Diary of an suddenly unstable x670e - Asus experience sought

ginjaninja

ginjaninja

ginjaninja

Associate
Joined
30 Jul 2007
Posts
1,281
Asus TUF X670e-plus
Corsair 2x32GB CL-30 configured to Expo 1
7800 X3D
NH-D15 cooler
Asus 4090 OC
Corsair NVME

System ran fine for 1.5 years....slightly undervolted CPU/GPU,
Started to get game instability Mid January...came and went...some days fine some days the odd game crash..
Stress testing with F95 prime and Uniengine, Furmark never showed an issue.
Temps are all good

Updated Mboard Version 3201 Beta...possibly not sensible in hindsight
Had a week on normal behaviour......some days fine some days the odd game crash.

Today the system was completely borked and refused to boot with a permanent orange dram light.
Within a lot of unknowns and guestimates the following was certain

1)I was getting a bios post message saying "New CPU installed, fTPM/PSP NV corrupted or fTPM/PSP NV structure changed...Y to accept new or N to keep old"
N kept giving me the same message on each boot, so i opted for Y and now post has settled down....I hadnt touched the CPU at this stage so no idea what that was about...And if new bios related why it took a week to show up.

2) At one stage, With only cpu and memory plugged in, the bios would not get passed initial checks. permanent orange light (waiting a few minutes in case training related)...which i understand is DRAM related. Tried swapping memory and slots to isolate but in the end have everything back in and cant reproduce the problem. Suspect the gremlin is somewhere between CPU<>MBOARD<>RAM. (and a complete guess that the issue is temperature related in the physical expansion/connection sense...not the over heating sense)

3) At one point the BIOS UI was corrupted..text supposed on text, responsive (i could press f10 and i would get the save and exit prompt ungarbled). Repeatable after a reboot. started to feel things were getting away from me..reseated everything... UI corruption problem never came back.

Current state
Everything reseated and back in....back to normal......things generally working fine...cyberpunk 2077 path tracing been running for a few hours.

Would love to know how the fTPM message came about and whether its indicative of a CPU issue?
There definitely is a hardware problem somewhere....but it comes and goes.
I feel the only way im going to track this down is with replacement parts. May take 'opportunity' to upgrade to RAM then 9800X then eventually new mboard perhaps..

thanks for any experience/advice.
 
ginjaninja said:
Would love to know how the fTPM message came about and whether its indicative of a CPU issue?
There definitely is a hardware problem somewhere....but it comes and goes.
I feel the only way im going to track this down is with replacement parts. May take 'opportunity' to upgrade to RAM then 9800X then eventually new mboard perhaps..
Click to expand...
A few possibilities I can think of:
1. The board did some kind of automatic reset due to a failure to boot.
2. There's some kind of BIOS corruption (rare, but it does happen).
3. There's a mounting issue, maybe due to the cooler shifting.
 
Through more testing any issues/the issue is masked/fixed by turning off expo 1 and running memory at default ..4800 i think...

Any ideas what cpu/ram voltages to nudge up to try and achieve 6000Mhz stability? If memory clock is the root cause then it seems something degraded over time which is a bit worrying*, so maybe i would just be buying time deferring the inevitable and the cpu/ram silicon just isn't up to it. Thats said game performance between 4800 and 6000 isnt noticable (on cyberpunk at least).

On that note...cyberpunk was exhibiting an odd behaviour which i attribute to running at 6000Mhz, namely significant screen stutters every 30 seconds or so. No issues at 4800.

edit *maybe when theres a new non beta bios i can try that again on Expo 1...could just be the bios
 
Last edited:
A couple things I’d try myself

-Replace the CMOS battery
-Run Testmem5 using the ANTA777 extreme preset to check that your RAM is fine
-Manually set the voltages for RAM, CPU, SOC, etc (this is less likely an issue though if you updated your BIOS)
-Updated your BIOS to a non BETA version - if that doesn’t work, change the BIOS again (sometimes a BIOS version can be borked)
-email support and see if it’s a wider issue that they can assist with.
 
ran each stick in each slot (1 at a time) for a few days no problem detected..i dont think the memory is bad and running 1 stick at 6000 seems to be no problem for cpu...thanks for tm5 recommendation @JollyJamma.
then i found a new bios 3207 released with notes "This BIOS enhances system performance and improves high-capacity memory compatibility. "
I dont know if this applies to 2 x 32GB dual rank memory but thought it couldnt hurt...

