Question How can I tell where air is entering my water-cooling loop?

Jan 18, 2021
5
0
10
I've got air entering my watercooling loop, causing the cpu block (which is the highest point that air can get trapped in my loop) to be constantly half-full as the water flows through the bottom half of the block only. When I tilt the case I can get rid of the air, but gradually within a few minutes of running normally, the air comes right back.

There's no water leaking at all. The only clue I have that the air isn't entering via the cpu block itself is that I can see the tiny air bubbles gradually flow up through the tube leading from the gpu block up to the cpu block instead of bleeding out through the open cap at the top of the reservoir as you'd expect if the air was being circulated all the way around the loop.

I've tried running the loop without the cap on the reservoir to let the air escape several times, but this hasn't worked which makes me think the air must be getting in somewhere between the pump and the cpu block. I have absolutely no idea how I can check where though.

I also wondered if anyone knows whether or not running the system with a half-full cpu water block could damage the cpu by not prodiving water all the way around as seen in the photos?

FYI the pump is running at 100% speed permanently. Thanks in advance for any help!

pojYsLrxj


poT8ioVOj
 
I've had it for just over 2 years, and it's still within a 3-year warranty including collection, parts & labour.
Yeah you've lost some water over that time. Evaporation still occurs in these systems. and as the above poster also mentioned you can see that the tank isn't filling up as much as it should, so the air is getting trapped in there and just pushing through your loop.
 
Yeah you've lost some water over that time. Evaporation still occurs in these systems. and as the above poster also mentioned you can see that the tank isn't filling up as much as it should, so the air is getting trapped in there and just pushing through your loop.
But there's enough liquid in the reservoir that the pump isn't pulling air down, and plus i've tried leaving the cap off so theoretically the air coming through the loop would escape up through the top of the reservoir instead of being pulled down through the pump?
 
But there's enough liquid in the reservoir that the pump isn't pulling air down, and plus i've tried leaving the cap off so theoretically the air coming through the loop would escape up through the top of the reservoir instead of being pulled down through the pump?
When the Water gets forced through the radiator to cool down Oxygen (aka Air) gets pushed in through the water that needs to be cooled down. That process of cooling will cause evaporation within the radiator over time thus allowing the oxygen in the water (H20) which is 2 parts oxygen to possibly get trapped into the tubes and keep pushing through. Its what happens over times. It is part of the reason I avoid Water cooling Loops. one leak and things get ruined and they only last for so many years. If you are still within warranty I would just get it fixed and or replaced.
 
When the Water gets forced through the radiator to cool down Oxygen (aka Air) gets pushed in through the water that needs to be cooled down. That process of cooling will cause evaporation within the radiator over time thus allowing the oxygen in the water (H20) which is 2 parts oxygen to possibly get trapped into the tubes and keep pushing through. Its what happens over times. It is part of the reason I avoid Water cooling Loops. one leak and things get ruined and they only last for so many years. If you are still within warranty I would just get it fixed and or replaced.
That's interesting and makes sense. Thanks for your help, I think i'll have to get it sent off to be repaired as you say.
 
Just curious, did you get this fixed?
The angle of the tubing coming out of the GPU to go up to the radiator is quite severe and looks like that might be where a little air is getting in. I would redo this in serial instead of parallel so that coming out of the CPU goes to the radiator instead of back to the GPU. Or else get an angle fitting so that bend is not putting pressure on the connection.
 
Just curious, did you get this fixed?
The angle of the tubing coming out of the GPU to go up to the radiator is quite severe and looks like that might be where a little air is getting in. I would redo this in serial instead of parallel so that coming out of the CPU goes to the radiator instead of back to the GPU. Or else get an angle fitting so that bend is not putting pressure on the connection.
Hi,

This is now fixed. As it turns out there wasn't any new air getting into the loop, but each time I tried to flush it out it was getting sucked right back into the pump. In the end what worked was releasing some of the liquid from the reservoir to begin with, then super-carefully tilting the case all the way back to make the reservoir the highest point in the loop, while also keeping the cap on the reservoir loose to let the air escape. Very fiddly to perform. I also found that adding in a plastic anti-cyclone device from EK to replace the mesh that had desintegrated inside of the reservoir helped the air avoid getting immediately sucked down into the pump.
 
I must not have seen this original thread, but what you are explaining is very common and happens frequently. This usually ends up with the pump cavitating and creating micro bubbles which collect over time and then move through the loop and collect.

Overall, the biggest issue is the pump pulling air which is collecting in the reservoir back into the cooling loop while also trying to dislodge the other air bubbles. This is best achieved by filling the reservoir as full as possible and capping completely prior to moving/tipping to dislodge air to help prevent more of it moving back into the system.
 
It looks like your loop is configured incorrectly. Your GPU cooler has an in and out port. You have both ins and outs being used. It should be into GPU, out of GPU into CPU, out of CPU into radiator.....etc.

Good point - looks like the coolant would take the route of least resistance, which is directly through. Some coolant will route through the block, but not moving full speed through the block. Those ports are straight through - and likely just a trickle moving through that GPU block, but would like to see if there is any detail otherwise.