Question Could this be a dying PSU?

[ARCADDER]

Hi, so I've been using this configuration for the past 11 years or so:

GTX 1080ti Strix -- Intel i74770K -- 16GB DDR3 RAM -- Maximus VI Gene Motherboard -- Corsair GS700 PSU.

I've had zero problems with this setup since I built this PC in 2012, up until this week, when I noticed that every game end up either crashing, freezing the entire PC but not the audio, or straight up shutting down the PC.

Just to name a few, HITMAN Absolution throws a "DXGI_Error_Device_Removed". Pathfinder WOTR, freezes the video but not the audio and I have to kill the game via Task Manager. Frtnte Battle Royale throws a "D3D device removed" error and LEGO Frnte just shuts down the PC and leaves it in a weird state in which it won't even boot the BIOS screen and the restart button doesn't do anything. And while I can turn it off by leaving the power button pressed, I have to unplug the power cable from the PSU, wait a few seconds, and only then the PC will reach BIOS and then post to Windows 11, otherwise it just stays in this black limbo.

These crashes can happen 1 minute in, or 1 hour in. It happens without rhyme or reason.

This is everything I've tried:

1. I've reinstalled Nvidia's drivers with the solution I used for years: NVCleanstall with the minimum recommended modules.
2. I've formatted the PC (I do this every 5 or 6 months anyway).
3. I use a perpetual Windows Pro license that I got directly from Microsoft. Every time I reinstall, I use their latest image.
4. I am not using any kind of overlay or extra software for peripherals.
5. I am not overclocking any of my devices.
6. BIOS is using its latest version.
7. I have restored the BIOS to its factory defaults.
8. I've combed every BIOS setting to check nothing is the correct one, not being overclocked or using any extra "performance/gaming" settings.
9. I've completely disassembled my PC to clean every contact both from the MOBO, GPU, PSU and RAM with 99.9% pure Isopropyl alcohol.
10. I've vacuum cleaned the motherboard and GPU.
11. I just used Thermal Grizzly thermal paste to replace the old one on the GPU and CPU water cooled block.
12. I've made sure every contact is in good shape and firmly connected. Same with the RAM, GPU and HDDs.
13. I've tried with brand new electrical cables for the PSU.
14. I've tried swapping RAM to older ones I had lying around, and also swapping my usual ones to Channel A.

And well, I'm lost here since I've also:

• Stress tested the PC via Furmark and OCCT. It's funny because even on combined stress tests with the GPU and CPU running at 100% of their power, the temp readings are completely fine (-60°C), and even after hours of these stress tests, the PC doesn't reproduce the crashes or the freezes.
• I don't think it's a dying GPU since I see zero artifacts or any kind of clue as to when it's going to crash, plus the audio also comes out from HDMI, and that doesn't freeze, stutter or gets cut once the crashes occur.
• Tried with Manjaro Linux and I face the same issues when gaming.
• Whocrashed's minidump readings conclude it's a WHEA_UNCORRECTABLE_ERROR from genuineintel.sys.
• Microsoft Event viewer claims these crashes are Kernel Power 41 critical errors.

 

[zinkles]

even if it's not a PSU problem, I'd replace that PSU. 11years, and the time has come.

did you try checking your PSUs voltage rails? to check this, go to your BIOS and you will be able to see a section called voltages. (different from mobo to mobo)
you'll be able to see different sections like:

CPU -
3.3v -
5v -
12v -

in front/below this, you'll see the actual voltage of it. if that voltage is lower than it's name, e.g. the 3.3v shows as 2.9v, then the PSUs voltage rails arent supplying the correct power for the components to run.

its worth checking this even if youre not facing such issues, which can be used to identify potential PSU failures.

 

[ARCADDER]

even if it's not a PSU problem, I'd replace that PSU. 11years, and the time has come.

did you try checking your PSUs voltage rails? to check this, go to your BIOS and you will be able to see a section called voltages. (different from mobo to mobo)
you'll be able to see different sections like:

CPU -
3.3v -
5v -
12v -

in front/below this, you'll see the actual voltage of it. if that voltage is lower than it's name, e.g. the 3.3v shows as 2.9v, then the PSUs voltage rails arent supplying the correct power for the components to run.

its worth checking this even if youre not facing such issues, which can be used to identify potential PSU failures.

Thanks! How much margin of error do you think there should be?

For 3.3V, it shows 3.216.
For 5V, it shows 4.960.
For 12V, it shows 12.000 or 12.096.

But, yeah for the age alone it might be time to send this PSU on retirement.

 

[zinkles]

the voltages seem fine to me.

The PSU may still be going bad even though the voltages are correct.