[SOLVED] Mini PC still crashing

Jeff Drew

Contributor
Joined
Jan 22, 2023
Posts
64
Location
Switzerland
I sent the mini PC in for repair, and it came back yesterday. I reinserted the RAM and M.2 drive, reconfigured it back on the network, and let it run for a few hours. It crashed again overnight. I ran the file collection app this morning and it crashed while running. I was able to get a successful run and have attached the results. Maybe all the crashes also screwed up something in the software. Can you see anything in this?
 

Attachments

From the dumps this looks as though this could be RAM related. I can see that your Kingston Impact DDR4 RAM is overclocked to 3200MHz. Whilst the RAM is supposed to handle that speed it's still an overclock. Please disable the RAM overclock in the BIOS (via XMP/DOCP or whatever) so that the RAM runs at its native speed. See whether it's stable there.
 
OK, so, two things -
1. I can't find a way to set the memory speed. The BIOS is very limited. I tried all the F keys and CTRL combos to get it into advanced mode, but I didn't come upon a combination that works. I searched for how to get the BIOS into advanced mode, but haven't come up with it yet. Still trying and I will let you know.
2. In the meantime I got a BSOD with KMODE exception not handled. I re-ran the collector.
 

Attachments

You don't see anything in the BIOS labeled XMP? If you do, disable it.

It could be under M.I.T. or Tweaker. If you don't see MIT or Tweaker look under Advanced or System Configuration. XMP could also be called EOCP
 
Laptops and mini PCs often have very limited BIOS options.

Why do you have Driver Verifier enabled? It wasn't enabled in the dump of 26th April, but it was enabled in the dump of 7th May (and in this latest dump). This latest BSOD was caused by a page fault in a Driver Verifier module...
Rich (BB code):
NOTE: The trap frame does not contain all registers.
Some register values may be zeroed or incorrect.
rax=0000000000000000 rbx=0000000000000000 rcx=fffff805846b78d0
rdx=0000000000000000 rsi=0000000000000000 rdi=0000000000000000
rip=fffff805158f5a67 rsp=ffff9306f513a470 rbp=000000000381a800
 r8=0000000000000013  r9=ffffdf86a6406c58 r10=fffff80583e62840
r11=ffffa5fd92200000 r12=0000000000000000 r13=0000000000000000
r14=0000000000000000 r15=0000000000000000
iopl=0         nv up ei pl zr na po nc
VerifierExt!DsIncrementTracker+0x43:
fffff805`158f5a67 8b08 mov ecx,dword ptr [rax] ds:00000000`00000000=????????
The ???????? indicate that the memory location being pointed to (by the RAX register) is invalid. This is likely because the RAX register is all zeros. Unfortunately there is no indication anywhere else in the dump whether Driver Verfier was faulting a third-party driver - that's what Driver Verifier is for. It does need to be enabled properly however.

For now disable Driver Verifier by entering the command verifier /reset, then reboot and see how things are.
 
It looks like I neglected to turn off the drive verifier before I sent the machine out for repairs. I turned it off and rebooted. Let's see what happens.
 
Another crash. I have run sfc and dism and no problems were found. I ordered 2666 RAM. It should be here tomorrow. Is there any indication that the RAM is the issue? This is new RAM. I still have the old sticks, but they are 3200 also.
 

Attachments

There were no new dumps in that latest upload. I can see a crash on 3rd June that wasn't caused by a BSOD, that suggests that something external to Windows was at fault (and Windows wasn't able to catch it and BSOD). There are other general indications in the other logs that bad RAM could be a possibility here, and I note that you installed RAM and an M.2 drive when it came back from repair...

I'd like you to remove the M.2 drive and re-insert it fully. I have seen issues caused by improperly seated M.2 drives - make sure the contacts are clean too.

I'd also like you to remove and re-seat the RAM sticks. Make sure the contacts are clean here too - but don't blow on them with your mouth. If it still BSODs or crashes afterwards I'd like you to remove one RAM stick and run on just the one for a while - check your manual to ensure that one RAM stick is in the correct slot. Swap the RAM sticks after a few days, or if you get a BSOD/crash, and run on just the other stick for a few days. That will indicate whether one stick is flaky.
 
I removed and reinserted both the M.2 and the memory. I cleaned the contacts. I had a crash last night. This one was weird in that the screen was blank this morning, and I couldn't get the machine to respond. I had to force a reboot. When it came back, it didn't know the PIN was set or that it was joined to a domain. I reset the PIN and rebooted. It knew the PIN. I disabled virtual machine support. This machine doesn't need it. I haven't popped one memory yet, but I will do that next. Attached are the latest scan results.
 

Attachments

DxDiagx86

Windows Error Reporting: 1,2, 3,6,7,8,9,

All crashes are connected with MicrosoftEdgeUpdate.exe and OnLogonLaunchError

0x80070002

What does 0x80070002 mean?

1. What does error 0x80070002 mean? You might get this issue of Error(s) found: Code 80070002 Windows Update encountered an unknown problem when you're trying to install some Windows 10/11 updates, such as KB3200970, as well as other numbers. It means that Windows Update can't find the updates.
 
Your logon issues are because the system is setup to use a domain controller, there are regular errors in the System log for this...
Code:
Log Name:      System
Source:        NETLOGON
Date:          06/06/2025 08:50:02
Event ID:      5719
Task Category: None
Level:         Error
Keywords:      Classic
User:          N/A
Computer:      NormaDesktop.DREW.local
Description:
This computer was not able to set up a secure session with a domain controller in domain DREW due to the following: 
We can't sign you in with this credential because your domain isn't available. Make sure your device is connected to your organization's network and try again. If you previously signed in on this device with another credential, you can sign in with that credential. 
This may lead to authentication problems. Make sure that this computer is connected to the network. If the problem persists, please contact your domain administrator.  

