Question about heatpipe efficiency

[suszterpatt]

Looking at the pics of the new Zalman VF900 cooler in the cooling forum, something struck me. All heatpipe coolers have the fins positioned so that their plane is perpendicular to the axis of the heatpipe. Leaving some room for air to flow between the fins, a cross section of the cooler looks like this:

http://img62.imageshack.us/img62/8780/hp16wz.jpg

(the blue circle is the heatpipe, the grey stuff is the body of the fin)

Note that because of the spacing and thinness of the fins, the heatpipe makes contact with a fin only in a small % of its length. In other words, picking a random spot on the heatpipe, you're much more likely to get an "empty" cross section than a "full" one.


If I'm not mistaken, heatpipes work by the principle that heat doesn't spread as fast inside a metal then hot steam can spread through air (or vacuum). In light of that, it's obvious that the outer areas of a "classic" cooler's fins don't get as hot, so they don't help as much at cooling as the parts that are closer to the heatpipe.


This leads me to my main question: would a cooler be more efficient if the heatpipe was surrounded by a thin metal coat, and the fins were sticking out radially, running down along the entire length of the heatpipe? Basically, if the cross section of the cooler would look roughly like this no matter where you cut it:

http://img58.imageshack.us/img58/7659/hp20ct.jpg



It seems to me that this would be more efficient, because you can think of it as taking the outer, less useful areas of the classic fins and putting them closer to the heatpipe. In this new model, the heatpipe could transfer heat to the fins in its entire length instead of only a few spots. Would this make up for the smaller radius of the fins (i.e. the reduced cooling capability of a single cross-section)?




Am I even making sense?

 

[CycloWizard]

Your design may indeed be more efficient. However, such a setup would be much more difficult to manufacture. In addition, vertical fins have the additional benefit of being nearly ideal in encouraging natural convection, which is the actual mechanism that removes the heat from the fin and transfers it to the surrounding air. So, while your design might give very slightly better performance, the differences would probably be negligible, so companies build what is easier - the vertical fin design.

To understand why, one must understand generally how these things work. First, heat is transferred from the fluid within the pipe to the pipe itself via convection. The pipe transfers this energy to the fins via conduction. Generally, convection is much faster than conduction, unless the conducting media are highly thermally conductive metals, such as copper or silver (or, to a lesser extent, aluminum). Once the energy reaches the fin, it is conducted within the fin and leaves the fin via natural convection. Natural convection can be a very powerful heat transfer mechanism if the design of the surface is correct, and it just so happens that the fin is a pretty optimal design to encourage natural convection. Natural convection occurs when a thermal gradient exists such that density gradients also exist in the heat transfer medium (in this case, air). The density difference causes the air to flow, causing recirculation and increasing heat transfer rates. By having vertical fins, the air can flow very easily, increasing the rate of heat transfer.

 

[suszterpatt]

So in laymen's terms, vertical fins are better because the air that they transferred their heat to can easily rise, allowing cooler air to take its place, right?

My design would fit this criteria if the heatpipe was placed vertically then, no?



Btw, you do have a point with the manufacturing costs. Guess we're not going to see this anytime soon.

 

[CycloWizard]

Originally posted by: suszterpatt
So in laymen's terms, vertical fins are better because the air that they transferred their heat to can easily rise, allowing cooler air to take its place, right?

My design would fit this criteria if the heatpipe was placed vertically then, no?



Btw, you do have a point with the manufacturing costs. Guess we're not going to see this anytime soon.

Basically, the vertical fins compensate for their other shortcomings (those that you mentioned) by facilitating airflow. For a vertical heatpipe, your design would be more efficient, though by how much I can't really say, as it is very dependent on the length of the pipe and the temperatures we're talking about.

I do know for a fact that radial heat transfer fins are used in some places. Typically, these applications are very specialized and can justify the high design and manufacturing costs. Design costs are higher because the engineers have to do a lot more math to solve the more complex radial heat transfer equations than the relatively simple cartesian equations.

 

[Vee]

The heat pipe design itself is born from manufacturing considerations.
If you only wanted the most efficient cooler design, it would be some very complex fractal design. Where 2-3 very thick fins emerging directly from the base, progressively branches into thinner fins, like a tree.

 

[icarus4586]

Your design would increase the rate of transfer between the heatpipe and the fins. However, that's not what slows down cooling, for the most part. The reason heatsinks have fans is to force convection, because convection is what's slow when you have as much energy to dissipate as CPUs do. Vertical placement allows for more fins, more surface area, and some benefit from natural convection. That's why they're used.

 

[Vegitto]

Originally posted by: Vee
The heat pipe design itself is born from manufacturing considerations.
If you only wanted the most efficient cooler design, it would be some very complex fractal design. Where 2-3 very thick fins emerging directly from the base, progressively branches into thinner fins, like a tree.

You mean the Pythagoras fractal, right? I suppose that could work, if it weren't for the sheer awesomeness of that fractal causing the cooler design to break instantly .

I think it could work, but who in their right minds would spend 150-300 bucks for an AIR-cooler? With OCZ's new innovation, you can get phase-change for that kind of money, damnit!

 

[suszterpatt]

Originally posted by: icarus4586
Your design would increase the rate of transfer between the heatpipe and the fins. However, that's not what slows down cooling, for the most part. The reason heatsinks have fans is to force convection, because convection is what's slow when you have as much energy to dissipate as CPUs do. Vertical placement allows for more fins, more surface area, and some benefit from natural convection. That's why they're used.

Obviously, having an ample supply of air is crucial. I was kind of assuming that it's a given.


Let me refine the question: between a classic, horizontal pipe cooler, and a radial, vertical pipe cooler (for both to get the max out of natural convection) of equal dissipation surfaces, which one would cool better?


Though I guess that this still is pretty vague, since (if I understand correctly) natural convection's bonus depends on the "area" over which the fins extend.

 

[CycloWizard]

Originally posted by: suszterpatt
Obviously, having an ample supply of air is crucial. I was kind of assuming that it's a given.


Let me refine the question: between a classic, horizontal pipe cooler, and a radial, vertical pipe cooler (for both to get the max out of natural convection) of equal dissipation surfaces, which one would cool better?


Though I guess that this still is pretty vague, since (if I understand correctly) natural convection's bonus depends on the "area" over which the fins extend.

Like I said, it depends on how everything is arranged and constructed. All things being equal, the radial fins would probably give you a very slight edge (pulling a number out of my arse, in the neighborhood of 2-3% higher efficiency).