast as it used to. Is the CPU rad an intake or exhaust?

 

[lsevald]

Jesus christ... but is it because You have two 360 rads?

I'm sure it helps a little, what are your temps when just starting the computer and the coolant is still cold? When I first mounted my GPU water block,one of the mounting studs wasn't screwed all the way in from the factory (1.5turn left), and I got results similar to yours...

 

[Lorec]

Case airflow, perhaps? Reduction on CPU load = rad is not pushing or pulling as fast as it used to. Is the CPU rad an intake or exhaust?

Could be. Inwin 303 is kinda hotbox.
360 rad is topside and blows hot air through PSU on one side.
Ive set fans at 1000rpms and to jump ~1800 @75C which doesnt happen often.

I'm sure it helps a little, what are your temps when just starting the computer and the coolant is still cold? When I first mounted my GPU water block,one of the mounting studs wasn't screwed all the way in from the factory (1.5turn left), and I got results similar to yours...

Around 35-38C on GPU, so basically Your temps at full load... its stock too...

 

[Vya Domus]

Its was under full load though! 99% usage!
Also its sharing the coolant with a cpu so I assumed its just what it is.

I know but it still seems a lot. A 360 rad will dissipate upwards of 600-700W, you have plenty cooling, 63C for a 1080ti just seems a lot. I have a 1080 with a 120 rad and it barely touches on 50c, of course there is no CPU and it's a 200W card but still, you have roughly 3 times the cooling.

 

[Regeneration]

Why do you assume the card is dying? Perhaps wrong configuration in the BIOS, VRM overheating, PSU failing / overheating.

 

[lsevald]

Around 35-38C on GPU, so basically Your temps at full load... its stock too...

My cold, idle GPU temp is 21°C, so a deltaT of 33-21=12°C. Same method to your temps: 63-36.5=26.5°C, and you got overclocked vs stock also.

Edit: stock xoc bios clocks&volt, temps:

 

[nguyen]

Yeah your thermal paste probably gone bad, 63C is quite high for watercooling.
Buy a new tube of thermal paste, re-apply the thermal paste and hope that it fix the problem , oh and take some picture when you take out the WB too.

 

[Caring1]

Check for hotspots on the card, if you don't have an infra red heat gun, use a kitchen thermometer as used in ovens, or you could risk using your finger and burning it.
You mentioned conductive tape on the vRam, but what about the power circuitry?
That section is more likely to get hotter.

 

[ruff0r]

My cold, idle GPU temp is 21°C, so a deltaT of 33-21=12°C. Same method to your temps: 63-36.5=26.5°C, and you got overclocked vs stock also.

Edit: stock xoc bios clocks&volt, temps:

View attachment 156894

idle 26c° load 41c° shared with cpu loop, 63c° way to high.

 

[Caring1]

This thread - https://www.techpowerup.com/forums/threads/asus-5700xt-strix-adventures.267566/
Indicates their issue was the EK water block. Yes it was on a different card.

 

[Lorec]

Why do you assume the card is dying? Perhaps wrong configuration in the BIOS, VRM overheating, PSU failing / overheating.

My 360 rad exhaust partially goes through PSU. I have postitioned so it helps remove hot air.
360 rad exhaust -> PSU intake.
Otherwise, PSU would block 1/3 of exhaust and heat up from backside.
Could it shorten PSU lifespan?

Check for hotspots on the card, if you don't have an infra red heat gun, use a kitchen thermometer as used in ovens, or you could risk using your finger and burning it.
You mentioned conductive tape on the vRam, but what about the power circuitry?
That section is more likely to get hotter.

This thread - https://www.techpowerup.com/forums/threads/asus-5700xt-strix-adventures.267566/
Indicates their issue was the EK water block. Yes it was on a different card.

EK waterblock for 1080ti has at least 5 standoffs and nothing was moving really,
though this post gave me an idea
Might have found the culprit.


GPU stand. Once I removed it, I got 10C less in GTAV under same settings.
The way how it was holding the card, it could`ve warped something.

Last night`s GTAV tests gave me serious hot spots around vram, now its the vrm thats been hotter.
Also, my PC didnt shut down during gameplay so thats a plus.

I will disassemble today and recheck contacts/repaste etc.

Thanks to all for awesome ideas.

EDIT:
When i play GTAV for couple minutes.
Also I noticed that my card developed coil whine, reacting to graphics under load *sigh*

 

[lsevald]

Those Superposition scores seems fine. Ran the same presets on mine @xoc bios stock clocks&volts, which I believe are slightly higher than yours (1999 vs 1873MHz). But I still think temps can be improved (down to 45-50°C), but it might not be worth the trouble?

 

[Vya Domus]

How can you tell there are hotspots on the VRAM and VRMs ? That card doesn't have sensors for those things.

 

[Lorec]

@nguyen

Kinda looks like bottom part (about 1/5) was not having contact properly.

Anyhow, repasted my card and screwed it back carefully.
My after boot temps improved, now its ~24C

Those Superposition scores seems fine. Ran the same presets on mine @xoc bios stock clocks&volts, which I believe are slightly higher than yours (1999 vs 1873MHz). But I still think temps can be improved (down to 45-50°C), but it might not be worth the trouble?

It sure was worth it to spend couple hours resetting the loop (cleaned waterblocks and all)

Max temp 40C sooo worth!

How can you tell there are hotspots on the VRAM and VRMs ? That card doesn't have sensors for those things.

I literally touched it and it burned my finger. oh it did.
Maybe I shouldn't have written it in such a dramatic fashion haha

Anyhow, this repair was a success.
Thanks a bunch guys!

 

[Vya Domus]

I literally touched it and it burned my finger. oh it did.
Maybe I shouldn't have written it in such a dramatic fashion haha

That's an inaccurate way of determining these sort of things, VRAM chips can sits at 80C and VRMs at 100C just fine. They are going to burn your finger.