[SOLVED] Computer black-screens and reboots while playing vr-games

[SilkyGreenBox]

I've got a computer issue I am entirely stuck on. I have no idea what to do next

The computer is 4-5 years old, but the issue started this year.

The computer reboots semi-randomly, mostly when playing vr-games. There is no BSOD, just black screen and reboot. This happens everywhere between 5-40 minutes of gameplay, sometimes immediately. Some games are worse than others, Phasmophobia is the worst and will always crash within 30-40mins, while other games like VRC seem fine. It has crashed once on a non-vr game(Deep Rock Gamactic) in the same fashion.

The crash does not create minidumps. I have found no events in event-viewer other than events regarding the reboot.

I have run stress tests with Prime95, FurMark and Unigine Superposition. Individually and all together. The CPU ran a little hot (95C). This was remediated by reapplying the thermal paste taking the max down to 85C . I did not experience any crashes during the tests(before or after applying new thermal paste).

Ran Memtest on a stick, which gave some errors. Replaced the RAM, but to no avail.

I would greatly appreciate any input on this

Specs below:
VR: VIVE Original with Index Knuckle Controllers
CPU: AMD Ryzen 5 3600
GPU: EVGA Nvidia 1080Ti
RAM: Corsair Vengeance 32GB 3600Mhz
Motherboard: B450 Tomahawk Max
SSD: Samsung 970 EVO Plus
PSU: Corsair CX750M

 

[TommyTwoTone66]

How are you running VR with no graphics card? Or did you just forget to list it?

95C suggests you are overclocking the CPU pretty hard. That's not a great idea for stability. I'd reset it to Stock while you're troubleshooting if you have.

 

[peptsu]

Sounds like a PSU problem honestly

 

[SilkyGreenBox]

How are you running VR with no graphics card? Or did you just forget to list it?

95C suggests you are overclocking the CPU pretty hard. That's not a great idea for stability. I'd reset it to Stock while you're troubleshooting if you have.

Yes, sorry, I forgot to list the GPU. I will add that to the original post.

The CPU is not overclocked. Mind that this is the max temp running tests on artificially high load. It is a Ryzen running on the stock cooler as well, so not that surprising. Applying new thermal paste took the max down to 85C as well.

 

[SilkyGreenBox]

Sounds like a PSU problem honestly

That was my first thought as well, but would that not show in a pressure test? I ran several stress tests, testing the GPU, CPU and RAM both with the monitor and VR running simultaneously, without crash. I would like to avoid buying and replacing parts unnecessarily.

 

[peptsu]

That was my first thought as well, but would that not show in a pressure test? I ran several stress tests, testing the GPU, CPU and RAM both with the monitor and VR running simultaneously, without crash. I would like to avoid buying and replacing parts unnecessarily.

Think of it like this, you've stressed your procesor/graphics and nothing happened, so your processor is not the problem, could be the RAM? don't think so, if the rams were faulty the system will not even boot, SSD? they can give you black screen but they will not reboot your system

So my best bet is PSU, motherboards don't do that kind of stuff unless they're already too fried.

The best way to test a faulty PSU is well, trying a new one on your current setup

 

[SilkyGreenBox]

Think of it like this, you've stressed your procesor/graphics and nothing happened, so your processor is not the problem, could be the RAM? don't think so, if the rams were faulty the system will not even boot, SSD? they can give you black screen but they will not reboot your system

So my best bet is PSU, motherboards don't do that kind of stuff unless they're already too fried.

The best way to test a faulty PSU is well, trying a new one on your current setup

It very well might be that it is the PSU, but I don't feel like I have any damning evidence for it. The stress tests did test the individual components, yes. Running them all together, over time, i think would be some pretty demanding consumption for a faulty PSU. It could be that I have been unlucky and that the PSU was achieving well just during those tests, but I would like just a few more pointers going in that direction before spending more money to fix this.

Do you know of any tools or tests I can do to give more evidence towards a faulty PSU, or to eliminate the other components?

Appreciate the input!

 

[TommyTwoTone66]

You still haven’t mentioned what GPU you have…

I agree, running all of those stress tests at the same time for 1-2 hours should have thrown up any PSU issues. I don’t think it’s your PSU or an overheating issue.

VR is still a very young and immature technology. I get frequent crashes in most VR apps since they’re just janky as all hell, and so is Steam VR etc.

Admittedly, not crashes so severe they actually restart the PC, but then I am running on Intel, not AMD, maybe that’s more common for AMD users idk.

I’m assuming you’re on Windows 10 and not anything dumb like 7 or XP? Got your bios, motherboard chipset drivers, graphics drivers etc all up to date?

 

[SilkyGreenBox]

You still haven’t mentioned what GPU you have…

I agree, running all of those stress tests at the same time for 1-2 hours should have thrown up any PSU issues. I don’t think it’s your PSU or an overheating issue.

VR is still a very young and immature technology. I get frequent crashes in most VR apps since they’re just janky as all hell, and so is Steam VR etc.

Admittedly, not crashes so severe they actually restart the PC, but then I am running on Intel, not AMD, maybe that’s more common for AMD users idk.

I’m assuming you’re on Windows 10 and not anything dumb like 7 or XP? Got your bios, motherboard chipset drivers, graphics drivers etc all up to date?

It's an EVGA Nvidia 1080 Ti.

Windows 10 20H2
Unsure about BIOS updates(had some difficulties last flash)
Up to date graphics drivers
Up to date chipset drivers(from motherboard manufacturer website, going to check official AMD website)

Edit:
I will try updating to the latest Win10 build, check bios ver. and check if there are any newer chipset drivers available directly from AMD.

 

[SilkyGreenBox]

Writing this for anyone with similar issues that find this post.

The crashes worsened over time, crashing in more VR games and also whenever running Unigine Superposition's VR-mode with Prime95. This was instant.
The PSU was the culprit, the crashes stopped immediately after replacement.