Possible design?

craneage2001 · Jan 25, 2004

Now, all I have is a block for my CPU. However, for those with multiple blocks, I was wondering about the following:

Have a setup with a res. and not a t-line; then have two pumps, with one driving a small loop consisting of res->pump->rad->res and another pump driving all of the blocks. What do you think?

Korndog · Jan 25, 2004

how about getting a Y splitter, haven't a line going from the res, split off and go to two pumps and then merging back into another Y and then to the rad. haven't pumps running in series, like your example, only your max head which would be good if you had a lot of resistance in your system. running parallel, like my example, would double the flow rate instead of the head. But it all matters on what you have in your system, you may need more pressure or you may need more flow.

neilmon2 · Jan 25, 2004

best design would have the block be after the rad. that way you recieve the "freshest" water possible.

i assume you are using the same res for both loops
Have one loop cool the water and one cool the blocks is not effiecent. depending on the location of your res, you water will heat up while sitting in the res.

plus what is to stop the block line from taking the warm water instead of hte cold water.

you could use 2 res.

res1-pump1-rad-res2-pump2-blocks-res1

but still very unproductive. just have 2 loops running with 2 rads and one res. keep one loop with the CPU block only and the other loop with the other blocks.

neilmon2 · Jan 25, 2004

Korndog said:
how about getting a Y splitter, ....

you don't need a Y-splitter, just 2 lines out of hte res and 2 lines in. that way you don't have any change a pressure difference between the 2 loops.

craneage2001 · Jan 25, 2004

I don't have a system anything like this...the idea came into my head a few days ago and was just curious what you thought. I didn't think of the blocks getting warm water from the res. One could possibly minimize that problem by having the blocks get water from the bottom of the res, but overall I guess the system wouldn't function optimally. Theoretically, wouldn't the same amount of heat be removed from the system?

Muku · Jan 26, 2004

you could reduce the mixing water temp effect by having the rad loop input into the res just opposite the outlet to the block loop, and likewise with the outlet to the rad loop opposite the returning warmer water from the blocks, although the temp difference really isn't going to be enough to make a big impact.
I think this could work quite well, as you could use one pump optimised for the low resistance of the radiator, and another pump more suitable for the restrictive blocks.
Heh, this idea is just begging to use a bong though

Rokk1972 · Jan 26, 2004

Have a setup with a res. and not a t-line; then have two pumps, with one driving a small loop consisting of res->pump->rad->res and another pump driving all of the blocks. What do you think? [/B]

I have 2 different systems like this, the first is my abit system: ehiem 1250 - thermolchill 120.2 - RBX - dual bayres (shared) danner mag3 - BIX II - maze4 gpu - z-chip - dual bayres.

asus system: L30 - Thermochill 120.2 - S-RBX - Dual bayres (shared) L30 - maze4 gpu - NV59 ram block - z-chip - dual bayres.

running 2 loop systems insures that the CPU gets the max flow and lowest temps. It is not the cheapest way to watercool, but it is very sweet looking when you set it up!! Looks like you have 2 miles of tubes in the rig, but really you have about 10 feet total (5 ft per loop) Then I run both the pumps off the same PCI relay card.

craneage2001 · Jan 26, 2004

Then I wasn't so lost...that's good to know. I thought it would make a pretty sweet rig as you could have a pretty good sized radiator due to it being driven by its own pump.

neilmon2 · Jan 26, 2004

craneage2001,

i don't think the setup you are thinking of is like Rokk1972's. he has one loop dedicated to the CPU and one loop for the rest of the blocks.

correct me if i'm wrong, but it sounds like you want one loop for just your blocks and no rad. and one look for just your rad and no blocks?

Rokk's design is optimal since the CPU has a dedicated line which will allow for maximal cooling of hte CPU.

craneage2001 · Jan 26, 2004

Oops, I misread his post...(it was early)

deathBOB · Jan 26, 2004

Wait... Arent you all wrong? He just wants a system to pump the blocks and a system to pump the rads, then he will mix them in a res. right?

craneage2001 · Jan 26, 2004

deathBOB said:
Wait... Arent you all wrong? He just wants a system to pump the blocks and a system to pump the rads, then he will mix them in a res. right?

REMEMBER: This was just a thought in my head...I have no system even close to this.

However, the above quote is what I was getting at. Provided there was massive turbulence or very little turbulence in the res, I think it would perform respectably. Basically, I was thinking this:

Big giant res with two loops; one pump to the blocks and back to the res and then one pump to a large rad and back to the res.

neilmon2 · Jan 26, 2004

i would have to stand by the thought that you will get best preformance if the blocks are right after the rad.

If you want to parellel loops i would y-split after the rad then have both line leading back to the res.

i don't think there would be a way to control the water coming out ot the res to make sure you have just cool water in your block line.

just an opion though. could be wrong.

very interesting idea, its good to keep thinking new things, maybe you will inspire someone else.

craneage2001 · Jan 26, 2004

Just wondering, does anyone have any idea just how much warmer the water is after exiting the cpu block (on ave., figuring around 100w for an OC Barton).?

Possible design?

craneage2001

Member

Korndog

Member

neilmon2

Registered

neilmon2

Registered

craneage2001

Member

Muku

Member

Rokk1972

Member

craneage2001

Member

neilmon2

Registered

craneage2001

Member

deathBOB

Member

craneage2001

Member

neilmon2

Registered

craneage2001

Member

Similar threads