How to Solve Kernel Powerl Failure / Event ID 41 permanently?

[BlackCastle]

I bought my PC in January 2020 and till November 2023, there has been a case of 'Kernel Power Failure', aka 'Event ID 41', a total of 4-5 times. However, in December 2023 alone, there have been multiple occurrences in a single day. I have exhausted all the options I could find after searching with the Even ID. I checked all cables and connections. I checked all drivers. I checked Windows updates and even installed preview cumulative updates, hoping that maybe that would fix it.

 

I even changed my Power Supply, and bought a new Corsair HX850 80+ Platinum. I tried the Checkdisk command, DISM commands, and SFC commands- these resulted in no error. But nothing was working, so out of frustration, I installed Windows again. This seemed to fix it for about a week or two, and then it started again. It occurred in the last month and the latest is today. What else can I do? During the time it occurs, the whole PC just freezes and stays like that until I force shut down it (in my last chassis, there was no restart button) or press the restart button and force restart it. If this keeps up, my components are at the risk of getting damaged, specially the hard drives.

.

Note: My BIOS is up to date with version 5003; The RAM speed is set to 3200 MHz from the BIOS, I didn't change anything else except the CPU cooler fan curve. The D.C.O.P in the BIOS is untouched as well. No TPM is enabled either.
.
My specs___

.

• CPU: AMD Ryzen 7 5700X

• Motherboard: ASUS TUF GAMING x570-PLUS WiFi

• RAM: Corsair Vengence LPX 16 GB 3200 MHz x 2

• GPU: Zotac GeForce GTX 1660Ti Non-AMP

• Case: Corsair Graphite Series 760T

• Storage: Seagate Barracuda 4 TB 5400 RPM x 1; Seagate Barracuda 2 TB 7200 RPM x 1

• SSD: GIGABYTE M.2 PCIe SSD 256GB

• PSU: Corsair HX 850W 80 Plus Platinum

• Display(s): LG 24” 24MK600M-B

• Cooling: DeepCool AK620

• Keyboard: A4 Tech

• Mouse: A4 Tech

• Sound: Microlab TMN1 4:1

• Operating System: Windows 10 64-bit 22H2 (OS Build 19045.3930)

[AbydosOne]

3 minutes ago, BlackCastle said:

case of 'Kernel Power Failure', aka 'Event ID 41'

All that log entry means is that Windows didn't record a valid shutdown reason before it started again. There's nothing about it that's actually an issue. You're going to have to debug further than that unfortunately.

 

6 minutes ago, BlackCastle said:

During the time it occurs, the whole PC just freezes and stays like that until I force shut down it

When I've encountered those symptoms, it's been a hardware issue (RAM error, faulty GPU, dying NIC).

[DoctorNick]

4 minutes ago, AbydosOne said:

All that log entry means is that Windows didn't record a valid shutdown reason before it started again. There's nothing about it that's actually an issue. You're going to have to debug further than that unfortunately.

 

When I've encountered those symptoms, it's been a hardware issue (RAM error, faulty GPU, dying NIC).

I agree. I would try without DOCP enabled running ram at auto speeds. I had this issue on my old PC. Turned out it was faulty memory. So maybe try that next. You can also test one ram stick at the time. Remember to use slot A2 and B2. So if testing with one stick. A2. 

Also got error 41 one time and it was some issue in windows. A clean install fixed it.   @BlackCastle

 

 

[Tetras]

9 minutes ago, BlackCastle said:

The RAM speed is set to 3200 MHz from the BIOS, I didn't change anything else except the CPU cooler fan curve. The D.C.O.P in the BIOS is untouched as well. No TPM is enabled either.

To be clear, are you saying that you just manually set the DRAM frequency to 3200 and nothing else?

 

I believe Corsair LPX is generally rated to do 3200 @ 1.35v and if you didn't use XMP/DOCP, your memory voltage (and other voltages/timings that XMP/DOCP is liable to change) will still be running at stock. I don't know if that's your problem, but I wouldn't be surprised if the result was an unstable PC.

 

You can buy memory that will run at 3200 @ stock settings/voltages (e.g. Crucial Pro & Kingston value), but LPX is not rated to do this.

[will0hlep]

15 minutes ago, BlackCastle said:

I installed Windows again.

Did you do a "reset" or did you reinstall with a new installation media?

[bal723]

i have same problem but i think its my ram because its clocked at 4400mhz and 18-24-24-44 timings instead of 4000mhz 20-26-26-46

[BlackCastle]

4 hours ago, AbydosOne said:

When I've encountered those symptoms, it's been a hardware issue (RAM error, faulty GPU, dying NIC).

I think my GPU and NIC are okay, if GPU had been causing issue, I would have noticed display issues, and as for NIC, I can't really confirm anything on its side. Does Windows memory diagnosis tool work for checking RAM issues? @AbydosOne

[BlackCastle]

4 hours ago, DoctorNick said:

I agree. I would try without DOCP enabled running ram at auto speeds. I had this issue on my old PC. Turned out it was faulty memory. So maybe try that next. You can also test one ram stick at the time. Remember to use slot A2 and B2. So if testing with one stick. A2. 

Also got error 41 one time and it was some issue in windows. A clean install fixed it.   @BlackCastle

 

 

Should I stress test the RAMs to see if they're causing kernel power issues? @DoctorNick

[BlackCastle]

4 hours ago, Tetras said:

To be clear, are you saying that you just manually set the DRAM frequency to 3200 and nothing else?

