Puzzling Problem

Associate
Joined
8 Feb 2014
Posts
9
Recently i had to replace the XSPC 750 pump/res in my system after it died on me.

My loop goes Pump/Red > CPU Block > 240x120 rad > gpu block > 120 rad > Pump/Rad

Since i replaced the pump/res and tubing i have been getting large pockets air forming at the output barbs of both the cpu and gpu blocks that never seem to move with the flow and they seem to grow faster when they system is under full load although the cpu is running well within temperatures (even under full load) although the gpu is running a little warmer than i would like when pushed to its limit, but the temperatures rapidly drop when no longer under load. (Neither is overclocked currently until i can fix the problem)

At first i thought it was a loose hose letting air in so i replaced tubing and replaced all the barbs with compression fittings but the problem has remained.

There are no noticeable leaks anywhere in the loop I can find and no matter how many times i move the air pockets to the res the water level never seems to drop significantly although the loop seems to be ever so slightly pressuring itself (there is an audible hiss when removing the filler cap).

Any help or suggestions would be gratefully received.
 
If the hosing was letting air in, it would also let coolant out.

Pockets of air are much harder to get rid of if your loop components were dirty/greedy or even a little dusty. I would periodically tilt the case in all different directions to get the air moving along. i tend to pinch the tubing a little bit if it gets stuck anywhere but avoid doing it near the barb. Make sure you let the air out by unscrewing one f the fill ports from time to time.

Pictures would make it easier to spot a problem if there is one, but it sounds like you are in need of a good bleed and that's it.

As long as there isnt a ton of air in the loop or major air bubbles in the blocks, you wont see any performance difference. If you have loads of little bubbles, that is normal and you will likely see more when the loop warms up but will disappear completely over the next few weeks with regular bleeding.
 
Thanks for the reply the issue is that i get two different types of "bubble" the ones that stick to the sides of the tubing after filling usually work their way out within a few days or with some flicking of the tubing they move with the flow as you would expect.

The problem bubbles look like this

DSC_00071.jpg


and start to appear about 1 week after refitting the system (this happened both times) and once they appear visibly grow larger when the graphics card is under load but there are no leaks anywhere i can find, and no matter how many times i move it back to the pump by rotating the components it always returns, there are no visible bubbles entering the block and the coolant level never seems to drop. I'm completely at a loss as to what to do.
 
Is there enough coolant in the reservoir to make sure that it's not pulling air through the system? Or if it's a tube res, are you sure that the flow isn't creating a vortex and pulling air through the system like that?

Generally, following a rebuild of your loop, it's inevitable that there will be air bubbles in the system, but over time, as you cycle your pump, the air bubbles should get pushed out and get "released" in your reservoir. This would cause the coolant level in your reservoir to go down, so it's important to check the level after your rebuild.
 
With an air bubble like that I would be wondering if your fluid is circulating properly since that kind should get forced round if your pump is moving it fast enough
 
With an air bubble like that I would be wondering if your fluid is circulating properly since that kind should get forced round if your pump is moving it fast enough

My thoughts too - air like this should be able to be moved by the pressure of the pump.Perhaps try moving the case to different angles to force the air round to the res, where it will escape.Try squeezing the tubes to force it round?
 
Are you forcing air back to the pump or forward to the reservoir to get rid of it?

I would try and pass it past all the blocks, rather than back in the loop, or itl just end up there again.
 
I always pass them forward through the loop to the return port on the res then open the fill port, i have probably repeated this procedure 100 times by now, and the bubble always returns and grows when the gpu is under load. It also seems to retreat into the block as it cools.

I too suspect some issue with the flow through the system but if that is the case i'm baffled as to why temperatures get pulled down so rapidly once no longer under load. And if this is the case is it a pump problem or a issue with the block itself?
 
This may sound like a stupid question so forgive me... Are you totally sure your blocks are connected with the water going the right way? We had similar on here not long ago and turns out his cpu block was wrong aka in port should have been out
 
Is it posssible to run the pump for a little while with the fillport open? Just wondering if there's a mini air lock there
 
After the bubbles are pushed to the reservoir, where do they come from to build up at the block again?

Have you turned the case so that the GPU block is the lowest point in the loop and ran the pump?
 
Whenever i move the air to the res i always run the loop with the cap off for an hour or so.

Certain the cpu block is the right way around it is clearly labeled, and think the gpu is the correct way(its the same as when i first got the system but i will try to find a spec sheet and double check for that. The block is a XSPC Razor GTX580.

And in the bubbles seeming come from nowhere and seem to originate inside the gpu block and yet there are no leaks(at least detectable ones).

Thanks for the suggestions so far guys this has been driving me crazy.
 
Last edited:
This may sound like a stupid question so forgive me... Are you totally sure your blocks are connected with the water going the right way? We had similar on here not long ago and turns out his cpu block was wrong aka in port should have been out

I thought that when I looke at the gpu fittingss but without knowing the block first hand can't say if it matters or not.
 
Only matters with CPU blocks, rest are compatible no matter which port. For CPU blocks, it is the more central port for EK blocks.

What is your block?

What fittings are those?

Possible that the thread on the fitting is a tad long for the GPU blocks. EK provide spacers in case the thread is more than 4mm long, as that's when it starts to block input/output to the block.
 
The fittings used are XSPC G1/4" to 7/16" ID, 5/8" OD Compression Fittings but the problem still occurred with the original barbs which i replaced when i assumed air was coming in at the barb, still glad i replaced them though at least i gained some adaptability and style ;) Block is a XSPC Razor 580
 
Last edited:
XSPC fittings should have the correct thread length. Not familiar with their 580 blocks. You could remove all the blocks and test using de-ionised water, whether there is a flow problem or not.

First would unscrew the compression part and see where the bubble 'cuts off'. You will find it easier to see if the air is coming from the block this way or from the other direction. It isn't much but at least itl give you a bit more certainty of where the air is coming from.

obviously do this with the PC off.
 
I'm pretty certain now it originates inside the gpu block its almost as if as the block heats up something is closing off restricting the flow and causing a gap. So i guess the next step is to try to find a new block compatible with a GTX580 as these XSPC blocks are now discontinued.

My other speculation is the pump may not powerful enough although this seems unlikely as it manages to cycle the bubbles out perfectly fine when filling and bleeding. And i replaced it with the same res/pump that blew originally just an updated version.
 
If this is reference 580, i know of one guy willing to let go of an EK block in the Surrey area ;)

What is the flow of coolant like when it comes out the GPU?

Obviously if its down to an aggressive trickle, you have a restriction problem but i doubt it will be that bad if the system was successfully running for a week.

What pump is it?
 
Normally the flow seems fine you can't detect it visually but can feel the warm water being pushed out of the block in the pipe and cool water moving out of the radiators. When the bubble forms in the pipe though you can see a trickle of fluid running down the inside of the pipe past the bubble.
 
Normally the flow seems fine you can't detect it visually but can feel the warm water being pushed out of the block in the pipe and cool water moving out of the radiators. When the bubble forms in the pipe though you can see a trickle of fluid running down the inside of the pipe past the bubble.

Definitely a flow issue then. A loop with no flow issues will feel the same pretty much all round the loop, since there is enough flow to keep water temperatures constant.

Maybe the new pump to the job either. If the blocks had no flow issues before, no reason why they would gain them now in a clean new loop.
 
Back
Top Bottom