System powers down when GPU usage spikes, triggering POST code 97
System powers down when GPU usage spikes, triggering POST code 97
Hello reader, I'm facing some persistent GPU problems after years of troubles. Building my PC has led to repeated boot failures and the POST code 97 appearing twice in the past year. This caused two replacements and several RMA requests. Everything functions normally for a while until the annoying POST 97 shows up. Recently, after about six months of stable performance, I experienced screen flickering and shutdowns during gaming. Testing with Heaven Benchmark confirmed this issue. I've tried several solutions: using a spare 1060 in both PCIe slots, using extension cables, testing on another PC, and even removing the GPU to see if it works. The problem persists across all attempts. Here are some details for reference:
- No overclocking has been applied
- XMP II profile is active
- Running Windows 10 64-bit
- BIOS updated to version 1402
- Latest drivers installed
- GPU updated with current drivers
- PC operates without GPU installed
Additional notes:
1. All components are at their default settings.
2. The system hasn't been modified beyond standard updates.
Thank you for your reply. I've already attempted to run without the extension cables, but the issue continues.
I've begun the return process. By the way, while checking the PCIe power cables for the GPU, I found one pin missing. I assumed that was the cause, reinserted it and retested, but the issue persisted! I’ve also switched to using separate PCIe cables from the PSU straight to the 2x8 pin GPU, thinking this setup might handle higher power cards better (previously I used a single long PCIe). Right now, I can turn on the PC and see display output, but when the PC is under load the display disappears and the PC stays on until it restarts. Occasionally I run the Heaven benchmark without problems, but games like Warzone crash as soon as I load them. I hope the RMA resolves this, and I appreciate any other ideas you might have.
The Corsair RMA process worked well and they provided a replacement PSU without any questions. After installing it, the issue continued. It wasn’t the POST 97 error anymore; instead, the display would turn off and I had to manually restart the PC. Once back in Windows, it would load again each time (unlike before). I’m still thinking it might be related to power consumption, especially under GPU load. Since I replaced the GPU already, I undervolted my card using MSI Afterburner and ran Heaven Benchmark. Before undervolting: GPU temp 53°C, draw 247W; after: 54°C, draw 2215W. Testing with reduced voltage showed a drop in GPU temperature and power usage. In-game performance dropped noticeably—about 30-40% lower FPS. I noticed the GPU was set to BIOS OC mode instead of Silent, which might be contributing. Any advice would be appreciated.
The RTX 3070Ti can experience sudden power surges, possibly leading to the PSU tripping and causing the system to power off. With a peak GPU draw around 247W, a 2.5x spike brings you to roughly 617W. Adding the CPU, motherboard, RAM, and other components could push the total to about 200W. Such spikes often trigger protection circuits on the PSU. Your issues seem consistent with reports from 'Tech Gezus'.
Adjust the settings to silent mode and see if the issue persists. It seems like the problem might be related to the motherboard, as the GPU worked fine on another system. Have you enabled the resizable bar? You can verify this using GPU-Z. Try moving the GPU to a different PCIe 3.0 x16 slot and test again.
Thanks for the feedback and the video. I'm sure this is the issue. It's hard to decide next steps except to upgrade the PSU and try a different one. The problem is that my case only supports SFF PSUs, and high-quality 1000W units are scarce and costly. I set it to silent mode today, but the issue reappeared right away. The resizable bar is active, yet the problem continues in the other PCIe slot—though it would be impractical to move it there since it's just a few millimeters from the bottom case fans.
Updated: Changed PSU to Silverstone SX1000W. No major impact except longer display time by 1.5 minutes before output failed. GPU now in second PCIe slot – adds another minute before failure. Reached out to Asus for RMA support and considered swapping the GPU while I'm here. At this point, it seems like a full PC replacement is necessary. *sigh* Check out the video: https://drive.google.com/file/d/1mCj244z...sp=sharing
Hope someone can help interpret the footage and figure out the problem.