Question PC keeps rebooting back in to Windows without any BSODs ?

[Kez0121]

Friend suggested I create a post here as I've been having issues for around ~6 months and we've tried to test each component and can't really find anything wrong.

Basically regardless of whether I am gaming or just sitting idle my PC will just freeze, sometimes my speakers begin screeching and then my monitors go off and my PC begins rebooting itself back into windows. This happens completely randomly but is happening a lot more lately (Used to be maybe once every few months, and it has already happened around 5-6 times this month alone)

My PCs RGB remains on at all times, I've done some CPU and GPU Stress tests and everything appears to be fine.

I originally suspected it was the ram, but I have replaced it and it still occurs, I secondly checked the PSU and just from checking the fan, the fact that it's cool and using HWInfo to log it, as far as I can tell it appears to be doing it's job.

The only thing I haven't been able to test is the motherboard, because I am unsure how exactly I would be able to test it without swapping it out and I don't have a spare.

tl;dr information on what happens:
PC Freezes
(Sometimes) Speakers begin screeching
Monitors go black (Best way to explain this is probably that they go into Standby?)
PC Begins rebooting (All RGB, Fans, etc all remain on at all times)
Happens randomly

PC Specs:
CPU - Ryzen 9 5900X
GPU - MSI 3060Ti Gaming Z Trio LHR
Motherboard - MSI B550 Gaming Carbon Wifi
PSU - 850W Corsair RM850X ATX
Storage - 2x 1TB Samsung 980
RAM - 32GB (4x8GB) DDR4 3600MHz Corsair Vengeance RGB Pro Ryzen Tuned
Case - Be Quiet! PureBase 500DX
Cooler - Corsair H150i Elite Capellix

Additional Information:

I have a HWInfo log of 9-10 hours of logging in which this happens (Around the 10 hour mark - the end of the logs) Within these 9-10 hours I watch videos, play games and benchmark the CPU and GPU to try and get this issue to trigger
I have benchmarked the CPU & GPU using AIDA64, Furmark and some other applications attempting to trigger the "reboot" without any luck
Originally this only ever happened when playing games, as of late it has been doing it regardless, even when sitting idle doing nothing.

Apologies if I am repeating the same information or have explained it badly, if you require any information I'll do my best to answer and if someone can direct me on the best way to upload this HWInfo log for someone to check over that would be appreciated

Edit: When this first began happening I was running windows 11, I have since downgraded to Windows 10, I have also disabled Window's automatic reboot

Edit #2:

Event Viewer Information:
"The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly."
Source: Kernel-Power
Event ID: 41
Task Category: (63)
Level: Critical
Keywords: (70368744177664),(2)

 

[Kez0121]

@op this is a good idea. Manually change the voltage only for the ram. maybe 1.40 (this is safe). You could also bump the SOC up 0.05mv. It should be set about 1.05 IIRC. With 4 DIMM slots populated this may help.

You should be able to set XMP, save setting, reboot, then manually change voltage and save again and exit. If not just set primary timings, leave everything else on auto and set voltages as suggested.

I've just taken a look in bios and my ram is sitting at 1.37v

If the extra 0.03v really matters then I guess it is something I can try out if they tell me my motherboard is fine before attempting to buy new ram/etc

 

[Vic 40]

Try the xtra voltage (would test with two matching sticks first) and if it doesn't help rma the motherboard as you say, might as well be that.

 

[Ralston18]

Also take a look in Reliability History (Reliability Monitor).

Much more user friendly than Event Viewer plus the timeline format can be very revealing.

For example: look for some errors, etc. that stopped for a week (per Post #9 above) and then started happening again.

Pay attention to error codes, warnings, and even informational events just before or at the time of the reboots.

Another place to look is in Update History: any failed or problem updates?

 

[Kez0121]

Also take a look in Reliability History (Reliability Monitor).

Much more user friendly than Event Viewer plus the timeline format can be very revealing.

For example: look for some errors, etc. that stopped for a week (per Post #9 above) and then started happening again.

Pay attention to error codes, warnings, and even informational events just before or at the time of the reboots.

Another place to look is in Update History: any failed or problem updates?

I took a look at this before I reinstalled windows, it was pretty low rated (Like pretty much sitting at a 2) but I asked a friend to check theirs too and theirs was ranked low too.

