Water speed logic

[Ghost]

This is a bit of logic question.

​Okay imagine this setup

As watercooling theory works water should pass through a block as fast as it can and go as slowly as it can through a radiator. This narrowing down of the tube would increase the pressure inside the block as the tubing would narrow down as the water would have to flow much faster for the same flow rate through the thinner tubes.

In theory you could go pump =(thick)=> block -(thin)-> rad =(thick)=> block -(thin)-> rad =(thick)=> res LOOP

This would also mean the water would slow down in your radiators which is what you want, because you want the water to remain as long as it can in the radiator.

What do you think, is there a reason no-one does this or has no-one thought of it before?

[MoeBrowne]

Assuming your pump can maintain the same flow rate at the increased pressure you would get the effect you're after.

Although it is likely that this is already happening in your radiator, this is why the input to the radiator is (usually) split in to multiple channels that flow in parallel (they of course also spread the heat).

You're probably thinking slowing it down even more with narrowed tubes would be better?

It depends on the difference between ambient (or rather the temperature of the air going into your radiator) and the temperature of the coolant flowing into the radiator (ΔT).

Newtons Law of Cooling describes: As the temperature of a body increases, the amount of energy dissipated from the body to its surrounds is indirectly proportional to the difference between the temperatures. (a graph showing the temperature of an object as it cools over time gives a similar same shape to that of the activity of a decaying isotope)

So unless your radiator or water block are highly inefficient you're going to see little difference unless ΔT is large.

[Ghost]

.

The longer the water stays in the radiator the more time it has to be cooled down. The faster the "cold" water is replaced in a waterblock then the better the cooling for the component will be.

[hiyayhi]

I suspect that it is the same kinda thing as if you tried putting your rads between blocks to try and get better temps (it doesn't do anything as the temp in the liquid is within point one of a degree margin in the loop). Wouldn't it just equalize the pressure around the whole loop, slowing all of your liquid down?

[Ghost]

I suspect that it is the same kinda thing as if you tried putting your rads between blocks to try and get better temps (it doesn't do anything as the temp in the liquid is within point one of a degree margin in the loop). Wouldn't it just equalize the pressure around the whole loop' date=' slowing all of your liquid down?[/quote']

Flow rate inside your loop must remain constant so no. Water cannot be compressed like air so the liquid must be moving at the GPH whatever it passes through and constantly throughout the whole loop until it reaches a res where it can compress the air if it must.

[Mentalguy]

wouldn't make much of a difference if at all.

[Whaler_99]

Not sure about this... when water goes into a block, there is already some restriction in the block itself as the water follows the channels. Wouldn't this be doing the same thing already that you are trying to achieve?

[EYCEthebest003]

Rather than restricting flow after the cpu block, you might try this. Instead of having, say, one double rad, have two single rads in their place (same for a quad rad and two doubles). If you were to then have the the tubing split off after the water block with a T fitting, effectively reducing the flow rate by half after the water block, you could have the two tubes go through the radiators and then return into your reservoir. This could yield the desired effect.

[Ghost]

Rather than restricting flow after the cpu block, you might try this. Instead of having, say, one double rad, have two single rads in their place (same for a quad rad and two doubles). If you were to then have the the tubing split off after the water block with a T fitting, effectively reducing the flow rate by half after the water block, you could have the two tubes go through the radiators and then return into your reservoir. This could yield the desired effect.

But it wouldn't look good. Which is half of what watercooling is.

[Ghost]

Not sure about this... when water goes into a block, there is already some restriction in the block itself as the water follows the channels. Wouldn't this be doing the same thing already that you are trying to achieve?

I'm not sure. Thats why I asked. I'm sure it would change something.

[Whaler_99]

Not sure about this... when water goes into a block, there is already some restriction in the block itself as the water follows the channels. Wouldn't this be doing the same thing already that you are trying to achieve?

Guess going to have to spend to time at Skinnee :)

[Ghost]

Not sure about this... when water goes into a block, there is already some restriction in the block itself as the water follows the channels. Wouldn't this be doing the same thing already that you are trying to achieve?

Someone at OC3D game me very clear and long responses so I'm fine :D

[Whaler_99]

Not sure about this... when water goes into a block, there is already some restriction in the block itself as the water follows the channels. Wouldn't this be doing the same thing already that you are trying to achieve?

Post or link please? :)

[Ghost]

Not sure about this... when water goes into a block, there is already some restriction in the block itself as the water follows the channels. Wouldn't this be doing the same thing already that you are trying to achieve?

http://forum.overclock3d.net/showthread.php?t=51440

[Salmiakbal]

Rather than restricting flow after the cpu block, you might try this. Instead of having, say, one double rad, have two single rads in their place (same for a quad rad and two doubles). If you were to then have the the tubing split off after the water block with a T fitting, effectively reducing the flow rate by half after the water block, you could have the two tubes go through the radiators and then return into your reservoir. This could yield the desired effect.

Nah this wouldn't make a difference. While the water would go slower through the single ones, the surface area will be less as well. If it were to go through a double one it would go twice as fast, but the surface area would be doubled as well. It wouldn't make a difference.

I'm busy with thermodynamics now and I haven't had fluids yet. It could prove useful :D The guy at OC3D seems like a pretty intelligent guy though, very good explanation.

[Whaler_99]

One thing about getting into water cooling... after a while you start to ask questions like this and you really start to see the sheer amount of science and data that goes into all this and designing, etc. Makes you think about some of the people companies like swiftech, aquatunning, XSPC, etc. must have working for them.

Good discussion...

[EYCEthebest003]

Rather than restricting flow after the cpu block, you might try this. Instead of having, say, one double rad, have two single rads in their place (same for a quad rad and two doubles). If you were to then have the the tubing split off after the water block with a T fitting, effectively reducing the flow rate by half after the water block, you could have the two tubes go through the radiators and then return into your reservoir. This could yield the desired effect.

@ Salmiakbal the point isn't to increase surface area, the point is to move the water through the radiators slower.

@Ghost I'm sure there is a way to do it and make it look good, there almost always is. as long as you have short runs and don't have tubes crossing one another, it should look pretty good.

[EYCEthebest003]

Not sure about this... when water goes into a block, there is already some restriction in the block itself as the water follows the channels. Wouldn't this be doing the same thing already that you are trying to achieve?

What did we learn? Throw more radiators at the problem! Also he has a point. The water going through your loop is going so fast already, slowing it down shouldn't really make a difference.