What are the downsides?

If water cooling my CPU, GPU, and Chipset all in one system. I thought about different ways to set the tubing up, and I finially came up with the setup that I drew in the linked picture. (pic is worth 1000 words)

Here is my Q: There will be a lot more restriction in the GPU, chipset line, and it will also be a lot longer. I'm not a physics major, but won't that cause some unwanted effects? Will the CPU line's water slow down to match the GPU/ chipset's water when they combine again? Will the water just go at different speeds?

I don't know how this will work out. I know that on a airplanes wing, the air is split before the wing, and combines after the wing. This is cond of like my setup. Just like my water lines being longer, the top of a air plane's wing is longer and curved. The air that goes over the longer path must meet with the same air that it split from, and thus (due to pysics) travels faster. Faster air moving over longer path = same as air moving on bottom. This is how planes fly!

But how will my water cooling be? I would think if would just be the same as with an air plane. But that means that the water going to the GPU, chipset will travel MUCH faster to meet up with the cpu water. If that was true, then why aren't people exploiting this, and making thier CPU line VERY long and another line very short. After all, then the CPU line's water would travel VERY fast to catch up.

I don't know what to think and this makes me hate physics more.

Thanks to anyone that A) can understand what I am even talking about. B)can actually provide me with a right answer.

The picture's not working....

The line that has more restriction will see less flow. Due to there being less flow, the velocity will be slower. If you want to prove this to yourself, take a Y connector, and put a ball valve on 1 line after the Y. As you close the ball valve, you will see more flow, and higher velocity, through the unrestricted line.

Hope this helps. Also, I cannot see your pic either.

pic should work now.

Pic works now but I'm afraid the idea doesn't.

I am no math expert and have no education in fluid dynamics other than what I have learned here with water cooling, but-

I think the flaw is comparing aircraft wings to a closed loop system. Completely different "math type stuff" must apply.

In water cooling the fluid will tend to take the path of LEAST resistance far more than the one with more resistance, so you will get little flow in the high-resistance line. In your pic that is the longer line with two blocks unless you have an extremely flow resistant cpu block....A Cascade with its die impingement jets might do it but I doubt it.

The upshot is this: unless you have perfect balance between the flow restriction in both lines, only one of them will get proper flow.

I'm no expert on water flow science, but would something like this work?

My guess is the major problem would be restriction from the gpu and nb. But how much?

agree,, the flow to the gpu chipset would suffer, just run them all together inline ....
YES<,, that picture above is the best way,,,

Be worth trying to see what happens, I kinda like the Idea. The GPU/NB really doesnt need extreme flow, and using 1 line in serial would restrict flow to the most important part the CPU.

I dont kow, but it would be interesting to see what happens.

well........ I don't know if I have enough tubing..wait, yes I do.

Doing this will waste precios destilled water, but it's only $1.19 for a gallon.

I will run the sytem using Y's and seperate lines. Then I will put them in a series and record temps.

Sound good? I'll have to buy a temp monitor of some kind to get NB and GPU temps, but I found one that goes in front of the case, has two temp probes, and is $20. No biggy for the INPRESSIVE SCIENTIFIC DISCOVERY we shall make!!!

Hmm... I'm envisioning an experiment to play with the paralell setup originally described...

Basically, you might start with a setup with fairly unrestrictive GPU and NB blocks, so that normally the flow to both branches of the cooling circuit would be close to even. Then you could put some kind of throttling valve on the GPU/NB branch, to optimize the flow each component gets. I'd imagine that the NB really doesn't need much cooling, and the GPU doesn't need as much as the CPU (with some exceptions *cough* geforcefx *cough*).

Something else to tweak...

I ain't buying no geforce fx. I've gone from Nvidia Guru to Nvidia hater. Radeon all the way! lol

it's kinda like electricity and how current travels path of least resistance. Same idea. My phys professor used water as an example to try to explain the idea.