Hi All,
 
So I am one of the unlucky ones to get a new 1080 Ti FTW3 with either a faulty GPU temperature sensor or a warped heat-sink due to the decrease in quality control it seems. My RMA was denied because my card is performing "within spec" according to the rep that looked at my screenshots I provided. All three of my fans are running full blast and my temperature is getting 84-85c all the time no matter what even with the case side-panel opened up with a fan blowing at it. 
 
I have already replaced the thermal paste with a better paste (Noctua NH-T1) and the temperature did not change at all. Now I read in some threads people were having the same issue and of course their replacement cards were getting 20c lower temperatures or so. Of course yesterday my game was crashing and freezing up my computer while I see the temperature spiked to 90-95c which I just mentioned to the rep through the ticket but I find it very disappointing that my initial RMA request was denied and the card is performing within spec while sounding like a rocket taking off inside my case.
 
Any advice or is there someone else that can help me from eVGA on this matter?

legcramp89
Hi All,
 
So I am one of the unlucky ones to get a new 1080 Ti FTW3 with either a faulty GPU temperature sensor or a warped heat-sink due to the decrease in quality control it seems. My RMA was denied because my card is performing "within spec" according to the rep that looked at my screenshots I provided. All three of my fans are running full blast and my temperature is getting 84-85c all the time no matter what even with the case side-panel opened up with a fan blowing at it. 

(Assuming your case has good air flow and your ambient-temperature isn't too high)
 
I could post about 5/7 links all pointing to other 1080ti FTW3 users all getting much...MUCH lower temperatures then what your getting, and yes their cards are in closed cases with good air flow and normal ambient temperatures.
 
The issue you will run into on these forums is people immediately telling you not to look at reviewer samples like Jaystwocents etc because their cards are in an open air environment and his air conditioning is cranked. But I bet if you crank your air conditioning and yank the side off your case and put an industrial size fan pointing towards it you would still be in the high 70's...which obviously isn't normal in comparison.
 
The reason I bring all this up is because I myself am going through something very similar, mind you my temps aren't as severe as yours are. I'm on my second FTW3 now because my first FTW3 that I bough new in a retail box turned out not to be stable all of the time at stock clocks.
 
The thing is, that first card never ran past 61c at 55% fan speed with an ambient temperature ranging from 23/24c. My first card lived up to the hype, it just turned out to be more of a FTW3 "DT" model with lower clocks. I figured I paid all that money, I want a card stable out of the box, so I RMA'ed my first card.
 
Now onto my second card, it runs at 78c at 55% fan speed with no OC, but...It is actually stable! I have yet to have 1 driver crash in all the same areas in Star-Wars Battlefront that my other card would have crashed on. But the fact remains that my second card runs a full 16c hotter then my first card did with the exact same settings.
 
And of course when I posted my concerns about this I had people jumping on the "open air/ambient-temperature" bandwagon, not actually considering that my previous card in the EXACT SAME CONDITIONS ran a full 16c cooler...noooo, it's my case or room temps (that never changed!).  
 
I've already changed the paste to MX-4 and seen no temp drops what so ever. I recently purchased some Thermal Grizzly Kryonaut (the non conductive stuff), will see if that brings the temps down but I doubt it. If I OC this second card it needs 80% fan speed in order to stay below 80c, consider that 80% fan speed would have had my first FTW3 running in the high 50's!!!
 
I think it comes down to the silicone/cooler lottery, you may have a super leaky chip, or you may have a cooler with a not so perfectly flat base. Either way your temps are even worse then mine and quite franky how is your card any better then a founders addition with an aggressive fan profile? And if you need to crank your fans to 90%+ all the time then I think that should be grounds for RMA.
 
 
 
 
 
 
 
 
   

Moving thread to warranty questions subsection.

Sajin
So your rma has been approved now right?

I have not received a response yet from the service rep (response time seems to be two days unless I bug them asking if there is any updates?) after finding out my system will crash with a black screen + fan speed went full blast sometimes and I can see the temperature hit 94-95 one time while playing Overwatch. 99% of the other times it will hover around 84-85c even if I set the fan curve to full or aggressive mode. I am actually really happy that the card ended up causing crashes after playing for a while because the only other solution to throw an EK block on the 1080 Ti if I can't get a replacement due to temps alone. The whole point of getting this card was to step-away from water-cooling the GPU because the air-cooling on this card is supposedly top-notch. 
 
