[SOLVED] [Hang/Freeze/Crash] Is my 10 years old PSU dying? "KERNEL-POWER ID 41" mixed with "Event ID 14 nvlddmkm"

bornot

Distinguished
Dec 31, 2009
19
0
18,510
Im coming from here https://forums.tomshardware.com/thr...-nvlddmkm-amd-nvidia.3594431/page-9?view=date as they said me my issue is quite different.

Im having this issue only while playing demanding games.

Bought a 3600 on August paired with a GTX980, no problems. Replaced the 980 for a 3060Ti, on December 20th. Been playing Cyberpunk ever since with no issues.

I installed Red Dead Redemption on January 14thh, and since I got an error 14 nvlddmkm after 40 mins of gaming, screen goes black, then on, I can move around the mouse no problem but a few seconds after again black screen, so I reset by power button. Then I play for about 3 hours no problem. AND, a lot of coil whine, with Red Dead Redemption and with Cyberpunk. No Coil Whine with, GTA IV in example, or Shadow of the Tomb Raider with everything on ULTRA+++.

Then I play Cyberpunk with everything in ultra, no RTX. No errors for 50 hours of Cyberpunk. Then, I start to play it with RTX ON; no errors also.

Yesterday, Cyberpunks updates to 1.1, it seems like they are improving some quality thing with the shaders, I dont know if that means now the game uses more PSU power, but now I get error 14 with Cyberpunk also, with RTX off even.
I had the PCIE on Windows Power Options fix ON; so I tried that.
Now Im doing the Nvidia Control Panel power mode so lets see.

I also played Shadow of the Tomb Raider last week for an hour and no problems.

This is too damn random, and I cant find exactly where the issue is I dont know what to do, is pretty deppresing.

I noticed that when I power on the PC, and then I play for about 40 mins I get the error, but then if I restart (not power off, but restart) I can play the whole day without an issue. Its like it only happens after powering the PC off, not restarting it. I dont know anymore.

Drivers are 460.89 since day 1. My PSU is a Corsair TX750W V2 from 2011. Motherboard Asus B550 PLUS.

I think that my issue is power related, it has to be. No crashes with Cyberpunk for 50 hours and when they update to 1.1 (maybe making the game to consume more power) starts the issue. Issue that already have only with RDR2 wich I play on uber Ultra, a game that to my understanding is very very power hungry. And my PSU is 10 years old.

Im sorry about my english.


Edit: So, i put the Maximun Power Comsumption for Cyberpunk and after 5 mins the game start to stutters and freezes, sound ok but with noise, and it forced me to hard reset?? This never happened before. Im going to try to revert the power comsumpttion to "normal" but this should be the fix you guys talk about right? Whats happening in my case?

Edit 2: So, after 5 mins of playing and with the power comsumption back on Nvidia Control Panel back to "normal", this time a very fast and complete freeze, pc reboots it self, and in event viewer says KERNEL-POWER ID 41, same thing happen with some of the previos issues I had, some times was "nvlddmkm" and sometimes "KERNEL-POWER ID 41"

Is my PSU (Corsair TX750W V2 from 2011) dying? Its my GPU faulty? Why its starting to fail after a month of upgrading the GPU, instead of instanly when I bought it?

EDIT 3 : I just did a OCCT stress test for GPU PSU CPU etc and for 20 mins no issues.

EDIT 4: And now, Ive been playing Cyberpunk for half an hour (power consumption on normal, as before) with no issues. I just dont understand.
 
Last edited:
Solution
Did you look at this - https://forums.tomshardware.com/threads/hang-freeze-crash-event-id-14-nvlddmkm-amd-nvidia.3594431/ (nvm, I see you posted in it and they said its not the same... )

Is the motherboard on latest BIOS? it might help.
removing afterburner is also known to fix it.

have you tried running DDU and using newer or older drivers for GPU? https://forums.tomshardware.com/faq...n-install-of-your-video-card-drivers.2402269/

have you got 980, can you swap it in and see if you keep getting errors, one way to see if new GPU might be cause. Not conclusive as might also just mean PSU doesn't mind the 980. Testing GPU in another PC would be best idea.

bornot

Distinguished
Dec 31, 2009
19
0
18,510
Im coming from here https://forums.tomshardware.com/thr...-nvlddmkm-amd-nvidia.3594431/page-9?view=date as they said me my issue is quite different.

Im having this issue only while playing demanding games.

Bought a 3600 on August paired with a GTX980, no problems. Replaced the 980 for a 3060Ti, on December 20th. Been playing Cyberpunk ever since with no issues.

I installed Red Dead Redemption on January 14thh, and since I got an error 14 nvlddmkm after 40 mins of gaming, screen goes black, then on, I can move around the mouse no problem but a few seconds after again black screen, so I reset by power button. Then I play for about 3 hours no problem. AND, a lot of coil whine, with Red Dead Redemption and with Cyberpunk. No Coil Whine with, GTA IV in example, or Shadow of the Tomb Raider with everything on ULTRA+++.

