Water Cooling Routing help!!!

deFiniLoGy · 2 May 2013 at 14:37

So... I am now planning the route for my watercooling rig,
but quite stuck as my case is a pretty small one and I am basically stuffing it up

Here is my spec:
Asus M5E
i7-3770K
MSI 7970 lightning x 2 CF
Corsair Dominator Platinum 4*4GB
Seasonic X-1250
NZXT Phantom 410 Midi Tower

Watercooling blocks:
Koolance 380i
EK FC7970 lightning * 2
EK FC Bridge CSQ Serial
EK X3 150 Reservoir
XSPC EX240
XSPC RX240
Koolance PMP-500

At the moment, my route plan is like this:

But I don't know whether having the pump directly before the Radiator will be a good idea or not?
As I think pumping coolants directly to a block might be better?
And how can I route that really

charic · 2 May 2013 at 14:41

I know VERY little about these things but wouldn't putting the res in one of the front bays be both tidier and save some space?

Ofcourse I may just be acting like a pillock

deFiniLoGy · 2 May 2013 at 14:59

charic said:
I know VERY little about these things but wouldn't putting the res in one of the front bays be both tidier and save some space?

Ofcourse I may just be acting like a pillock

My front bay is full

andybird123 · 2 May 2013 at 15:02

looks fine... loop order makes no difference to flow or temps

fan position might cause you problems... pulling cold air over a rad and then dumping this hot air over the 2nd rad means that the 2nd rad's efficiency will be severely compromised

you could put a 140mm fan on the side panel on exhaust and turn the rear upper fan in to an intake which might help,
or put the side panel and rear on intake and all your rad fans on exhaust (which is the way I would do it)
or all rads on intake and the 2 other fans on exhaust (which would be the best for loop temps as long as your other components don't mind being in a hotbox)

CaptainComedy · 2 May 2013 at 15:08

by putting the pump after the graphics cards, isn't that giving the graphics cards the worst end of your pressure / flow? and pumping straight into the rad the fastest end of the water, wouldn't that affect the time the water has to pass on the heat in the pump?

deFiniLoGy · 2 May 2013 at 15:13

codec said:
by putting the pump after the graphics cards, isn't that giving the graphics cards the worst end of your pressure / flow? and pumping straight into the rad the fastest end of the water, wouldn't that affect the time the water has to pass on the heat in the pump?

I've absolutely no idea...
That's why I need all the pro's help

ubersonic · 2 May 2013 at 15:24

Looks pretty good, as said loop order is irrelevant as long as the pump is after the res, you just go with whats easier to route. Personally looking at that I would do it as you have planned but turn the rear fan into an intake, gaining positive airflow and helping to offset the hot air coming in from the front rad.

andybird123 · 2 May 2013 at 15:46

codec said:
by putting the pump after the graphics cards, isn't that giving the graphics cards the worst end of your pressure / flow? and pumping straight into the rad the fastest end of the water, wouldn't that affect the time the water has to pass on the heat in the pump?

water isn't compressible, so it's flow speed will be the same through the whole loop no matter what order you put things in - the same size piece of tubing will have the same flow rate whether it's near the "beginning" or the "end" of the loop, though being a loop there is no beginning and end strictly speaking

the only reason to put the res just before the pump is because it makes priming the pump easier

once full, the pump will suck water through a component just as much as it pushes it through another

mjones1 · 2 May 2013 at 15:56

Loop order pretty much makes no real difference long as res before pump

If you can flip the rads i would do something like this.

yellow = res
pink = pump
The pump would be pushing into the gpus.

Hagar · 2 May 2013 at 17:12

Water is not (very) compressible but air is and the most likely trapped air is in a rad not a block.

I would put the high pressure end of the pump (before pipe losses) to a block and the return from a rad.

Hence loop looks like this Res|Pump|Block|anything else|Rads.

andybird123 · 2 May 2013 at 19:15

nkata said:
Water is not (very) compressible but air is and the most likely trapped air is in a rad not a block.

I would put the high pressure end of the pump (before pipe losses) to a block and the return from a rad.

Hence loop looks like this Res|Pump|Block|anything else|Rads.

If you have air trapped in your rads then you're doing it wrong

deFiniLoGy · 2 May 2013 at 19:34

andybird123 said:
If you have air trapped in your rads then you're doing it wrong

how do u know that u have got air?

CaptainComedy · 2 May 2013 at 19:36

agreed, the res is shut off and air is pumped out as the loop is pushed, the reservoir is the only place with air in.

deFiniLoGy · 2 May 2013 at 19:37

codec said:
agreed, the res is shut off and air is pumped out as the loop is pushed, the reservoir is the only place with air in.

This makes me wonder - do pumps work better in push or pull? (does that even apply to pumps?

I think pumps pull water from res and push to the others right :confused:

CaptainComedy · 2 May 2013 at 19:41

yeah sorry I misread the guys post above (with diagram) thought he was trying to push into the res (not a good idea), I didn't read that he said into the gpus

andybird123 · 2 May 2013 at 19:46

deFiniLoGy said:
how do u know that u have got air?

You can hear it, and it comes out real easy if you tilt your rads port side up and back

deFiniLoGy · 2 May 2013 at 19:50

andybird123 said:
You can hear it, and it comes out real easy if you tilt your rads port side up and back

right...
have been watching tutorial videoes

Hagar · 2 May 2013 at 20:46

Flow is not the only measure. The pump provides the motive force to the fluid. The pump intake has a fluid head equivalent to the height of the reservoir or possibly the height of the case. The output side of the pump has a head equal to the pump capacity ie 2-4m or 6-12 feet.

If the pressures were equal across the pump there would be no need for a pump at all. It makes sense for the higher pressure to do the most work, pumping across the restrictive blocks.

The overall flow rate is a constant agreed, but the pressure will drop (head loss) around the loop at each point away from the pump with losses at each component part, pipe, fitting, block, bend and radiator. Basic fluid dynamics.

Any slight air bubble will increase the compressibility of the fluid and it is at the slowest point in the circuit where it will have the most effect.

deFiniLoGy · 2 May 2013 at 20:49

nkata said:
Flow is not the only measure. The pump provides the motive force to the fluid. The pump intake has a fluid head equivalent to the height of the reservoir or possibly the height of the case. The output side of the pump has a head equal to the pump capacity ie 2-4m or 6-12 feet.

If the pressures were equal across the pump there would be no need for a pump at all. It makes sense for the higher pressure to do the most work, pumping across the restrictive blocks.

The overall flow rate is a constant agreed, but the pressure will drop around the loop at each point away from the pump with losses at each component part, pipe, fitting, block, bend and radiator. Basic fluid dynamics.

Any slight air bubble will increase the compressibility of the fluid and it is at the slowest point in the circuit where it will have the most effect.

So the best option will be routing from res to pump to block,
But does it matter which block it is going through first?
display card vs CPU

Hagar · 2 May 2013 at 20:54

deFiniLoGy said:
So the best option will be routing from res to pump to block,
But does it matter which block it is going through first?
display card vs CPU

Not really on a single loop, the total heat picked up on the last block will be the same each way. Tradition is CPU|GPU but after several circuits, the fluid will reach its working temperature anyway.