[SOLVED] PC randomly resets, can't pinpoint cause at all ?

[MangoYango]

PC has been randomly resetting all month, I'll be running image editing software or watching videos or just leaving it idle, and it will reset with the fans still spinning and the lights still on--no BSOD or any useful errors in Event Viewer. Sometimes the machine refuses to post after reset and can't be turned off from the power button. This current build is about 3 years old and has worked perfectly fine until recently. Nothing I've done seems to solve my problem and I'm at my wits end here at things to try. I recently paid for a new motherboard thinking it was the problem, only for it to go 5 days before it reset itself just like before, including being unable to be turned off sometimes.

Things I have tried:

• - memtest86 checks on both sticks individually (passed)
• -checked drive health with chkdsk and crystaldiskinfo, with no issues on either. have ran sfcscannow with no errors reported.
• -updated graphics drivers
• -checked temps on all components, CPU currently sits at 40-50c at idle
• -turned off fast startup (this caused my machine to completely forget my SSD and would only boot to my old windows install on my old HDD, from which the SSD is a clone of)
• -updated bios

At this point I took it to a local shop. There they:

• - swapped out motherboard (msi x570-a) for an identical one
• -swapped PSU from a Corsair RM750 for an 850

As mentioned before, it worked fine for a few days before randomly resetting. The only clues I can find are errors about bad blocks "Harddisk1" which after searching, seems to be on a drive I don't even use or is a system reserved--I feel like if that was the culprit I'd see blue screens.
I really don't know what to try now. I feel like issues with the CPU would be giving me BSODs , but the whole thing feels totally random.

Mobo: MSI X570-A Pro
CPU: AMD Ryzen 5 3600
Cooler: Noctua NH-C14S
GPU: Nvidia GeForce RTX 2060
PSU: Corsair RM850
RAM: Corsair 8GB DDR4 (x2)
Storage: 1 SanDisk SSD (1tb), 1Western Digital HDD (1tb)
Nothing overclocked.

 

[MangoYango]

Today I swapped the currently listed RAM sticks for 2 4gb sticks that I've been assured work correctly. The machine has been operating relatively normally before and after the swap. Waiting to see if it resets.
Additional information: This morning, I turned on POST beep in my BIOS settings and when I saved and exited, the machine went to reset and didn't post. Turning it off with the power button button caused it to reset, then pressing it again actually turned it off. Strange power button behavior was present on the last motherboard+PSU as well.

 

[MangoYango]

Today I swapped the currently listed RAM sticks for 2 4gb sticks that I've been assured work correctly. The machine has been operating relatively normally before and after the swap. Waiting to see if it resets.
Additional information: This morning, I turned on POST beep in my BIOS settings and when I saved and exited, the machine went to reset and didn't post. Turning it off with the power button button caused it to reset, then pressing it again actually turned it off. Strange power button behavior was present on the last motherboard+PSU as well.

Machine reset while I left it alone doing nothing but playing music. After getting back into windows to try and check event viewer, it reset twice, and now it won't post, nor can I turn it off with the power button. Tried unplugging the power pins to the GPU and unplugging the drives, but still doesn't post. At least I can rule out RAM as the issue...again. Doubt I have two failing mobos in a row so all I can imagine is the CPU, GPU, or some mystical software/windows hiccup.

 

[MangoYango]

Got it to post and boot to Windows, but it shortly resets just minutes into logging in, just like it did before switching out the mobo and PSU. I don't have a USB flash drive big enough to do a Windows bootable drive from. I don't even know what the issue is anymore and why it decides to randomly reset or why it sometimes boots and sometimes doesn't make it to post. I'm no closer to solving this issue after two weeks of tinkering and I really need some help.

 

[MangoYango]

Ended up finding a USB stick to install Windows onto--with both my hard drives disconnected, the system reset while looking for my product key... doesn't seem likely the SSD or HDD are responsible either.
I've taken my machine back to the repair place mentioned in my opening post alongside a 5 year old tripplite UPS that had a dying battery for further testing. Its only cleared for 480w iirc and I have a feeling that it might be possibly damaging a component due to lack of proper surge protection, but that's difficult to gauge. Outside of the CPU and GPU, I can't imagine what would cause this sort of behavior.

