Is my card on a timer? GTX 1080 FE total shutdown, not PSU, all temps below 73c

[airsickace]

I built a new computer last week. It won't stay running during gaming or a GPU stress test. Like clockwork, about 20 minutes into gaming (PUBG's and CS:GO) or 26-30 minutes into the GPU stress test (Furmark) the computer hard crashes into an off state instantly. The computer ONLY crashes during gaming or GPU stress testing, it is guaranteed to happen around 30 minutes in.

I thought this was a classic case of a bad PSU, but I've tried 3 different PSU's now and the crash still occurs. I've also tested other components with stress tests and monitored temperatures the entire time. The weirdest part is that I can crank the fans during Furmark to 90% and the max temp of the GPU will reach 73c and the test will run stable for over 4 hours. If I leave the fans on auto they reach a max of 55% and the temp reaches 83c and the crash happens around 30 minutes in. As far as I am aware 83c is a safe temp for the GTX 1080 FE especially for only 30 minutes. Even after returning the GTX 1080 for a identical brand new one, the crash still occurs. My other components stay under 65c.

Temps when crash occurs
-CPU -60-65c
-GPU 73c during gaming, 83c during stress test
-ambient 32c
-ssd 41c

PC specs
-Windows 10 64bit home
-i7 7700k
-Asus Maximus ix code
-Samsung m.2 960pro
-Nvidia GTX 1080 Founders Edition
-Gskill TridentZ 3200mhz
-EVGA 850 BQ (brand new)

Things I have tried
-reinstall windows
-reinstall drivers
-updated bios
-Whocrashed returned nvlddmkm.sys error before I reinstalled windows and that seems like a pretty common error even when the PSU is the cause. There are no dump logs after the windows reinstall even with the crash happening about 15 times.
-3 different power supplies (1200w, 850w, 750w)
-re-seating all the components and motherboard
-prime95 cpu test for 4 hours stable
-memtest86 test for 4 hours no errors
-returning my GPU to the store for a brand new 1080
-running gpu test with only 1 stick of ram, interchanged in case a stick was bad
-contacted Nvidia, they took down my system info and said they would send me an e-mail

Sorry for all the info I just wanted to be as thorough as possible. Any help or ideas are appreciated.

 

[Dark Lord of Tech]

Just so you know the 850w BQ is a poor power unstable supply made by HEC.

 

[airsickace]

Just so you know the 850w BQ is a poor power unstable supply made by HEC.

Thanks for the heads up. It's brand new so it should work for now, and judging by the other PSU's doing the exact same thing I can rule this one out.

 

[drakehotbling]

hey , im having the same problem right now , how did you fix please help me

 

[airsickace]

hey , im having the same problem right now , how did you fix please help me

Turned out to be a temperature issue for me. Whatever you are doing, even if you think you have good air flows, put a box fan on the thing and see if it crashes. The problem also wasn't with the graphics card, because I have since gotten a new one, and it would still happen. I believe it is with my mobo, but since I have tweaked my fans and stuff, it doesn't happen anymore.