Tm5 testing away now..back on 2x32Gb at 6000....so far so good
 
hmm...i think there are/were multiple things going on / my testing was invalid....highlighting the need for trust worthy test tools/more careful testing methodology..
(Sooner i get off beta Asus bios the better.)
tm5 extreme seems happy running 2x32GB @ 6000 all day.
One of the games i was using, Foundation, is still crashing after a few hours but i now realise the game is unstable for many/acknowledged by the developer so that crashing is possibly a red herring.
I don't know whether i can trust/rely on cyberpunk 2077 pathtracing transformer model results, maybe that crashes too from time to time; they are fairly recent patches.

Any more / other recommendations for test tools out there for ensuring 100% [game] stability please?
thanks
 
Update drivers for both your motherboard and GPU.

Also try running your RAM at a slower speed - some games don’t like high speed RAM and can crash.

Something like 5200 MT/s with latency settings on auto should work.

Hard to tell without testing each and every factor.

Some games are a pile of garbage tho.
 
ginjaninja said:
Asus TUF X670e-plus
Corsair 2x32GB CL-30 configured to Expo 1
7800 X3D
NH-D15 cooler
Asus 4090 OC
Corsair NVME

System ran fine for 1.5 years....slightly undervolted CPU/GPU,
Started to get game instability Mid January...came and went...some days fine some days the odd game crash..
Stress testing with F95 prime and Uniengine, Furmark never showed an issue.
Temps are all good

Updated Mboard Version 3201 Beta...possibly not sensible in hindsight
Had a week on normal behaviour......some days fine some days the odd game crash.

Today the system was completely borked and refused to boot with a permanent orange dram light.
Within a lot of unknowns and guestimates the following was certain

1)I was getting a bios post message saying "New CPU installed, fTPM/PSP NV corrupted or fTPM/PSP NV structure changed...Y to accept new or N to keep old"
N kept giving me the same message on each boot, so i opted for Y and now post has settled down....I hadnt touched the CPU at this stage so no idea what that was about...And if new bios related why it took a week to show up.

2) At one stage, With only cpu and memory plugged in, the bios would not get passed initial checks. permanent orange light (waiting a few minutes in case training related)...which i understand is DRAM related. Tried swapping memory and slots to isolate but in the end have everything back in and cant reproduce the problem. Suspect the gremlin is somewhere between CPU<>MBOARD<>RAM. (and a complete guess that the issue is temperature related in the physical expansion/connection sense...not the over heating sense)

3) At one point the BIOS UI was corrupted..text supposed on text, responsive (i could press f10 and i would get the save and exit prompt ungarbled). Repeatable after a reboot. started to feel things were getting away from me..reseated everything... UI corruption problem never came back.

Current state
Everything reseated and back in....back to normal......things generally working fine...cyberpunk 2077 path tracing been running for a few hours.

Would love to know how the fTPM message came about and whether its indicative of a CPU issue?
There definitely is a hardware problem somewhere....but it comes and goes.
I feel the only way im going to track this down is with replacement parts. May take 'opportunity' to upgrade to RAM then 9800X then eventually new mboard perhaps..

thanks for any experience/advice.
Click to expand...
I had a Asrock x470 tachi where all the non-cpu pice lane devices and slots just died, took me a while to work out that what was going on.

System still booted an ran though.


Said motherboard now hangs on my wall as decorative part.
 
Last edited:
I upped the VDDSOC voltage by 0.06 to 1.3v.....so far so good with other games....no crashing for a few days..if it lasts a week i may have made progress.
 
You must log in or register to reply here.
Back
Top Bottom