[SOLVED] PC restarts repeatedly while booting ?

[urchine]

I built my PC nearly 2 years ago, and it has been working fine up until about 3 weeks ago, where it has started to crash while i am using it. It initially crashes after about 15-30 minutes of usage, after which it crashes while restarting repeatedly. I can't find any pattern to the crashes, and it can happen even when I leave it at an empty desktop.

When it first restarts, there is no BSOD, and an audible click sound comes from the PC. When it starts to restart loop, it doesn't even go to the manufacturer logo screen before it crashes and restarts again. The CPU, DRAM and VGA lights on the motherboard light up every time it restarts.

I have managed to do a fresh installation of Windows in between crashes, but it does not seem to help. I have also brought the PC to a local repair shop, but they didn't get the crashing in the 2-3 days there.

Specs:
Motherboard: Gigabyte B450 Aorus M
CPU: Ryzen 5 2600 3.4Ghz
RAM: T-Force Vulcan Z 8GB x2
GPU: Zotac Gaming Geforce 1660 Super
PSU: Cooler Master 550w Gold
Storage: Western Digital Blue 1TB SSD and Seagate Barracuda 2TB HDD
Case: Fractal design Focus G Mini

 

[Ralston18]

What did the repair shop do? Did they just plug the computer in and let it sit once successfully booted?

For example: did they do any cleaning, re-seat cards, connectors, RAM, and jumpers. Maybe a fresh CMOS battery?

What about testing the PSU?

A two year old PSU is a likely suspect especially if there is a history of heavy gaming use. Or perhaps video editing and mining.

Take a look in Reliability History and Event Viewer. Either one or both may be capturing error codes, warnings, or even informational events related to the crashes.

Start with Reliability History. Reliability History uses a timeline format that can prove very helpful in diagnosing problems. Look for happenings that began about 3+ weeks ago.

Varying errors and increasing numbers of errors are a sign (to me) of a failing/faltering PSU.

Another thought: At home check the power source/path serving the PC - could be a problem somewhere along that path.

Sudden, unexpected shutdowns can and do corrupt system and program files. Often making things worse.

Try the built in Windows troubleshooters along with "sfc /scannow" and "dism".

https://www.lifewire.com/how-to-use-sfc-scannow-to-repair-windows-system-files-2626161

How to use DISM command tool to repair Windows 10 image | Windows Central

Quite a bit I agree. But take your time and work through it...

 

[urchine]

What did the repair shop do? Did they just plug the computer in and let it sit once successfully booted?

For example: did they do any cleaning, re-seat cards, connectors, RAM, and jumpers. Maybe a fresh CMOS battery?

What about testing the PSU?

A two year old PSU is a likely suspect especially if there is a history of heavy gaming use. Or perhaps video editing and mining.

Take a look in Reliability History and Event Viewer. Either one or both may be capturing error codes, warnings, or even informational events related to the crashes.

Start with Reliability History. Reliability History uses a timeline format that can prove very helpful in diagnosing problems. Look for happenings that began about 3+ weeks ago.

Varying errors and increasing numbers of errors are a sign (to me) of a failing/faltering PSU.

Another thought: At home check the power source/path serving the PC - could be a problem somewhere along that path.

Sudden, unexpected shutdowns can and do corrupt system and program files. Often making things worse.

Try the built in Windows troubleshooters along with "sfc /scannow" and "dism".

https://www.lifewire.com/how-to-use-sfc-scannow-to-repair-windows-system-files-2626161

How to use DISM command tool to repair Windows 10 image | Windows Central

Quite a bit I agree. But take your time and work through it...

Ok so I checked Reliability History and Event Viewer, and for the crashes RH just says that Windows did not shut down properly for all of them, with no other info. EV has 3 critical events, all of which are Kernel Power 41s.

The Windows troubleshooters don't find any problems. SFC /scannow does find some files but even after fixing them, the crashes still continue. DISM check health doesn't find anything, and the PC crashed while I was running scan health.

Before I let the repair shop take a look I had already cleaned and reinstalled most of the hardware, other than separating the CPU from the motherboard. I'm trying to contact the person that was working on my PC but its taking a while.

