ROG x570 Crosshair Inconsistent crashing/freezing unable to troubleshoot. Need advice

[BbyDckDddy]

I am absolutely at a loss here.

Short Version:
Computer keeps freezing, but afterburner, performance monitor, and essentially any software used for troubleshooting freezes before the issue can be identified.

Longer version:

PC specs:
Case: Corsair 5000D Airflow
Fans: 10x Lian Li AL120 Uni
MOBO: ROG x570 Crosshair VIII Hero
OS: Windows 11
CPU: Ryzen 9 5900x
GPU: Radeon PowerColor RedDevil 6900XT
RAM: 64 GB (4x16) T-Force 3600
Cooler: Asus ROG Ryujin 360
PSU: Asus ROG Thor 1200w
TEMPS: In general, everything stays below 40*, GPU tops out at around 55* under stress tests.
I don't think it matters, but I have 1Gigabit internet (ethernet), on a Nighthawk AC1900 router, and surfboard SB8200 Modem. House is ran with Cat8.

Symptoms:

*Crashes during regular use (Youtube, facebook, spotify) and gaming(no specific game, anything will cause it to freeze), nothing points towards anything being the sole cause.
*I get something similar to screen tearing when it is about to freeze. It will just be white pixelated lines going across the screen, randomly flashing.
*Task manager won't come up during freeze.
*Letting it sit and waiting for upwards of 30-45 minutes doesn't relieve the issue.
*Power switch will not turn off or reset PC during freeze.
*Only way to reset during freeze is to flip PSU switch.
*No temperature spikes on any components that I can tell. But again, it's frozen so I really can't check.
*Airflow is optimal, thermals in general don't seem to be an issue and fans/AIO pump continue to run during freeze.
*Only once has it completely crashed, lost power and went to a black screen with "No input detected". Every other time has been a freeze.
*Prior to the issues at hand, I noticed one stick of RAM had been kind of off color, no longer the case.


Attempts at fixing and/or diagnosing in no particular order(But everything was attempted again post clean windows installs):

Clean install windows 10
Clean install windows 11
Ran DDU and did clean install of new Graphics drivers(Also tried rolling back drivers prior to this).
Removed one channel of RAM at a time trying to eliminate one set/stick. Shuffled them around and attempted again. Was unable to diagnose.
Put GPU and CPU under various stress tests via programs such as 3D mark.
Ran MemTest64 for 30ish minutes - no errors.
BIOS and Task manager both recognize every component, and running at appropriate frequencies, voltages, etc.
Completely took apart PC, thoroughly cleaned and replaced all thermal compound/pads.
Replaced MOBO. Old MOBO is ASUS Prime X570 Pro, new is ASUS ROG Crosshair VIII - Issue persists.
New sticks of RAM on the way, should be here tomorrow - Direct replacements.

I am at my wits end with this thing. Admittedly, I am not a professional when it comes to diagnosing computers. I feel pretty comfortable digging around in BIOS and what not. But I simply cannot figure out the issue I'm having. I am open to all suggestions here. I've been trying to figure this out for a few days now, and have already thrown a motherboard at it. Now I have new RAM coming. Before I throw any other parts at it, does anyone have any suggestions? Any software I can run to help diagnose if any particular component is an issue? If you need more info, please let me know and I will get it to you.

[Silent Scone@ROG]

Hello,

The white pixelated lines and screen artefacts point towards a GPU issue.

First and foremost you need to make sure there is no overclock whatsoever.

This means Optimised Defaults. No XMP enabled, and no PBO.

Are you using a GPU riser cable?

Are you able to use an alternative GPU to test?

[BbyDckDddy]

Hello,

The white pixelated lines and screen artefacts point towards a GPU issue.

First and foremost you need to make sure there is no overclock whatsoever.

This means Optimised Defaults. No XMP enabled, and no PBO.

Are you using a GPU riser cable?

Are you able to use an alternative GPU to test?

Hi,

I replied to this yesterday. I'm not sure why it didn't save. Regardless, No overclocks, no riser cables, I'm really just trying to eliminate anything else before dumping money into a new GPU. I haven't been able to swap my wife's in yet. She uses hers for work quite frequently, but will at the first chance I get. I believe my warranty -JUST- (conveniently) lapsed. Any other ideas? Thanks for the reply.

[STARRAIN@ROG]

Hi BbyDckDddy,
Does the issue also occur on ROG CROSSHAIR VIII HERO BIOS 4201 and PRIME X570-PRO BIOS 4403 with bios all default settings(no OC)?
Do you have the issue since first build or after any settings or components changed?
Have you tried new sticks of RAM and another SSD to install windows?
Thank you.

[RedSector73]

T-Group Ram, you mention, its config 4x16 (is that two kits of 2x16) with 'TF4D432G3600HC18JDC01' as that is 2x16 kit.

From what I read of your posts, you have a memory issue going on. Try using one 2x16 kit (DOCP 3600) (on its own from CPU ram slots 2 & 4) and see if the issue goes. Just be aware, your install of windows was most likely corrupted if you had memory issue when install is done.

