The laptop powers off abruptly without any notice or blue screen.
The laptop powers off abruptly without any notice or blue screen.
I verified the situation. The only way to eliminate the battery is by taking off the backpanel. There isn't a direct method available. This gave me a suggestion. I attempted to run the game on battery power and proceed with the same tasks. No crashes were observed during three benchmark tests when power was disconnected. However, upon reconnecting the power, the game crashed immediately after the benchmark, just as described in the first post. I haven't conducted sufficient tests yet to confirm, but it suggests issues with power management and charging. I'm planning to repeat the benchmark on battery a few more times. If it remains stable, we're making progress. As mentioned before, the PSU might be unstable. Although OCCT didn't trigger any errors, another factor could be involved. Perhaps the game rapidly switches power usage during session exit or SSD read while the CPU is throttling. This could lead to undervolting. At least I'm narrowing down the possibilities, which gives me some hope in identifying the cause.
I repeated the benchmark multiple times without power and observed stable performance. Once power was restored, the system crashed. This suggests a problem with power delivery management when power is connected versus battery. I ran the OCCT test again and saw no errors in the PSU check. It appears the crash occurs only with power, not battery. The OCCT results show no faults, and the PSU stays cool even during extended tests. It seems a software setting or chipset feature might be altering the system's response to power supply changes. A possible cause could be a power configuration issue, such as CPU boost or other settings being triggered. Without a separate PSU, I’m unsure of another reliable test method beyond OCCT.
Updated since previous update. Ran the system on battery only a few times, no crashes observed. When power is connected, it crashes frequently, usually during loading. No crashes during gameplay. Running smoothly for hours with full GPU usage and 50% CPU, no interruptions. It appears the power supply isn't the issue, but a configuration setting might be. Possibly related to the SSD—BIOS lists it as Bus lane 6, which seems incorrect. I’ll reinstall the SSD and reset the BIOS next. Fingers crossed that helps. If not, I’ll be stuck again.
So I finally had the chance to work on it. I took out the backpanel and the battery. The laptop booted up a bit slower after the change. Still, crashes keep occurring. They seem more unpredictable now—like it crashed mid-boot while loading the game. It looks like games might fail when connected to power, though only in a few cases. I’m thinking about reinstalling Windows from the beginning and just installing this one game to test.
I reinstalled the windows entirely, drivers and the game again. It still crashes—either during main menu loading or after a session. This suggests a driver problem remains even after the fresh install. The issue seems tied to switching between games and power connections. I’m unsure why it only happens in certain games or during loading. No crashes appear in other games or even when I keep the session running. It’s likely a combination of game menu transitions and power supply issues. I’ll take the laptop to a repair shop for a CPU liquid metal replacement and test with a new power supply.
Same problem with another power source as well. After resetting the BIOS and swapping the liquid metal on the CPU, it occurred in a fresh game I installed. It only happened when loading a level, not during extended play sessions.
There appears to be an issue with the GPU power delivery. Install gpu-z and execute it in logging mode. Perform a powered test to trigger a crash. Share the log file here. The crash during the loading screen seems to occur immediately when switching GPU modes.
I used the GPU-Z utility and included the logs. It appears GPU-Z isn't detecting fan speeds even though they're active. Manual checks confirm fans are running normally and Armoury Crate displays correct readings. I logged data at 0.1 seconds to capture the crash moment. No unusual patterns noticed, though I might have overlooked something. Edit: The issue doesn't appear during session loading—only after the game starts. Same behavior in Football Manager, especially when switching to low power mode instead of high. I added logs for Zelda, showing no crashes even after extended use. Edit: Included crash log from main menu during initial game launch, right after the internet check. Navi 22.rom GPU-Z output files are available for reference.
I discovered another discussion about the vbios problem leading to similar crashes in other laptops of the same model. I attempted a workaround by updating the vbios to the latest version mentioned here. Also, I performed a fresh installation of the GPU drivers. Still, the issue persists—after restarting everything and changing drivers, it crashes again. I’m considering reverting to the original vbios that came with the system, as it functioned properly before. I plan to test that version next and observe the outcome. It appears the problem isn’t hardware-related but stems from software or firmware conflicts affecting how the GPU receives power. So far, I don’t have a clear solution, though I’m thinking about possible next steps.
I used liquid metal recently to test it, and the temperatures stayed consistent across all games or GPU-heavy tasks. Despite this, no crashes occurred in other titles. In the game itself, I didn’t experience any issues when keeping the session open for hours. I completed a 10-lap benchmark with full RT enabled, where both GPU and CPU were under heavy load to achieve around 10fps, yet no crashes happened. When reviewing logs for two laps, the same pattern appeared. Crashes typically appear after closing the GPU-intensive task or when switching to menus. If temperatures were the cause, it should happen in every session. However, I’ve seen similar crashes in FM, which runs like a simple spreadsheet without needing a dedicated GPU. Additionally, there were no crashes on the integrated GPU or when using battery power, and none in OCCT runs. Overall, changing liquid metal didn’t resolve the problem.