Should I just change the PSU first and see how things proceed from there? If so, any recommendations for any specific model? Unlikely to upgrade anything, maybe the motherboard in the next few years.

Sorry if I take a while to reply. I can only try to check things on the PC at night, if it doesn't crash as soon as I turn it on.

 

[Ralston18]

If at all possible try swapping in another known working PSU. Do not mix and match PSU cables when you do so.

Another option is to test the current PSU. Doing so requires that you have a multi-meter and know how to use it. Or know someone who does.

FYI:

https://www.lifewire.com/how-to-manually-test-a-power-supply-with-a-multimeter-2626158

Not a full test because the PSU is not under load. However, any voltages out of tolerance make the PSU the likely culprit.

PSU models - start here:

Best Power Supplies of 2022 - Top PSUs for Gaming PCs | Tom's Hardware (tomshardware.com)

 

[DRagor]

The nature of your crashes (1st at half an hour, next immediately after) suggests PSU failure.
Repair shop failing to replicate problem could mean two things: either they were completely incompetent; or you have problem with electricity grid at home.
Have you tried booting to BIOS and leave it there indefinitely? Will it work or also crash about same time as in Windows?

 

[Satan-IR]

As said above it seems like a PSU issue to me too.

Does the PSU fan spin when it turns on? Does it spin after a while and keep spinning? If not it might be PSU over temperature protection. As DRagor said, the time frame points to that.

It might be that it takes half an hour for it to reach the temps where protection kicks in and after that it won't turn on at all. That audible click you hear might be the relay. If it can start after a while maybe its because it cools down.

I would check with another known good PSU. Just keep in mind to use each PSU with the cables it came with if either is semi/fully modular and cables can be detached.

 

[DRagor]

One more thing - check CPU and motherboard temps. It potentially could be a overheating problem too.

 

[urchine]

Update:

Got a multimeter and tested the PSU, seems to be working within tolerances, so not sure what the problem is. The click sound is from the PSU turning off, as it happened when the shorting wire dropped out. Can't really test my house power but I'm pretty sure its stable, never had any kind of issue with it before.

PC can sit in BIOS for an extended time, and if it boots properly can also sit in Windows without anything running.

It seems like the crash causes the PSU fan to stop for a moment and the rest of the PC to fully lose power, as the case LEDs turn off for a moment before it tries to boot again. When it boots, the motherboard error LEDs for VGA, DRAM, and CPU all turn on for a bit each, only turning off if it boots properly.

I managed to turn it on for a decent amount of time, installed OpenHardwareMonitor to check temps. CPU was stable at around 50C, and everything else was sitting around 30-40C when not doing anything. Running a bit of stuff it got up to CPU 70C and others 40-50C without crashing, but it did crash twice, once when I first launched steam and again for genshin.

Ran sfc /scannow and DISM again, no problems same as last time.

 

[DRagor]

When it boots, the motherboard error LEDs for VGA, DRAM, and CPU all turn on for a bit each,

Tha LEDs turning on and off during boot is perfectly normal. Only when it stays lit it means problem.

It seems like the crash causes the PSU fan to stop for a moment and the rest of the PC to fully lose power

More proof for power problem - but again hard to tell if that's faulty PSU or your home grid.

Running a bit of stuff it got up to CPU 70C and others 40-50C

Those are fine, no sign of overheating problem

PC can sit in BIOS for an extended time, and if it boots properly can also sit in Windows without anything running.

So it looks like problem manifests itself only under load. This usually means either overheating or power issue. Since your temps are fine the only logical conclusion is power issue.

 

[Satan-IR]

I agree with DRagor. What you said points to an over temperature situation. But it seems CPU and GPU don't get unreasonably high.

Multimeter shows output are at OK nominal values. As Ralston18 said above, a basic test that shows PSU is apparently working within standard ranges, with no load. Once the transients hit any PSU the whole dynamics of its performance, heat dissipation and error margins change. Faulty bit operates "OK" with light load but when going gets tough doesn't behave as it should.

I too would check with another known good working PSU.

 

[urchine]

Update: Replaced the PSU with a new Seasonic Focus GX-650 Gold and it seems to be working fine so far, able to run games normally. Put in another fan for good measure too. Thanks for the help with this guys.