Temperatures in coolant are unstable because of GPU activity.
Temperatures in coolant are unstable because of GPU activity.
Recently upgraded to a 3080Ti which required my radiator for the AIO cooler to be positioned at the top of my case (previously on the side). When running games that load the GPU, the radiator becomes warmed by the hot GPU air. This leads to coolant temperature reaching over 60°C after 30–45 minutes, even with minimal CPU usage (around 10%). Two 120mm fans draw in fresh air from the front, and two 120mm radiator fans on top serve as exhaust. I understand the obvious fixes are a new case or cooler, but are there other ways to ease the problem without immediately replacing parts? Thank you.
60+C meets the requirements well. If your build allows 2x140mm or 3x120mm front placement, I’d consider an upgrade. A top-mounted AIO offers a smooth airflow path from the upper-front and middle-front fans, directing air to the GPU from the bottom-front. Adding a rear fan can help dissipate heat more effectively, preventing reliance solely on the radiator.
You're moving too fast. An under 85°C is acceptable—I'd adjust your fan settings so the exhaust flows away from the radiator, just as @Agall recommended.
The 60°C water is extremely warm; it’s unlikely your GPU is responsible. Are the temperatures you mentioned accurate?
The coolant temperature is unusually high, not related to the CPU. The GPU generates 400W of heat and sends it straight to the radiator. This explanation fits because GPU-heavy tasks clearly raise coolant temps significantly, unlike CPU-heavy ones. I don’t think the coolant sensor is faulty. When the CPU stays around 60°C with no workload, it’s likely the coolant is warming the CPU. The AIO LED starts blinking red and the system enters a fail-safe mode, which probably checks both CPU and coolant temperatures. The radiator feels excessively hot to the touch.
The system is running smoothly with no overheating issues reported.
I found another 120mm case fan and will check if it fits in front. Probably won't work, though. Just noting the coolant temperature—specifically not CPU. The manufacturer says 60°C is the max allowed. @ItTakes2ToMango
Yes, I can feel the pump working. I've experienced a failing corsair pump before, and once your CPU gets hot during idle after about 30 minutes, that’s usually a sign. It’s not the pump itself, but past issues with corsair AIOs have made me cautious. Installing a new 240 might help a bit. The problem likely isn’t the AIO, since it ran fine when placed at the front of the case.
I plan to keep the GPU benchmark active once the coolant temperature stabilizes. After that, I’ll share the HWinfo logs; I verified the coolant readings match what iCUE reported.