Strix 1080 ti OC Crashing Unless I have Fan Manual Full Blast 100%

[CpCrunch49]

I need some help from the Community, I think have a Lemon Asus Strix 1080 ti OC 11gb ?


This is crazy,
I have done extensive trouble shooting and Found the Cause.
its a combination of High GPU usage and Power 90 to 100 % and have the Fan Curve Default.
Basically, OC mode, gaming Mode and event Silent Mode will crash with the Default Fan curve in Asus GPU tweak, in a Open Case setup , Open air flow pretty much.


1080p Fortine crashes around 60c. You can see the gpu get 100% and then drop.
Forza 3 4k crashes at 47 c. Same thing.

I can get fortine 1080p running ok with FULL manual fan 100% or a super aggressive Fan Curve. maxing out fan around 45 to 50c.

Forza 3 4k,, not even the full FAN can make it stable. I have to add external house fan blowing at the whole PC, to get it to 43c , then it runs ok.


I tough this cards can get up to 80c ?
I disable the GPU fan, in gpu tweak.. let the PC idle.. GPU temps kept climbing and climbing till it reach 81c. i got scared and turn on the fan again.
like wtf. is the heat sink not enough to hold the temp at idle ?




My Specs
Node 202 case: shell only. Open air. for testing.
SF Corsair 600w PSU
i7 7700k,, NO over clock . CPU temps around 70c.. when gaming. CPU not at 100%
16gb Corsair Vengance 3200, no xmp enable, default bios settings.
m.2 256 Kingstone Drive.
1 usb xbox RF receiver.
Asus Strix 1080 ti OC 11gb
Lates Nvidia drviers. from Nvidia website.
win10 Pro 1803


If I RMA with Asus or BHphoto vendor ? will the swap it out with another feburbish unit. or probably just Re-do the thermal paste on my own unit and send it back ?

This is frustating, I remmeber My card would be stable at around 75 to 80c. with the case cover on. Only had it for about 5 months.


Thx in advance for the help
CpCrunch.

[Silent_Scone]

Hi, I would install a dummy OS / reinstall current OS and test the GPU there. Alternatively, test the GPU in another system if possible.

Failing that, return the card to the retailer and ask them to test the card for you before processing an RMA.

[CpCrunch49]

Unfortunately the Retailer will not take it back,, their cut out time its 30 days.
After that its Asus Warranty Pretty much,

I'm going to Replace the Thermal Paste and hope for the best. .if NOT then, I'll have to RMA

I'll post results later.,

[Vlada011]

I believe changing thermal paste maybe is not solution if no overheating.
And if they make you problem because you remove cooler and change thermal material...???

Your GPU is unstable on 45-60C.
I would rather downlclock on exactly FE specification, (default 1080Ti frequency) if that stop crashing card is fault 100%.

First move should be installation of previous working and older driver and downclock GPU on specifications set by NVIDIA for same graphic processor.

[CpCrunch49]

I believe Asus changed their policy wear they now allow changing Thermal Paste .

Yeah. I can downclock the card and the problem goes away.
Or I can Max out the Fan 100% and Play for Hours.

But can play 1080p Fortnite on Game mode with default Fan curve, because it crashies around 60c.
I then simply Max out the Fan.. Re-launch the same game. same settings. and No crashes.. Temps around 40c with fan Maxed out.

IDK, If the temp sensors are screw up.. or the Transistors are BAD and just cant handle even low temps.
This has to be a bad Card.

I need a friends System to test it on

[Silent_Scone]

Have you tried on a fresh operating system as suggested?

[CpCrunch49]

I did not try a Fresh OS, because I'm fairly confident thats not the issue.

I was able Directly relate the Crash to Temps. All same, I simply max out the Fan, and NO crash.. change the Fan to Default and it crashes soon as Temps starts rising.
I can reproduce the Crash, the issue is replicatable. I been a tech for 10+ years.
THis is not a driver PSU or OS issue.

UPDATE:
So I took apart GPU, removed the SInk, TO my Surpirse , sure enough, 1/3 of the GPU thermal paste was gone or completely dried out.

This corner of the Chip was Toasting causing the crashes.. While the rest of the CHIP operating at normal Temps.
This explains why My Temp readings were low or Normal. Yet still F$%#$% crashing.

where do they place this Thermal Censors ? If the temps of that corner of the chip wear properly reported , I would have known right away if was a bad TIM issue or Part of the Chip Overheating causing the crashes, Or uneven temps ?

So the card thinks temps are ok around 60c. so it never throttles, it never ramps the fan. etc. Yet 1/3 of the chip is Frying.. so it fu$%% crashes.
I hope all those crashes did not cause any type of permanent damage on my 900$ chip.. all because its not BAD TIM application from the Factory,

I re-apply new thermal Paste, Closed it up. fire it up.. same test... OC mode, default fan, Fortine 1440p. 99% gpu load, 72C steady.. no crashes .

are you serious. Can they put better Thermal sensors on this chips, damm, that pissed me off. put 1 thermal sensor on every corner of the CHIP,
Now I have a potentially Permanent partially damaged Chip ? at 5 months old

[Vlada011]

If card crash on fabric clock and stop to crash after downclock she is unstable 100% and ready to RMA.
Except if card have overheating and temps cause that. But's even high temp will rather downclock than crash.
You are on edge and wall of your graphic processor.

I had such card, EVGA asked what happen I show video clip, she crashed in some games and 3D Mark Sky Diver in moment when it's inside cave and show one pyramid up side down.
That was worse moment for unstable cards, and I saved video clip, immediately when they saw that card crash to desktop with driver message on default and pass on 50MHz downlock immediately recommend RMA and sent me other stable. One video clip described everything.