Cooling New watercooling, advice needed!!

Okay I have got all my kit now and I was just wondering if everyone thought this would be allright:


Its a dangerden supercube rad with two 120mm panaflo H1As, and all the blocks are OCPC with a 1048 pump on relay.
All stuffed (lol) into a PC-70.

Now I was wondering if this is the best way without a res, or should I have another t-piece next to the other one so that I have one pipe to fill from and one for air to escape from, tho I suppose I would have to do it with everything out of the case so that I could move the rad around to shake out the air.

Helpful advice and suggestions most welcome

 

New design with the t-piece at the top instead, I think this is amuch better design

 

I was going to propose that... but you need to turn the flow around. The T should not only be at the top, but also the last item in the circuit before return to the pump.

 

How about this???

 

Or this one:

 

I think Nexxo means like this



Notice the T before returning to the pump at the top.. Not sure if you need that one between the pump and rad..

Your pics with the splitter going the CPU and NB may restrict your flow and you may risk one side with no pressure. Thoughts anyone else?

 

l33t is right: splitters are a pain (because you have to observe several principles to actually make it work) and offer no advantages over putting things in series. The T needs to be the last thing _before_ coolant returns to the pump: so it's T --> Pump Inlet, NOT Pump Outlet --> T. The pump ideally needs to be the lowest item in the circuit, so it shouldn't be up there.

 

It looks to me like your setup could be improved. The water coming out of the pump will be moving at its fastest. Going to the rad first, this will cause 2 problems. First it will restrict flow right off the bat. Second it might be moving too fast to remove all of the heat.
I know it seems like you'd want coolest water (coming out of res) going right to cpu. But in real life, you get better results with fast water going thru cpu, then as its restricted by the other blocks it will be going slower thru the rad and able to dump more heat to the air. Also there is a lotta debate over the merits of water cooling the NB. Most seem to think its not worth the restriction and opt for passive(zalman) cooling on NB.
Something like this might do better. You'd prolly need to put you bleeder/fill T between the pump and rad, although it would be better to have it before the rad.

 

NO, NO, NO, NO, NO! Put the darn T AFTER the rad and BEFORE the pump inlet! Like a res, you CAN'T just put it anywhere!

 

I don't understand how water can flow faster at one part than another if you're using the same size tubing.

 

Flow is decreased by friction *walls of tubing and blocks for example cause friction* and until the water gets another boost from the pump it will slow down....think of it this way: u press the gas on ur car....car goes forward....but eventually ur car will stop unless u press the gas again because of the friction with the tires on the road....

 

down boy.....calm down and post....everybody was a beginner at one point *yes even me *

 

Yes, but it always has the force of the water behind it pushing it along. Unless the pressure is different there's no way the speed can change because you'd end up with a vacuum before the pump. If there's resistance then surely the water flow in the whole system will slow down.

 

Thats true but the pump has a certin "head"....the more it pushes the less the flow rate becomes....when it goes up the flow rate drops....when it comes back down the flow rate increases a lil....when the water enters a block it slows down because of the twist and turns.....Have to remeber the bends and such in the system. What your saying is correct if the system was just one big circle with no bends or twists....

 

The speed of the water up a vertical pipe is constant at any point in the pipe. If the speed increased as it changed direction then you'd get a vacuum because the water would be moving away faster than it was getting there, which isn't possible. The speed changes in a waterblock because the volume of water that can flow in a given time is increased because of the waterblocks increased volume in comparison to the tubing not because of the "twists and turns".
As soon as the water gets back into the pipe after the waterblock the speed is going to be the same as it is when it entered.
Its true that waterblocks etc. are going to create resistance, but this will decrease the flow rate throughout the system, not just after it passes something that causes resistance.

 

You can put the T anywhere in the loop as long as its positioned properly. The reason I suggest having it before the rad is to prevent air from getting in the rad, as it's trickiest to remove it from there.

To the other guys, there is a difference between flow and velocity. If you have 100gpm going thru 3/8" and 1/2" tubings, the flow will have much greater velocity in the 3/8" tubing. The waterblock prefers hi flow and hi velocity. The rad on the other hand prefers hi flow and low velocity.

 

And there's a difference between velocity and speed. We're not talking about a vector property here.

 

Depend on the design of the block.....The gemini block series for example doesnt require high flow...it acutally preforms better with low flow *depending on which version you get*....Now ive yet to seen a raditor that was designed for high flow personally, so i would have to agree that they like semi slow flow

 

err... no, really. This is because any items behind it (except the pump) will offer resistance and water when pushed will, as you know, take the route of least resistance. So it'll prefer to go up the T. What you'll find then is that the air in the T piece will start to act as a pressure buffer, and that keeping it sealed becomes pretty critical (to stop it from overflowing). Trust me, it's best to make it the last item before the pump inlet. Same as with a res.

 

well here's how i have mine setup right now:

pump->cpu->res->rad->pump...
I did it like this so the res would trap the air and not the rad. In my experience the rad is the hardest item to remove air from. I did haveta put more liquid in the res than I planned. I was having trouble with whirlpool/bubble effect much like the DD res. The liquid goes to the bottom of the cap leaving about 3/4" of air. The interesting thing is I can remove the cap while its running. If the rad were restricting it like you suggest, it should overflow with the cap removed. There are 2 reasons I'm able to do this. First the res is highest point in the system (only by an inch or so). Second the heater core must be less restrictive than my water block.