Yes, I only set the DRAM frequency to 3200 MHz as the sticks supported 3200 MHz, but by default, they were running 2133 MHz. @Tetras

[BlackCastle]

4 hours ago, will0hlep said:

Did you do a "reset" or did you reinstall with a new installation media?

Reinstall with a new installation media. I downloaded the ISO with the media creation tool and created a bootable pendrive and installed it from there by cleaning C drive. @will0hlep

[--SID--]

6 minutes ago, BlackCastle said:

Should I stress test the RAMs to see if they're causing kernel power issues? @DoctorNick

Disable DOCP and try again. Ryzen has issues with lots of Corsair RAM SKUs and versions.

[BlackCastle]

An update on the case: a friend of mine suggested that it might also be a CPU issue, that because of the motherboard's own limit on power draw, whenever there is a spike in power needs and because of the limit, the CPU may not be getting enough power. Also, on Microsoft's forum, someone suggested that I turn off power saving for the CPU from BIOS, so I added (+) an offset value of 0.15 to the original volt. After that, I tried a stress test; now the CPU is averaging on 2.xx GHz to 3.xx GHz; mostly staying on 2.88 GHz with 70% to 90% usage, but on idle, it goes up to 4.5 GHz and sometimes a bit more with up to 10% usage.
.
When I first installed the CPU, I ran a stress test to test my CPU cooler and it went 100% usage with 4.5 to 4.6 GHz on everything set default. Not really sure, what could be the issue here. @AbydosOne @DoctorNick @Tetras @will0hlep

[BlackCastle]

3 minutes ago, --SID-- said:

Disable DOCP and try again. Ryzen has issues with lots of Corsair RAM SKUs and versions.

I can afford to buy another brand of RAM if needed, Should I? and if I should, which brand would be best? I am not at all interested in RGB RAMs, personally, they're waste of money. Suggest me some 3200 MHz 16GB RAMs. These are the brands that are available in my country__ Corsair, Kingston, GSkiLL, Patriot, PNY, Antec, Team, Gigabyte, Thermaltake, Addlink, Walton, Hikvision, GeIL, HP, AITC, Kimtigo, Revenger, Adata, KingSpec, TwinMos, Apacer, Transcend, Crucial, LEVEN, RAMSTA, UltraDisk, OLOy, ZADAK, Silicon Power, Lexar, Netac, Colorful, Anacomda, OCPC, Biwintech. @--SID--

[Tetras]

1 hour ago, BlackCastle said:

Yes, I only set the DRAM frequency to 3200 MHz as the sticks supported 3200 MHz, but by default, they were running 2133 MHz. @Tetras

In that case, I would put them back to 2133, OR if you want them at 3200, use DOCP/XMP and ensure the DRAM voltage is set correctly.

 

It might not be your issue, but your sticks are currently operating outside of spec and for troubleshooting purposes that is bad news.

[BlackCastle]

8 minutes ago, Tetras said:

In that case, I would put them back to 2133, OR if you want them at 3200, use DOCP/XMP and ensure the DRAM voltage is set correctly.

 

It might not be your issue, but your sticks are currently operating outside of spec and for troubleshooting purposes that is bad news.

Is it strictly a RAM issue, or because the RAM is LPX type? Would changing the RAM fix this? I am willing to change the RAM if necessary. But, I had this issue occur even when DOCP was enabled in the past. And recently, I was executing 'free space optimization' with Auslogics Disk Defrag and it was going on for about 37 hours straight before the whole PC froze and gave me the kernel power failure. That time, DOCP was enabled. @Tetras

[Tetras]

12 minutes ago, BlackCastle said:

Is it strictly a RAM issue, or because the RAM is LPX type? Would changing the RAM fix this?

LPX does have known issues with Ryzen systems, but not everyone has had problems with it.

 

12 minutes ago, BlackCastle said:

But, I had this issue occur even when DOCP was enabled in the past. And recently, I was executing 'free space optimization' with Auslogics Disk Defrag and it was going on for about 37 hours straight before the whole PC froze and gave me the kernel power failure. That time, DOCP was enabled.

All that I can tell you for certain is that your memory is rated for 3200 @ 1.35v, so if you're running 3200 @ 1.2v, that's outside of spec. Running memory (or the CPU's memory controller) outside of spec can definitely cause instability (e.g. BSODs, data corruption, freezes and resets).

 

Enabling DOCP/XMP usually sets the voltages correctly, but not always, so it is still worth checking.

 

Once you have either downclocked to 2133, or enabled DOCP/XMP for 3200 (and checked the voltages), then you can go from there, onto the next troubleshooting step. Right now, if you leave it, there's a possible cause staring us in the face.

[BlackCastle]

10 minutes ago, Tetras said:

LPX does have known issues with Ryzen systems, but not everyone has had problems with it.

 

All that I can tell you for certain is that your memory is rated for 3200 @ 1.35v, so if you're running 3200 @ 1.2v, that's outside of spec. Running memory (or the CPU's memory controller) outside of spec can definitely cause instability (e.g. BSODs, data corruption, freezes and resets).

 

Enabling DOCP/XMP usually sets the voltages correctly, but not always, so it is still worth checking.

 

Once you have either downclocked to 2133, or enabled DOCP/XMP for 3200 (and checked the voltages), then you can go from there, onto the next troubleshooting step. Right now, if you leave it, there's a possible cause staring us in the face.

Just checked the BIOS, the DRAM voltage is set at 1.35v by default with and without the DOCP. @Tetras

[Tetras]

18 minutes ago, BlackCastle said:

Just checked the BIOS, the DRAM voltage is set at 1.35v by default with and without the DOCP. @Tetras

Cool, then we can discount that as a problem.