Initial Custom Water Loop (High CPU Usage)
Initial Custom Water Loop (High CPU Usage)
liljamerz :
Biglizard :
I took off the protective film from the CPU block and attached it with Thermal Grizzly Hydronaut. I placed the suggested water jet inside, with the inlet positioned in the middle; the insert I used was the one already installed. This build is three days old. I live in an apartment, and the air conditioning unit is about two meters to my left, so everything remains quite cool. I didn’t remove the CPU at all, and the setup worked perfectly before. I’m not sure how a different CPU block would impact the cooling, but it seems unlikely.
I’ve noticed occasional air bubbles, though all the microbubbles have disappeared. Still, I keep moving the computer around to try removing as many bubbles as possible. Could a stuck air bubble cause such a large effect if there’s a temperature difference between the radiator and the CPU block?
Hope Rubix_1011 sees this.
Yes, I’ve experienced air bubbles affecting performance.
How did the TIM spread pattern look? Were there any caps or other issues with the mount?
Could you share a picture of your setup?
My TIM spread was just a single dot in the middle method, which I’ve used for years without problems. I’ve reinstalled the block several times in hopes of finding something, even added a bit more paste, but still no drops. Everything fits perfectly and everything worked smoothly. There’s a link in my original post with an IMGUR album of the build—I’ll include it again here.
Build Pictures
Beautiful build.
Geez, I just don’t know.
Maybe the pump isn’t powerful enough, but I’d expect your setup to keep up with your old AIO.
Biglizard shared his experience, removing the protective film on the CPU block and installing Thermal Grizzly Hydronaut. He placed a water jet inside with the inlet centered, using the recommended one that came with the unit. The build is three days old. Living in an apartment with air conditioning about two meters to my left kept things cool. I didn’t take out the CPU, and everything worked perfectly before. I’m curious how a different CPU block would impact cooling, but it seems unlikely.
I’ve noticed occasional air bubbles, though all microbubbles are gone now. Still, I keep moving the computer to clear as many as possible. Could a stuck bubble cause such a big difference if there’s a temperature variation in the radiator or CPU area?
Hope rubix_1011 sees this.
Yes, I’ve experienced air bubbles affecting performance before.
The TIM spread pattern was just a single dot in the middle, which I’ve used for years without issues. I reinstalled the block several times, added more paste, but still no drops. Everything fit well and everything worked smoothly. There’s a link to an image of my build on Imgur that I’ll share again.
Build Pictures
Beautiful setup.
I’m still unsure. Maybe the pump isn’t powerful enough, but I’d expect your system to keep up with an old AIO.
I’ve been shifting the case around and still dealing with air bubbles. Last time I did it, a few bubbles floated up into the reservoir. I’m hoping over time the air clears and temperatures normalize. I’m also hoping someone can review the loop order to ensure everything is correct.
The sequence of the loop is generally unimportant, and your pump works well for your configuration. Because your CPU is delidded and your AIO performed better, I concentrate on the CPU block. I recommend confirming you're using the appropriate mounting hardware and that the correct jet plate and insert are installed.
Biglizard summarized the situation, explaining the changes made and observations about performance.
I altered the protective film on the CPU module, secured it with Thermal Grizzly Hydronaut. I placed the suggested water jet inside, positioning the inlet in the center; the recommended unit was already installed. This assembly is three days old. Living in an apartment, the air conditioning is about two meters to my left, so temperatures remain quite comfortable. I haven’t removed the CPU entirely, and everything functioned perfectly before. I’m uncertain how a different CPU block would influence cooling, but it seems unlikely.
I’ve observed occasional air bubbles, though all microbubbles have disappeared. Still, I’ve moved the computer around to try removing as many bubbles as possible. Could a trapped air bubble cause such a large effect if there’s a temperature difference between the radiator and CPU block?
Hope this helps rubix_1011.
Yes, I’ve experienced air bubbles impacting performance.
Regarding the TIM spread pattern—are there any caps or other components interfering with the mounting?
Could you share a picture of your setup?
Indeed, I did see it.
I was also curious about whether full seating of the CPU block was achieved. The temperatures listed don’t seem logical, especially if your overclock hasn’t changed since switching cooling solutions. Have you checked the CPU block for proper thermal paste application and complete contact?
You confirmed you’re seeing flow and turbulence in the reservoir. Would opening the cap allow air to escape? Keeping it capped would block air from exiting the loop. You might also try removing the jet plate to see if that improves flow. Sometimes these jets actually restrict rather than enhance cooling.
I’ve re-seated the block three times so far, and it appears to spread evenly across the IHS. I can feel the pump running. Regarding the jet plate, since I changed it to J2 for the 1151 socket, should I remove it entirely and run the insert? Would that increase flow? Not sure if the jets are meant to focus water onto a specific area or serve another function.
I’ll need to extend the hose leading to my pump and possibly install a fitting for easier drainage. When I mount the pump, I might have to detach the CPU block and remove the jet. Will this cause any problems?