So to update -

After thinking about the ram some more I decided to remove the sticks and inspect them again, turns out I have 2 different versions, two were Ver 3.32 and the other two were 8.31

I emailed the store I bought my parts from back and discussed it with them and they said that they don't believe the ram to be the issue, they cancelled my motherboards RMA and told me to test the ram. (They also stated that they don't believe it to be the ram because the issues began a month after using the ram, they said the ram would have shown issues immediately)

So with their advice I downloaded MemTest86 and tested each stick by themselves @ 3600MHz, each stick took around 2 hours to test, by the end of the night all sticks had came back with no errors and passed.

So with that I figured I'd stick them all back in and run a test on them together. (Which took about 4 hours 30 minutes)

It threw 2 errors and showed that the ram is vulnerable to high frequency row hammer bit flips and failed the MemTest

Took this information back to the store, they advised me to test the sticks by themselves again.. so I did, again no errors.. so they are adamant the ram is not the cause and have now decided they want to go back to the original idea and have decided to RMA my motherboard again, so it is being collected on the 31st August and they again said it'll take around 3 days to test and determine if they can find a fault.

I haven't ran into it force rebooting since I reinstalled windows again but I haven't had my PC on for long amount of times lately, which I also guess doesn't say a lot because it has previously rebooted itself an hour after booting up and at other times 12+ hours after

Oh and also to add -
They said I couldn't rule out that my GPU/CPU/PSU were bad and asked me to stress test my CPU & GPU at the same time using their suggested settings, so I did and left it for 3 hours, no forced reboot or anything, temps were stable at all times, CPU hit a max of 60c and the GPU hit a max of 68c with a max hot spot of 80c

Sooooooooo.. I shall see what they say about the board when they test it, if it's not the board then I'm kind of stumped on what is causing it because I am pretty sure I've tested every part


Edit: Just decided to take another look at the reliability report since I reinstalled windows and it actually has multiple critical events stating windows was not shutdown correctly.

But one day it shows the critical error 3 times, once at 10:13am again at 10:16am and then again at 10:58am but I don't recall shutting down my PC at all on these days, typically I turn my PC on in the morning and it will remain on until I am finished with everything and I turn it off to sleep

Same again for yesterday, it's claiming it didn't shutdown correctly at 10:40am yesterday, but my PC was already on before this time and I had not restarted it all day.

Out of the 7 days since I reinstalled windows, 3 days have critical events saying windows was not shutdown correctly while the others have critical events stating apps stopped working (But they didn't and I know they didn't because I used them)

Here is a list:
21/08/22 - Samsung Magician & Xbox App (Not tech details)
22/08/22 - Windows Audio (No tech details)
23/08/22 - The 3 windows not correctly shutdown noted above
24/08/22 - Rocket League (Details just state the app crashed, I don't recall it ever crashing)
25/08/22 - Another windows not correctly shutdown noted above
26/08/22 - Another windows not correctly shutdown noted above

 

[Ralston18]

Increasing numbers of errors and varying errors suggests, to me anyway, that the PSU may be starting to falter and fail.

One thing that happens with "not shutdown correctly" is file corruption.

My suggestions:

Power down, unplug, open the case.

Clean out dust and debris.

Verify by sight and feel that all connectors, cards, RAM, jumpers, and case connections are fully and firmly in place. If unsure, disconnect and reconnect a couple of times. Something may just suddenly feel more secure and solid.

Inspect for signs of damage: bare conductor showing, melted insulation, blackened or browned components/areas, missing or loose screws, pinched or kinked wires.

When that is all done, then boot up again.

Then run "sfc /scannow" and "dism" to clean up corrupted files.

Reference:

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

Also: if you have already not done so be sure to backup all important data at least 2 x to locations off of the problem computer. Verify that the backups are recoverable and readable.

 

[Kez0121]

Increasing numbers of errors and varying errors suggests, to me anyway, that the PSU may be starting to falter and fail.

One thing that happens with "not shutdown correctly" is file corruption.

My suggestions:

Power down, unplug, open the case.

Clean out dust and debris.

Verify by sight and feel that all connectors, cards, RAM, jumpers, and case connections are fully and firmly in place. If unsure, disconnect and reconnect a couple of times. Something may just suddenly feel more secure and solid.

Inspect for signs of damage: bare conductor showing, melted insulation, blackened or browned components/areas, missing or loose screws, pinched or kinked wires.

When that is all done, then boot up again.

Then run "sfc /scannow" and "dism" to clean up corrupted files.

Reference:

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

Also: if you have already not done so be sure to backup all important data at least 2 x to locations off of the problem computer. Verify that the backups are recoverable and readable.

