Question Random Reboots (No BSOD) on new system using PSU from previous build. Advice?

Dec 9, 2019
6
2
15
Hello friends.

I recently built a new PC with the following components:

Intel Core i7-9700K CPU

ASRock B365 Phantom Gaming 4 Motherboard

Corsair Vengeance LPX 16GB (2x8B) (CMK16GX4M2B3000C15) RAM

EVGA GeForce RTX 2070 Super GPU


The PSU I am using with these new parts is from my previous build: Corsair HX750W Silver

I should note that the PSU was purchased used from a friend and might be 5+ years old by this point.

The PSU seemingly worked fine with my old build until I built my new PC on 12/05, but I am now suspecting that it might be the culprit, though I'm not really sure. As far as I could tell, there were no random restarts for the first few days, though they may have occurred without my noticing while I was away from the desktop. From what I can tell, it appears to happen maybe 1-3 times a day, in and out of games. There is no BSOD and I also couldn't find any dump files when using the Windows Debug. Windows simply says "Restarting".

I tried the following to diagnose/resolve the issue:

I monitored my CPU and GPU temps under load (while playing a demanding game). The CPU cores never exceeded 57 °C and my GPU hovered around 80-90°C. This leads me to believe that overheating isn't the issue, especially because the restarts happen even when the system is idle.

I turned off automatic update restarts, as well as automatic restarts in the system recovery tab. I even created a registry key to prevent automatic restarts.

I ran the Windows memory test (though perhaps I should use Memtest86 instead?) and it found no issues.

I reverted to an older Nvidia driver.

I updated my BIOS to the latest version.

I changed minimum process state to 0% and changed power plan to high performance.

The next steps I plan on taking are:

Purchasing a PSU tester to try and diagnose the PSU.

Removing my GPU and using onboard graphics for a while to see if the GPU might be faulty.

Doing a clean install of Windows or using a restore point.

...

Do you have any advice as to how I should proceed troubleshooting? I'm sure there are plenty of other steps I haven't yet taken or thought of. For example, diagnosing the CPU or Motherboard would be helpful, but I'm not sure what the best way to go about it is. For what it's worth, my PC performs very well in demanding games and appears to function perfectly fine, until the spontaneous reboots rear their ugly heads. I also have multiple hard drives connected to my PC. What would be a good way to test my drives?

Thank you in advance and please let me know if further information would be helpful.
 
Last edited:

PC Tailor

Illustrious
Ambassador
Welcome to the forums my friend!

Purchasing a PSU tester to try and diagnose the PSU.
These aren't really that effective unless you go for the high end units. Ultimately you can achieve the same results most PSU testers have with a multimeter. They usually only monitor voltages when there is low load, and it might be that the PSU faults when there are spikes or changes in load for example.

Removing my GPU and using onboard graphics for a while to see if the GPU might be faulty.
Good shout, just remember that even if the restarts stop, it could still be mean that it is either the GPU or the PSU. Being as your GPU is the biggest draw on power, it could be the PSU handles the system without hat power draw much better.

Doing a clean install of Windows or using a restore point.
This would certainly cover if it is indeed hardware or software, but usually random restarts tend to be hardware. Sometimes firmware can be the issue, but not too often. But this would be worth doing if you're able to verify if it is potentially some instability.

Having said that, did you clean reinstall windows with the new build?
Are you running any overclock/XMP at all?
 
Dec 9, 2019
6
2
15
Hey PC Tailor, thank you for the warm welcome and for the prompt reply!

All of your feedback is well-taken and I will keep it in mind when performing further tests. As for overclocks, I haven't modified my hardware and kept all settings on default.

When removing the GPU, if the restarts cease, I will try replacing the PSU first. If I try the new PSU with the card and the restarts return, then perhaps I could safely assume it's the GPU.

Regarding the PSU tester, I am aware that it won't catch all potential issues. If it happens to show no irregularities, I know that this still shouldn't be enough to disqualify the PSU as the culprit.

Do you think I should troubleshoot my hard drives? If so, what is the ideal tool to use? I think I may also run Memtest86 overnight if it's more thorough than the standard memory diagnostic.

The annoying thing about this issue is that it could take hours for a restart to occur and this makes troubleshooting more lengthy and complicated. I'm paranoid about surpassing the 30-day mark for Amazon in case returns are needed, but oh well.

Thank you!
 
  • Like
Reactions: PC Tailor

PC Tailor

Illustrious
Ambassador
You're welcome my friend!

