F5F Stay Refreshed Hardware Desktop Water cooling keeps the CPU at 100°C, but coolant remains at 30°C.

Water cooling keeps the CPU at 100°C, but coolant remains at 30°C.

Water cooling keeps the CPU at 100°C, but coolant remains at 30°C.

J
jonhub
Junior Member
32
01-06-2022, 03:54 PM
#1
Over the last weeks my system has experienced some temperature problems I initially thought were due to dried thermal paste or a poorly secured waterblock, even though the loop hasn’t been touched for about a year. When idle the CPU stayed around 80-90°C, but under load it reached nearly 100°C. The coolant would sit at about 35°C when idle (fans off or barely spinning) and rise to 45°C during load (fans at full speed). Today I opted to remove it and repaste, using the PTM7950 I’d gotten earlier. Right after the work, temperatures fell by roughly 20°C—great! But after a few hours things changed. Idle temps jumped to almost constant 100°C, while the coolant stayed surprisingly cool at around 25°C. I ran some checks: drained the loop to see if the pump failed, tightened the waterblock more, but nothing helped. The GPU also got unusually hot, hitting a 70°C (110°C hotspot) under load—a situation I’d never seen before. When I ran a benchmark to test if the GPU could warm the coolant, it didn’t. I noticed the coolant flowing from the CPU to the first radiator (and even the radiator itself) was extremely hot, possibly just under 40°C. But it cooled down significantly when going to the second radiator and back into the reservoir, even though the temperature should be uniform. What’s happening? My specs are: 12700k (thermalright ILM + PTM7950), iChill Frostbite 4090 waterblock, Corsair XC7, XD5 pump/reservoir combo, Mayhems Eco X1 UV Blue soft tubing. Any clues?
J
jonhub
01-06-2022, 03:54 PM #1

Over the last weeks my system has experienced some temperature problems I initially thought were due to dried thermal paste or a poorly secured waterblock, even though the loop hasn’t been touched for about a year. When idle the CPU stayed around 80-90°C, but under load it reached nearly 100°C. The coolant would sit at about 35°C when idle (fans off or barely spinning) and rise to 45°C during load (fans at full speed). Today I opted to remove it and repaste, using the PTM7950 I’d gotten earlier. Right after the work, temperatures fell by roughly 20°C—great! But after a few hours things changed. Idle temps jumped to almost constant 100°C, while the coolant stayed surprisingly cool at around 25°C. I ran some checks: drained the loop to see if the pump failed, tightened the waterblock more, but nothing helped. The GPU also got unusually hot, hitting a 70°C (110°C hotspot) under load—a situation I’d never seen before. When I ran a benchmark to test if the GPU could warm the coolant, it didn’t. I noticed the coolant flowing from the CPU to the first radiator (and even the radiator itself) was extremely hot, possibly just under 40°C. But it cooled down significantly when going to the second radiator and back into the reservoir, even though the temperature should be uniform. What’s happening? My specs are: 12700k (thermalright ILM + PTM7950), iChill Frostbite 4090 waterblock, Corsair XC7, XD5 pump/reservoir combo, Mayhems Eco X1 UV Blue soft tubing. Any clues?

S
spoodkra
Member
55
01-06-2022, 08:37 PM
#2
It makes sense to me that the temperature decreases after a radiator.
S
spoodkra
01-06-2022, 08:37 PM #2

It makes sense to me that the temperature decreases after a radiator.

J
jedimonkey1001
Junior Member
15
01-06-2022, 08:46 PM
#3
The coolant temperatures remain consistent throughout the loop, yet the water after the radiator is noticeably cooler than before. How are you tracking these readings? The phrase “possibly” suggests I’m making an educated guess. If you’re uncertain, this detail won’t be helpful. Your GPU is likely the main source of heat, drawing more power and generating significantly more warmth. As the biggest heat producer in your PC, it will raise the water temperature the most. If you could measure the water temperatures accurately before and after each stage, you’d see this pattern clearly. Do you know the pump’s speed setting? How long has the loop been running, and have you cleaned it? From what I hear, either the pump isn’t spinning fast enough or the loop is blocked. After draining, did you clean and flush it? Usually I perform a few alternating hot water and vinegar washes with the hoses and pipes, followed by fresh water rinses. Then I inspect the components to ensure no debris has accumulated on the fins.
J
jedimonkey1001
01-06-2022, 08:46 PM #3

