Flawed from the start?

iB CoMpL3XiTy

iB CoMpL3XiTy

iB CoMpL3XiTy

Associate
Joined
10 Sep 2010
Posts
1,533
Location
Cornwall, England
Hello :)

I recently split my watercooling loop into two sections (not separate loops) So basically I have two pumps, aload of blocks and one res.

This Pic hopefully explains it better
lalalajl.png


So basically Res > Pump1 > CPU > 120.3 > Res
and Res > Pump2 > 120.4 > GPU's > Res

The problem is that the CPU loop has loads of Air left in it, the pump is making a lot of noise and I can see air bubbles in the tubing and a big air lock in the CPU block. Its been running for 3 days now, there's plenty of water in the res. I've tried tipping the case in all directions which does move the air around but theres still a air lock in the CPU block which wont budge.

Any Idea's?
 
You'll probably find that the pump with the least resistance through its loop will rob the water from the other pump.

You'll need 2 reservoirs, they could possibly be linked though.
 
Have you tried pulsing the loop. By this i mean pinching the tubing to restrict flow and then letting go - this sometimes helps. Another option is laying the case on its side (detech the res so it still stays vertical).
 
@ Danm54,Two Res's would be a real PITA as I'd really want another the same style (FrozenQ T Virus) and I dont really have anywhere to put it.

If there's enough water in the loop then surely both pumps will get enough coolant?

@w3bbo, I've tried pinching the tubing and the pump then makes a real racket and air bubbles move around the loop, but still no movement of the air in the CPU block, Not tried laying the case down with the res vertical


Will the fact that mavity wont be forcing the coolant to Pump1 cause a problem? (because of the angle of the tubing going from the res to pump inlet)
 
How about disconnecting everything but the cpu loop and run that only. Would need to short out the 24 pin PSU for it to work
 
Ok alittle update

I first turned off the GPU pump (obviously when power was off and shorted out the 24pin) and just ran the CPU loop, first pic here shows the air in the CPU block

http://imageshack.us/photo/my-images/703/secondu.jpg/

Next image shows the CPU pump off, notice the gap of coolant at the top of the block.

http://imageshack.us/f/607/secondl.jpg/

Last image shows the res

http://imageshack.us/f/221/thirdcj.jpg/

Its worth noting that with either of the 2 pumps on there is no pump/bleeding noise, but when both are on its like there fighting for coolant.

P.S Sorry guys image tags not working so please use the links. Cheers
 
Last edited:
Well the CPU block looks like its the wrong way up, outlet should be at the top.


If you only have one outlet on the res the a tee piece you're trying to pull enough water for 2 pumps through one bit of tube, the stronger pump/the loop with least resistance will always be favoured.

Is it possible to have 2 outlets and 2 inlets?
This way both will share the res but will each balance their flow out as they'll be pulling the same amount of water they're returning.

So seperate outlets and seperate inlets and all should work fine once its bled. If you get an issue then unplug the pump thats bled fine.

Also the pump below the res wont help for bleeding, make sure the tubing falls to the pump whilst bleeding, once its running it'll pull small air bubbles out but it wont bleed itself that way.

:D
 
Danm54 said:
Well the CPU block looks like its the wrong way up, outlet should be at the top.


If you only have one outlet on the res the a tee piece you're trying to pull enough water for 2 pumps through one bit of tube, the stronger pump/the loop with least resistance will always be favoured.

Is it possible to have 2 outlets and 2 inlets?
This way both will share the res but will each balance their flow out as they'll be pulling the same amount of water they're returning.

So seperate outlets and seperate inlets and all should work fine once its bled. If you get an issue then unplug the pump thats bled fine.

Also the pump below the res wont help for bleeding, make sure the tubing falls to the pump whilst bleeding, once its running it'll pull small air bubbles out but it wont bleed itself that way.

:D
Click to expand...

Wrong way up? there isn't a Top port, both inlet and outlet are on a horizontal plane.

The Res only has 3 ports, hence the use of the Q fittings. I think what I'll do is: Res > Pump1 > GPU1 > GPU2 > 120.4 > Pump2 > CPU > 120.3 > Res

Hopefully that will solve the issue.
 
Yes, I certainly wouldn't use a Y splitter for the return to the res line at least. With a y splitter there, the pressure at that point has to equalise so if you had separate loops and say the CPU loop had a lower pressure than the GPU loop at that point then when you connect them up you will get back pressure in one of the lines which will reduce flow to that side. You want them emptying their water into the res completely independantly.

Edit: And if you're going to do a single loop with pumps in series, which is the best idea anway assumig the pumps are identical, then you want one pump straight after the other.

You other option is to run one of the loops without a res at all, just with a T line.
 
Last edited:
Yea sorry, rotate the cpu block 90°.

Here's an option for you to try. Run the cpu loop with just one connection to the res.

Using 1 res outlet take a short bit of pipe down to a tee piece, it must go down to allow air to travel up it, it must also be higher than the pump.
Use one side of the tee to feed the pump and return your loop to the other side, it will be a pain to fill and you may have to disconnect the return from your loop and run it into a container untill most of the air is out. Obviously block the open end of the tee.

Once that loop is connected and full and you've filled your res, turn that pump off and fill your second loop that returns to the top of the res.

Once you're done the cpu loop wont run through the res but will be kept full and any air will find its way to the res. The gpu loop will run through the res as normal.

Oldskool plumbing for the win!
 
Whilst the above would appear to work on paper there is no difference between that and a T line really, and a T line would be vastly easier to deal with. I don't think you have to rotate the block as we have pretty much established I think that the cpu loop is being starved of flow and therefore there is not enough oomf to force the air out which is why it is staying there.

Also with that particular res, are those double helix tubes part of the loop or are they purely decorative? If they are decorative then it means the single inlet at the top is just dumping water into the res through an open port, which means whenever you switch the loop off there is nothing stopping water draining out of the top rad letting air into the rad, which may or may not get stuck in there the next time the loop starts, which is something I personally wouldn't like.

And another thing, how have you got a 120.4 in the bottom of an 800D? Where is the psu? Pics? Final solution, get a swiftech microres and stick it in the bottom compartment.
 
Last edited:
My bad, you'd also need to put a tee piece above the res and have a fill port, no way to fill it otherwise!

It works on more than paper as well, its simply a combined feed and expansion system, same as older domestic heating systems.

It'll work if you don't want to use another res.
 
One of the pumps is a DDC350 10W and the other is a DDC355 18W so I'll have to run the 10W to the 18W rather than the other way round.... This probs explains why the CPU loop (with the DDC350) is struggling atm.

I'll do a semi build log with final pics in a new thread :)

Oh and its a modded TJ07, the Obsidian is my old case (sig hasnt been updated)
 
iB CoMpL3XiTy said:
One of the pumps is a DDC350 10W and the other is a DDC355 18W so I'll have to run the 10W to the 18W rather than the other way round
Click to expand...

Pumps have to be indentical to put in in series in a single loop. Just eveything in one loop with the 18W.
 
You must log in or register to reply here.
Back
Top Bottom