Device restarts unexpectedly while playing, but the cause isn't clear
Device restarts unexpectedly while playing, but the cause isn't clear
This problem has been occurring over the past eight months, and it’s really frustrating. I’m no longer able to enjoy competitive games with my friends. The frequency has increased since upgrading my RAM and SSD eight months ago (previously it happened about 3 to 4 times in two years). The main issue is that when I play a game, everything freezes and my PC restarts—no blue screen appears. It mainly affects Overwatch 2, where I can’t even launch a game; it crashes within 5 to 10 minutes. In Apex Legends and CS:GO, it crashes between 30 minutes to an hour. For non-intensive games (like CS:Source or L4D2) or just browsing the web, I haven’t seen this problem. Temperatures are normal: CPU reaches 75°C and GPU 67°C during demanding games. I’ve tried several fixes: restoring BIOS settings to defaults, disabling DOCP updates, reinstalling Windows 11 multiple times, reinstalling on a different SSD, updating graphics drivers, uninstalling RGB software (Bitwit confirmed this). My PC specs are: Ryzen 5 5600X, Asus B550-I Rog Strix 2x16GB/32GB, Corsair Vengeance LPX RAM, EVGA XC3 Ultra 3070, Corsair SF600 Platinum, Samsung 980 PRO 1TB, Noctua NH-L12S case. My observations so far suggest the problem might be due to RAM, SSD, or GPU. For RAM, disabling DOCP didn’t help. With the SSD, I reinstalled Windows 11 on an older drive and it worked initially, but the issue returned after a week. The new SSD (980 PRO) has good cooling thanks to its direct airflow cooler, so I doubt it’s overheating. For the GPU, I borrowed it from a friend a year ago; he broke the HDMI cable, rendering the port unusable. The port looks fine externally but doesn’t work. I’ve checked for loose components, but nothing seems to be the cause. Any suggestions or help would be appreciated. Thanks for your time.
I’d prefer a 650W non SFF power supply for a 3070, especially with all the sudden spikes. Choosing SFF means you end up needing more capacity than necessary. Even 600W is just the bare minimum for this model, and Nvidia recommends a 650W unit. Since this is during gaming, it only makes sense the PSU would be overwhelmed by the GPU’s transient spikes.
Refer to the previous advice by reducing the card's voltage to avoid larger power surges. Consider removing GPU drivers using DDU and reinstalling them. Use a single RAM stick in each slot.
I suspect the issue isn't related to the PSU, as others ran 3080 with a more demanding CPU. I've used this setup for over two years without any problems before. My card also runs at an undervolt, but that didn't work. I found a video from Optimum Tech that encouraged me to consider the SF600. Since I haven't tried it yet, I plan to check it out this weekend.
From what you've addressed, it seems the problem likely lies with the hardware. The clearest sign points to the power supply unit. To be certain, you should replace each component one by one and see where it fails, starting with the PSU.
It might be coincidence, yet the 30 series cards are famous for their strong power surges.
Today I tested Furmark on a QHD preset, pushed the GPU to its limit in afterburner (112%), and my 3070 drew just over 270 watts. I also ran the CPU burner with 11 threads (PBO enabled), adding another 100 watts. The system stayed up without shutting down, and I ran the tests for an hour. Would a low wattage PSU still be a concern in this situation?
The main concern isn't overall power consumption, but the brief surges in draw that can disable the PSU's OCP protection. This causes immediate shutdowns, typically during demanding tasks like gaming or when CPU and GPU work together. Over time, components wear down, so unexpected failures can occur even after years of reliable use.