Question Hard System Crash with RTX 2080TI

ronin026

Prominent
Aug 26, 2017
26
0
530
0
Hi there,

Recently updated to a ROG-STRIX RTX 2080TI, everything went fine for the first couple of weeks. I stress tested everything under load, ran Furmark as well as a silverlight CPU simultaneously and everything seemed stables. Then recently I began getting system crashes, specifically it shuts off the PC while under load. No BSOD or error messages, no tearing or artificating, just shuts off. I have checked all temps and nothing is overheating. It's well cooled, with Corsair H100i Pro for GPU and 3 fans equipped on the 2080ti to air cool the GPU, 4x 200mm case fans to help. GPU temps were not rising above 67c and GPU was running in the 70s. I can not even get the system to heat up under stress tests before it crashes now. I just stress tested the system prior to this post, running silverlight and Furmark and it crashed within 5 secs, well before anything got enough heat to cause any issues. I read that it may be a PSU going bad, which could be the case as the PSU on this computer is about 5-6 years old, but I'm running a Thermaltake 850w PSU so it's certainly not suffering from lack of power.

I don't have any logs or error codes to post as its completely powering off the system. I still have my old GTX 1070ti to use, and am about to replace the GPU and run stress tests with that to see if I can recreate the problem. The system was completely stable since I upgraded it for about 6 months prior to switching GPUs, and was stable for several weeks after upgrading GPU's before I ran into this problem.

Any help would be much appreciated.

TL;DR : System hard crashing under load, well before any components heat up enough to be the culprit. Please help.
 

ronin026

Prominent
Aug 26, 2017
26
0
530
0
So i just replaced my RTX 2080 TI with my old GTX 1070ti and was able to recreate the system crash. So obviously this isn't a GPU problem like I originally had thought. Secondly sorry for putting it in this thread as the only recent hardware change was the GPU and I suspected that as the culprit. Nonetheless the system crashed out on the 1070ti which also had been running stable for about 2 years on this computer, so I no longer believe the GPU is the culprit. I am increasingly more suspicious the PSU is on its way out, is there any way to tell before I go out and spend money on one?
 
Naturally, any brand of PSU can fail...

Given crashes or hard power-offs with two different GPUs, naturally most would opt for a known good PSU of sufficient wattage at this point. (DO NOT REUSE ANY MODULAR CABLES FROM OLD PSU WITH THE NEW ONE!)
 
Reactions: ronin026

ronin026

Prominent
Aug 26, 2017
26
0
530
0
So I believe I have fixed the issue. The PSU I have is a modular PSU, and I had an unused PCI-E slot on the PSU. While tinkering with it yesterday I tried using a different connector on the PSU, and this seems to have fixed the issue. I have stress tested it all night and was unable to reproduce a system crash.
 

ASK THE COMMUNITY