Question MSI 4080 suprim x high temps after repaste (multiple times)

[the0caesar]

So I saw this JayzTwoCents video about Kryosheet from Thermal Grizzly and I thought I could give it a try for CPU and GPU

So CPU was a success, but GPU was a flap, temps went very high and I assumed its because the sheet is soo thin that it's not making good contact with the cooler or it had moved from the die, I checked and tried again, same results, so I said, I will just repaste and be done with it, but surprise surprise, same high temps again, tried repasting multiple times, checked if VRAM pads had moved and caused a bad contact for the whole cooler, but they are all in the same location with the same imprints, tried different paste, same results.

I have checked the screws being fully screwed, I made sure I go X with screwing the bracket, I even loosened it at one point thinking it could be because of too much mounting pressure, but nothing worked, more paste, less paste, X application, pea size, spread paste, fans are all working, I started to suspect that the problem is with the cooler but then checked the VRAM temps and they seem ok (the same as before the repaste 60c)

anyone has any idea what could be the problem because I'm about to get Alphacool Eiswolf 2 AIO - 360mm RTX 4080 Suprim with Backplate but I'm worried it will all be the same.

 

[JohnBonhamsGhost]

anyone has any idea what could be the problem because I'm about to get Alphacool Eiswolf 2 AIO - 360mm RTX 4080 Suprim with Backplate but I'm worried it will all be the same.

you don't mention why you wanted to tear this card apart and void any possible warranty to begin with.

were temps really that bad to start? what type of range from idle to high usage?

you also don't mention the rest of the system setup.
what type of system-wide cooling do you have in place, in what case, and what are the ambient temperatures in the room/etc?

in regards to replacing it with another model;
just read multiple user reviews from multiple retailers.
if temps are higher than expected many will have mentioned it and the card will have a lower rating.

 

[the0caesar]

you don't mention why you wanted to tear this card apart and void any possible warranty to begin with.

were temps really that bad to start? what type of range from idle to high usage?

you also don't mention the rest of the system setup.
what type of system-wide cooling do you have in place, in what case, and what are the ambient temperatures in the room/etc?

in regards to replacing it with another model;
just read multiple user reviews from multiple retailers.
if temps are higher than expected many will have mentioned it and the card will have a lower rating.

Although the reason is irrelevant, but I noticed the temps creeping towards the high 60s low 70s recently in the same setup and same ambient temperature while they used to be in the low 60s and sometimes if it's lower usage it dips below 60, and I have experienced multiple gpus in the past having the original TIM drying and cracking causing higher than normal temps and big deltas between core and Hotspot temps, I was planning to repaste anyway.

To cut things short, testing superposition in static scene, temps used to be 91c Hotspot 70c core 60c VRAM @1.075v 2985Mhz

After repaste temps reach 103c Hotspot 83c core 60c VRAM @1.045v 2925Mhz

And I'm not replacing it with another model, I'm replacing the cooler with a water cooling kit from Alphacool.

 

[JohnBonhamsGhost]

I noticed the temps creeping towards the high 60s low 70s recently in the same setup and same ambient temperature while they used to be in the low 60s

wow, anything near 60° @ idle is definitely an issue with the card itself
unless you have it cooking in a closed mITX case with terrible intake/exhaust.

i definitely would've returned this thing before trying to tackle it myself.
i guess if you've got money to blow then a project like this is a good learning experience and can be a interesting project.

I have experienced multiple gpus in the past having the original TIM drying and cracking causing higher than normal temps

out of hundreds of cards that i've personally worked with, i've never seen such a thing with heavily used cards under ~5 years old at the youngest.

replacing the cooler with a water cooling kit from Alphacool

hopefully this would just prove flawed manufacturing on this particular example and allow at least normal operating temps.
also a possibility that the card has just malfunctioned and cannot properly regulate power so is burning up.

have you also tried different PCIe slots and other PCIe power cables?

sucks you've already voided the warranty so whatever the outcome you're stuck with it now.

you also don't mention the rest of the system setup.
what type of system-wide cooling do you have in place, in what case

include the complete system specs

 

[the0caesar]

wow, anything near 60° @ idle is definitely an issue with the card itself
unless you have it cooking in a closed mITX case with terrible intake/exhaust.

I'm sorry this was not clear, those temps were with load from superposition in static scene, but used to idle 30c- and now its 32c+ (ambient ranges from 25 - 27 c) I live in a hot climate

out of hundreds of cards that i've personally worked with, i've never seen such a thing with heavily used cards under ~5 years old at the youngest.

