where do air bubbles come from

I was running fine, and then bam, mass air bubbles came out of nowhere, more and more keep coming up.

Usually a sign of a leak somewhere. It could also be that some trapped air has just found it's way out of where ever it was stuck. How long has the loop been running? Tee line or res? More info would be helpful.

Over time, you actually lose water through the tubing. It's such a small amount it doesn't drip but evaporates away. In it's place is of course, air. At least that's been my experience. Move the computer and the air will move and create bubbles. Just add more water and bleed the loop again.

Edit - my slow typing got me again lol. Or what nikhsub1 said!

its only been running for a couple hours. but once the bubbles started forming, it was like it multiplied afterwards, more and more

im really suprised how speedfan is not showing any real increas in temp. its either 30 or 36 c. My T line had about 2 inches of emptiness, now its water to the top of the t-line. oh well.

Where is the res/T-line in your loop? If you don't have one or the other it will never get better, nowhere for the bubble to get trapped and you let them out by filling back up. I have a T-line and a fresh install I still get some air come out of the exchanger.. Mine for the most part cleared up after a day or so, adding my additive made it foam up just enough that it helped break up the air masses and move it to my T-line..

T-line is right after the radiator.

I am going to go with Nik and say that you most probably had some air trapped somewhere. The thing that I am wondering about is how the level in your T-line is rising. Could you give us a pic that specifically shows your T-line and how it is connecting to the rest of your system?

I see the T now, just outside the top of the rad.. I agree the water going up is confusing to me also. water in my T would slowly drop as micro bubbles got caught up in it.. Can you clarify like Malpine said?

Its pretty much already been said with the bleeding issue and how bubbles can get trapped when you fill the loop. the best way to get rid of em is to shake the $^&* out of your rad, since thats where most end up trapped. also, your pump could trap a few but it looks like they would bleed out fine in your loop.

I think i have heard some people mention their pump draws air in from the t line, although since your's is after the block, i dont see this being an issue.

Add some more water though, looking at your picture, you will run out of reserve soon from everyday evaporation.

AlucardCasull said:

its only been running for a couple hours. but once the bubbles started forming, it was like it multiplied afterwards, more and more

Click to expand...

theres your problem. you should always leave your loop running for 24-48 hours with ONLY the psu in the case (to power the pump) this is what we like to call "leak" testing.


the air thats coming out is simply air stuck somewhere inside your loop. if you just let it run for 24 hours and keep on topping up the t-line as the water level goes down you should be good to go.



but from now on. you should ALWAYS let it leak test otherwise you could have a spill everywhere.

From Experience.. It can take a few days of running to get almost all the air out.

I usually leak test mine for at least 24 Hours. And normally after that most of the bubbles are gone, but during next day or two afterwards, when comp is running there are a few bubbles that do rise up.

With my T setup it will completely bleed in an hour. Then I'll take the case and turn it every way possable to pull any bubbles out. I'll wait another hour or two then do it again until no more hidden air pockets exist.

probably trapped air indeed, the only other things i can think of are:
-dishwasher liquid used as additive iso waterwetter/antifreeze
-some chemical reaction. what did you use as additive? if any
-pump cavitation. what pump are you using

btw, its actually not a bad idea while leaktesting to run the pump slower (7volts) if it allows it, it tends to bleed the air out faster

If you do happen to have a trapped air bubble and it gets to the pump you've got a bazillion smaller bubbles all of a sudden, so much so the water looks like foam. When it gets that way it's sometimes better to turn the pump off for a few minutes to let things settle - and hopefully some of the bubbles will decide to settle in the T-line ...

Rolls said:

probably trapped air indeed, the only other things i can think of are:
<snip>
-pump cavitation. what pump are you using

Click to expand...

Well, cavitation is caused by a drop in pressure across a surface (the pump vanes in this case) that causes the water to boil at room temperature. As such, they are composed of water vapor and they will be reabsorbed later in the loop when the pressure normalizes. The major concern of cavitation is that the bubbles can collapse faster than the speed of sound (in water) and the resulting shock waves can eat away at your components over time, eventually to cause leaks.

That much being said, I am still concerned about the rising water level in the T-line. Unless water is being created out of nothing, it is being displaced from somewhere else in the system. Basically, the air bubbles are collecting at some hidden high point in the system. My best guess is that the radiator is the culprit here.

Now in that the tubing ID is going to be pretty close to the OD of the barbs (which I assume is similar to the ID of the radiator tubing), that could add up to a fairly long length of radiator tubing with trapped air in it. If the pump head is sufficient to push through it then great. However, if that were the case, then why would it not clear the air pocket as well?

I could, of course, be wrong here but run with me for a moment and you see a possible restriction in the coolant flow. It should be easy enough to fix with the usual tricks such as shaking the radiator or pulsing the pump. Then, instead of not seeing an increase in your temp, you could see some few degrees of decrease.

Turn off computer.

Unplug mobo and peripherals from PSU. Start pump, run 20 secs. Stop. Wait for bubbles to form bigger bubbles(1-2 minutes). Run pump another 20 secs so larger bubbles reach T-line and bleed out. Repeat. Once in a while tilt the case with pump running to force out air pockets. They like to hide in the rad
and waterblocks.

This is why you use prime the pump and fill the entire loop with distilled water first, and then introduce additives.

I had a leak. Now I have a fried pump.

Thats Y leask testing is a important step . I still dont see how a Leak will cause your water in your line to rise

Bubbles take up more space than just water.

Where was the leak?