I need to have a different tubing size for one run in my loop

clayton006

[H]ard|Gawd
Joined
Jan 4, 2005
Messages
1,089
I run 1/2 ID tubing to all parts of my Loop. My loop order is as follows:
Res -> Dual pump top RAD 1 -> Heatkiller XL Mutli Block (4 GPUs) -> RAD 2 -> CPU -> Res

I'm adding a motherboard chipset block and a vreg block. The chipset block is the problem. The tubing just barely fits between two titans. I would like to go with something smaller for the chipset blocks so I can use quick disconnects so that I can easily get the motherboard out. I won't be able to go with quick disconnects if I stick with 1/2 ID.

Since I'm running two pumps I think I'll be able to overcome the restriction problem. Thoughts? I cannot return the chipset block.
 
I haven't started messing with watercooling yet, still waiting for parts to come in, but I'm thinking you could run the chipset and vreg block in parallel (via a 'y' or 'T') with one or both having smaller tubing and still maintain a similar flow rate. The determining factor for the tubing size of each being how restrictive the respective blocks are: if it's more restrictive, it should get the larger tubing.

As far as running them in series, I couldn't tell you. Too many factors, and I'm new to this myself. Good luck. ;)
 
Parallel is a bad idea. If you run it in parallel you run the risk of having the majority of your flow going through the chipset / vreg block instead of whatever else it is in parallel with. This will give your motherboard great cooling, but screw over your actual important parts. Parallel is typically only recommended for like blocks, for example putting multiple identical GPU blocks in parallel. With like blocks you will get more or less equal flow going to each component. If you have ever taken an electrical engineering course and studied Ohm's law, water loops behave in exactly the same way as the flow of electricity. Restriction = Resistance, Flow = Current, Pressure = Voltage.

You will be just fine using smaller diameter tubing. It will not add very much restriction to the loop. I wouldn't worry about it.
 
3/8" ID tubing has been shown to not have more noticeable restriction than 1/2" ID at the typical flow rates computer systems use. You would have no issues whatsoever using 3/8" ID tubing for that. The QDCs themselves will probably add more restriction than the tubing.
 
I use 3/8" tubing in my custom loops for the very reasons you mention. I don't generally recommend going lower than that unless you just have a single block or 2 low restriction blocks in your case. Also instead of parallel, you can do separate loops or put both pumps in serial on the same loop. I have done both and generally found either does the job fine depending on your pump and overall restriction. Here is how some of my loops look:

Double Loop
Res->Pump->Rad->CPU->RAM
Res->Pump->Rad->GPU->GPU

Single Large Loop
Res->Pump->Pump->Rad->CPU->GPU->GPU

My buddy's system I planned out
Res->Pump->CPU->CPU
Res->Pump->Pump->GPU->GPU->GPU

I generally like the double loop better myself with CPU and GPUs on different loops, but that is just my preference. I also happen to like double pumps as well, although that can be overkill depending on the setup. The double pump just gives a bit more overall head to the loop as well as some redundancy in case one pump fails. Although, personally, I have never had a pump fail.

Now if you want to get really freaky I have seem someone do this before:

Res->Rad->Pump->CPU
---------------->Pump->GPU
---------------------------->GPU
---------------------------->GPU


Where the pumps are in parallel, and the GPUs with the same block are in parallel, and they all come from the same Resevoir and use the same Radiator. I wasn't personally a fan of that design and I think they had some on and off issues with flow. Although their temps most of the time looked normal.
 
Last edited:
If you have ever taken an electrical engineering course and studied Ohm's law, water loops behave in exactly the same way as the flow of electricity. Restriction = Resistance, Flow = Current, Pressure = Voltage.

I have, and understood that, but I wasn't really thinking it through I guess. I was thinking if you use smaller tubing on the block that has less restriction, then less water would flow through it. But I guess if the restriction is lower on one block than the other (and different sized tubing has minimal difference in restriction), it would only (or mostly) increase the rate of flow through the block with less resistance, rather than redirect it through the other block.
 
I have, and understood that, but I wasn't really thinking it through I guess. I was thinking if you use smaller tubing on the block that has less restriction, then less water would flow through it. But I guess if the restriction is lower on one block than the other (and different sized tubing has minimal difference in restriction), it would only (or mostly) increase the rate of flow through the block with less resistance, rather than redirect it through the other block.

Generally larger tubing allows more flow because of the "bandwidth" capability of the tube allowing more area for the water to flow through. Restriction comes into play where that water's pathway is suppressed by having to flow through a smaller opening, flow over or around obstacles, or go through a number of twists and turns. There comes a point of limited gain increasing tube size when the water is still being forced through a small opening. In that case it doesn't make much sense to have a huge tube when it goes through so many smaller openings. Generally your tube size is dependent on the fittings that come with the blocks you are using. You should align your tube and fittings to the largest inner threaded block in your loop. IE, if your largest block only uses 3/8" inside threads, using a 1/2" tube size doesn't really make sense. Using 3/8" will better align with the overall flow, providing hopefully with the least amount of restrictions or obstacles for the water to deal with.
 
Back
Top