 

[MangoYango]

I've been trying to diagnose an issue with my PC randomly resetting for about a month now--ive tested temps and internal components and swapped the PSU, RAM and mobo with no luck. With my options running out, I'm starting to suspect my old 480w Tripplite battery backup might be the culprit (took it to be tested Friday and the battery was confirmed to be on it's way out),but recently I took my entire machine to a shop and it ran there for 72 hours hooked to the backup battery with a constant stress test --not a single reset the entire time.
Even if I plug the machine directly into the wall it still has the same reset behavior...which leads me to my question: could my backup battery have been protecting my computer from power fluctuations in the wall? Could those potentially cause my PC not receive adequate power to trigger a reset? Now that the UPS's battery is dying, could it be failing to protect my machine from them, causing it to restart?

After checking my parts on PCPartPicker (https://pcpartpicker.com/list/btWgDc) I learned that after adding in the watts drawn by my Cintiq 16, the total watts required by the machine itself is around 400w; not counting the 3 other monitors, several other peripherals, and USB chargers daisy chained through power strips, is it possible I'm putting too much on a backup cleared for only 480w?

 

[vov4ik_il]

While your UPS is waaaay too weak for the system, it would be in bypass mode while the power is on. It might be VRM or the PSU. Try changing "power loading" option in your BIOS and/or limiting the current draw of the CPU.

 

[jonnyguru]

Line Interactive PSUs are passive unless they're on battery power. So, if your mains power is working while the PC reboots, it's not the UPS.

 

[MangoYango]

Line Interactive PSUs are passive unless they're on battery power. So, if your mains power is working while the PC reboots, it's not the UPS.

So basically there's no way an issue with the UPS , even if it's too weak, could trigger a restart or keep it from reaching post, then. I'm totally at a loss if that's the case, I've tried a brand new mobo and PSU along with testing two sets of RAM and have been assured the CPU is running fine. I can't imagine anything in either drives triggering it, or the GPU either.

 

[MangoYango]

6 days later and still no solid leads. Leaving the PC at the same shop running stress tests on a Linux nmve for 72 hours with my old UPS connected, reported no restarts the entire time. Completely unhelpful.

Does nobody have any idea what could cause this at all? Am I going to just have to swap out the CPU /GPU at this point?? Surely SOMEBODY has had a similar situation as mine, there needs to be a way to check for this.

 

[jonnyguru]

So basically there's no way an issue with the UPS , even if it's too weak, could trigger a restart or keep it from reaching post, then. I'm totally at a loss if that's the case, I've tried a brand new mobo and PSU along with testing two sets of RAM and have been assured the CPU is running fine. I can't imagine anything in either drives triggering it, or the GPU either.

Troubleshooting: Plug it directly into the wall.

 

[MangoYango]

Troubleshooting: Plug it directly into the wall.

Tried that, even with no other drives attached and booting from a windows usb stick, it still reset with a wall only connection.

 

[MangoYango]

While your UPS is waaaay too weak for the system, it would be in bypass mode while the power is on. It might be VRM or the PSU. Try changing "power loading" option in your BIOS and/or limiting the current draw of the CPU.

If the UPS isn't the issue, what could possibly remain? Aren't resets and the machine not posting a sign of faulty power? I made no changes to the power draw outside of when the machine turns off the screen/sleeps for the 3 years I've had this build. What software problem could possibly cause it to reboot with no relevant cause in event viewer logs? There has to be some reason it was fine for a few days upon switching mobo/psu then suddenly resetting again.

 

[vov4ik_il]

I had a similar issue, was driving me crazy, turned out to be a motherboard problem but could be other stuff too. You can follow my footsteps here (would be different on the CPU tuning side as you are on AMD).

 

[MangoYango]

I had a similar issue, was driving me crazy, turned out to be a motherboard problem but could be other stuff too. You can follow my footsteps here (would be different on the CPU tuning side as you are on AMD).

Sounds like getting another motherboard solved your issue, ultimately? I replaced my motherboard with another of the exact same model and the reset issue persists. The place I bought it from didn't find anything abnormal with it from their tests it seems, either.

 

[vov4ik_il]

Sounds like getting another motherboard solved your issue, ultimately? I replaced my motherboard with another of the exact same model and the reset issue persists. The place I bought it from didn't find anything abnormal with it from their tests it seems, either.

In my case changing the make/model was the only thing that did. Insufficient VRM (or more likely improper decoupler circuitry that cannot accommodate abrupt current changes) is a design flaw, not a broken part. If I am right, it would perfectly handle a high load but load variations (sharp transitions from idle to full load and back to idle) will cause the system to fail.
Another thing that may cause such a failure is a similar problem in PSU circuitry and it would be a design flaw as well.
Since it was common in low end PSUs, motherboard vendors added an option of putting a dummy load onto the PSU (thus softening abrupt changes by keeping some load on the PSU when CPU load drops to a low).
Not sure what the name for this option in MSI, Gigabyte called it "Power Loading".