Already done all of that, nothing is visually wrong and scanning for corrupted files finds nothing

+ the issue followed between Windows 11 and Windows 10, and I've reinstalled Windows 10 three times now

 

[Vic 40]

After thinking about the ram some more I decided to remove the sticks and inspect them again, turns out I have 2 different versions, two were Ver 3.32 and the other two were 8.31

So with their advice I downloaded MemTest86 and tested each stick by themselves @ 3600MHz, each stick took around 2 hours to test, by the end of the night all sticks had came back with no errors and passed.

So with that I figured I'd stick them all back in and run a test on them together. (Which took about 4 hours 30 minutes)

It threw 2 errors and showed that the ram is vulnerable to high frequency row hammer bit flips and failed the MemTest

That shouldn't happen, might well be that each stick on it's own works fine, but seems that mixing gives issues. Did you test each kit by themselves as well? So kit of 3.32 on it's own and 8.32 on it's own.

 

[Kez0121]

That shouldn't happen, might well be that each stick on it's own works fine, but seems that mixing gives issues. Did you test each kit by themselves as well? So kit of 3.32 on it's own and 8.32 on it's own.

No I didn't, I would've tested it but they're coming to collect the motherboard already tomorrow so I may as well let them do that first.

I decided to recheck the reliability report and found a hardware error on 27/08/22 - I've looked into it via event viewer and the only thing to have logged at that time was a warning for bluetooth (it's built into the motherboard) I have previously had issues with the bluetooth upon further thinking, it has resulted in games crashing (Mainly PUBG, I have to disable bluetooth to even play the game or else I crash ever single game) and with the bluetooth being built into the motherboard maybe it is what is causing these issues, hopefully when they receive my motherboard and test it they can confirm it.

Information below

Spoiler: Reliability Report

Source
Windows

Summary
Hardware error

Date
‎27/‎08/‎2022 08:51 PM

Status
Report sent

Description
A problem with your hardware caused Windows to stop working correctly.

Problem signature
Problem Event Name: LiveKernelEvent
Code: 144
Parameter 1: 1020
Parameter 2: ffffac0191ebf960
Parameter 3: 0
Parameter 4: 0
OS version: 10_0_19044
Service Pack: 0_0
Product: 256_1
OS Version: 10.0.19044.2.0.0.256.48
Locale ID: 2057

Extra information about the problem
Bucket ID: LKD_0x144_INVALID_TRANSFER_EVENT_PTR_ED_0_DUPLICATE_USBXHCI!TelemetryData_CreateReport_VEN_1022_DEV_43ee_REV_0000
Server information: b8039d77-fea3-4d5b-8719-2fd7a6b9a9bf

Spoiler: Event Viewer

Log Name: System
Source: BTHUSB
Event ID: 18
Level: Information
User: N/A
OpCode: Info
Logged: 27/08/2022 08:51:29 PM
Task Category: None
Keywords: Classic

Windows cannot store Bluetooth authentication codes (link keys) on the local adapter. Bluetooth keyboards might not work in the system BIOS during start-up.

 

[pwnunow]

I am having the same exact issue and have tried literally everything you’ve tried. I also tested RAM with memtest86 and win mem diag and got no errors. I replaced my motherboard (didn’t want to wait for RMA) and it didn’t work. Now I got a perfectly good MOBO sitting at home that I can’t sell lol.
However, when testing out CPU via prime95 I get my crash also when testing out CPU,FOU and cache all together via AIDA64 I get a hardware error. But when I test CPU via cinebench or CPU on its own via AIDA64, it passes with flying colors.
I wonder if it really is the RAM, or could it be CPU?

 

[Kez0121]

I am having the same exact issue and have tried literally everything you’ve tried. I also tested RAM with memtest86 and win mem diag and got no errors. I replaced my motherboard (didn’t want to wait for RMA) and it didn’t work. Now I got a perfectly good MOBO sitting at home that I can’t sell lol.
However, when testing out CPU via prime95 I get my crash also when testing out CPU,FOU and cache all together via AIDA64 I get a hardware error. But when I test CPU via cinebench or CPU on its own via AIDA64, it passes with flying colors.
I wonder if it really is the RAM, or could it be CPU?

I've still never been able to successfully trigger it yet, I've literally tested everything and stressed the GPU and CPU using Prime95 + FurMark for 3 hours and it never triggered at all

Edit: I've also tried AIDA64 and stressed the CPU/FPU/Cache & Memory - It never triggered it

Edit #2: AIDA64 has thrown a Hardware failure this time, after around 26 minutes of stress testing

 

[Vic 40]

they're coming to collect the motherboard already tomorrow so I may as well let them do that first.

