Solved TPM related BSOD?

Losing my mind with this seemingly TPM related crash.

Started about two weeks ago when the system would freeze when waking my system up from the monitors being off. It would hard lock with no BSOD displayed. It has now progressed to freezing between every few minutes to every few hours during normal usage .

Running W11 Pro 23H2

Event viewer would show a variety of errors related what appeared to be TPM, as well as a bunch of generic crashes.

Things I have tried since the problem started.

• Reinstalled Windows 11 Fresh. (Same issue)
• Swapped motherboard for an identical one I had on hand. (Same issue)
• Ran Memtest from bootable USB for 4 days (multiple passes)
• Updated Bios to most recent firmware. (03/12/2024 bios)
• Disabled TPM in bios, and reinstalled Windows 11 with TPM requirement disabled (using Rufus created media) (This changed the error, but I was still getting hardlocks)
• Changed out 3090 for a brand new 4060 Ti (No Change)
• Reset Bios, re-enabled TPM, disabled virtualization. (Event viewer errors returned to previous)
• Reset TPM in bios. (No effect)
• Reset TPM in Windows Troubleshooter. (No effect)

I have ordered a replacement dTPM module, but it has not arrived yet.

Here is the logs collected by the V2 Log collector Logs

Any help would be greatly appreciated.

Rel said:

I have ordered a replacement dTPM module, but it has not arrived yet.

Click to expand...

Hello and welcome to ElevenForum.

I think you solved your own problem.
You've swapped everything out. Pretty much all that's left is the TPM module.

You even reinstalled Windows, so Windows isn't the cause.


Note: When they first announced Windows 11, the market was flooded with flaky TPM modules.
Most are gone, but there are probably a few still floating around. You may have just gotten one of those.

I was hoping that was not the case, it's such a weird thing to fail randomly.

The only consistent error I have gotten throughout the whole troubleshooting process is the "GetCACaps Not Found" error which I have already tried the posted solutions for, I guess I will report back when the TPM module arrives.

Oh to clarify on the last bit, the system is currently using fTPM (part of the Ryzen 5900x), not in add in module currently. The replacement I ordered is a hardware dTPM, but its still a week out.

Rel said:

Oh to clarify on the last bit, the system is currently using fTPM (part of the Ryzen 5900x), not in add in module currently. The replacement I ordered is a hardware dTPM, but its still a week out.

Click to expand...

I'd still wait for @zbook to go through the logs. Especially with fTPM still having the problems.

Yes... I know about fTPM... I'm running on it as we speak.

@Rel

Here's a similar issue... he got a dTPM and that solved the problem?

Run Tuneup plus > post a share link






Run Furmark for 30 minutes> report the continuous test duration and tmax

Furmark test is running now.

Tuneup results is here.

Note: The system hard locked about 1 minute after finishing its run of Tuneup.

Tuneup plus results were good.

Download and install one of these temperature monitoring software:
Speccy
HWMonitor
Speedfan


Run Prime95 for three hours > report the continuous test duration and Tmax > post share links

30 Minute Furmark result can be found here.

Core temp did not go above 68C, Hot spot peaked at 78.5C

Crashed again a couple minutes after I ended the furmark run.

Its starting to form a trend about it locking up when returning from idle or going to idle.

Doing the Prime95 now.

Hi there
Another possible solution :

Copy your Windows install to a VM
Install the TPM as "an Emulated TPM"
Copy back to real drive.

The emulated TPM will be used instead of the physical one and will work -- note the emulated TPM is a Windows driver -- it's often the IBM emulated driver.

You can also get more info by selecting "Add an optional feature" from Windows settings and look for TPM diagnostics utility.

Cheers
jimbo

I will see about trying the emulated TPM this weekend.

I let Prime95 run for 16 Hours, no crash. I even unplugged the fans I have on my radiators to see if it would crash from thermal throttling (The hottest CCD got to 89C with the fans unplugged)

Here is a screenshot from HWMonitor.

I still can't understand why the crashing does not occur under heavy load, but randomly under light load/idle. Unless this has something to do with when windows tries to do tpm operations.

I also ran TPM diagnostics but did not get much out of it, I have also attached a screenshot of the results.

Have you thought of trying another B550 (or X570) motherboard?

Also, remember to disable CSM! Everytime the BIOS gets updated, you lose required settings!

I already swapped it for another B550 Tomahawk, (I have two systems roughly the same, except the other one has a 3900XT cpu in it) the issue did not transfer with the motherboard.

Per my understanding the fTPM is part of the CPU. I am just trying to figure out at this point if the whole CPU is Fubar, or if its just the fTPM.

@Rel

As mentioned above... try something like a Furmark run, with CSM enabled and with CSM disabled.
Ove the past couple years, I've noticed that some motherboards like CSM enabled (like mine), and some like CSM disabled.

Please post a new V2 share link into the newest post.

Run OCCT multiple one hour tests > post images or share links

Sorry for the late response but I ran out of time to work on this on the weekend. I tried with both CSM options but it made no difference, OCCT did not crash until after completing the benchmark and the system idled for a few minutes.

That said, I was able to track down a TPM module locally first thing this morning, and I have had zero issues since installing it. The idle crash behavior seams to have disappeared as have all the event log errors related to TPM.

If the issue reappears I will update this post but for now it appears solved. TPM Module used was this one. (Sourced from a local computer place)

Rel said:

Sorry for the late response but I ran out of time to work on this on the weekend. I tried with both CSM options but it made no difference, OCCT did not crash until after completing the benchmark and the system idled for a few minutes.

That said, I was able to track down a TPM module locally first thing this morning, and I have had zero issues since installing it. The idle crash behavior seams to have disappeared as have all the event log errors related to TPM.

If the issue reappears I will update this post but for now it appears solved. TPM Module used was this one. (Sourced from a local computer place)

Click to expand...

Wonder if you have a defective Ryzen 9 5900X?