BSOD even after replacing all parts

Ihaveworms

Ihaveworms

Ukfay Ancerkay
Joined
Jul 25, 2006
Messages
4,758
A friend has been having his PC BSODing for months while playing games. Some games BSOD sooner than others. I have been helping him in the troubleshooting process and I am at a loss now. We have tried pretty much everything.

The CPU, RAM, Motherboard, GPU, PSU, and even case has been replaced at this point - pretty much an entire new build. The only thing that was carried over was storage. We even tested using a different SATA SSD vs his NVMe and it was still BSODing. Anytime a part was swapped, we tried doing a fresh windows install. In addition to that, I have made sure that the bare minimum is plugged in during testing - so no usb headset or mic. Just a mouse, keyboard, and monitors. Even tried new HDMI/Displayport cables. Still crashing. Checked the power outlet to make sure voltage was good, at it was.

He has even gone down the rabbit hole of making registry tweaks and lower level adjustments that aren't helping and likely are not the cause.

These are the current specs:
Gigabyte X870 Gaming X WiFi 7 (BIOS up to date)
AMD 9800X3D
Crucial 2x16GB DDR5 6000MHz C36
Zotac 4090
be quiet! Pure Power 13M 1000W
Antec Flux Pro case

And again, the crashing was happening before he replaced all the components with the ones one by one above. I am at a loss at this point, especially since all components have been replaced and with a fresh Windows 11 install. Looking for any possible solutions at this point.
 
1) USB devices. USB thumb drives left plugged in. USB water pump controllers reused? Swap the keyboard, mouse and even try a different monitor if the kb and mouse don't solve it.

2) For blue screens in general in this scenario most common cause is RAM somehow. There's always a Chance you swapped a bad or incompatible set with another that is also unstable with that particular board and bios. Make sure the new mb bios is fully updated to maximize RAM compatibility.

Run ycruncher passes on it and see if it's stable.

3) Try a video card driver from a few months ago.
 
4. Also check for dirty or unstable AC power. Including not always on other devices that make line noise while on, and happen to be on or near the same circuit.
 
Advil said:
1) USB devices. USB thumb drives left plugged in. USB water pump controllers reused? Swap the keyboard, mouse and even try a different monitor if the kb and mouse don't solve it.

2) For blue screens in general in this scenario most common cause is RAM somehow. There's always a Chance you swapped a bad or incompatible set with another that is also unstable with that particular board and bios. Make sure the new mb bios is fully updated to maximize RAM compatibility.

Run ycruncher passes on it and see if it's stable.

3) Try a video card driver from a few months ago.
Click to expand...

1) Already tried removing all unneeded USB devices. Just keyboard and mouse is plugged it. Also tried different keyboard and mouse.
2) Already tried two sets of RAM. Even the first set passed memory tests, but it was the last thing we swapped out since nothing else has worked. This new kit we are still getting BSOD. Motherboard BIOS is up to date.
3) He tried both his old card (6750 XT) and new card (4090), still crashing.
 
MikeTrike said:
4. Also check for dirty or unstable AC power. Including not always on other devices that make line noise while on, and happen to be on or near the same circuit.
Click to expand...
He bought a tool to test for power "noise" and his outlet used for his PC was considered the cleanest of the ones he tested.
 
Ihaveworms said:
He bought a tool to test for power "noise" and his outlet used for his PC was considered the cleanest of the ones he tested.
Click to expand...
And nothing on or near the circuit occasionally runs? Meaning the line is clean until it kicks on?

Also, is he using wired or wireless peripherals?

If wired, is he a big ol husky guy? I've had end users BSOD their machines when they are fatties, where their pants rub together when they walk, building up static electricity, crashing their PC when they get back to their PC and touch the mouse.

Even if not fat, maybe feet dragging on carpet.

Long story short, don't rule out static.
 
MikeTrike said:
And nothing on or near the circuit occasionally runs? Meaning the line is clean until it kicks on?

