[SOLVED] Sudden Reboots During Gameplay

Jun 4, 2020
5
0
10
0
My PC suddenly shuts down and reboots unexpectedly during gameplay (Star Wars Jedi Fallen Order on Steam). Kernel 41 critical failure I think it's called (unexpected shutdown). Only way I've been able to recreate this is running FurMark GPU Stress Test. Had this same problem, but less often, a few years ago with a different graphics card running Far Cry 3 on Steam. With Far Cry 3 the problem was rare enough I could beat the game with some minor irritation. With Jedi Fallen Order it's unbearable, happening every time within 30 min or as fast as 4 min. I think a slight burning smell usually starts just before it crashes too, maybe dust? I tried vacuuming everywhere I could inside the case, out all the vents, and the GPU is very clean. Even having the A/C on, ceiling fan on, and a floor fan pointed at the front of the case didn't seem to help. This first crash seems to take 20+ min but then happens faster 2nd/3rd/4th times.

At one point I tried a complete clean boot, and it appeared to fix the problem, but then was never able to recreate. Continues to crash during clean boot with all non-windows startups disabled. I have the latest Nvidia driver although previous version driver had same results. Lowering the game graphics didn't help either. The last game I played on this machine (with no problems) was Total War: Rome 2 (Steam). Everyday computing and playing through the full Mass Effect (Origin) series aren't a problem either.

Here are some quick failure data points with full dataset attached:
Running AIDA64 System Stability Test (all boxes checked except GPU) concurrently with FurMark GPU Stress Test:
  • Reboot within 10-30 min (even FurMark running by itself)
  • GPU temp constant at 86C, with fan speed slowly rising
  • Reboots before fan speed hits 100%
  • 12V drops as low as 11.592V
HWiNFO logging every second during gameplay reboot:
  • 12V dropped as low as 11.74V
  • GPU temp constant at 84C
  • GPU Fan % max at 52%
  • GPU Power max 220W
  • GPU max memory usage 85%
  • Motherboard max temp 45C
  • Physical memory max usage 38.3%
  • CPU max temp 56C / Core max temp 65C
Non-failure data points:
  • Neither AIDA64Extreme nor Prime95 cause reboot unless running FurMark concurrently, although I haven't tried running more than 30 min
  • Passed MemTest86 with no errors
Specs (built late 2012 unless otherwise noted):
Gigabyte GA-Z77-HD3
Intel Core i7-3770 Processor (4x 3.40GHz/8MB L3 Cache)
Liquid CPU Cooling System, 120mm Fan
GeForce GTX 980 TI, 6GB GDDR5 (recent upgrade from GTX 660 2 GB)
24GB DDR3-1600 Memory Module - G.Skill Ripjaws X (4GB x 2 old / 8GB x 2 new)
120 GB ADATA S510 SSD
AZZA alpha 500W PSU (model: PSAZ-500S12)
AZZA Armour Gaming Case
Windows 10 Home

I'd really like to be sure what the problem is before dropping more money on a fairly old build. Only ideas I have that I haven't tried are buying a new PSU, or de-clocking the GPU.

HWiNFO Sensor Log every second from game boot to failure.
It's a ton of data, so I've been using GenericLogViewer_v5.2 tool to try and analyze it, but hoping you guys see something I'm missing.

AIDA64extreme screenshot one minute prior to my PC crashing (running FurMark concurrently to stress GPU).
 

Phaaze88

Splendid
Ambassador
I'm betting on the psu.
-The OEM is Andyson; not known for making quality psus.
-It had a 1 year warranty. That speaks volumes of what the company thought about it's quality...
-It doesn't have any kind of 80+ rating - dang, they could've at least tried to lie about it...
-It's not even a true 500w psu; it's just 372w on the +12v rail - the main one used by the cpu and gpu.
 

Phaaze88

Splendid
Ambassador
I'm betting on the psu.
-The OEM is Andyson; not known for making quality psus.
-It had a 1 year warranty. That speaks volumes of what the company thought about it's quality...
-It doesn't have any kind of 80+ rating - dang, they could've at least tried to lie about it...
-It's not even a true 500w psu; it's just 372w on the +12v rail - the main one used by the cpu and gpu.
 
Jun 4, 2020
5
0
10
0
Wow, I’ll be honest, I didn’t realize you’re supposed to look solely at the 12V number. I knew the PSU might need to be upgraded when I went to the GTX 980 TI, but it had happened even with the GTX 660. Plus I figured it should’ve crashed instantly when running Prime95 + FurMark simultaneously.... yet it lasted as much as 30min. Maybe heat reduces PSU output?

What would you recommend for this setup?
 

Phaaze88

Splendid
Ambassador
Prime95 is a cpu and ram only stress test. You can't push a gpu with it, and a gaming graphics card is the most stressful part of a gaming PC for a psu to deal with.
That's why nothing happened when you ran P95. If you had ran Stress Gpu on Aida64, it would've crashed too.

Maybe heat reduces PSU output?
Even psus have a recommended operating temp for max output, though one has to go out of their way to find out what it is, as it's usually not listed in the usual spec sheets.

550-650w range:
Corsair CX, TX, RM, RMX
Antec Neo Eco Zen Gold
Thermaltake Toughpower
Seasonic Focus, Core, Focus Plus
EVGA Supernova G3
be quiet! Dark Power Pro
Superflower Leadex 3
Due to the pandemic, a number of these may be sold out, so I listed a bunch.
 
Jun 4, 2020
5
0
10
0
New PSU came today, installed it, and played Jedi Fallen Order a couple hours without any trouble. Definitely seems to have fixed the problem. It pulls less power from the outlet and produces less heat as a bonus. Maybe I’ll try FurMark again to really test it. Anyways, thanks for the help!
 

ASK THE COMMUNITY

TRENDING THREADS