Problem getting air out of my WC setup

[Wolf2000me]

Hey,

I have my watercooling gear mounted in a 2nd case of the same size as my PC case. I am only watercooling my CPU at this time.

My pump is on the bottom of the case, which is connected to the radiator (120mm) which is mounted where the ATAPI devices should normally be situated.
The radiator is connected to the CPU WC block (Maze) which is in turn connected to the reservoir. The reservoir is attached to the pump.

Pump: Eheim 1048
Res: aqua computer aqua inlet (bottom)
CPU block: Maze
regular 120mm radiator cooled with 120mm fan.

To the problem. Whenever I need to power down my watercooling setup or when it gets powered down because of a power fail in the house I spend at least 30 minutes getting the pump reasonably quiet. It always makes noise because there is always air in the setup. It makes quite a rattle. So far I have been able to get it completely quiet only once. More often it will make a slightly audible noise. To get it at that state I need to move and tilt the case housing the WC setup but it doesn't always help. It seems more like a matter of luck anyway. And leaving it run thinking it would stop making noise and the air would be out doesn't work at all. If it makes a noise it will keep doing that until I start tilting the case. There are no leaks anywhere. If there were to be a component that would take on air it would lose water, right?

Any thoughts are appreciated. Certainly with this warm weather coming up i will need all my cooling capacity

[robbie]

Shake shake shake...shake shake shake aowhhhh shake your setup!!! shake your setup!!! (to the tune of shake your boogy)

You'll need to shake your Rez to get ALL the air out of it. Do this with the power on.

[Samuknow]

In addition, a res is better is situated at the top so air can be trapped there. As Rob said, shake it up and rotate the radiator as well. Also a good additive (water wetter) has a suffasctrant that will help collect the air together and move it to your res.....

[cadaveca]

Res should always be the highest part of your loop. as well, it sounds like you have either really short hoses, or not enough water behind the pump. My pump will do the same thing when cranked to full speed, unless i use a res, and I'm only using a t-line to fill the loop.


As the pump cranks up, it compresses the water...and hence the lack of water might be a problem. Pure distilled water will compress less it seems, but can be quite corrosive to your block and rad.

[Wolf2000me]

Well I am using distilled water with an additive against alagae and such. I did shake the setup quite some as well. The reservoir is refilled completely. There's no way that the pump runs out of water.
I thought my hoses were too long actually. I'm using a bit more than 1 meter for hosing. Those hoses are 6mm/8mm. Maybe that's it?

Thanks for the replies so far

[cadaveca]

just because the system is full does not mean the pump has enough water. that's why i said short hoses.


the pump partially compresses the water as it pumps it...this can create air bubbles if the pump doesn't have enough water to force into itself. You wouldn't see this most likely..it would be right behind the impeller. the fact that your pump is noisy kinda bolsters my thoughts. either that, or your additives are foaming the water when it compresses, which might happen if your mix is a little strong.

this is the same for any pump...not just pc water-cooling ones. the sound the pump makes should tell you what's hapopening.

[QSDT]

The effect is called cavitation.

[cadaveca]

The only way to truly avoid this, and to prolong pump life, is to have the flow the same at the inlet as it is at the outlet, which, by using a res, definately isn't happening....never mind how restrictive a block is, and how this will slow flow getting back to the pump.

An easy way to offset it, is to put more water behind the pump, but by using a res to do this, you lose the pressure of the coolant, and therby some of it's motion.

[Aedan]

Originally Posted by Wolf2000me

If it makes a noise it will keep doing that until I start tilting the case. There are no leaks anywhere. If there were to be a component that would take on air it would lose water, right?

In reverse order here. No, it is possible for a seal to be watertight but not airtight. Some seals are designed to cope with positive pressure, but will leak with negative pressure. I know that on a couple of pumps I have, they leak air when faced with negative pressure.

What orientation do you have with your pump? Ideally, you want the outlet pointing vertically upwards. Where is the pump in the system? My understanding is that generally, the pump is best left as the lowest item in the system.

Cavitation is possible, but you'd have to have a pump that's vastly overspeced for the system. Cavitation occurs when conditions inside the pump allow localised pressure drops below the vapour pressure. Effectively it means that you get small bubbles of vapour in the liquid on the suction side as the low pressure allows the coolant to boil. However, I would be a little surprised if that were occuring inside the pump you have!

[Wolf2000me]

I really need to get myself a digital camera

The pumps water output is facing upwards and indeed when I turn the pump 180° around it does seem to take on air and doesn't leak water. That's no problem of course.

The pump is the lowest component of the loop and the radiator is the highest part.

The hose from pump to radiator is roughly around 40 cm. From rad to CPU block is around 80cm and from the block to the res is 50 cm. I'd say that's more than long enough but it allows me to shake and turn the thing as I want.