Getting rid of bubbles

elmuzzy66.

elmuzzy66.

elmuzzy66.

Soldato
Joined
24 Aug 2006
Posts
10,113
Location
Gibraltar
Whats the best way to get rid of bubbles. My loop has a whole load of them, really small, very vissible. Thinking of bleeding the loop and starting again, unless there is a easier way. If it helps loop order and componets are.
dcc pump/res top, gpu,cpu, dual rad. 7
 
leave the top of your res off and run for a while then move the case everyway up up you can (with res closed) and shake about a bit...best off having your set up at stock so mothing overheats while bedding in water!!!! then like magic the bubbles go!!
 
I found with the res top that th bubbles were just being forced back into the loop again. What I did was bend a spare PCI backing place and put it in the res to change the direction of the water so rather than it being forced straight back into the pump it went into the res first.

I got some pictures of what I did from an old thread.

IMG_0128.jpg


This is not so clear, so I went over it with paint to show where the pci backing plates was etc. you can just about see the pci plate blocking the pump inlet at the bottom so it forced the water coming into the res away from the hole. Obviously the hole is still open to complete the loop.

IMG_01282.jpg


This got rid of all the bubbles in about 10 mins. Don't forget to take the pci plate out after you are done though, you don't want it to start rusting!
 
I wouldn't recommend putting anything in the res over the pump inlet to speed up bleeding. At one time it was fashionable to put a bit of foam in the Alphacool reservoir top because that helped until it was realised that the foam slowly dissolved over a long period and destroyed everything.

The issue with setups like this is people buy the most powerful pump available (usually the 18W DDC) and then the fluid shoots around the loop so fast the bubbles take forever to clear or worse, a semi-permanent vortex appears and the bubbles go on for a very long time indeed. People running 10W DDCs have far fewer problems.

Another issue is people's expectations. Lots of people tend to run their PCs for 2-3 hours in an evening and they expect any modification to be completed in that timescale. Bleeding a high flow loop can take 48 hours of continuous running with much tilting and shaking of the components required to ensure all the blocks aren't airlocked.

Just leave it running. The loop will be a little bit less efficient than if it was 100% filled with coolant and it will be very noisy, but it will work.
 
WJA96 said:
I wouldn't recommend putting anything in the res over the pump inlet to speed up bleeding.
Click to expand...

There is nothing wrong with this, the inlet isn't blocked, all that is happening is the direction of flow into the res is deflected away from the pump inlet so the bubbles have chance to escape ( you can just about see in the top picture that the inlet is not obstructed at all). Without it the bubbles are forced straight through the res and back into the pump inlet.

The Pci plate was only left in long enough to get the bubbles out, so around 10 -15mins.
 
I don't quite get how that changes the direction of flow in such a way that it stops the air coming back in, which you said was the problem you were having.
 
Superewza said:
I don't quite get how that changes the direction of flow in such a way that it stops the air coming back in, which you said was the problem you were having.
Click to expand...

The reason being instead of the bubbles being forced in to inlet of the pump and thus circulating the loop again indefinitely the are forced into the side of the res by the pci blanking plate and so rise to the surface of the res and out of the liquid. I can't think of a simpler way of explaining it than that.
 
Slinkie's idea would certainly work. It'll leave some iron in solution which isn't great but probably doesnt matter. Instead you can take a length of tubing, cut a notch out of the end and slide this between inlet and outlet. It'll kill the flow rate, which is good, and it'll stop any bubbles being driven straight into the pump inlet. Plus it wont dissolve any worse than tubing does anyway.

The simpler approach is to just put a clamp somewhere down the line, a spare jubilee clip or the like. The restriction slows down the flow rater and allows bleeding. Use several clamped down a little bit instead if you like.
 
Slinkie, you're idea is sound, but WJA96 is correct in a way. The issue albeit small is the vortex created by the pump/res. This is where the PCI plate is a good idea...just not right above the inlet. And then there is the issue of galvanic corrosion.
 
Thats a good idea with the pci plate, i think my innovotek pump/mini res gets most of the bubbles out if you leave it with the top open for a while, maybe also having the res at the very top of my case helps.

I have never tried the clamps on the pipes so might give that a go next time i have to refill.
 
Bit odd that one. However if you leave it sealed for a while the pressure inside drops, and you hear a little hiss when you open it. Can't explain this one, but it does suggest leaving the cap off to me.
 
I think most of it is gone. i can't see the bubbles but i think theres a few in the rad as everytime i start the pc i can hear the coolant flowing through the rad. It goes away in 1 second. Its also very silent so must have got rid of most. Thanks for the help guys.
 
You must log in or register to reply here.
Back
Top Bottom