Question PC randomly reboots to BIOS ?

[LordHuffnPuff]

So I built a new PC and it runs great, except one problem...

The computer will randomly reboot to BIOS, with the M2 NVMe drive not detected. If the PC is fully powered down and then powered back up, the drive is detected and it boots as normal. This only occurs when the computer is idle (though maybe this is a coincidence). It might happen once every 3-4 days, it's frequent enough to be annoying but not frequent enough to interfere with day to day use. I thought at first it was one of the other drives (and this did help me weed out a dying HDD) but after disconnecting them the problem has not abated.

I have tested the M2 drive with chkdsk, crystalmark, hwinfo and HDDscan and every single one returned perfect health reports.

I have updated the BIOS to the latest version and checked the voltages on the RAM. I'm beginning to run out of ideas about what it could be beyond RMA'ing the Motherboard itself with the suspicion that it might be a bad NVMe slot. Does anybody have any ideas??

Build:
AMD Ryzen 7 7800X3D 8-Core Processor
MSI MAG X670E Tomahawk WiFi Gaming Motherboard
NVIDIA GeForce RTX 4080 SUPER
SAMSUNG 990 PRO SSD 4TB
G.Skill Trident Z6 RGB Series DDR5 RAM 32GB (2x16GB)
Windows 11 Pro 64x

Some technical stuff:
No abnormal temperatures (HWiNFO)
sfc /scannow - 100% verified, no integrity violations
dism /online /cleanup-image /restorehealth – completed successfully
chkdsk – no issues on any drives
wmic memorychip get manufacturer, capacity, partnumber, speed, memorytype, devicelocator, formfactor
Capacity DeviceLocator FormFactor Manufacturer MemoryType PartNumber Speed
17179869184 DIMMA2 8 Unknown 0 F5-6000J3636F16G 4800
17179869184 DIMMB2 8 Unknown 0 F5-6000J3636F16G 4800


EDIT: I noticed that this is about a decade exactly since my last build gave me enough trouble to check for help here, which is perversely neat.

 

[Lutfij]

I have updated the BIOS to the latest version and checked the voltages on the RAM.
Please state the BIOS version that you've updated to. Did you clear your CMOS after verifying your BIOS was successfully flashed to the latest version? If you haven't, disconnect from the wall and display, remove the CMOS battery, press and hold down the power button for 30secs and then replace the battery after 30mins.

You forgot to mention the make and model of your PSU. If the unit is recycled from an older build, please specify it's age and what it'd powered during it's tenure.

How are you cooling the processor? Perhaps try and relieve some pressure off the CPU's cooling mount and see if that alleviates the issue.

The computer will randomly reboot to BIOS, with the M2 NVMe drive not detected. If the PC is fully powered down and then powered back up, the drive is detected and it boots as normal.
Do you feel a mild tingling sensation when you touch the metal part of your case and your bare feet are making contact with a tiled(non-wooden, non-carpeted floor)? If so, you might be having a grounding issue.

 

[LordHuffnPuff]

I have updated the BIOS to the latest version and checked the voltages on the RAM.
Please state the BIOS version that you've updated to. Did you clear your CMOS after verifying your BIOS was successfully flashed to the latest version? If you haven't, disconnect from the wall and display, remove the CMOS battery, press and hold down the power button for 30secs and then replace the battery after 30mins.

You forgot to mention the make and model of your PSU. If the unit is recycled from an older build, please specify it's age and what it'd powered during it's tenure.

How are you cooling the processor? Perhaps try and relieve some pressure off the CPU's cooling mount and see if that alleviates the issue.

The computer will randomly reboot to BIOS, with the M2 NVMe drive not detected. If the PC is fully powered down and then powered back up, the drive is detected and it boots as normal.
Do you feel a mild tingling sensation when you touch the metal part of your case and your bare feet are making contact with a tiled(non-wooden, non-carpeted floor)? If so, you might be having a grounding issue.

I updated to AMI BIOS 7E12v1E (2024-08-13), the latest non-Beta build available (there are two newer builds that remain in Beta state). The CMOS should have cleared. I suppose I can clear it again?

The PSU is brand new for this build. It is a MSI MEG Ai1300p 80+ Platinum 1300W module.

The processor is cooled through a dedicated heatsink (it comes with one attached). Temperatures on the PSU have not spiked in any meaningful way in any of the monitoring I did (and in any case it has never shut down under load - only when it has been idle for hours). There is no notable pressure on the cooling mount.

There is no tingling issue that I can detect.

Thank you for your response!

 

[LordHuffnPuff]

Some addition information...

RAM passes a memtest fine (and is new)
There are two monitors, one connected to the GPU via HDMI, one connected via DisplayPort

I guess I could buy another drive, clone this one to that one as a backup, do a long format on this drive and see if that solves it, because maybe the factory format is bad somehow? That seems like a stretch but it's possible.