Also, is he using wired or wireless peripherals?

If wired, is he a big ol husky guy? I've had end users BSOD their machines when they are fatties, where their pants rub together when they walk, building up static electricity, crashing their PC when they get back to their PC and touch the mouse.

Even if not fat, maybe feet dragging on carpet.

Long story short, don't rule out static.
Click to expand...

Yeah I asked about the some sort of electrically noisey things turning like a vacuum or something else when it crashes, but no pattern to that. It's been running in the same spot from Jan 2023 to Sept 2025 without issue. No new things introduced on that circuit around the time the issues started. And his crashing isn't happening on contact with the mouse. It takes a few minutes, length of time it takes to crash also depends on what he is playing. Some games BSOD faster than others.

His floor isn't carpet. His thickness in normal. Wired keyboard and mouse.
 
Last edited:
Does he always run it in the same spot or has he tried running it in another side of the house?

Getting it completely off the circuit, as well as moving it away from potential emi. Sounds very strange...
 
"He has even gone down the rabbit hole of making registry tweaks and lower level adjustments that aren't helping and likely are not the cause."

^This is troubling.

What is the error code from the BSOD?

Ram in the correct slots?

What was the hardware previously? Was it exactly the same just replaced with new?

Was AC power tested while the pc was gaming and drawing higher current or was it just tested when idle?
 
MikeTrike said:
Does he always run it in the same spot or has he tried running it in another side of the house?

Getting it completely off the circuit, as well as moving it away from potential emi. Sounds very strange...
Click to expand...
Haven't tried a different circuit yet.
Darunion said:
"He has even gone down the rabbit hole of making registry tweaks and lower level adjustments that aren't helping and likely are not the cause."

^This is troubling.

What is the error code from the BSOD?

Ram in the correct slots?

What was the hardware previously? Was it exactly the same just replaced with new?

Was AC power tested while the pc was gaming and drawing higher current or was it just tested when idle?
Click to expand...
As for rabbit hole shenanigans, he tried changing in registry the TdrDelay (and related) values, and disabling the MultiPlane overlay. We did a fresh windows install afterwards since that fiddling hasn't helped.

BSOD errors are most commonly a 0x133 DPC_WATCHDOG_VIOLATION, a 0x119 VIDEO_SCHEDULER_INTERNAL_ERROR, or occasionally a DXGI Device Hung error reported by UE crash reporter.

RAM are in the correct slots - I double checked that.

Previous hardware was the following. An overall upgrade compared to the old:
ASrock B650 PG LIGHTING AM5
AMD 7700X
Corsair Vengeance DDR5 5600
6750 XT GPU

We did not put the 9800X3D in the Asrock board and went straight to putting that in the new board. Just stating in case there were concerns of the Asrock cooking that CPU.

As for the AC power, this is the device he is using https://www.tingfire.com/ . I believe he has it plugged in 24/7. So this wasn't an inline meter like kill-a-watt.
 
For power things to look at; power cord, outlet or power strip it is connected to. Need to check voltage as close to the pc as possible to get accurate information while it is under full load.

What versions of video drivers have you tested.
 
Last edited:
He doesn't have a tankless water heater or any other such inductive system behind the wall of the pc does he?

Neutrons are the other one I am gonna blame.
 
The friend, here. Thanks for all the feedback already!

Nothing in the walls to cause interference, and nothing special about the chair.

I have tried different cords and display cables, both DP and HDMI, dropped to a single monitor, turned down refresh rates, etc.

The new platform has not seen the 6750 XT card yet. That's my task for this evening. My best guess now is that the 4090 (loaned by a coworker) is bad and the old RAM was also bad/incompatible.

The fact that it all started in September and hasn't gotten worse makes me think it's something inside of Windows, but there's nothing particularly special about my build, so I would think the PC building community would be far noisier about something in the Windows kernel causing timing issues with GPUs.
 