Before drawing conclusions, I would try increasing the Load Line Calibration value, or set/change a current limit for the CPU in BIOS, as the defaults might not be suitable.

 

[MangoYango]

In my case changing the make/model was the only thing that did. Insufficient VRM is a design flaw, not a broken part.

I guess I'm a little confused as to why it was stable for up to 3 years and only recently has it began giving me troubles despite not overclocking or doing anything different than usual. Recently it was left running for a 3day stress test (what it was tested with I don't know) but somehow didn't reset at all despite constantly doing so days beforehand. The only variables that differ from the shop environment and my home are the UPS+power strip combo listed earlier and my SSD+HDD which I removed beforehand.

 

[MangoYango]

In my case changing the make/model was the only thing that did. Insufficient VRM (or more likely improper decoupler circuitry that cannot accommodate abrupt current changes) is a design flaw, not a broken part. If I am right, it would perfectly handle a high load but load variations (sharp transitions from idle to full load and back to idle) will cause the system to fail.
Another thing that may cause such a failure is a similar problem in PSU circuitry and it would be a design flaw as well.
Since it was common in low end PSUs, motherboard vendors added an option of putting a dummy load onto the PSU (thus softening abrupt changes by keeping some load on the PSU when CPU load drops to a low).
Not sure what the name for this option in MSI, Gigabyte called it "Power Loading".

Before drawing conclusions, I would try increasing the Load Line Calibration value, or set/change a current limit for the CPU in BIOS, as the defaults might not be suitable.

Tell me what you make of this:
Got a another reset earlier while leaving it to play music and came back to a "Windows didn't load correctly" screen, and changed the voltage from auto (was .4 something) to .175 in BIOS. Ran chkdsk twice but never found a report in event viewer. Backed up all my data in the event this was being caused by a bad windows install or harddrive, ran Uningine Heaven for about 30 minutes and it reset while I was away.
Reset period lasted for what had to have been 15 or so minutes before getting to another"Windows didn't load correctly" screen: told it to turn off, but it simply reset instead. After getting to boot I couldn't even click the option to power down before it triggered another restart.
Flicked the power off the PSU, then proceeded to remove my drives and try the USB windows stick once more, which restarted before I could verify the language/region. Removed the stick so there were no drives left to boot from.
Here's where the weird part comes in: I got a reboot just on the BIOS screen; something I've never seen before, even with the old mobo. I have no clue why the techs I've sent this to didn't see these resets over the weekend while I'm getting them with no drives attached at all. This is starting to sound more like CPU failure than anything--it and the GPU are the only parts that haven't been swapped or removed.

 

[vov4ik_il]

You worsened the power starvation situation by lowering the voltage. Should be set the Auto back or increase it instead. I would change the load line calibration value instead of voltage.

 

[MangoYango]

This is almost embarassing to admit, but the issue seems to have been a faulty reset button/header. Unplugged it and it's been running stable for over 2 weeks. What tipped me off was catching it reset in the BIOS with no drives or sticks attached. Not sure what could bhave caused it to malfunction, maybe it's being pressed in weird by the front panel? The case is pretty old so it's anyone's guess why the reset button/header with act up. I'll write a summary of my experience here for anyone who's facing the same problems and hopefully save them the time and money I burned to figure this out.

Machine reset and remained in an unresponsive state where the fans spin and lights stay on. Holding the power button will not turn the machine off, but causes it to reset. Machine resets even with drives detached.
Things I've tried include:
-Checking event viewer for any relevant shutdown causes
-Adjusting the minimum processor amount in Windows' power settings
-Checking Windows install with sfcscannow and drive health with chkdsk, and crystaldiskinfo
-Using memtest86 to check both ram sticks
-trying a different outlet
-Swapping the PSU
-Swapping the motherboard
-had the CPU benchmarked for a total of about 4 days
-Swapping the RAM
-ran benchmarks on the GPU
-replaced UPS battery and swapped around the plugs to minimize the devices plugged into it
-lowered voltage in the bios from default to .175
None of this worked, and the resets still triggered randomly. Only when it reset in the bios did I think the reset button itself could be the culprit. Holding down the reset button when plugged in causes the machine to halt until it's released, perfectly replicating the behavior described previously.
I'm admittedly a little upset the tech who checked my machine didn't catch this, but it's a really weird issue to begin with. Hopefully this helps somebody facing the same problem.