For x4 DDR4 kit, you really need to be sure it is approved for use, you can get 2 X nearly anything to work these days (even this had issues back in the day), but not 4 X Anything https://www.amd.com/en/products/ryzen-compatible-memory

I doubt these kits were ever approved or made to be used 4xRAM. Be interesting to know who is the manufacturer of the actual ram chips, Hynix or Micron (at a guess) CL 18 @ 3600

[BbyDckDddy]

Hi BbyDckDddy,
Does the issue also occur on ROG CROSSHAIR VIII HERO BIOS 4201 and PRIME X570-PRO BIOS 4403 with bios all default settings(no OC)?
Do you have the issue since first build or after any settings or components changed?
Have you tried new sticks of RAM and another SSD to install windows?
Thank you.

Hi, thanks for replying.
Yes, issues with both MOBOs and bios with no OCs.
Only had issues fairly recently, about the last two weeks at this point, I believe. Have had this variation of the build for at least six months, but I believe right around eight months.
No to both the RAM and SSD. Ram is on the way though. I am really hoping the ram fixes it.
Thanks again!

[BbyDckDddy]

T-Group Ram, you mention, its config 4x16 (is that two kits of 2x16) with 'TF4D432G3600HC18JDC01' as that is 2x16 kit.

From what I read of your posts, you have a memory issue going on. Try using one 2x16 kit (DOCP 3600) (on its own from CPU ram slots 2 & 4) and see if the issue goes. Just be aware, your install of windows was most likely corrupted if you had memory issue when install is done.

For x4 DDR4 kit, you really need to be sure it is approved for use, you can get 2 X nearly anything to work these days (even this had issues back in the day), but not 4 X Anything https://www.amd.com/en/products/ryzen-compatible-memory

I doubt these kits were ever approved or made to be used 4xRAM. Be interesting to know who is the manufacturer of the actual ram chips, Hynix or Micron (at a guess) CL 18 @ 3600

Interesting. I could've swore I checked compatibility when I bought everything, but now it doesn't appear so. However, yes, it is two kits. If that is the issue, I wonder why it ran solid for about eight months. Think it just gave finally?

[RedSector73]

Interesting. I could've swore I checked compatibility when I bought everything, but now it doesn't appear so. However, yes, it is two kits. If that is the issue, I wonder why it ran solid for about eight months. Think it just gave finally?

Eight months of 'ok' is really a good result, but I'm wondering if you maybe were not using DOCP profile and was loose & slow enough at default to just not be present.

The system, if set DOCP will use DOCP profile (read from slots 2 & 4 then assumed 1 & 4 you could try swapping the matched kits and see if this helps) for the base timings and voltage, but sub timings can change. A Bios update or reset and since that you have been unlucky, and just keeps trying an unstable setting, rather than one you got lucky with, maybe ?

From my experience, video card gone bad is artefacts then death, death of video output, or non-stop driver stop responding to the system errors and restarted. I've had one that would freeze the system, but that was like 15 years ago.

T-Group Ram is good, used it myself, just never used 4x16 in Ryzen system from them. To be honest, on first read, I was surprised it worked at all.

[BbyDckDddy]

Hi everyone. Sorry this is such a late response. I honestly forgot that I had made this post. It's been a long (and regretfully expensive) journey. However, I've learned a lot.

You won't believe me when I tell you what the fix was.

So. Lets get down to it. It was NOT the RAM. I swapped out two motherboards, two compatible RAM kits, and two CPUs. Currently using the original of each of those except the CPU. Now I have a Ryzen & 5800x 3D. That was not the fix though. I also bought the new ROG THOR II 1200w PSU because I didn't have a spare laying around to test, or a PSU tester; and I figured I should have a spare for a testbench anyway, so I pulled the trigger. That also was not the fix. I now have an ASUS TUF Gaming 3090 TI. Guess what. Not the issue.

You want to know what the fix was? Wanna know?

My keyboard. It had an internal short. Never would have guessed. We tallied up the cost of all the parts we ended up throwing at this thing and wasting money; instead of just taking a step back and doing some basic troubleshooting. It ended up coming to around $1,500 by the time we found the issue.

Before anyone asks, GPU was covered by microcenter warranty as they ripped me off for the 6900xt (something like $2400), it actually had damage caused to it by this issue.

So yeah. Lots of lessons learned. PC is up and running great now. Got a new keyboard. No issues. Pretty shocked at the fact that a keyboard was downing my whole computer. We plugged it into my wife's PC and it shut off hers as well. That's how we found out. Should've tried that from the beginning.

[Lizziezilla]

I am having similar issues when I first got the Hero VIII with the 5900x. I went through 7 kits of 3600mhz ram, G.Skill samsung B dies, corsair micron E dies, either didn't take the 3600 speed or was unstable as all hell. I finally got 1 kit to work out of them all and that was Team Group T Create 3600 CL 14 samsung B die.

However, it was still unstable, but wouldn't crash during games, it would hang or semi power down when idle. I ended up going in complete OC mode and setting everything manually for voltages. Same problem, manual tuning of CPU/ Ram voltages didn't work. So I reset bios and tried something else, I manually tuned the ram, bumped it to CL 16, but tightened the sun timing so tight that spandex would fall off. Set voltages on SoC 1.1 and set ram's ohms to 43.5. BIngo, I got more stability and even got a better 3D mark Time Spy score CPU side ( above 15000 )

However..... if left idle it will send the board into safety mode ( semi power off state at code 93 )

This tells me the bios is a buggy mess, something in the bios is not fully supporting the 5900x's and of course ASUS will not admit it or try to fix it like other issues I have called out in the past on their products ( looks at the ROG strix G15 advantage edition )