Question EVGA GTX 1080 very high temp spikes

codewill

Prominent
Aug 5, 2020
25
0
540
I recently bought a used GTX 1080, and immediately upon booting up to windows 10 I checked MSI afterburner, and the temps were rising, only stopping at 94c. It then lowers down to around 30-40c, and intermittently has sudden spikes back up to 94c. All of this is just idling, under no load. Under any load whatsoever the card thermal throttles at 94c. Also, I should note that whenever the temperature spikes occur, the GPU utilization still remains within under 5-10%. Not sure what could be the issue, I've tried reapplying the thermal paste, reseating the heatsink, and using a different heatsink altogether. No luck thus far, here are the pc specs:
i5 8600k
GTX 1080
16gb ram
Asus z370 a
650w 80+ gold PSU
h100i watercooler
 
I recently bought a used GTX 1080, and immediately upon booting up to windows 10 I checked MSI afterburner, and the temps were rising, only stopping at 94c. It then lowers down to around 30-40c, and intermittently has sudden spikes back up to 94c. All of this is just idling, under no load. Under any load whatsoever the card thermal throttles at 94c. Also, I should note that whenever the temperature spikes occur, the GPU utilization still remains within under 5-10%. Not sure what could be the issue, I've tried reapplying the thermal paste, reseating the heatsink, and using a different heatsink altogether. No luck thus far, here are the pc specs:
i5 8600k
GTX 1080
16gb ram
Asus z370 a
650w 80+ gold PSU
h100i watercooler
tried to total clean the gpu heatsink, heatfins, and fans? How are the airflow in the case?
 

Phaaze88

Titan
Ambassador
Post some pics - use imgur.com.
-a pic of the inside of the chassis - please remove the side panel.
-take the gpu out, and take a pic of the backside without the backplate(if a backplate is present.

From the first post, it sounds like you got shipped a loose cooler...
 

codewill

Prominent
Aug 5, 2020
25
0
540
Post some pics - use imgur.com.
-a pic of the inside of the chassis - please remove the side panel.
-take the gpu out, and take a pic of the backside without the backplate(if a backplate is present.

From the first post, it sounds like you got shipped a loose cooler...
okay well good news is I took heed your suggestion about the loose heatsink, and turns out I apparently don't know how to properly install a heatsink. There was little to no contact whatsoever between the heatsink and the actual GPU. I readjusted the heatsink, tightened the screws as much as I could to ensure the maximum amount of tension between the heatsink and the GPU, and now I am not encountering the same spikes I previously was. I hit about 75c max in games, however when I run anything intensive like ethermine, my card is still thermal throttling immediately at 94c, so I imagine I am still not achieving proper contact between the heatsink and the GPU. Any particular reason this could be happening? I can provide pictures afterwards if still needed
 

Karadjgne

Titan
Ambassador
I'd not use Noctua for paste. It works excellent for cpus which have a very rough surface compared to the glass like surface of a gpu direct die. What can invariably happen is the paste tends to get run out of the edges leaving very little actually on the die, making hotspots. Gpus tend to fare better with a thicker, gummier paste such as Kryonaut/Hydronaut or Arctic MX-4.

Also check the thermal pads, make sure they still have some bounce to them, if they seem to be permanently 'crushed' they may not make good firm contact with vram/VRM's and that can create further issues when stressed under high loads.
 

Phaaze88

Titan
Ambassador
Don't over-tighten screws... when they 'stop', don't keep going, that's an easy way to warp the PCB, and cause the cooler to not mate correctly - not to mention possibly crack the PCB.

75C is a normal result for games. Can't say anything about that mining one.

I'd still like to see those pics.


NT-H2 works fine on gpus; I've been using it on a 1080Ti. Can't say the same about NT-H1.
 

codewill

Prominent
Aug 5, 2020
25
0
540
Alright, I made sure not to over-tighten any of the screws. I checked the thermal pads and they seemed fine. I'll order some MX-4 although I believe for this specific scenario the thermal paste brand is not the issue. What happened is actually the opposite due to (I believe) poor contact between the die and the heatsink. Here are pictures of the back of the PCB, and the die:
View: https://imgur.com/a/zNZ6PN3

View: https://imgur.com/JUPTkF4
 
What happened is actually the opposite due to (I believe) poor contact between the die and the heatsink.
I do think of that, because if the air blow is not as hot like what it should be then it maybe that the paste and heatsink and gpu die aren't attached and reach the contact wit the supposed pressure. Used MX4 and Gelid GC Extreme now, my gpu blows hot air much when it is stressed.
 

Phaaze88

Titan
Ambassador
Post some pics - use imgur.com.
-a pic of the inside of the chassis - please remove the side panel.
-take the gpu out, and take a pic of the backside without the backplate(if a backplate is present.
I wanted to see pics of both. That's one out of 2.


In that 2nd picture, there's a lot of paste towards the center of the die. The cooler cold plate isn't mating with the die correctly.