Question Random "Kernel Power 41 (63)" errors, even after a fresh Windows install ?

jaskowski

Distinguished
May 3, 2016
19
0
18,510
So i reinstalled Windows 10/11 bunch of times, sometimes i get kernel power error literally after 5 minutes in clean system, sometimes doing nothing, sometimes when gaming/high load.
I have no idea, it was ok before the reinstall and now suddenly happening without any sense. Randomness is absurd.

PC specs:
CPU: i5-11400F
Mobo: MSI MAG B560 TORPEDO
GPU: Asrock Radeon RX 6950 XT Phantom Gaming 16GB OC
RAM: DDR4 GOODRAM 32GB (2x16GB) 3600MHz CL18 IRDM PRO
SSD: Silicon Power 1TB M.2 PCIe NVMe XD80
Soundcard: Asus Xonar DGX
PSU: SilentiumPC Supremo FM2 Gold 650W


Newest BIOS flashed.
CPU tested by OCCT
GPU tested by Furmark/3DMark Time Spy benchmark
RAM tested by memtest
M.2 tested by Crystaldisk

Everything passed, its ok.

Tried different ISO images of Windows, different versions of it, different pendrive.

No BSOD dump created, i have only event logs. I get seconds of lag, freeze and then BSOD kernel power, after it restarts, PC goes to BIOS because cant detect harddisk. After full power off/on its ok, disk detected - booting like normal. This is kinda new 1TB cooled by MSI mobo shield, always like 30-35'C (different slots 1 or 2 the same). Nothing is on OC except XMP profile on RAM but without that it was the same. Tried without Xonar - the same.

Now its fairly okay (i mean no BSOD when load is close to 0) but dying after playing game (Palia) for a while. The strange part is it was ok before the reinstall.
I did Power Supply test on OCCT but after CPU was reaching 95'C i turned it off - 2 minutes of test - AND THEN after ending stress - 2 minutes later bang, Crash. Then literally after entering windows another one.
For me it makes no sense, help ;)
 

Lutfij

Titan
Moderator
i reinstalled Windows 10/11 bunch of times
Where did you source the installer for your OSes?

SilentiumPC Supremo FM2 Gold 650W
How old is the PSU in your build?

Newest BIOS flashed.
Please mention the BIOS version. On that note, did you clear the CMOS after you'd verified that the BIOS was successfully flashed to the latest version?
 

Misgar

Respectable
Mar 2, 2023
1,926
517
2,590
Is the "DDR4 GOODRAM 32GB (2x16GB) 3600MHz CL18 IRDM PRO" in the motherboard QVL list? If not, it might not be a stable combination. You may need to buy new RAM, preferably on the QVL.

I would avoid invoking any XMP profiles during Windows installation on a flaky system, expecially at 3,600MT/s. If overclocked RAM causes file corruption during Windows installation, it will not recover.

Disable XMP and switch back to the JEDEC default speed of 2,133 or 2,400MT/s. Boot the computer from USB and run MemTest86+ for several complete passes, to ensure your memory is good at JEDEC speed.

Wipe the drive and reinstall Windows without XMP. Test to see if Windows is stable.

Only then enable XMP. If it's unstable at 3,600, try 3,000MT/s and work you way up until the system hangs. Reduce the XMP speed or loosen the CL (CAS) setting by one or two clock cycles (make CL number bigger). Check again for stability.
 

jaskowski

Distinguished
May 3, 2016
19
0
18,510
Official Windows ISO from Microsoft site.
PSU is like 3-4 years old.
BIOS version E7D15IMS.A90 (7D15v29 MSI site). I didnt clear the CMOS.

I dont see any GOODRAM sticks on MSI supported memory list. It was ok when i changed it, couple months from then 0 issues, problem starts after Windows reinstall.

I will test memory without XMP on.

Reinstalling Windows on old SSD drive now, M.2 out from PC for now.
 
Last edited:

jaskowski

Distinguished
May 3, 2016
19
0
18,510
Is the "DDR4 GOODRAM 32GB (2x16GB) 3600MHz CL18 IRDM PRO" in the motherboard QVL list? If not, it might not be a stable combination. You may need to buy new RAM, preferably on the QVL.

I would avoid invoking any XMP profiles during Windows installation on a flaky system, expecially at 3,600MT/s. If overclocked RAM causes file corruption during Windows installation, it will not recover.

Disable XMP and switch back to the JEDEC default speed of 2,133 or 2,400MT/s. Boot the computer from USB and run MemTest86+ for several complete passes, to ensure your memory is good at JEDEC speed.

Wipe the drive and reinstall Windows without XMP. Test to see if Windows is stable.

Only then enable XMP. If it's unstable at 3,600, try 3,000MT/s and work you way up until the system hangs. Reduce the XMP speed or loosen the CL (CAS) setting by one or two clock cycles (make CL number bigger). Check again for stability.


RAM tested on XMP and without on 2133, 0 errors.
When i changed hard drive to the old SSD it was all good untill full powerdown and couple restarts, mobo stopped at RAM yellow led.

That may suggest this can be a memory problem so i used OCCT to stress the memory 80% load.
3200/3600 crashing the test quite fast, 2133 gets 30 minutes cycle without problem.
Anyway system failed soon after with BSOD displaying UNEXPECTED STORE EXCEPTION.
Now im just using default 2133 speed and waiting for crashes...

Any idea after those clues? Is it really memory or just coincidence to crash like that, or what about that new BSOD content.
 

jaskowski

Distinguished
May 3, 2016
19
0
18,510
After hour of gaming Critical_process_died again. Default RAM values.
Literally cant diagnose any part because its dying the same way even after pointing on something like RAM.