Initially, you need to monitor your pump's RPM when connected to a fan/opt header on your motherboard. It should increase as the CPU uses more power, provided the header is set up correctly in your BIOS. Do you notice any changes during load?
Regarding loop order, there’s a standard approach and yours seems reversed. It’s better to supply components directly from the pump, then let the hot liquid flow through the radiator for cooling before returning it to the reservoir. This method is simple and aligns with EK’s advice. You can adjust this by swapping the tubes at the pump/res and CPU block to maintain correct in/out connections.
Finally, are your radiator fans increasing their speed under load as expected? Are the fan headers correctly set up?
The main issue with comparing current temperatures to your AIO before is that you’ve already changed a lot. Your CPU has transformed significantly. The only accurate way to compare now would be to run the AIO on the CPU and observe its temperatures, allowing a fair apples-to-apples comparison.
marko55 :
First, you should be able to watch your pump RPM assuming its connected to a fan/opt header on your motherboard. It should be spinning up higher when your CPU comes under load, assuming you've got that header configured properly in your BIOS. Do you see it varying during load?
Second, in regards to loop order, there is a best practice and yours is actually backwards. Its recommended to feed your components directly out of the pump and then have the hot liquid go through the radiator to cool it before it goes back to the res. Yes, your loop is small and not super complicated so it may not make a huge difference but it could, and that's straight from EK's recommendations. To achieve this you could just switch the tubes at the pump/res and at the CPU block to ensure the proper in/out connections are being used.
Lastly, are your radiator fans spinning up higher when under load like they should be? Are those fan headers configured properly to do so?
The problem with comparing your temps now to your AIO before is you really can't for one big reason, you've delidded. Your CPU is completely different now than it was before. The only real way to do that comparison right now would be to put the AIO on the CPU right now and see what kind of temps you see. Then you'd actually be comparing the two different cooling solutions apples to apples.
I have the pump/fans configured to speed up with temps, and I have been able to check that in BIOS. I should have been a little bit more clearer, my 7700k was delided on the AIO previously, this is not a fresh delid just a new water cooling setup. I have a bracket for the pump and a T-fitting for draining coming tomorrow/early next week, so maybe I will try that swap like you recommended.
If I swapped the order, then it would be Pump -> GPU Block -> CPU Block -> Rad -> Pump -> repeat, is this correct?
The way I set it up now was because I had it in my head that I wanted the cool fluid to exit the radiator and hit my components first, but if the other way around is what EK recommends, I might have to try that out.
You should be able to monitor your pump's RPM when it's connected to a fan/opt header on your motherboard. It should increase as your CPU uses more power, provided the header is set up correctly in your BIOS. Do you notice any changes while under load?
Concerning loop order, there’s a standard approach and yours seems reversed. It’s better to feed your parts directly from the pump, then let the hot fluid pass through the radiator to cool before returning to the reservoir. Your current setup is simple and won’t drastically affect performance, but it might help. You could adjust the connections at the pump and CPU block to ensure proper in/out flow.
Also, are your radiator fans increasing their speed when under load as expected? Are the fan headers correctly set up?
The issue with comparing current temperatures to your AIO before is significant. Your CPU has changed considerably since then. The only accurate way now would be to install the AIO directly on the CPU and observe the temperatures. This would allow a true apples-to-apples comparison.
I’ve configured the pump/fans to increase speed based on temperature, and I can check this in BIOS. I should have been clearer—my 7700k was previously delidded, not just swapped. This isn’t a new problem; it’s a fresh water cooling setup. I have a bracket for the pump and plan to install a drain fitting soon, so perhaps trying that swap would be worthwhile.
If I reversed the order, it would look like Pump → GPU Block → CPU Block → Rad → Pump again. Is that correct?
The setup I have was based on the idea of letting coolant exit the radiator first to reach components. If EK suggests otherwise, I might need to experiment.
Ah, got it. I didn’t realize it was still delidded under the AIO. The recommended order is common for most cases with the pump at the bottom. I usually start cooling the liquid to the CPU first, so it doesn’t warm up by the GPU. With a single GPU, that’s less of an issue.
I also noticed your CPU block is at a 90-degree angle, which is typical in standard layouts. Studies show this can slightly affect cooling. I’ve flipped the same block before, and the temperature change would be minimal (<5°C), but a 90-degree mount could noticeably impact it. Remounting it in the correct orientation—so 'in' and 'out' are clear from left to right—would still keep things clean and EK’s logo visible at the bottom-right. This should work well with your new loop configuration.
I value all the responses received so far.
When I set up the T fitting for the drain and mount the pump, I'll turn the CPU block 90 degrees and run a few lines to ensure components are feeding first before the radiator. This will help determine if temperatures drop.
I'll keep sharing updates as I progress and add more photos once it's complete!
If you aim to enhance the cooling effect, adding another 120mm fan on the top or inside of your radiator in a pull direction might help.