PC Processors

[Stormrz7]

A few months ago I upgraded my CPU, motherboard and RAM and while everything runs fine, I occasionally get a random stutter out of nowhere like in this video: https://www.youtube.com/watch?v=qBtwHudKwDA

It usually happens at complete random, although on a few occasions it has occurred when opening a program. It occurs on average about once every 2-3 days, and the longest time I went without one was 10 days after reflashing the BIOS.

Here’s a list of my specs:

Ryzen 7 3700X (Not overclocked)
MSI GTX 1070 Gaming X

16GB (2x8GB) DDR4 @3200Mhz Corsair Vengeance PRO RGB RAM

Asus Strix 450F gaming II motherboard

Seasonic 550W Gold PSU

Samsung 950 Evo 150GB SSD (OS)

Western Digital 1TB HDD

Sandisk Ultra II 1TB SSD

TP-Link Archer T6E AC1300 Network card.

Originally I had a Samsung 850 EVO as my main drive, and I still had to issue there, so I did a complete clean install of Windows on the new NVMe drive.

I have tried may different things to fix this issue such as reinstalling windows, updating the BIOS, disabling/enabling D.O.C.P, reinstalling the AMD chipset drivers, making sure all other drivers are up to date. I am aware of the AMD SATA driver issue however I do not have any of those in my device manager.

I am at a complete loss here, before the upgrades I never had this issue, and I’m certain it is caused by one of the new components, however I’m not sure which. The CPU is a second hand one but it still performs perfectly well under stress tests and cinebench.

I have run memtest86 and the results did pass, and the computer has never fully crashed, only had these hiccups. And because it happens so infrequently it’s very difficult to determine what is causing them and it can take ages to figure out if anything has fixed it.

I have this custom power plan before: https://www.youtube.com/watch?v=QA84hwqLUWs, however I have still had the issue occur with one of the default AMD plans.

There are no errors in event viewer or reliability history when it occurs, however ever tiem I boot up the PC I do get a Trusted Paltform module error with Event ID 15 which says;

"The device driver for the Trusted Platform Module (TPM) encountered a non-recoverable error in the TPM hardware, which prevents TPM services (such as data encryption) from being used. For further help, please contact the computer manufacturer."

No idea if this has anything to do with it.

Update: My computer just became partially unresponsive and I got a hardware error in reliability history:

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

Problem signature
Problem Event Name: LiveKernelEvent
Code: 1000015e
Parameter 1: 25
Parameter 2: 23
Parameter 3: ffffba8816eb34a8
Parameter 4: d010310
OS version: 10_0_19044
Service Pack: 0_0
Product: 768_1
OS Version: 10.0.19044.2.0.0.768.101
Locale ID: 2057

This has happened a couple of times before but never occurs when the micro freeze does, perhaps this has something to do with it.

[johnnyenglish]

The first thing that pops out to me is minimum wattage PSU. I would've went for 650Watts at least.

It seems that you have done already a good troubleshooting with what you have, now maybe its time to check hardware components.
But before that, does the freezes happen on safe mode? Just to check out harware.
Keeping everything up to date, is sometimes the biggest offender. Try a clean system with basic windows native drivers to see if happens.

Then go after the harware.
Try removing one stick of ram, try running only with the NVMe drive and unplug the others. Try running windows offline.

One more thing, I had some weird crashes on DX11 games (only!) And the culprit was the Corsair memory modules, try fiddling with memory timings and voltages. I haven't gotten rid of the crashes but they are much much less often after I lowered voltages from 1.35v to 1.3v in DOCP.

Check if the memory is also present on the QVL.

Good luck

[Stormrz7]

Thanks for the reply, the PSU was what I had for my old build, and using PSU calculators I felt I didn't require a new one, but I will look into getting a 650W PSU, however the system still passes stress tests fine.

The biggest issue with my issue is that I can go days without it occurring, and I regularly use more than 8GB RAM, so having to use my computer for days at half ram may hinder what I can do, but I can give it a shot. As for the other drives, they were working fine with the old build, and they do contain all of my important files, so running the PC without them for days is not something that's feasible.