Let's see what that brings.

 

[pwnunow]

I've still never been able to successfully trigger it yet, I've literally tested everything and stressed the GPU and CPU using Prime95 + FurMark for 3 hours and it never triggered at all

Edit: I've also tried AIDA64 and stressed the CPU/FPU/Cache & Memory - It never triggered it

Edit #2: AIDA64 has thrown a Hardware failure this time, after around 26 minutes of stress testing

I get about the same… I think whatever solution you get will be the solution for me too. Did you test with the RMA’d mobo? Because I already replaced my mobo and it didn’t work.

 

[pwnunow]

@Kez0121 did the RMAd mobo work?

 

[Kez0121]

@Kez0121 did the RMAd mobo work?

Well, they've returned it and said there was no fault, they tested it with a 5800X and 2 sticks of 8GB DDR4 @ 2133Mhz

They've basically gave me no support since.. seems like their support don't really communicate with one another because they told me to MemTest the ram again, told them I already did that and that they were all fine but when put together they failed and they've said nothing since.

Since it's return I purchased a new AIO (Corsair H170i Elite LCD) because I discovered my previous AIO was faulty and have RMA'd that (doubt it was the issue though?)

It's been okay since rebuilding it, hasn't rebooted itself so far but there has been one hardware malfunction reported in the reliability report which appears to have been reported the day I rebuilt the PC which seems to direct towards "USBHUB3" but I don't have a USB Hub so not too sure on that one. I checked the event viewer and it's within there too and says it has dumped logs but when I went looking for them they're not there and clicking on the directory within the report for "attached files" states they can't be found.

Here is some info on that:

Spoiler: Event Viewer

Fault bucket , type 0
Event Name: LiveKernelEvent
Response: Not available
Cab Id: 0

Problem signature:
P1: 144
P2: 3003
P3: ffff910e06835670
P4: 40010001
P5: 0
P6: 10_0_19044
P7: 0_0
P8: 256_1
P9:
P10:

Attached files:
\\?\C:\WINDOWS\LiveKernelReports\USBHUB3\USBHUB3-20220908-1635.dmp
\\?\C:\WINDOWS\TEMP\WER-396328-0.sysdata.xml
\\?\C:\ProgramData\Microsoft\Windows\WER\Temp\WER17FF.tmp.WERInternalMetadata.xml
\\?\C:\ProgramData\Microsoft\Windows\WER\Temp\WER1810.tmp.xml
\\?\C:\ProgramData\Microsoft\Windows\WER\Temp\WER180F.tmp.csv
\\?\C:\ProgramData\Microsoft\Windows\WER\Temp\WER182F.tmp.txt

These files may be available here:
\\?\C:\ProgramData\Microsoft\Windows\WER\ReportQueue\Kernel_144_f843447367781f4430985eee7901a89a7a54_00000000_cab_0090ab3b-2db9-4fea-9307-6d9714407a02

Analysis symbol:
Rechecking for solution: 0
Report Id: 0090ab3b-2db9-4fea-9307-6d9714407a02
Report Status: 4
Hashed bucket:
Cab Guid: 0

Spoiler: Reliability Report

Source
Windows

Summary
Hardware error

Date
‎08/‎09/‎2022 04:35 PM

Status
Report sent

Description
A problem with your hardware caused Windows to stop working correctly.

Problem signature
Problem Event Name: LiveKernelEvent
Code: 144
Parameter 1: 3003
Parameter 2: ffff910e06835670
Parameter 3: 40010001
Parameter 4: 0
OS version: 10_0_19044
Service Pack: 0_0
Product: 256_1
OS Version: 10.0.19044.2.0.0.256.48
Locale ID: 2057

Extra information about the problem
Bucket ID: LKD_0x144_USBHUB3_DEVICE_ENUMERATION_FAILURE_UsbHub3!TelemetryData_CreateReport_VID_0000_PID_0000_REV_0000
Server information: 43bb49df-44fb-429e-98e0-c575b48f75ec

There has also been 2 reports of windows not shutting down correctly since I rebuilt the PC, one the day I rebuilt it and another 2 days later on the 10th September, as with previous reports of this, they show no information just literally "Previous system shutdown was unexpected" but every time I shutdown my PC I do so by shutting it down and nothing like hitting the power button or anything so I'm unsure about that.

But so far up to now while reliability report and event viewer are stating info, I haven't had it force reboot on me so far and hopefully it won't continue to happen but I'll see.