It is definitely not my case or airflow, I have already ruled that out by putting the case fans on max (triple 120mm in front, 140mm rear, and 2x 140mm at the top). I also took off the side panel, blasted my room's AC to 60c, and pointed a large fan blowing cold air directly to the video card and no change in temperature at all. 

I'll give that a try later today. I also want to mention all my temperatures beside the GPU core are between 50-65c. 

Well I just had similar results as the OP, I changed my thermal paste again from MX-4 to Thermal Grizzly Kryonaut and I went from 74c at 70% fan speed down to 67c at 64% fan speed.
 
I cant believe the difference, I remember hearing somewhere that MX-4 wasn't the best for direct die applications and seeings as I just seen a -7c drop (with even less fan speed) by moving to Kryonaut, that seems to be the case.
 
All in all it's still a tad annoying having to change the thermal compound on a 1000+ dollar card, but in the end I'm just happy to have reasonable temps again.
 
The OP going from 85c down to 65c is incredible, seems these 1080ti's are very sensitive to paste quality and application methods. 
  
 

bcavnaugh

Stackermodder777
Well I just had similar results as the OP, I changed my thermal paste again from MX-4 to Thermal Grizzly Kryonaut and I went from 74c at 70% fan speed down to 67c at 64% fan speed.
 
I cant believe the difference, I remember hearing somewhere that MX-4 wasn't the best for direct die applications and seeings as I just seen a -7c drop (with even less fan speed) by moving to Kryonaut, that seems to be the case.
 
All in all it's still a tad annoying having to change the thermal compound on a 1000+ dollar card, but in the end I'm just happy to have reasonable temps again.
 
The OP going from 85c down to 65c is incredible, seems these 1080ti are sensitive to paste quality and application methods more so then ever. 

Funny because I went from MX-4 to Grizzly Kryonaut and my temps went up 3C returned back to MX-4 and went back down 4C.
It hard to tell now a days because each time you do you well not get the same reading.

It seems a big part in it all too might just be the luck of how the paste spreads itself as you're tightening the heat sink screws down. Your results from what I've just experienced and all the reviews I read on Kryonaut is a bit puzzling though, maybe you just got a bad batch or something, who knows. 

imjune
Hi I've read this thread since I also got temp issue with my card. But it's only a cheap one EVGA GTX 1060 6GB single fan. It used to be runnning max around 70-75 C during full load, which I consider normal for a single fan GPU. But lately (maybe a couple of weeks now) it gets hotter and hotter. And right about yesterday I run 3D Mark and Unigine Valley, and the card maxed out at 94 C with almost 100% fan speed. I've mailed EVGA regarding this matter and as usual, they're still on the airflow bandwagon, which is not my issue at all (I'm using an open air case). I actually want to try to replace the old thermal paste using Thermal Grizzly, but I found one of the screw is having an EVGA sticker, which I assume will void the warranty if removed. But I see you guys are replacing your thermal paste without losing your warranty. What should I do?

Try replacing the thermal paste. Breaking/removing the sticker that says evga all over it won't void the warranty.

I am bumping my own thread since I am about to sell my trusty 1080 Ti FTW3 which initially had the high temperature issue. The actual issue is simply not just re-seating the heatsink or re-applying the thermal paste. The issue is the four screws that seats the cold plate to the GPU will have some play to it either out of the box like I did or over-time. The fix to this issue is to add some plastic/nylon spacers to these four screws or like what I did... I doubled up on the springs because I didn't have any spacers on-hand. 
 
Default fan curve out of the box I am hitting 72c max running superposition benchmark with my boost speed locked around 1960Mhz. With max fan speed, my GPU temp never goes above 57c in the same benchmark. The initial "fix" I found was just re-seating and re-applying but that one lucky time fixed my issue. Now that I removed my waterblock and put back the original heatsink fan I ran into the issue a couple times but adding the springs/spacers 100% fixes the issue. Hope this helps anyone that runs into this issue in the future... I am sure this not only applies to 1080Ti FTW3. 
 
Forgot to mention that the first reseat without the springs I got 84c and the boost clocks were much lower due to the higher temps... if anyone is seeing these kind of temps even at 100% fan speed this fix will probably work for you. Never settle for high temps especially if you're seeing it using a high fan speed.