It seems my RAM (CMW16GX4M2C3200C16) is not listed on the QVL list, however a very similar ram is (CMW16GX4M2C3600C18).

Ok as I was writing this opening a program caused the issue to occur, and it seems this is a likely cause of it, it has happened with any random programs, even portable ones. I will see if I can cause it to occur in a clean boot, and if so, then perhaps try with 1 RAM stick.

[ThreeDee]

BIOS up to date on your motherboard? (you said reflashed, but didn't indicate which version)

Have you tested with PBO and CPB disabled in the BIOS?

How are your CPU/GPU temps? .. and case air flow? You have to keep those VRM's and memory cool too

I have read that some have had to up their voltage on their Corsair RAM from 1.35v to 1.36v ..etc .. (but that shouldn't be an issue with DOCP disabled .. unless it's setting RAM voltage to 1.20v)

Are you using any cable mods inside your case?

Is fTPM enabled in your BIOS? .. Or are you running an actual TPM module plugged into the motherboard? maybe disable it and see what happens

[Stormrz7]

Re: Ryzen 7 3700x random micro-freezes (Windows 10)

(I have been trying to send this message for the last half hour but the spam system keeps preventing it)

I do not have a TPM connected to my motherboard, only the one in the CPU, but even with a clean boot I still got the TPM error in event viewer on start up:

Log Name: System
Source: TPM
Date: 22/01/2022 13:15:15
Event ID: 15
Task Category: None
Level: Error
Keywords:
User: SYSTEM
Computer: Ryzen1070
Description:
The device driver for the Trusted Platform Module (TPM) encountered a non-recoverable error in the TPM hardware, which prevents TPM services (such as data encryption) from being used. For further help, please contact the computer manufacturer.

I flashed the latest version of the Bios from the motherboard page (https://rog.asus.com/uk/motherboards/rog-strix/rog-strix-b450-f-gaming-ii-model/helpdesk_bios) (Version 4602), but I was also running that same version before.

CPU temperatures are around 40C idle, and never go about 70C under full load, I have closely monitored cooling and it doesn't seem to be an issue, and the memory and motherboard temps never go far above 40C.

I have not tried disabling PBO and CPB options yet, I could try and see if disabling those two options has any effect soon.

I am currently running in a clean boot waiting for the issue to occur again, remember it seems to happen when I open a program, though it has sometimes randomly occurred, but lately it has been mostly opening one, plus it's incredibility random.

I do not have any cable mods installed.

[johnnyenglish · Accepted solution]

Looks like you should be better with fTPM disabled, if you can.

Lots of people are having troubles, even on Windows 10.

https://hothardware.com/news/bugs-in-ryzen-ftpm-reportedly-tanking-windows-11-performance?fbclid=IwA...

[Stormrz7]

@johnnyenglish

I believe this was the issue, I found out from reddit and disabled my fTPM about a week ago, and so far so good, not a single stutter (Fingers crossed)

As for that other error message, seems it was the new driver from my network card, I've reverted it back to the previous one and it seems fine for now.

Thanks to everyone for your help.

Also I've upgraded my PSU to a Seasonic Focus GX 650W, because yeah 550W is a little on the low side, and it gives me more head room in the future.

(Also note my main SSD is a Samsung 970 Evo 250GB SSD, just a typo in the first post that I'll correct here)

[ThunderBeaver]

I would get 2 more sticks of the same RAM so your running dual and not single channel. You could be experiencing a RAM bottleneck. 

You may want to look into a bit beefier PSU say a 750 or 850 watt.

I can't recall if the Sea Sonic PSU's have had power load/distribution problems but you may want to research that. Inadequate or uneven power can cause a lot of PC issues.

Also check the power connectors from your PSU and all PC hardware make sure they are seated properly. Best way to do this is unplug and reseat each one. I would focus on the CPU 8 or 4 pin power cable and its connection points and the MOBO 12 18 or 24 pin cable. Make sure to power down and unplug PSU power cord from the back before doing this to reduce static discharge risk.