- Mar 4, 2025
- 2
- 0
- 10
I am struggling with a Kernel Power 41 error that is causing BSODs, reboots, and soft crashes (Desktop) while gaming. The crashes occur more often when *closing* games. I have monitored voltages, temperatures, and GPU/CPU loads during crashes and they occur at low utilization. I have not even tried a high-intensity game (Bannerlord got to about \~40% utilization). Temperatures look good while gaming (40C-60C). Voltages stable. Not sure what wattage my system *should* be using so I do not monitor it (But I can try!). I seem to crash while gaming or when closing a game only, the mid-game crashes are at low-intensity/usage points: *sitting in my half-built house in valheim* 🙁
**Current setup:**
**Board: Gigabyte B650M AORUS ELITE AX**
**Processor: AMD Ryzen 5 7600X**
**RAM: Corsair Vengeance DDR5 RAM (16GB x 2) \[CMK32GX5M2B6000C30\]**
**GPU: NVIDIA GeForce RTX 4070 SUPER (Zotac Gaming)**
**PSU: SEASONIC VERTEX GX 80+G 1000W**
**SSD: KLEVV CRAS C910 2TB**
**Peripherals: TP Link cheap usb wifi adapter and Corsair Void Wireless USB headset**
This setup has been modified/replaced since the original. I have replaced the RAM first, the board second, and the PSU most recently. I previously had some Silicon Power 16GBx2 RAM which I thought to be the culprit since RAM/PSU usually is for BSODs to my knowledge. I got the new corsair RAM at this point. I tested with memtest and gaming and kept crashing, so I replaced the board from the original MSI Pro B650M-P to the Gigabyte board above. Both brands of RAM passed 4 rounds of memtest and some OCCT testing (Corsair passed on both boards, SP on the original and not tested on Gigabyte). The corsair ram is on the supported list for both boards, and the silicon power on the list for the MSI board if memory serves. The drive is on the list for the Gigabyte board, but I don't even think this matters. I checked that the CPU and GPU are compatible on both boards. I did have memory management BSOD on the old RAM and board (Silicon Power x MSI), but I recall this being on XMP and that it stopped without overclocking (I may have exceeded the CPU limit?). I then replaced the ThermalTake ToughPower 80+G 850W PSU with the Seasonic listed above. Still crashing. I have tried every configuration of hardware, windows 10/11, and monitoring software (Eliminating potential conflicts) that I can think to do and observed crashing at all points. I also tried a new outlet.
I am using the correct drivers from the correct sources, I have yet to try an old BIOS version (I am using the most recent) on the Gigabyte board, although I did try many BIOS versions on the MSI board while varying the available hardware. I am running no overclocking at all, RAM (Corsair) at 4800 MHz with default timings. I have tried every BIOS option you can conceive of on the current build. It should be noted that the GPU made it through an hour of stability testing with OCCT's variable load settings (Not proof-positive of stability), and Furmark runs as long as I want it to. The monster hunter wilds benchmark runs \~200 fps all the way through. I wanted to run the OCCT CPU test, but it seems that it pushes your CPU temp up to the maximum by design (It got to the 95C danger zone a while in and I stopped it).
I occasionally have sfc /scannow repair files, but chkdsk doesn't ever seem to find anything. I don't know what this does or means. Windows has wiped itself a couple times when trying to use the safe mode troubleshooting. SEATools drive test says the SSD is ok. On a previous windows install, I tried to DDU clean install GPU drivers and when logging back in windows said "something happened" to my PIN and I had to go to safe mode to repair startup leading to an unwanted system wipe (Thanks windows). I had to turn off my wifi adapter to prevent any chance of automatic installs/updates for the clean install for context.
I do have some crashes that were not accompanied by a Kernel 41 (soft crash in Valheim), with event viewer faulting module name: d3d11.dll. I believe this points to GPU? I can get more detail on this if wanted, whatever event viewer gives me. I have clean installed NVIDIA drivers many times and messed with power management (Performance/Normal in Nvidia control panel) with crashes on any setting. I have not looked into power limiting.
My GPU is and has been powered from the native 12VHPWR cord from the PSU, without using the included adapter from the card (The splitter to combine two PCI-Es into the 12VHPWR GPU connector). I have not been sharing cords between the power supplies.
My suspicions are the following, listed from least to most costly/painful:
1) My WIFI adapter or headset is imploding and causing power instability. I hear frequent crackling in the headset while using it, and the WIFI adapter takes a long while to come back online after a Kernel 41 crash. This behaviour has been consistent across hardware.
2) Broken CPU
3) Broken GPU
My next steps:
Attach the included antenna to the motherboard to assist the onboard WIFI and use that instead of the adapter. Take out the 12VHPWR cord and use the NVIDIA adapter with two PCI-E connections. Unplug the Corsair headset and use bluetooth headphones if my board has that, otherwise I will game silently. Will configure this tomorrow and post results.