 

[Ralston18]

This motherboard?

https://download.msi.com/archive/mnu_exe/mb/MAGX670ETOMAHAWKWIFI.pdf

Do verify that I found the applicable manual.

How was the RAM installed?

The motherboard requires that the first physically installed RAM be placed in DIMMA2 (Reference English -Page 28).

Also review the other Important Notes posted there.

 

[LordHuffnPuff]

This motherboard?

https://download.msi.com/archive/mnu_exe/mb/MAGX670ETOMAHAWKWIFI.pdf

Do verify that I found the applicable manual.

How was the RAM installed?

The motherboard requires that the first physically installed RAM be placed in DIMMA2 (Reference English -Page 28).

Also review the other Important Notes posted there.

Thank you for responding - that is the correct manual. The RAM is installed in DIMMA2 and DIMMB2 as recommended.

 

[Ralston18]

Next I will ask about Reliability History/Monitor and Event Viewer.

Is either or both tools logging any error codes, warnings, or even informational events just before or at the times of the BSODs?

 

[boju]

Possibly it may be physical connection between Nvme drive and M2 slot. Try reconnect/reseat the Nvme drive, the M2 screw or latch properly fastened at the rear of the card not sitting on it.

Try another M2 slot afterwards if symptoms persist. Don't need to reinstall Windows moving drive location.

 

[LordHuffnPuff]

Next I will ask about Reliability History/Monitor and Event Viewer.

Is either or both tools logging any error codes, warnings, or even informational events just before or at the times of the BSODs?

There are no BSODs as far as I can tell - it's simply a loss of power (though I've never been awake to witness it). Event Viewer and the rest show nothing whatsoever. There are no log files.

Possibly it may be physical connection between Nvme drive and M2 slot. Try reconnect/reseat the Nvme drive, the M2 screw or latch properly fastened at the rear of the card not sitting on it.

Try another M2 slot afterwards if symptoms persist. Don't need to reinstall Windows moving drive location.

This was my next plan - I will report back, though it may take a few days for the problem to resurface as it's pretty inconsistent.

 

[boju]

Not sure if you've mentioned updating Nvme firmware with Samsung Magician, that's another thing id recommend doing as well. Also no need to reinstall Windows after updating firmware.

 

[LordHuffnPuff]

Not sure if you've mentioned updating Nvme firmware with Samsung Magician, that's another thing id recommend doing as well. Also no need to reinstall Windows after updating firmware.

I had not done that, but it turns out that it has the most recent firmware already.

I swapped the M2 to a different slot (M2_4 -> M2_1) so we'll see how that goes.

 

[boju]

Good luck. All slots should function properly but yeah, M2_1 should be used first anyhow as the SSD communicates directly to cpu whilst others are chipset driven, better performance from cpu.

 

[LordHuffnPuff]

Well, here we are about two full days later and the problem has reoccurred, so changing the slot did not fix things. As usual it occurred during a period of full idle, when I had been out of the house for hours.

Given that none of these crashes are producing a log file (noted by the Event Viewer, "Dump file creation failed due to error during dump creation.") my next suspicion is to replace the drive itself even though all diagnostics return that it is in perfect health (and, of course, it's brand-new...) unless somebody has a better interim step. An expensive next step.... these M2 drives are a lot.

 

[LordHuffnPuff]

Okay, I'm back with a potentially new piece of data.

I've been experimenting with putting my computer into sleep mode when I'm not actively using it, as the issue only occurs when the computer is idle. I have to date never had the computer restart without hard drive detected while in sleep mode.

What this indicates to me is that this is NOT a power issue, because the power flow is either on or it's off - it the power were cutting out sleep mode should be irrelevant, right? This means it's either the M2 or the Motherboard almost certainly.

If I'm wildly off-base please let me know - I feel like I'm grasping at straws a little bit.

 

[boju]

I be grasping at straws as well.

So when you come back to your computer do you still find it sitting in bios still?

Could try another SSD or motherboard but if everything else is running correctly, just a change of operation suggestion, maybe power off completely when leaving house for lengthy times. You might feel better

 

[LordHuffnPuff]

I be grasping at straws as well.

So when you come back to your computer do you still find it sitting in bios still?

Could try another SSD or motherboard but if everything else is running correctly, just a change of operation suggestion, maybe power off completely when leaving house for lengthy times. You might feel better

Only sometimes - this is an intermittent problem. I can leave the PC on for two or three days without issue and then sit down, use it for twenty minutes, go eat lunch and when I get back in an hour it's at BIOS. There doesn't seem to be any pattern to it.

I've been testing sleep mode because electricity doesn't care if you're in sleep mode. If the power is cutting out, then it will cut out regardless of what the software is up to. So far, it has never gone to BIOS while in sleep mode. This isn't a perfect solution of course, and it totally precludes remote access.

 

[boju]

I don't think it has anything to do with sleep states either. Motherboard or 990pro, SSD would be easiest to replace first. Do you have a spare hdd to test with? whilst having SSD removed temporarily.