# Gelid's Tries Rearranging Heatpipes in New GX-7 Cooler for Gamers

5

### RGCook

OK, the principles of thermodynamics are pretty clear here folks. Pardon me, I am a chemical engineer by day to pay my computer building ways at night.

So packing more heat transfer area as the end game. So why not use more, smaller pipes to wisk away those pesky thermal gremlins. Give them more lanes to travel, more area to depart. Kind of like more HOV lanes.

And dont forget you water cooling folks, turbulence is your friend. Laminar flow will prevail at typical rates that these cooler pumps use so start thinking ways to reduce viscosity and get that fluid working for you.

Idea - get the air moving IN the pipes as well as the heat through them. Duct the fan on the backside into a plenum that reroutes the air down through the heat pipes for maximal thermal tranfer. Might even quell the SPL of the fragistat.

### kiaghi7

Many smaller heat pipes used to be the way of high end coolers, until they realized that bigger is indeed better.

Because more, yet smaller, heat pipes is actually less efficient than fewer and larger heat pipes.

Think of it like this, three half inch inside diameter (ID) pipes still don't equal the volume of a single 1 inch ID pipe:

Just comparing the cross-sectional area of them:

.5" pipe:  .5^2 * 3.14 = 0.785 in^2

Three of those pipes = 2.355 in^2

1" pipe:  1^2 * 3.14 = 3.14 in^2

It takes four times as may half inch pipes to equal the same throughput and volume as a 1 inch pipe, and air, like any fluid, only flows but so efficiently without being pressurized to go faster.

### RGCook

OK kiaghi7,

First, you are using the equation for the area of a circle but you really should be using circumference times length to calculate the effective surface area available to heat transfer. Another mistake in your example to me is that you used the diameter in your calculation, not radius. That is moot. Hear me out now;

The effective heat transfer area is not cross section but surface area of the tubes. Therefore, you will want to calculate the surface area of three small pipes that will fit inside one large pipe.

It can be shown that for three small pipes of radius = 1, the smallest large diameter circule that will contain them is 2.154. Units are not important.

Let's assume that the length of the pipes is unit for the sake of convenience. Therefore, the surface are of the large pipe, unit long is 2 x pi x (2.154/2) = 6.77 sq units

The surface area of the three smaller pipes that would fit inside would be 3 x 2 x pi x (1/2) = 9.42 sq units.

We can see that three smaller tubes affords an areal transfer area that is (9.42 - 6.77) /6.77 = 39% higher.

But that's not the whole story. The packing density of the three small pipes is higher than the larger pipe. If you draw a circle with three small ones inside it, you will see that you can actually overlap the large circles to make the small ones pack together. I won't do the math but you effectively pick up another 30% area by doing this. But I will leave it out under the pretense that it would be more effective to not tightly pack the smaller tubes but rather leave a gap bewteen them to optimize the conduction away from the heat source.

I know this stuff because in my industry, when you buy a heat exchanger to cool something, generally, what happens is the vendor will quote you a big exchanger that contains a million small tubes inside it. That way, you maximize your transfer area and keep the exchanger size down. But in cases wihere the fluid could plug or scale up, having all those small tubes is not a good thing when it comes time to clean them out.

I can send you a diagram that helps explain this better. In the end, you are wrong but your challenge is accepted and appreciated.

### iceveiled

Interesting concept, looking forward to seeing some benchmarks on this. I'm torn between staying with air cooling or trying one of those all-in-one water CPU cooling kits that are all the rage now.

### ducis

Thank heavens its designed with gamers espeacially in mind. To bad its no 'Xtreem' though, would prolly buy like seven if it was trolol.

sounds cool though, I love how heatsink design involves the most low-tech improvements that can actually improve performance by like 20% (via improved OCing potiential); while chip manufactures are busting their butts with weird crystals and stuff for an extra 5%.