ADDITIONAL INFO 
If this computer is a domain controller for the specified domain, it sets up the secure session to the primary domain controller emulator in the specified domain. Otherwise, this computer sets up the secure session to any domain controller in the specified domain.

I can see the crash on the night of 5th June in the System log, but there are no errors preceding it that might indicate a cause. I looked in your Application log at the same time period and there are no obvious entries in their ether that might suggest an imminent failure. This still looks very much like a hardware issue, since Windows has had no opportunity to catch the failure, the hardware just failed and the rug was pulled from under Windows.

I'd like you to test your RAM a bit more thoroughly now by removing one stick for a few days (or until you get another BSOD or crash). Then swap the sticks over and run on just the other stick for a few days (or until you get another BSOD or crash). This will clearly show whether one stick is flaky, it's a more reliable test than running a memory tester and it means that you can continue to use the PC.

I do have another suggestion, but let's se how the RAM test goes first...
 
Oddly, I have 2 domain controllers, so it should have found one of them. Anyway, I tried one stick out, and it crashed. I tried it with the other stick, and it crashed. I tried the new 2667 RAM, and that crashed too. It doesn't look like RAM. This machine has just returned from RMA, so I would expect that they have made some repairs, but maybe not. Worst case scenario, I guess I can RMA it again. But, before that, you have another idea?
 

Attachments

I can see the RAM is running at 2666MHz, I can also see the crash on 7th in the logs, but again there is nothing in the logs to indicate why. This still looks like a hardware problem to me.

Download Sumsung Magician and run the maximal diagnostic test on your 980 Pro. Also display the SMART data and see whether anything is indicated in there?

Although I think that this has a hardware cause, it is worth enabling Driver Verifier to see whether that can catch a misbehaving third-party driver. Follow the instructions at Driver Verifier Instructions - BSOD - Windows 11, 10, 8(.1), 7 and Vista. If you've never enabled Driver Verifier before then have a read through these instructions before you start. Note that you may see some performance degradation with Driver Verifier enabled, this is normal and is expected.
 
I returned the 2666 RAM because it crashed as well, and I had a few days left to get a refund. I already ran the maximal diagnostic. I was reviewing the Samsung Magician and noticed it has a setting to power down the drive. I disabled that. So far, the machine has not crashed. The SMART data shows everything is green (well, they use blue, but all good). I'll reenable the driver verifier, but I only have a few days left before I go on vacation until September. So, I will close this and reopen it in September if it is still crashing. I feel like they didn't do anything when I sent it in for repairs. I have some time left on the warranty, so I may send it back for repairs again.
 
Another crash and then lock-up. I tried to remote in and wake it up, but neither worked. Attached is the latest run. I thought the setting on Samsung Magician would do it, but it didn't. I am leaning towards sending it back for repairs again. I am pretty sure they didn't do anything when I sent it the first time.
 

Attachments

That last crash was a BSOD and produced a dump. One thing I notice in that dump is that the failure happened as the processor came of of the idle state - I've seen this before.

Can you go into the BIOS setup and disable all C-States for all processors. This will stop the processors entering a low power state when they go idle, and that means no power-up transition is needed when they have new work. The CPU will run a tad warmer when idle and you will consume a fraction more power, but neither are significant in a desktop.

See whether that helps?
 
Hey Ubuysa, thanks again for your work. I don't have a function for changing C-states in my BIOS. It is very limited. I can adjust the PPC, but it was already set to P0. I noticed the NX mode was enabled and disabled that. As a reminder the drive verifier is running.
 
+++ WER4 +++:
Fault bucket AV_VRF_nt!KiCheckForTimerExpiration, type 0
Event Name: BlueScreen
Response: Not available
Cab Id: dd9d86e8-0caa-441d-8130-582f0c78c6df

Problem signature:
P1: a
P2: fffff7800000001a
P3: ff
P4: 4c
P5: fffff802a0336a69
P6: 10_0_26100
P7: 0_0
P8: 256_1

+++ WER8 +++:
Fault bucket 1641377438867682891, type 5
Event Name: MoAppCrash
Response: Not available
Cab Id: 2156577870405239723

Problem signature:
P1: AppleInc.iCloud_15.3.152.0_x64__nzyj5cx40ttqa
P2: praid:iCloud
P3: 475.10.0.1
P4: 6805decd
P5: dwmcorei.dll
P6: 10.0.22621.1006
P7: f4b45ccd
P8: c0000005
P9: 00000000001967ef

+++ WER9 +++:
Fault bucket 1852923122151023629, type 5
Event Name: RADAR_PRE_LEAK_WOW64
Response: Not available
Cab Id: 0

Problem signature:
P1: jabra-direct.exe
P2: 6.22.11401.0
P3: 10.0.26100.2.0.0

.................

MoAppCrash:

What does MoAppCrash mean?

Notice the Problem Event Name is “MoAppCrash.” This means a Modern App (aka UWP app, usually an MS Store App of some kind) has crashed.

.................

RADAR_PRE_LEAK_WOW64

can indicate issues with your system's hardware configuration. You run the Windows Memory Diagnostic Tool (mdsched) to help troubleshoot your issue.
 

Has Sysnative Forums helped you? Please consider donating to help us support the site!

Back
Top