3/8" to 1/4" tubing, choices...

[pyroman1]

Hello,

I am hoping someone can give me some advice on going from 3/8" tubing to 1/4" tubing, as some of my blocks only accept 1/4".

Bear with me here I'm going to do some brain-storming. If I am wrong about any of this please let me know.

Alledgedly, going from 3/8" to split 1/4" will reduce flow. Makes sense as the cross sectional area goes from 9*pi/256, roughly .1104", to 2*pi/64, roughly .0982". This means you lose about .0122" in cross sectional area which makes it harder for the water to flow.

What if two splitters were used? Since only three 1/4" tubes are desired this would be the configuration:
Use a Y or F-type splitter to split the single 3/8" into two 3/8". Immediately split one of the new 3/8" into two 1/4" while using a 3/8" to 1/4" adapter on the other one. This changes the total cross sectional area of the three 1/4" tubes to 3*pi/64, roughly .1473". Now there should be no reduction in flow through the system.

If this looks right and makes sense then I would do the following:
Res->Rad->SB->NB->CPU->MOSFET1->MOSFET2->Splitters
Y1->GPU->RAM1->Y-splitter to 3/8"->Res
Y2->RAM2->Y-splitter to 3/8"->Res (allows for 2nd future GPU cooling)
Y3->HDD1->HDD2->Y-splitter to 3/8"->Res

Of course if someone makes a 1-to-3 splitter already then that would be the better choice.

Does this make sense? Could it work?

I also found this article which states:

Presumably, the hose reducers used in the V10 loop had enough of an impingement effect to significantly increase the cooling performance of that block. As the water is forced from the 1/2" tubing into the 3/8" tubing it increases in both velocity and turbulence (like using your thumb to spray a jet of water from a garden hose). We have seen the principle of impingement used in the design of several water blocks of late, including the RBX used in this review, the effect on cooling is clear. When we switch the V10 for the V13 with its large 1/2" fittings we lose that impingement and the resultant velocity/turbulence. I would have thought that the increased flow would have been more important, but the numbers don't lie.

Now I realize this article is rather old, but is it really going to be worth it for me to have a three-way splitter?

[Brians256]

Splitters are hard to balance. This means that components may get the lions share of the flow and other components have stagnant water.

[billbartuska]

How will you control how much water goes through each component...you won't, the water will follow the path of least resistance.

[Long Haired Git]

But not all components need the same amount of cooling, so as long as you align you will be fine (I'm a poet and I know it).
In your example, however, you have a GPU block in one the three branches, and then a bunch of stuff that doesn't need to be cooled much. I'd much rather you have the MOSFETS and the SB and NB in the 1/4" loops than in the main arm. It will make it easier to plumb with the thinner tubing too.

Lastly, resistance is approx equal to the square of the velocity. So, if you have multiple parallel paths then the resistance of your blocks will be greatly reduced as the flow rate through them is greatly reduced. So, if you manage to equally balance all three branches, each will get 1/3rd of the flow rate of the main arm. For a CPU or GPU or any other large wattage temperature intollerant device, this would be a bad thing. For HDDs, mosfets, ram, SB's and NB's however, its quite sensible.
BladeRunner's silent rig comes to mind.

Me, I'd have:

Res->Rad->CPU->GPU->Manifold1_in
Manifold1_out->SB->RAM1->MOSFET1->Manifold2_in->Manifold2_out->Res
Manifold1_out->NB->RAM2->MOSFET2->Manifold2_in->Manifold2_out->Res
Manifold1_out->HDD1->HDD2->throttle->Manifold2_in->Manifold2_out->Res

Where a throttle is like a thumb-screw impedement to the flow to enable the flow to be reduced through this branch to the point where all components get the cooling they require.

If, as you say, you do get a 2nd GPU block, then you may look at more paths etc as required.

To reiterate: Having more flow to one branch than another is not a bad thing if the low-flow branch has relatively low wattage and temperature tolerant devices on it.

[Edge]

(I'm a poet and I know it).

Thanks git, I needed a laugh.

The points on balancing are good, but I already knew that from your other posts. Being a dumb software guy, I like to keep the engineering simple and try to avoid multiple paths in a single loop (though you get a plethora of pumps).

edge

[Brians256]

I was most concerned about anything producing more than 15W being after a splitter. The GPU is such a beast.

Basically, if you want to split the flow, split it after the CPU and GPU(s) and then it should be fine. Git has a pretty good flow regime suggested.

So, split it, and monitor temps. If it doesn't work, re-do the layout.

P.S. Glad to see another SW guy on the boards. Software rules!