New custom loop handles high GPU and water temperatures effectively
New custom loop handles high GPU and water temperatures effectively
Hi everyone.
I’m just starting out with this custom watercooling setup. I’ve been trying to create a loop for my PC for quite some time, and finally, I succeeded. After a lot of effort, using a faulty monoblock, my CPU temperatures stabilized.
Now the main issue is that my water temperatures are increasing a bit too much (probably too high?), while my GPU is heating up more than what others report.
When idle, here’s how my temps look:
CPU: 35°C
GPU: 30°C
Water: 34°C
Ambient: Around 25°C
During gaming, things change (as expected):
CPU: Runs steady at about 63°C, spikes to 72°C on the hottest cores.
GPU: Has reached up to 61°C in both games and benchmarks.
Water temperature drops to around 44°C (the fans are running much louder than I’d like).
Ambient temperature remains roughly 25°C.
Am I seeing these numbers correctly? I’m a bit surprised because my custom loop is getting noisier than my CLC + air-cooled graphics card.
Specs:
I7-8700k @ 5.0GHz, 1.33vcore, LLC level 6 (out of 7), delidded
Asus Z370-f gaming motherboard
MSI 1080 TI gaming X, with a small +50 clock and +100 MB/s memory boost
Case: Lian-Li O11 Dynamic
Watercooling:
Two 360 EK radiators, medium thickness.
EK EVO supremacy CPU block
Phanteks GPU waterblock
EK combo res/pump
I’m using six Noctua fans at 85% speed – three for intake and three for exhaust.
Flow was okay when I first filled the loop. Now the pump runs at about 45% speed constantly. The RPMs are getting hot under load.
Photo of the loop, no water visible:
The loop
To summarize, I noticed a 6-8 degree drop in water temperature without the panels. What does this indicate? 😊
To be precise, I'm referencing one of my earlier messages.
Your readings seem consistent with your own observations. I'm maintaining the same CPU and GPU series. In gaming, my CPU stays near 30°C and fluctuates around 60°C, while the GPU remains stable. Both water and air temperatures are lower than expected. I have probes for both, and currently the air is about 26°C and water at 27°C. During extended gaming sessions, water temperature rarely exceeds 35°C. The GPU also idles around 30°C, dropping to low 40°C during heavy load. It might be worth checking this one. Your build looks good.
I am also operating both CPU and GPU within a single loop. The sequence is Res>Pump>Rad1>GPU>CPU>Rad2.
My water temperatures come from a monsoon stop plug equipped with a temperature probe, positioned at the inlet of my GPU block. These readings should be sufficiently accurate. I don’t have access to ambient temperature data, though I can check it using the thermometer in my car. It seems unlikely to be significantly off.
I’ve tested running the pump at both lower and maximum speeds, but it has no effect. There’s no flowmeter available, but as mentioned earlier, the fluid flowed normally during loop filling.
I’m not sure how to diagnose this issue.
Initially, what issues existed with the original monoblock? The flow rate was likely one of the first concerns to address. It seems there should have been sufficient cooling capacity for the system. Fans probably needed to run at a lower speed, which I observed they were set to around 1200 RPM. The GPU block appears properly secured. There were no visible air bubbles in the system, suggesting an air pocket might have been resolved. No restrictions were apparent in the picture.
The monoblock I used initially didn't establish good contact with my CPU. Removing it left the TIM not even spread. I've tried multiple times without success. Temperatures would sit around 60°C idle and 95°C under load. The mounting pressure works fine with a regular CPU block, though. I attempted to shake the system slightly, which released some air from the top radiator and suggested more air might help. However, I think there should be a bigger air gap in the radiator to raise temperatures like this. It didn't improve the temperature readings either. After that, I let some liquid out of the loop so I could observe air movement within the system. The water flowed at a reasonable rate.
When you lift the lid of the container, does the liquid start flowing and proceed in the intended direction?
nice looking loop.
the idle temps are fine. i'm very curious where the liquid temp sensor is, as i can't find it on the pic.
the liquid temp goes too high for the amount of rads and the ambient of 25C.
so the questions are:
1. how are the fans set up on the vertical rad (intake or exhaust)?
2. have you tried to raise the pump speed to 100%?
3. did you remove the case side panels during testing?
if the answers to q 2 and/or 3 are no, try it and share the results.
as a side note, i hope you're aware that the current CPU block orientation isn't ideal. outer cores will be hotter than inner cores because of liquid channeling inside the block in this setup.
I really overlooked the image link—appearance is good, but those sharp 90-degree angles might cause flow restrictions and slow things down. (I think a smoother, gradually arcing bend would help more.)
D5 models tend to struggle with releasing air from tight loops because they have lower head pressure compared to many DDC pumps, so you might still see some air pockets, probably inside the radiators.