Is your fan malfunctioning on a second-hand GPU?
Is your fan malfunctioning on a second-hand GPU?
Hello and thank you for your support.
I recently purchased a second-hand Palit RTX 3090 GamingPro GPU, and I've observed that its fan behaves independently. It remains relatively quiet when idle, but during gameplay it increases and decreases rapidly every few seconds. For instance, while playing Overwatch, it would spike to 100% for short bursts at irregular intervals, then drop back down to nearly silent. In Yakuza 3 Remastered, it consistently ramps up during cutscenes and fights, only occasionally slowing during normal play.
The sound is bothersome but not overwhelming; however, the sudden changes in fan speed make it hard to ignore.
I've applied MSI afterburner and adjusted the fan curve settings, but the improvement has been limited (0% to 48°C, then a gradual climb to 70% at 80 degrees).
What adjustments or solutions might help resolve this problem?
I recently purchased a second-hand Palit RTX 3090 GamingPro GPU. You might consider taking it apart and swapping the thermal paste and pads for better quality. This could also be a reason to thoroughly clean the card. Did you run DDU in Safe Mode to remove all GPU drivers (Intel, AMD, Nvidia), then manually install the newest driver from Nvidia's support site using an elevated command, such as right-clicking the installer and selecting Run as Administrator?
I haven't tested the driver yet, but since it's an upgrade from a GTX 1080, I assumed it wasn't needed. I'll follow the suggestions.
I'm able to take apart the card, though I'm not sure if that's the real problem. I don't believe it's overheating; my concern is about the fan curve. I wondered if adjusting it to a quieter setting might help. My main question was whether such an adjustment is possible.
DDUed and repasted GPU as mentioned earlier, we had some MX-4 available, so I used it. Now my issue is mostly resolved—I no longer experience spikes in fan speed. However, when I open HWINFO, I notice there’s no variability in the fan speeds, and the GPU temperature remains steady at 38°C. After about an hour of gaming, the hotspot reached 100°C, along with the memory, but the main GPU temperature stayed consistent. This happened before I repasted, so it isn’t a new behavior. Since the GPU temperature didn’t change much, the fans didn’t start spinning and stayed quiet at around 1300rpm. The backplate was extremely hot; I couldn’t touch it until it cooled down. Is the temperature sensor malfunctioning? Are there any ways to try and restore it?
I also performed a furmark 2 stress test; memory and hotspot temperatures gradually increased (the hotspot didn't reach 100°C immediately, so I adjusted the settings properly), while the fans failed to activate. The program also displayed a GPU temperature of 38°C.
Edit:
Following a reset, GPU temperatures remain elevated at 42°C, with hotspot and memory functioning correctly now.