Sign in with
Sign up | Sign in
Your question

Science of heat sinks

Tags:
Last response: in CPUs
Share
April 28, 2009 7:31:56 AM

Just a little science project of mine. what if i made a heatsink setup. Picture this( wish i had a camera ) U shaped hollow tube. At the bottom with a hole cut out for the copper heat sink to sit inside the U tube. Now..both ends of the U lead outside of the case. One end of the u has a intake fan sucking in cooler air. The other end has a fan to blow out the heat. Juss scribble curiosites. humor me :]

More about : science heat sinks

a b à CPUs
April 28, 2009 7:39:15 AM

sounds like an idea with potential. I had a similar idea once about using guttering to port air from outside the case into the graphics card but couldn't be bothered in the end!
April 28, 2009 7:39:40 AM

there is a lot factors

contact with the cpu case is one, some are dished


air in the case is not much hotter then outside as is think antec 1200
plus its thermal dynamics, its heat removed, lowering the air temps does not increase cooling as much as you think moving heat is key

fins and heat pipes, they are hollow with water under a vaccum the water everaporates at low temps 70-80f or so
Related resources
April 28, 2009 7:42:11 AM

dragonsprayer said:
there is a lot factors

contact with the cpu case is one, some are dished


air in the case is not much hotter then outside as is think antec 1200
plus its thermal dynamics, its heat removed, lowering the air temps does not increase cooling as much as you think moving heat is key

fins and heat pipes, they are hollow with water under a vaccum the water everaporates at low temps 70-80f or so


So if heat movement is key then wouldn't this exhaust setup move it better than traditional pipe and fan setups? Considering its a smaller controlled area to move?
April 28, 2009 7:47:58 AM

Another non simple to answer question. Is a solid mass of heat conductive material( for example copper ) better for heat transfer than a fin setup like the zerotherm butterfly cpu? I know heat transfer is just the spacing out of moving heat molecules. So as far as spacing goes is it better to have solid mass or spaced fins.
a b à CPUs
April 28, 2009 9:00:52 AM

spaced fins. the heat can only be dissipated from the surface of any metal.
April 28, 2009 9:07:06 AM

Contact area is the key. Because heat transfer from metal to air is not great(because of low air density) You need very large surface area with air to move heat. I am afraid surface area of tube will be fraction of that what You have in normal heatsinks therfor it will not be nearly as effective.
a b à CPUs
April 28, 2009 10:28:53 AM

Interesting. Why not make a poor man's LN2 cooler? A four-sided funnel shaped copper box with one end closed and flat to match the CPU, the other end open and you just insert some icu cubes? Low tech, no tubes, no fins, no tubes, just a cool piece of copper drawing heat away? If I was back in 7th grade machine shop class I could probably build it for less than $10. I bet Newegg would sell a million of them in a week!
April 28, 2009 10:34:05 AM

U would need a tube to guide melted water away and constant feed of ice. I don't think home freezer would keep up freezing enough ice for it
April 28, 2009 11:24:05 AM

It is an interesting idea, but in the end it wouldn't work out. The main reason is that copper and aluminum are both much better conductors of heat than air. Air is more of an insulator than anything else. In this case, you would be relying on solely the movement of air past the contact of the CPU. That is a drastically reduced surface area for heat to be dissapated. Having a large surface area with copper for instance would allow much more air to pass by and take the heat with it than just the top contact with the CPU.
April 28, 2009 2:39:46 PM

Yea i would have the traditional copper contact conducting surface. My idea consisted more of fine tuneing the fan aspect of it. Not soley relying on air or cutting out heat dissipation but having a better working fan to circulate air through the heat sink. Well thank you all for all your replies i really do appreciate the replies. I have soo many farfetched and out of the norm ideas ..sometimes i just need to ask people and see where it goes. Cant deny the power of question.
a b à CPUs
April 28, 2009 3:51:43 PM

ainarssems said:
Contact area is the key. Because heat transfer from metal to air is not great(because of low air density) You need very large surface area with air to move heat. I am afraid surface area of tube will be fraction of that what You have in normal heatsinks therfor it will not be nearly as effective.


Right. Air actually makes a decent insulator due to its low heat conductance, which is why fiberglas insulation is mostly air. The fiberglas fibers just serve to greatly restrict convection and other air movement, so the only 2 remaining heat transfer mechanisms are conduction and radiation. And air is relatively poor at both of those.
April 28, 2009 9:10:57 PM

A better way would be to get some dryer ducting and use that to isolate air moving thru an aftermarket tower heatsink (IE Arctic Freezer or Xigmatech HDT-12xx) People have had good results using plexiglass to isolate the GPUs from the CPUs such that they each have independent air flows.
a c 87 à CPUs
April 28, 2009 9:39:20 PM

You can also use (washed) milk containers to make your ducting/funnels. Use some tape or even better hot glue to attach it to the case.
April 28, 2009 9:48:53 PM

Another random thought. sandwich theory. haha. from left to right. CPU fan blowing hot air away. Proper copper heatsink.ya know surface mastered and all that stuff. then intake fan blowing air through the finned or spaced heatsink. Yea i know water is better. yea air is less important than surface contact from cpu. Humor me.
April 28, 2009 9:55:24 PM

Depends, some HSFs are already designed that way. On some it helps, others makes no difference.
April 28, 2009 9:56:01 PM

I actually thought about this breifly once before...The problem as stated previously is efficiency. With proper heatpipes, some really cool custom heatsinks could be made ( I'm a machinist ). The problem is, there is alot going on inside those little pipes. They are either liquid filled, vacuum sealed, or pressurized with all sorts of gases. More importantly, they are not just little tubes...they consist of a fine honeycomb structure called wicking that promotes heat transfer, regardless of gravity or heats natural tendency to rise. There are places I found searching the net to purchase these, but because they are sealed under pressure, and the gases involved, they must be custom built per application. In other words, no buying a couple 6" lenghts and bending them up yourself.

I agree, that custom funneling and fan tubes are the way to go. The heatsinks themselves are just too complicated and expensive to custom make.
April 28, 2009 10:14:56 PM

Brushing up on my knowledge of heat. It rises. Or actually is less dense than cooler air so it "floats" about cooler air. Take this idea and picture this. Standing case. Regular as any. Having a fan setup on the top of the case. ( where the air would float ) For easy reduction of heat. Now imma dig a bit into pressure and such. Forgive me if its a bit of a jump. What if you were able to make some form of cooler or miniature fridgerator that aided in more cold air in a case. Would that build pressure or pressurize the hot air sitting at the top of the case? If so wouldnt that move heat or hotter air out fo the case quicker? Im Sure people have thought up some similar ideas. But the only way to do somthin new is to go to whats already been done.
April 28, 2009 11:50:27 PM

So long as the exhaust pulled enuf air out, it shouldnt cause any pressure issues.
!