just a quicky

agentx

New member
im prepping for a new build i'll be doing, i want to cool my 680i and a 98gtx. I'll be running a singel parallel loop powered by my modded resorater xt. My question is are these parts good enough?

2 black ice xtreme 240 rads

water blocks:

1 FuZion cpu block

2 koolance chc-125

1 ek-fc 98gtx blocks

1 ek 680i mosfet block set (1&2)

the 240s will have pull mounted fans. one last thing i changed my res xt's pump to this one koolance pmp-400 for extra flow rate thats needed with all the blocks and splitters, are thses parts sufficiant? :)
 
I know im not a veteran water cooler, but im pretty sure that one pump wont be enough. if you are having 6/7 rads and 6 waterblocks it will just be far too restrictive... so i would be inclined to go for 2 loops. but thats just me, wait for some more expert advice on whether it will work or not :)
 
well i do believe (and i could be wrong) but running a parallel loop is less restrictive than series loops. however i could always just look for a pump with more head pressure as i havent modded the xt yet.
 
yes, i suppose it will be less restrictive because it's in parallel, yes it prob could work then, but would be more efficient with 2 loops i reckon. again, i could also be wrong, but just thought i would let you know what i thought :)
 
i appriciate ur input the reason i assume this is the optimum setup is that the rads combined boast about 3kw of heat dissapation. since im running a parallel all the components are cooled effiecently and equaly. so i figured that i have more than enough cooling power its just a matter of making sure (to be safe and not waste money u know)
 
Im no watercooling guru but am i right in thinking your gonna have a thermaltake p500 pushing water through 6 rads and the cpu and gpu waterblocks and y-splitters ?
 
i've actualy changed the pump as i want one with a 3-pin connect not a molex. and theres 7 blocks plus rads and splitters
 
First of all, let me preface this with, I am a total water cooling noob.

I am just now building my first H20 cooling solution -- but my presumption when I designed my loop was that any split will drop pressure and flow across each split equally but only if the resistance in each split is exactly equal. Otherwise, flow will increase in each channel inversely to the resistance in that channel.

What I mean is, if you split your pumps output into two loops (A and B), if the resistance in each loop is identical then water will flow equally through each loop.

However, if loop A has twice the resistance of loop B, then only HALF as much water will flow through loop A as goes through loop B.

The water flow will follow the path of least resistance first, only diverting to more resistive paths when the constricted volume through that best path equals the resistance of the more resistive path(s). :confused:

It seems to me, that splitting your loop will create variable flow rates through each loop in your system and make troubleshooting problems difficult at best.

I have no idea if i am right or not; It just seems like the old "Keep It Simple Stupid" rule applies equally well to fluid dynamics as computer programming... = : )

-EazzyTarget
 
u hit the nail right on the head but me being the idiot i am didnt think the pressure drop wud be that great and ran a test loop. OH GOD. my cpu/mosfet had no flow, southbridge had very slow flow and gfx/northbridge had the most so now im looking for a high pressure pump.
 
I am not sure you are going to be alble to acheive your goals by increasing pressure. Yes, you will get more flow from the out port on your pump, but that flow will still split across the six blocks in your loop in the exact same ratios.

And it is ratios, (wow, my algebra teacher was right!) if your CPU is so restrictive that it really isn't getting ANY flow right now, I seriously doubt you can increase outflow enough to fix the problem.

But let us know how it goes!

-EazzyTarget

Ps. For what it's worth, what I did was serialize my radiators in my loop:

Res - Pump - CPU - Rad - T-Splitter - GPU (x2) - T-Splitter - Big Rad

It's extra hose and coolant to make those runs out to the radiators and back to the mobo, but hoses are cheap compared to CPU/GPU's :D
 
Again I'm pretty new to WC but I'm not to science being a BSc and Science teacher. I do believe the passives won't add much of a problem as I have looked at them myself. Eazzytarget has nailed it properly but I think the problem can be overcome with some setup fiddling to maximise flow and use the restrictions to collect backpressure and help flow problems.
 
i know that if i get enough pressure he flow will increse the only problem is finding a pump with enough head im hoping 2m of head is enough (the pump is rated at 500l/h) but havent ordered it yet as my mobo is on the fritz :( . my cpu block is the d-tech fuzion which is pretty flow friendly. i have the actual loop set up and the cpu gets more flow now but still not as much as i'd like and the mosfet blocks still dont get any. i just have to figure out how to streamline the kinks and i'll be good to go.
 
My Laing DDC 18W has a head height of 4.7m so surely your pump will beat that :O that thing is TINY.

You could try the split immediately after the pump then combine for the CPU and entrance to res that would give backpressure and help flow.
 
What is the actual restriction stopping you using a Molex?

Just not enough or reach or you just don't want to?

We might be able to DIY a solution for you so you can get a more powerful pump. 2m is not alot especially for your kit so your gonna have to be very canny to get it spot on and maybe buy loads of tubing to get the right flow as once cut you probably won't be able to use half of the tubing again.

I'm personally getting fed up with fan controllers so am wondering if I should DIY the fan connections so the controller will be forced to deal with what I want.
 
the reason i cant use a molex is b/c the pump has to plug into my reserator xt. my loop isnt exactly restrictive persay the splitters just cut too much pressure out of the loop. heres some pics of my loop. theres actually 4 splitters but they are pared in 2 groups of 2

Picture001.jpg


Picture005.jpg


Picture002.jpg
 
Thats mental piping mate, you must have hands of a midget to be that precise. :worship: Well done

I'm sorry i cant post help im trying to figure out how you have routed it before saying anything.
 
actually i have rather large hands but thanks :) . if u like a can take more pics for you just post bak and let my know. the splitter you see in front is the one that gives the cold liquid to the blocks the one behind it is the return. heres another pic to help show u the tubing layout.

Picture006.jpg


1 goes to the mosfet blocks

2 goes to the cpu

3 goes to the gfx

4 goes to the nb

5 goes to the sb
 
Nope thats much better. I thought you would set it up like that as its the logical way but could see all the connections well enough.

Good news

I think I know where your problem is and a possible solution.

If I'm correct your problem was created by simple ratio mistake but must be solved by fluid dynamics (which BTW if anyone has a program for I would love as I no longer have access to the one I used to use).

Bad news

I cant be sure as I can't see all you components in one shot for relative height and such.

It might cost if the first option does not resolve it.
 
ok first how dow i fix the ratio problem if thats actually the problem. as for the second, the following is a list of the blocks from lowest to highest.

sb block

gfx block

nb block

cpu block

then the mosfet blocks. the first block is above the cpu but the second (which is in series with the first) is parallel to the cpu, positioned just to the rite of the i/o ports running vertically / up and down the board.

edit: the reserator is the highest point
 
Back
Top