Then I play Cyberpunk with everything in ultra, no RTX. No errors for 50 hours of Cyberpunk. Then, I start to play it with RTX ON; no errors also.

Yesterday, Cyberpunks updates to 1.1, it seems like they are improving some quality thing with the shaders, I dont know if that means now the game uses more PSU power, but now I get error 14 with Cyberpunk also, with RTX off even.
I had the PCIE on Windows Power Options fix ON; so I tried that.
Now Im doing the Nvidia Control Panel power mode so lets see.

I also played Shadow of the Tomb Raider last week for an hour and no problems.

This is too damn random, and I cant find exactly where the issue is I dont know what to do, is pretty deppresing.

I noticed that when I power on the PC, and then I play for about 40 mins I get the error, but then if I restart (not power off, but restart) I can play the whole day without an issue. Its like it only happens after powering the PC off, not restarting it. I dont know anymore.

Drivers are 460.89 since day 1. My PSU is a Corsair TX750W V2 from 2011. Motherboard Asus B550 PLUS.

I think that my issue is power related, it has to be. No crashes with Cyberpunk for 50 hours and when they update to 1.1 (maybe making the game to consume more power) starts the issue. Issue that already have only with RDR2 wich I play on uber Ultra, a game that to my understanding is very very power hungry. And my PSU is 10 years old.

Im sorry about my english.


Edit: So, i put the Maximun Power Comsumption for Cyberpunk and after 5 mins the game start to stutters and freezes, sound ok but with noise, and forced my to hard reset?? This never happened before. Im going to try to revert the power comsumpttion to "normal" buut this should be the fix you guys talk about right? Whats happening in my case?

Edit 2: So, after 5 mins of playing and with the power comsumption back on Nvidia Control Panel back to "normal", this time a very fast and complete freeze, pc reboots it self, and in event viewer says KERNEL-POWER ID 41, same thing happen with some of the previos issues I had, some times was "nvlddmkm" and sometimes "KERNEL-POWER ID 41"

Is my PSU (Corsair TX750W V2 from 2011) dying? Its my GPU faulty? Why its starting to fail after a month of upgrading the GPU, instead of instanly when I bought it?

EDIT 3 : I just did a OCCT stress test for GPU PSU CPU etc and for 20 mins no issues.

EDIT 4: And now, Ive been playing Cyberpunk for half and hour (power consumption on normal, as before) with no issues. I just dont understand.
 
Last edited:

Colif

Win 11 Master
Moderator
event 41 is a response to the restart, not the cause of it. Windows runs a report every start-up and if it finds previous shutdown was unexpected, it creates event 41.

but then if I restart (not power off, but restart) I can play the whole day without an issue. Its like it only happens after powering the PC off, not restarting it. I dont know anymore.
try turning fast startup off and see if you still get the crashes - https://www.tenforums.com/tutorials/4189-turn-off-fast-startup-windows-10-a.html
its possible its an old driver on PC that is causing crashes. Old drivers not written to use Win 10 power plans have problems with Fast Startup

PSU might be cause , it might be the GPU power spikes might be more than PSU can handle.
 

bornot

Distinguished
Dec 31, 2009
19
0
18,510
event 41 is a response to the restart, not the cause of it. Windows runs a report every start-up and if it finds previous shutdown was unexpected, it creates event 41.


try turning fast startup off and see if you still get the crashes - https://www.tenforums.com/tutorials/4189-turn-off-fast-startup-windows-10-a.html
its possible its an old driver on PC that is causing crashes. Old drivers not written to use Win 10 power plans have problems with Fast Startup

PSU might be cause , it might be the GPU power spikes might be more than PSU can handle.

I forgot to say that I already tried to turn off fast startup and no luck. The thing that makes me think of faulty PSU is that the moment I turned power consumption in NVCP to "Max" I had the crash just 5 mins in the game, instead of the "usual" 30-40 mins in. But when I did the stress test with OCCT it pushed the GPU CPU and PSU i guess too to the max, for 20 mins, with no crash. Then again, there were no actual video or anything 3D running, and no Coil Whine. I have Coil Whine with both Cyberpunk and RDR2, the games that make my PC to crash/reboot.

Still I dont undestand then why at first I could play all day after having the "30-mins-after-powering-on-the-PC" crash. How that makes sense in the ecuation?
 