Edit: Also ran AIDA64 since rebuilding and where as previously it gave Hardware error around the 30 minute mark, this time it ran without issues and I left it for around ~2 hours (I stressed everything except the local discs)

 

[Vic 40]

That shouldn't happen, might well be that each stick on it's own works fine, but seems that mixing gives issues. Did you test each kit by themselves as well? So kit of 3.32 on it's own and 8.32 on it's own.

Looked at this since,

No I didn't, I would've tested it but they're coming to collect the motherboard already tomorrow so I may as well let them do that first.

 

[Kez0121]

Looked at this since,

Nah I haven't yet. I'm going to wait and see if the issue shows itself again first because so far it's been fine, if it happens again it'll be the first thing I test again

 

[Kez0121]

Nah I haven't yet. I'm going to wait and see if the issue shows itself again first because so far it's been fine, if it happens again it'll be the first thing I test again

My PC just died about an hour ago (this time with a BSOD at least)

Error was - "Page Fault in Non-Paged Area" which I believe is directed to a ram fault (and usually drivers, but my drivers are all fine and I have installed nothing new?)

Either way, I've contacted them back about the RMA and discussed this with them and they've agreed to let me return my ram for a refund, so once that refund is processed I'll buy some new ram (Unsure whether I want to go 4x 8GB or 4x 16GB)

Everyone suspected ram in the first place and hopefully this will finally be the end of it once the ram is replaced (again)

 

[Vic 40]

M
Either way, I've contacted them back about the RMA and discussed this with them and they've agreed to let me return my ram for a refund, so once that refund is processed I'll buy some new ram (Unsure whether I want to go 4x 8GB or 4x 16GB)

Everyone suspected ram in the first place and hopefully this will finally be the end of it once the ram is replaced (again)

Good they let you return the ram, hope that with new ram the problem will be gone.

What size is up to you, if mainly gaming is 32gb for me for the future fine, if doing content creation could the extra 32gb make a difference. Make sure when buying 4 sticks that they come as one kit, so no 2x8gb kits again. Any reason not to go for 2x16gb?

 

[Kez0121]

Good they let you return the ram, hope that with new ram the problem will be gone.

What size is up to you, if mainly gaming is 32gb for me for the future fine, if doing content creation could the extra 32gb make a difference. Make sure when buying 4 sticks that they come as one kit, so no 2x8gb kits again. Any reason not to go for 2x16gb?

Well just an update - Currently still running 16GB 3600MHz and my system has been 100% fine, they just received the ram I sent back today and have tested it and have confirmed that it was faulty and was causing BSOD along with other issues, so once that refund has been sorted I'll get to buying some new ram and I believe the issue from here on will hopefully be completely gone.

Thanks everyone.

 

[Vic 40]

That is good to hear.

 

[Kez0121]

That is good to hear.

Another update.. for some reason they reached out to me asking for my CPU and Ram speeds (Unsure why as I already gave them this information multiple times)

I stated 3600MHz & Ryzen 9 5900X

They responded by saying "Ryzen 9 5900X does not support 3600MHz" (Max support is 3200MHz but 3600MHz should still be fine from what I've read) and they're saying that because of that it is the reason the ram has "failed"

I disagree with what they are saying completely, I'm even currently running my ram at 3600MHz with no issues since removing their sticks and my PC is back to being stable and their testing of the sticks resulted in instant multiple BSOD so unsure why they're trying to suddenly blame the CPU's max memory speed on the issue.

Either way they have processed the refund and I've gone ahead and purchased 2 sticks of 32GB Corsair Dominator Platinum RGB 3600 MHz C18 (AMD Optimised apparently) and they should be arriving in about a week - Unsure why but Amazon delivery times on ram as of late seem to be long.

 

[Vic 40]

Another update.. for some reason they reached out to me asking for my CPU and Ram speeds (Unsure why as I already gave them this information multiple times)

I stated 3600MHz & Ryzen 9 5900X

They responded by saying "Ryzen 9 5900X does not support 3600MHz" (Max support is 3200MHz but 3600MHz should still be fine from what I've read) and they're saying that because of that it is the reason the ram has "failed"

That's nonsense.

I disagree with what they are saying completely, I'm even currently running my ram at 3600MHz with no issues since removing their sticks and my PC is back to being stable and their testing of the sticks resulted in instant multiple BSOD so unsure why they're trying to suddenly blame the CPU's max memory speed on the issue.

Agreed.

Either way they have processed the refund and I've gone ahead and purchased 2 sticks of 32GB Corsair Dominator Platinum RGB 3600 MHz C18 (AMD Optimised apparently) a