EVGA RTX2070 Super overheating!!! Need help :-)

[yannis76]

Hi,

Since September I have been experiencing PC shutdowns with fans at 100% whenever I try to play a game with graphics. I don't play very often, so I haven't been noticing so much. In the beginning I thought it was Troy, but it seems it is every 3D game! (https://forums.overclockers.co.uk/threads/tw-troy-keeps-crashing-my-pc-hard.18937429/)

The card is the EVGA 2070GTX Super and temperature reaches 86oC, works there for about 1-2min and then PC shutdowns with the fans at 100%. CPU temperature is around 50oC. Of course if I leave the case side windows open, pc works fine.

Case is the Corsair 280X with two fans in front entering air and a H115 radiator also entering air with its two fans (based on corsair advise). Graphics card is at the bottom of the case.

Any ideas? Maybe de-assembly the card and reapply thermal paste? Adding on more fan at the bottom? Thing is that that didn't happen in the past, so I feel something is failing on the cooling part of the card.

Thanks

 

[HazardO]

It shouldn't shut down at 86C. My first suspect would be the PSU, what make/model is it?

 

[yannis76]

Its a new Corsair RM750x. How can I test it?

 

[Game]

ran my amd r9 at 85 ish for 4 years no issues

it must be something else

do you have any exhausts 5 fans in seems a little over the top to me

you might end up with the air not going anywhere in a purposeful way

can you monitor the motherboard temps

oh and do you have filters>? they could all be clogged with dust

 

[willhub]

My 2070 laptop version still operating at 90 degrees, highly doubt it's an overheating issue.

 

[Aretak]

Whilst it shouldn't be outright shutting down at that temperature, that's still completely abnormal and well into thermal throttling territory. You may have two seperate issues. For the overheating problem, an easy way to check if it's a case airflow problem is to just take the side panel off and see how that fares. It won't be an ideal airflow setup, but it will certainly eliminate any hotbox effect. Other than that, repasting would be the next step.

 

[pc-guy]

For that 2070S to hit 86c it isn’t exactly normal. Those 2070s should be operating between 65-75c even in very poorly ventilated cases.

in itx where everything is next to each other then it will be a bit different.

start your stress test on each of your components to eliminate issues. Having the case open doesn’t automatically mean it is the GPU over heating.

 

[opethdisciple]

Do you have any thing overclocked in the PC?

If so put it all back to stock.

 

[yannis76]

Whilst it shouldn't be outright shutting down at that temperature, that's still completely abnormal and well into thermal throttling territory. You may have two seperate issues. For the overheating problem, an easy way to check if it's a case airflow problem is to just take the side panel off and see how that fares. It won't be an ideal airflow setup, but it will certainly eliminate any hotbox effect. Other than that, repasting would be the next step.

Thanks. I wrote also in my initial post. PC works fine when side panel is off.

 

[yannis76]

Do you have any thing overclocked in the PC?

If so put it all back to stock.

No overclocks

 

[yannis76]

For that 2070S to hit 86c it isn’t exactly normal. Those 2070s should be operating between 65-75c even in very poorly ventilated cases.

in itx where everything is next to each other then it will be a bit different.

start your stress test on each of your components to eliminate issues. Having the case open doesn’t automatically mean it is the GPU over heating.

I have done some stress tests like CPU and GPU. CPU doesn't increase temperature that much, around 50-55oC with the Radiator at full load. Plus it works fine with case closed at LR or anything that doesn't stress the graphics card.

 

[pc-guy]

Downland 3Dmark they have CPU and GPU tests I believe. Download furmark to see if there are any thermal issues with the card.

get GPUz to see what your core temp and vram temps are doing etc

get hwinfo64 to monitor your entire system especially your CPU/MOSFET/package power etc etc

 

[raralis]

As mentioned above, it's a process of elimination till you find out what's causing the issue.

I would try to undervolt by 15% to 20% and check that your FPS is capped at 60 or 120/144 fps depending on your monitor.

This should at least let you play a game without any major issues and you temps should be lower. Failing that, your thermal pads might need replacing/readjusting.

 

[subbytna]

Check the screws that are pushing the cold plate, of the cooler, on to the GPU die itself.

 

[yannis76]

Downland 3Dmark they have CPU and GPU tests I believe. Download furmark to see if there are any thermal issues with the card.

get GPUz to see what your core temp and vram temps are doing etc

get hwinfo64 to monitor your entire system especially your CPU/MOSFET/package power etc etc

Ok I ran Furmark with case open (default 1280x720) and max temp was 83oC with 97-100% GPU usage. When I closed the case with my hand for a few minutes, temperature raised to 86oC and when I stopped the test, it dropped very fast to around 60oC and later to 47oC.
Putting the front fans at full speed didn't make any change at all!!!
Left glass was very warm near the card

 

[pc-guy]

Ok I ran Furmark with case open (default 1280x720) and max temp was 83oC with 97-100% GPU usage. When I closed the case with my hand for a few minutes, temperature raised to 86oC and when I stopped the test, it dropped very fast to around 60oC and later to 47oC.
Putting the front fans at full speed didn't make any change at all!!!
Left glass was very warm near the card

Try repaste the GPU. Could be something up with the thermal paste.

Also have you tried DDU and reinstall driver?

 

[yannis76]

Try repaste the GPU. Could be something up with the thermal paste.

Also have you tried DDU and reinstall driver?

Hi I just repasted the GPU with Arctic MX4 thermal paste, just a drop as I do with the CPU and temperatures are higher!!!! I now reach more than 90oC (now 94oC!!!) with open case and lower ambient temps!! What did I do wrong??? Maybe I need more paste?? When I removed the heatsink there was much more paste there than the one I applied...

Also what is the DDU?

Thanks

 

[pc-guy]

DDU is display driver uninstaller. It is a special utility programme to remove display driver so you can install driver cleanly. Needs to be run under safe mode with admin access.

have you got some small washer that can go under the screws on the back plate? Sound like you may need some additional tensioning of the screws to get the heatsink to get proper contact with the core. Also did you remove thermal pads on vram chips at all and replace with new ones. Incorrect pad thickness can lead to improper contact with the core which lead to higher temps

 

[yannis76]

DDU is display driver uninstaller. It is a special utility programme to remove display driver so you can install driver cleanly. Needs to be run under safe mode with admin access.

have you got some small washer that can go under the screws on the back plate? Sound like you may need some additional tensioning of the screws to get the heatsink to get proper contact with the core. Also did you remove thermal pads on vram chips at all and replace with new ones. Incorrect pad thickness can lead to improper contact with the core which lead to higher temps

Hi there are some good news. The good news are that I had forgotten to put the fans cable, fans weren’t running and that is why temperatures were so high. Ran Furmark with case open and temperature didn't rise above 76oC which is much better than before... Now need to test with case closed...

 

[syks]

The temps do seem quite high but im sure a 2070 shouldnt thermal throttle until you hit 90/95 degrees, i have a rtx 2070 (not super) and temp is mostly around 70 to 76 degrees(depending on game) when fully stressed, normal working temps for a 2070 are between 70 and 80 degrees