My typical go to storage drive tests: https://www.sysnative.com/forums/threads/hard-drive-hdd-diagnostics-sea-tools-for-dos-ssd-test.4072/
Memtest is a good should too, if you want a guide on that one too: https://www.sysnative.com/forums/threads/test-ram-with-passmark-memtest86.24300/

I'm paranoid about surpassing the 30-day mark for Amazon in case returns are needed, but oh well.
I wouldn't worry about that, if the unit is faulty, you are covered, as they'll have a warranty period. It's just Amazon's quick return policy.
 
Dec 9, 2019
6
2
15
I just thought I would provide an update:

I ran the Intel Processor Diagnostic Tool and my CPU passed every test. I have the results file saved. Shall I share it? And if so, is there a preferred filesharing website I should use?

I ran Seatools for Windows. I did the S.M.A.R.T and long generic tests, and all of my drives passed. Is the USB bootable version much better and should I use that instead?

Although I can't find any trace of my random restarts in the Windows Event Log or in memory dumps, I did create a SysnativeFileCollection.zip file. I'm not sure if that would be helpful at all.

I ran the Unigine Heaven benchmark on my GPU for about 30 minutes or so as a stress test, and it went perfectly.

I did a malware scan and nothing was found.

I ran Prime95 with the Blend option (CPU + RAM stress) and I had no issues or overheating at all for 4 hours. It also said 0 errors or warnings for each core.
...

I plan on running Memtest86 overnight to see what happens and I'll report those results tomorrow (assuming they actually save and don't get lost because of an unexpected restart).

I may also try switching my RAM slots, and then trying one stick at a time if that doesn't work, even if the memtest turns out okay...or would this be a waste of time? I assume Memtest isn't always 100% when it comes to catching things.

I suppose I can also try making sure that each connection from the SATA and PSU cables isn't loose in any way.

My new PSU arrives on Wednesday, so at least then I'll know if this PSU is truly the problem.

I also read somewhere that the Reset SW from the case might cause this issue, so I may even try unplugging that to see if anything changes.

My Task Manager says that my computer's up time has been about 24 hours, which is much better than I expected compared to the last few days of restarts, even though I really didn't do anything as far as potential fixes are concerned.
 
Dec 9, 2019
6
2
15
The memtest came up with 0 errors after 4 passes (the max allowed by the free version).

I'm leaving my PC on Unigine Heaven and Prime95 while I'm at work so the GPU, CPU, and RAM will be under load for 8 hours or so. I'll see what happens.
 

PC Tailor

Illustrious
Ambassador
I assume Memtest isn't always 100% when it comes to catching things.
No software solution is. Always safely assume that ANY software you run for diagnosis, does not rule out the component, simply gives you more confidence that it probably isn't.

There are plenty of times where a faulty CPU passes Intel Diagnostic, faulty GPUs pass Furmark stresses, faulty RAM passes memtest etc.

The only 100% guaranteed way to identify the point of cause is to replace the component and retest.

I also read somewhere that the Reset SW from the case might cause this issue, so I may even try unplugging that to see if anything changes.
Absolutely, I have seen this a few times too, simply removing the reset switch from the motherboard will help verify that, you may also want to remove the power switch.
 
Dec 9, 2019
6
2
15
No software solution is. Always safely assume that ANY software you run for diagnosis, does not rule out the component, simply gives you more confidence that it probably isn't.

There are plenty of times where a faulty CPU passes Intel Diagnostic, faulty GPUs pass Furmark stresses, faulty RAM passes memtest etc.

The only 100% guaranteed way to identify the point of cause is to replace the component and retest.


Absolutely, I have seen this a few times too, simply removing the reset switch from the motherboard will help verify that, you may also want to remove the power switch.

Thanks for the feedback! I'll keep experimenting. I do have one question: The restart isn't just a crash and then reboot. Windows actually says "Restarting" (at least the one time I witnessed it occur in person). This is strange, also because I disabled the restart on system failure option in Windows. Would this perhaps point more toward a software issue since the reboot isn't an abrupt crash, but instead Windows begins the restarting process? It might explain why the Event Log doesn't show anything abnormal and why memory dumps aren't created. For this reason I'm thinking of doing a fresh install of Windows 10 before proceeding with further troubleshooting. Also, I think I'll use strictly the drivers from the motherboard website rather than let Windows automatically download them. Who knows if that could be the issue as well -- I am using whatever drivers Windows picked for me, minus the GPU driver.