Last edited:

Colif

Win 11 Master
Moderator
I dont undestand then why at first I could play all day after having the "30-mins-after-powering-on-the-PC" crash.
that was what made me think of fast startup. You would think if its hardware it wouldn't care if you restart or not. Its just the other behaviour that makes it look like PSU
max consumption of a 3060ti is 200watts.
750watt PSU should be enough but it is 10 years old, it could be degrading.
Its due for retirement soon even if its not cause here. Longest I have had one last is 6 years. Better to retire it before it takes part with it.

wonder if testing PSU would help
the paper clip method - https://forums.tomshardware.com/threads/what-is-the-paperclip-method-of-testing-a-psu.1336402/
or multimeter, best way as can see rails under load.
or in the BIOS to check the +3.3V, +5V, and +12V. - https://www.lifewire.com/power-supply-voltage-tolerances-2624583
 

bornot

Distinguished
Dec 31, 2009
19
0
18,510
that was what made me think of fast startup. You would think if its hardware it wouldn't care if you restart or not. Its just the other behaviour that makes it look like PSU
max consumption of a 3060ti is 200watts.
750watt PSU should be enough but it is 10 years old, it could be degrading.
Its due for retirement soon even if its not cause here. Longest I have had one last is 6 years. Better to retire it before it takes part with it.

wonder if testing PSU would help
the paper clip method - https://forums.tomshardware.com/threads/what-is-the-paperclip-method-of-testing-a-psu.1336402/
or multimeter, best way as can see rails under load.
or in the BIOS to check the +3.3V, +5V, and +12V. - https://www.lifewire.com/power-supply-voltage-tolerances-2624583

I checked the voltages and they are in the normal range, problem is the crash only happens under very high load on the GPU/PSU.
 

Colif

Win 11 Master
Moderator
Maybe you need a Repair store to check that out, as they should have a multimeter they can attach to PSU and see if its the problem under load.

You can use software like HWINFO to log sensor scores while you play but its not always accurate.
Download HWINFO = https://www.hwinfo.com/download/
when you open it, click sensors only and click run
in next window, click the gear icon and change the polling rate to 500ms and click Set. (this makes the temperature and other scores more accurate)
Click okay, and on button next to settings, that says "logging now", click it. It opens a File Explorer window that lets you create a log file in a location you will remember
file format is csv which can be read in Excel or Google docs.
It collects way more info than you need but it might show a deviation in voltages.
If you want me to read it, upload it to a file sharing website and show link here.
 

bornot

Distinguished
Dec 31, 2009
19
0
18,510
Maybe you need a Repair store to check that out, as they should have a multimeter they can attach to PSU and see if its the problem under load.

You can use software like HWINFO to log sensor scores while you play but its not always accurate.
Download HWINFO = https://www.hwinfo.com/download/
when you open it, click sensors only and click run
in next window, click the gear icon and change the polling rate to 500ms and click Set. (this makes the temperature and other scores more accurate)
Click okay, and on button next to settings, that says "logging now", click it. It opens a File Explorer window that lets you create a log file in a location you will remember
file format is csv which can be read in Excel or Google docs.
It collects way more info than you need but it might show a deviation in voltages.
If you want me to read it, upload it to a file sharing website and show link here.

Welp, after 2 hours in GTA IV, with normal temps and 40% GPU load, I had several Event ID 14 nvlddmkm. I was playing, the screen went black for a second like when you alt tab and are in fullscreen, and i was in the desktop, i could move mouse and all but after a few seconds, again, black screen, and then again in desktop, and again and again, so I reset (no hard reset) and it kept doing the same thing even when windows says "Restarting/rebooting". After that I checked event viewer and there is a "Event ID 14 nvlddmkm" for every black screen I got. This is totally new, never had the issue with GTA IV even after several hours of gameplay. Only with RDR2 and Cyberpunk, more power hungry games.

Should I start to think my GPU is the issue, faulty?
Or its just that my PSU is slowly gradually dying? Like for overheating problems? Again, is +10 year PSU but this "nvlddmkm" events and how to crashes are this time it makes me think of faulty GPU, but really I just dont know, im lost.
 
Last edited:

Colif

Win 11 Master
Moderator
Did you look at this - https://forums.tomshardware.com/threads/hang-freeze-crash-event-id-14-nvlddmkm-amd-nvidia.3594431/ (nvm, I see you posted in it and they said its not the same... )

Is the motherboard on latest BIOS? it might help.
removing afterburner is also known to fix it.

have you tried running DDU and using newer or older drivers for GPU? https://forums.tomshardware.com/faq...n-install-of-your-video-card-drivers.2402269/

have you got 980, can you swap it in and see if you keep getting errors, one way to see if new GPU might be cause. Not conclusive as might also just mean PSU doesn't mind the 980. Testing GPU in another PC would be best idea.
 
Last edited:
Solution

bornot

Distinguished
Dec 31, 2009
19
0
18,510
Did you look at this - https://forums.tomshardware.com/threads/hang-freeze-crash-event-id-14-nvlddmkm-amd-nvidia.3594431/ (nvm, I see you posted in it and they said its not the same... )

Is the motherboard on latest BIOS? it might help.
removing afterburner is also known to fix it.

have you tried running DDU and using newer or older drivers for GPU? https://forums.tomshardware.com/faq...n-install-of-your-video-card-drivers.2402269/

have you got 980, can you swap it in and see if you keep getting errors, one way to see if new GPU might be cause. Not conclusive as might also just mean PSU doesn't mind the 980. Testing GPU in another PC would be best idea.

[Be Quiet! Straight Power 11 Platinum 750W 80 Plus Platinum Modular] is on its way, e.t.a. one week.

https://www.bequiet.com/en/powersupply/1768