Those 3 BSODs seem very... video related. Not the ones you'd normally see for DDR5 RAM problems.

Where did the Zotac 4090 come from? It's clearly not new... I did find forum posts from elsewhere about DPC errors on a possibly failing Zotac 4090.

What if you've got two failing video cards for different reasons?

Or... here's an idea:

If you are desperate, I'd like you to try this specific Nvidia driver 566.36 on that 4090:
https://www.nvidia.com/en-eu/drivers/details/237728/

Make sure you do a DDU uninstall of the driver first, or reinstall Windows.

This is the last driver before 50 series support was added. I can't imagine why this could still be an issue, but I'd try it. If by some miracle it solves the problem, I'd go looking for a firmware update for the Zotac 4090 or contact Zotac.

I'm suggesting this for a reason. I've personally seen something like this before on customer systems, just not recently. It WAS a very real problem but I thought it was gone for good after mid last year as Nvidia sorted out the 50 series drivers and compatibility with the older cards.

Edit -fixed the link
 
Last edited:
Advil said:
If you are desperate, I'd like you to try this specific Nvidia driver 566.36 on that 4090:
https://www.nvidia.com/en-us/drivers/details/258745/
Click to expand...
I actually have tried that specific driver, and still saw issues. I had to pick a different test game (I can't remember what I used) since UE5 games are unhappy with something that old.

The Zotac was in a closet for a while, though inside the static bag inside the original packaging, for quite a while after he got two more specific for liquid cooling.

Something else in the middle of this saga that wasn't mentioned: I also bought and returned a 9070 XT prior to receiving the 4090 to test if my GPU was bad. I saw the same behavior with that card as I see with the 6750 XT: crashes mostly when idle/power state transitions. The 4090 mostly fails under game load. It passes OCCT 3D adaptive without issue.
 
Last edited:
Wow. You've got a classic "cursed system" going there.

The bsods still suggest video related issues somehow.
 
Ihaveworms said:
DPC_WATCHDOG_VIOLATION, a 0x119
Click to expand...
This suggests something became unresponsive to interrupts and system calls.

Ihaveworms said:
VIDEO_SCHEDULER_INTERNAL_ERROR, or occasionally a DXGI Device Hung error reported by UE crash reporter.
Click to expand...
This suggests the unresponsive device/driver is related to the video card or video subsystem.

I would try cleaning out the drivers as best you can, and reinstalling them. Also make sure the card is inserted properly, powered, and not damaged.

If that all checks out, I'd look at a new card, and return/rma that one if you can.
 
It could also be corrupted libraries related to UE, so maybe reinstall the games that are crashing too.
 
Advil said:
Wow. You've got a classic "cursed system" going there.

The bsods still suggest video related issues somehow.
Click to expand...
That's why I started where I did. Replaced the GPU, then the windows install and SSD on the M.2 slot, then tried removing everything but the GPU from the PCIe bus, then replaced the CPU (I did start seeing IOMMU errors at one point, but they are not consistent, even non-existent after the new RAM), and then the RAM. PSU was replaced after the 4090 came into the mix because I wanted to make sure I had enough power for it. Went from 850W Corsair to 1000W be quiet!.
 
Nobu said:
It could also be corrupted libraries related to UE, so maybe reinstall the games that are crashing too.
Click to expand...
All of this has persisted across multiple Windows installs on different drives, both SATA and NVMe.
 
Mostly nvlddmkm reporting error 14, but I also see 153 from the same source and 162 from volmgr. That last one is usually what writes the minidump which boils down to the Nvidia driver in the stack.
 
Volmgr has to do with disk (drive) volumes. Did you reuse cables when trying different drives? Do you have some good looking sata cables you can try swapping in?
 
I'm guessing you're on Windows 11?

Go in to C:\windows\system32 and search for nvlddmkm.sys. There should be one in the "Drivers" subfolder, and one in "DriverStore" subfolder. If one is missing in either, copy and paste it into the other folder. You may have to update the security settings for both files.
 
