I recently bought a brand new desktop setup, the Ultra 9 285k, which is liquid cooled and comes with a 750W Gold PSU and integrated graphics. It runs flawlessly at the shop for hours, but when I bring it home, it crashes within minutes during heavy file I/O tasks (I'm using Django with Uvicorn, which loads over 50,000 files in PyCharm). Interestingly, gaming works perfectly fine for hours.
In Event Viewer, I noticed several error codes, such as Kernel-Processor-Power (Event 37), Kernel-PnP issues, and Kernel-Power 41 (indicating the system rebooted without a clean shutdown). I've tried replacing the RAM, tested it in single and double channel configurations, reseated components, and confirmed that temperatures are great—maxing out at 60°C. My SSD is also new and NVMe. At the shop, I don't use a UPS, and everything runs solidly, but at home, the crashes happen consistently with that Python reload process.
I live in a place with a lengthy 50-meter cable run to the outlet, although my sockets are grounded. Could this be an issue of low voltage, brownouts, or unstable mains supply? Would using a line-interactive UPS with AVR help resolve this? Also, am I correct in thinking that the Kernel-Processor-Power errors are a sign of power starvation? I'm really hoping to figure this out since I'm tired of troubleshooting—any advice from others who experienced similar home crash issues would be greatly appreciated!
3 Answers
For that Kernel-Power 41 event you mentioned, try getting the bugcheck code and its parameters. That could give you more insight into the crash.
It's definitely worth testing your outlets for safety first. Just because the cable is long doesn’t automatically mean it’s faulty, but better safe than sorry! Have you tried plugging the PC into a different outlet?
I had an electrician check the outlets already, but it didn’t fix the crashes. I might try moving to a different location with a new power source.
One important thing to consider is whether you’re putting the same strain on the PC at the shop as you do at home. If that’s not consistent, you might have isolated the problem to your home power supply. What model is your PSU? Also, try disabling Turbo and other high-power features; your PSU might not handle bursts of power well if your home voltages are lower.
I kept the workload consistent as much as possible, and my PSU is a Deepcool 750W Gold, which I believe is in TIER-B. Lowering the load seems to help, but that sudden spike in I/O is causing those BSODs.
The bugcheck code was 59 (SYSTEM_SERVICE_EXCEPTION), pointing to an access violation. Since the same workload runs without issues at the shop, I’m thinking this is a memory corruption issue due to unstable mains.