The coolant temperatures remain consistent throughout the loop, yet the water after the radiator is noticeably cooler than before. How are you tracking these readings? The phrase “possibly” suggests I’m making an educated guess. If you’re uncertain, this detail won’t be helpful. Your GPU is likely the main source of heat, drawing more power and generating significantly more warmth. As the biggest heat producer in your PC, it will raise the water temperature the most. If you could measure the water temperatures accurately before and after each stage, you’d see this pattern clearly. Do you know the pump’s speed setting? How long has the loop been running, and have you cleaned it? From what I hear, either the pump isn’t spinning fast enough or the loop is blocked. After draining, did you clean and flush it? Usually I perform a few alternating hot water and vinegar washes with the hoses and pipes, followed by fresh water rinses. Then I inspect the components to ensure no debris has accumulated on the fins.

B
beichner
Senior Member
447
01-18-2022, 05:14 PM
#4
Certainly, but previously there wasn’t a significant variation between incoming and outgoing water in the rad. That’s not something I ever recorded; it’s just an educated guess because I lack a reliable method to verify. I do remember how the top rad typically warms up at certain temperatures. Usually, I set the pump to 80% of its maximum speed, which means around 4000rpm idle, with a curve that increases as the CPU heats up until it reaches 4900rpm. iCue confirms that speed right now, but I don’t have a way to measure flow accurately, so I’m unsure if the readings are correct. I didn’t attempt to clean the loop yet; I just drained it quickly to test if the pump was faulty and it appeared fine. The loop is approximately 1.5 years old, and I haven’t performed any cleaning before. I plan to reapply paste and re-seat the CPU, then clean the loop next, and possibly switch the pump if that fails too, which could affect the CPU block.
B
beichner
01-18-2022, 05:14 PM #4

Certainly, but previously there wasn’t a significant variation between incoming and outgoing water in the rad. That’s not something I ever recorded; it’s just an educated guess because I lack a reliable method to verify. I do remember how the top rad typically warms up at certain temperatures. Usually, I set the pump to 80% of its maximum speed, which means around 4000rpm idle, with a curve that increases as the CPU heats up until it reaches 4900rpm. iCue confirms that speed right now, but I don’t have a way to measure flow accurately, so I’m unsure if the readings are correct. I didn’t attempt to clean the loop yet; I just drained it quickly to test if the pump was faulty and it appeared fine. The loop is approximately 1.5 years old, and I haven’t performed any cleaning before. I plan to reapply paste and re-seat the CPU, then clean the loop next, and possibly switch the pump if that fails too, which could affect the CPU block.

G
grisu112
Member
170
01-18-2022, 07:02 PM
#5
When the pump is operating, it’s likely okay. The blocks don’t require replacement since there’s nothing that should deteriorate. It might only need to be taken apart and cleaned.
G
grisu112
01-18-2022, 07:02 PM #5

When the pump is operating, it’s likely okay. The blocks don’t require replacement since there’s nothing that should deteriorate. It might only need to be taken apart and cleaned.

B
Bartekdwarf
Posting Freak
791
01-19-2022, 03:17 AM
#6
I believe the issue is entirely with the pump, which is why I ordered a replacement soon after. The signs were clear: when I pressed the tubings and pushed water through, mimicking CPR while following the beat of "Stayin Alive," the temperature plummeted by as much as 40°C, dropping back down to around 60°C for the CPU. There were also large air bubbles that I couldn’t eliminate, likely due to pump malfunction. Fingers crossed everything works out fine.
B
Bartekdwarf
01-19-2022, 03:17 AM #6

I believe the issue is entirely with the pump, which is why I ordered a replacement soon after. The signs were clear: when I pressed the tubings and pushed water through, mimicking CPR while following the beat of "Stayin Alive," the temperature plummeted by as much as 40°C, dropping back down to around 60°C for the CPU. There were also large air bubbles that I couldn’t eliminate, likely due to pump malfunction. Fingers crossed everything works out fine.