I don't about the reason but so far my past 2 cards had the same issue (Asus 2080 and Gigabyte 3080) each after a year or less (could be of high ambient temps around here)

hopefully this would just prove flawed manufacturing on this particular example and allow at least normal operating temps.
also a possibility that the card has just malfunctioned and cannot properly regulate power so is burning up.

I'm just worried I have caused something to the cooler causing in to not cool the
die properly and I'm monitoring power and it does not seem to have any problem.

have you also tried different PCIe slots and other PCIe power cables?

I did, nothing different, also reinstalled driver just to rule out anything

include the complete system specs

alright,

AMD Ryzen 7 5800x
Corsair Icue H115I Elite Capellix Liquid Cpu Cooler (top mounted as intake)
Corsair VENGEANCE RGB PRO 32GB (2x16GB) DDR4 3200MHz
ASUS ROG Strix B550-F Gaming
CORSAIR RMx Series RM850x
All enclosed in a Lian Li O11 Dynamic XL
3x120mm fans at the bottom as intake
3x120mm fans on the side as exhaust

 

[JohnBonhamsGhost]

Corsair Icue H115I Elite Capellix Liquid Cpu Cooler (top mounted as intake)

this means that all air heated by the CPU is being pumped directly back into the system.

in this position the fans should definitely be pushing out through the radiator as exhaust.

Lian Li O11 Dynamic XL

3x120mm fans at the bottom as intake
3x120mm fans on the side as exhaust

that's a pretty nice case.

though a large portion of the cool air intake from the bottom may be sucked directly out through the side exhaust without having much effect on the rest of the system.

i would try;
2x bottom intake directly below the GPU,
1x rear exhaust,
2 or 3x side intake,
radiator exhausting out of the top, mounted towards the rear.

this should provide a steady current of cooler air flowing across all components while being pulled towards the top-left and properly exhausting.

 

[the0caesar]

this means that all air heated by the CPU is being pumped directly back into the system.

in this position the fans should definitely be pushing out through the radiator as exhaust.

that's a pretty nice case.

though a large portion of the cool air intake from the bottom may be sucked directly out through the side exhaust without having much effect on the rest of the system.

i would try;
2x bottom intake directly below the GPU,
1x rear exhaust,
2 or 3x side intake,
radiator exhausting out of the top, mounted towards the rear.

this should provide a steady current of cooler air flowing across all components while being pulled towards the top-left and properly exhausting.

I only have the fan set up like that because my previous gpu was watercooled and I had a bottom 360 rad so I wanted to make sure both CPU and GPU are cooled with fresh intake air.

But I don't think not having the optimal fan arrangement can effect GPU's temp that way, can it?

 

[JohnBonhamsGhost]

I don't think not having the optimal fan arrangement can effect GPU's temp that way, can it...
Cpu Cooler (top mounted as intake)

not saying that this would be exactly what is causing your current dilemma.

but, if that heated CPU air is streaming straight down to the GPU it would definitely affect the temp somewhat.
the current bottom intakes may even blow the majority of this away though.

i was just commenting on the best possible setup system-wide.
why not have everything as cool as possible all the time? can extend life, extend comfortability around the system, affect max clocks for components with adjustability algorithms, etc.

I had a bottom 360 rad so I wanted to make sure both CPU and GPU are cooled with fresh intake air.

this would seem a good idea, and is in many scenarios, but this way you've got all the heated air from both CPU & GPU flowing directly into the system.
with very strong exhaust in the side and the rear it may still end up okay.
but, i would definitely keep an eye on the VRM, RAM, NVMe, etc temperatures with a setup like this though.

with plenty of cool air intake and both radiators exhausting there would be no worries at all though because all of the warmest air is immediately leaving the system without affecting any other components.
then you could have both side & rear as intake and probably come out better overall with system-wide temperatures.

 

[the0caesar]

I have no idea why I'm getting notified with new replies but cannot see them in the post.. Anyway just to be clear, my temps were between 58c and 64c but recently they start to creep up to 65c to 72c within a month or two (while ambient temps are almost the same 25c-27c), that's clearly an indication that there's something wrong and will continue to get worse, and I know dusting could fix such increase in temp but I already dusted my PC clean a day before but no change, and given my experience with the last 2 Nvidia GPUs I assumed it's the paste so I primitively tried to replace it, and btw it was dry as a bone, I had to scrape it with the paper tissues for a long time while breaking it loose with alcohol to a point where I thought that's not made from the same material like other pastes cause they're usually easily removed once you use alcohol. but all I can think about is one of two, either the cooler got damaged somehow and lost its ability to dissipate heat or something is wrong with the card that is causing it not to have a proper mounting pressure but I could not find anything.