Power loss under load
Power loss under load
I’m noticing intermittent power drops while gaming at high settings. The PC shuts down completely, turning off all lights and fans; I have to manually switch the PSU on and off before it restarts. There are no visual errors beforehand, and no BSOD appears. Windows event logs show Event ID 41 related to kernel power. The system is brand new with updated drivers and BIOS, default BIOS settings applied, and memory tests passed. Stress tests using OCCT, Aida64 (with graphics, CPU, and memory), and ECO CPU stress all trigger the issue immediately. Testing via a different power source also causes it to fail. I’ve checked the cable connections and used a 12VHPWR connector for the GPU. I’ve also tried connecting the power directly to a wall outlet, battery UPS, and surge UPS to rule out UPS problems. Since testing was straightforward, I’ve also run CPU stress tests with ECO settings and lower memory speed. The estimated power draw is between 700-800W depending on the measurement point.
Welcome to the forums, newcomer!
Troubleshooting involves narrowing down possibilities. Have you tried using a donor(borrowed) PSU with slightly higher wattage for the entire system? This might help eliminate your PSU as the cause.
If you need to repeatedly turn the power switch on and off, it likely indicates a tripped protection circuit breaker in the power supply. It's hard to determine if the issue is with a faulty protection circuit or a significant power surge that caused it. This video is interesting, but a 1000 watt supply should suffice.
The video is really helpful; the demo at 26 seconds matches exactly what I'm seeing. I plan to replace the test with a more powerful PSU on Wednesday, but if the issue is GPU spikes, then 1500W should suffice even for the 4080 drawing 2-2.5x in this setup. The video also notes that some motherboards struggle with it, so it might just be the PSU causing the problem.
Determine if the PSU operates in 12v single rail or 12v multi rail configuration by consulting iCUE settings.
iCUE is labeled as Multi-Rail OCP.
Edited: Changed to Single Rail OCP, performed the same Aida64 test previously which caused immediate power loss. It stopped around 35 seconds before shutting down. Turning on the power switch on the PSU and trying a reboot triggers the case lights and fan for half a second before turning off again. Opened the case, rechecked cables. Unplugged the GPU first and tried to reboot, which worked. Plugging the GPU back in and rebooting was fine. Power loss happened a few minutes after restarting while idling, same problem where it refuses to power on again.
Didn't interfere once it began struggling to power on yesterday. Found the HX1500i this afternoon. Replaced the RAM, GPU, and all cables, checked the board while the case was open. Swapped the HX1000i for the HX1500i.
Tested the system through several non-patreon OCCT checks for a few minutes; it passed where it previously failed during the Power test. Ran Aidia64 stability test with all settings enabled and no issues arose. Luckily, the power loss appears to be resolved.
This is the sole issue I've encountered so far—any suggestions on what to inspect further that might have been impacted by the power fluctuations?
The HX1000i seems to be either running low on power during this build or possibly defective. It's unclear whether the connector placement was the problem or if another factor contributed, and I'm uncertain about next steps.
A note for others using the Lian Li O11D case and a comparable PSU: the layout left minimal space for cable management.
FYI
The BIOS version 1.11.AS03 is a beta release. Depending on your situation, I’d contact Asrock support to see if the issue might be related to the BIOS version. If that doesn’t resolve it, I think the problem could be with the HX1000i itself. However, considering the beta status adds another challenge to solving this.