Volmgr has to do with disk (drive) volumes. Did you reuse cables when trying different drives? Do you have some good looking sata cables you can try swapping in?
Click to expand...
I tried 2 different M.2 drives in 2 different slots and one SATA SSD, all in isolation. Can't be cables or drives. I did fall down the rabbit hole of an NVMe issue caused by a Windows update on Phison storage controllers back around the right time frame, but the SATA SSD ruled that out.
 
El_Capitan said:
I'm guessing you're on Windows 11?

Go in to C:\windows\system32 and search for nvlddmkm.sys. There should be one in the "Drivers" subfolder, and one in "DriverStore" subfolder. If one is missing in either, copy and paste it into the other folder. You may have to update the security settings for both files.
Click to expand...
I have one buried down in DriverStore:
Code: 
C:\Windows\System32\DriverStore\FileRepository\nv_dispi.inf_amd64_7840b4313191ae17\nvlddmkm.sys
, but nothing in Drivers. Is there a path it belongs there?
 
keek said:
I have one buried down in DriverStore:
Code: 
C:\Windows\System32\DriverStore\FileRepository\nv_dispi.inf_amd64_7840b4313191ae17\nvlddmkm.sys
, but nothing in Drivers. Is there a path it belongs there?
Click to expand...
Copy the one from DriverStore and paste it into the Drivers subfolder.
 
Back to basics. Rebuild the system with one ram stick only, no video card and run on the integrated video only. Plug in only one SSD and completely delete the partitions and install from scratch. Put absolutely nothing else on or in the machine other than minimum necessary to boot and install. Nothing else. Mobo, CPU, CPU cooler, one ram stick, one SSD, one keyboard, one mouse. Mobo power plugs. That's it. Make sure front panel usbs, audio, RGB, extra fans, everything else unhooked from the mobo. Narrow it down to just the bare minimum parts to post and install and prove that SOME of these parts work stably.
 
El_Capitan said:
Copy the one from DriverStore and paste it into the Drivers subfolder.
Click to expand...
Sorry to be thick, but do you mean

Code: 
C:\Windows\System32\Drivers\nvlddmkm.sys

or

Code: 
C:\Windows\System32\Drivers\FileRepository\nv_dispi.inf_amd64_7840b4313191ae17\nvlddmkm.sys

?
 
Advil said:
Back to basics. Rebuild the system with one ram stick only, no video card and run on the integrated video only. Plug in only one SSD and completely delete the partitions and install from scratch. Put absolutely nothing else on or in the machine other than minimum necessary to boot and install. Nothing else. Mobo, CPU, CPU cooler, one ram stick, one SSD, one keyboard, one mouse. Mobo power plugs. That's it. Make sure front panel usbs, audio, RGB, extra fans, everything else unhooked from the mobo. Narrow it down to just the bare minimum parts to post and install and prove that SOME of these parts work stably.
Click to expand...
Decent thought, and I've actually done that aside from using integrated graphics. The main problem with that is, the only reliable crash avenue is the 4090 in a game with the settings turned up. For the AMD cards, it happens at random, exacerbated by low-power states like display off. I never let it sleep/hibernate, though.

I can try the iGPU for a few days, though, and see how it goes with the other cuts.
 
If this happens after a clean install, you might want to put the PC on a UPS.
 
keek said:
Sorry to be thick, but do you mean

Code: 
C:\Windows\System32\Drivers\nvlddmkm.sys

or

Code: 
C:\Windows\System32\Drivers\FileRepository\nv_dispi.inf_amd64_7840b4313191ae17\nvlddmkm.sys

?
Click to expand...
The first one.
 
I added the file in Drivers as El_Capitan mentioned, and got a fresh minidump.
 

Attachments

  • arcraiders_minidump_260108-1955.txt
    13.7 KB · Views: 0
You must log in or register to reply here.
Back
Top