Any help is appreciated, and apologies for the infodump. I included everything I know.
All of my return windows have expired, and I am already \~$600 in the hole on RAM/Board/PSU. Whatever it takes, I am at wits end.
**Current setup:**
**Board: Gigabyte B650M AORUS ELITE AX**
**Processor: AMD Ryzen 5 7600X**
**RAM: Corsair Vengeance DDR5 RAM (16GB x 2) \[CMK32GX5M2B6000C30\]**
**GPU: NVIDIA GeForce RTX 4070 SUPER (Zotac Gaming)**
**PSU: SEASONIC VERTEX GX 80+G 1000W**
**SSD: KLEVV CRAS C910 2TB**
**Peripherals: TP Link cheap usb wifi adapter and Corsair Void Wireless USB headset**
This setup has been modified/replaced since the original. I have replaced the RAM first, the board second, and the PSU most recently. I previously had some Silicon Power 16GBx2 RAM which I thought to be the culprit since RAM/PSU usually is for BSODs to my knowledge. I got the new corsair RAM at this point. I tested with memtest and gaming and kept crashing, so I replaced the board from the original MSI Pro B650M-P to the Gigabyte board above. Both brands of RAM passed 4 rounds of memtest and some OCCT testing (Corsair passed on both boards, SP on the original and not tested on Gigabyte). The corsair ram is on the supported list for both boards, and the silicon power on the list for the MSI board if memory serves. The drive is on the list for the Gigabyte board, but I don't even think this matters. I checked that the CPU and GPU are compatible on both boards. I did have memory management BSOD on the old RAM and board (Silicon Power x MSI), but I recall this being on XMP and that it stopped without overclocking (I may have exceeded the CPU limit?). I then replaced the ThermalTake ToughPower 80+G 850W PSU with the Seasonic listed above. Still crashing. I have tried every configuration of hardware, windows 10/11, and monitoring software (Eliminating potential conflicts) that I can think to do and observed crashing at all points. I also tried a new outlet.
I am using the correct drivers from the correct sources, I have yet to try an old BIOS version (I am using the most recent) on the Gigabyte board, although I did try many BIOS versions on the MSI board while varying the available hardware. I am running no overclocking at all, RAM (Corsair) at 4800 MHz with default timings. I have tried every BIOS option you can conceive of on the current build. It should be noted that the GPU made it through an hour of stability testing with OCCT's variable load settings (Not proof-positive of stability), and Furmark runs as long as I want it to. The monster hunter wilds benchmark runs \~200 fps all the way through. I wanted to run the OCCT CPU test, but it seems that it pushes your CPU temp up to the maximum by design (It got to the 95C danger zone a while in and I stopped it).
I occasionally have sfc /scannow repair files, but chkdsk doesn't ever seem to find anything. I don't know what this does or means. Windows has wiped itself a couple times when trying to use the safe mode troubleshooting. SEATools drive test says the SSD is ok. On a previous windows install, I tried to DDU clean install GPU drivers and when logging back in windows said "something happened" to my PIN and I had to go to safe mode to repair startup leading to an unwanted system wipe (Thanks windows). I had to turn off my wifi adapter to prevent any chance of automatic installs/updates for the clean install for context.
I do have some crashes that were not accompanied by a Kernel 41 (soft crash in Valheim), with event viewer faulting module name: d3d11.dll. I believe this points to GPU? I can get more detail on this if wanted, whatever event viewer gives me. I have clean installed NVIDIA drivers many times and messed with power management (Performance/Normal in Nvidia control panel) with crashes on any setting. I have not looked into power limiting.
My GPU is and has been powered from the native 12VHPWR cord from the PSU, without using the included adapter from the card (The splitter to combine two PCI-Es into the 12VHPWR GPU connector). I have not been sharing cords between the power supplies.
My suspicions are the following, listed from least to most costly/painful:
1) My WIFI adapter or headset is imploding and causing power instability. I hear frequent crackling in the headset while using it, and the WIFI adapter takes a long while to come back online after a Kernel 41 crash. This behaviour has been consistent across hardware.
2) Broken CPU
3) Broken GPU
My next steps:
Attach the included antenna to the motherboard to assist the onboard WIFI and use that instead of the adapter. Take out the 12VHPWR cord and use the NVIDIA adapter with two PCI-E connections. Unplug the Corsair headset and use bluetooth headphones if my board has that, otherwise I will game silently. Will configure this tomorrow and post results.
Any help is appreciated, and apologies for the infodump. I included everything I know.
All of my return windows have expired, and I am already \~$600 in the hole on RAM/Board/PSU. Whatever it takes, I am at wits end.
Facebook
Twitter
Instagram