Heatsink: Thermal Resistance Vs Length

[vinay]

This relationship is not linear. I was plugging away some numbers at the Thermalloy site. A heat sink rated at 0.9 degrees/watt/3" is only 0.44degrees at 12 inches (and 0.62 at 6" and 0.5 at 9")

My thermodynamics is a bit rusty, but I thought that the longer the heatsink, longter the length between the fins, thus getting more of the chimney effect.

So, what are peoples experiences in this matter. I have a couple of 0.75 degrees/3" rated heatsinks that are 12" long. Should I cut them down to 6" and use them? I was leaning to stereo chassis (smaller footprint), but monos are OK.

Thanks,

Vinay

[roddyama]

Hi Vinay,

The longer heatsink has a couple things working against it WRT convection cooling. As the air takes heat away from the fins of the heat sink as it moves upward, the heated air has a smaller temperature differential with the heatsink and therefore become less effective as the heatsink become longer. Adding to this ineffectiveness is the fact that the heated air is less dense (fewer molecules / volume) making it even more difficult for the air to take heat away from the heatsink.

This being the case, you would gain nothing by cutting them to 6” and stacking them one atop the other. The decision to cut them will depend on how much cooling you need. Are you sure that the 12” length is too long for your cooling needs?

Rodd Yamas***a

[ThingyNess]

In most cases, I believe those numbers are also considered with a single point source radiating heat in the center of the sink.

If you had 30 heat sources all spread across the surface of the sink, the 0jc would be a lot better.

Example: if you have 4 transistors spaced equally along the sink and it's 12" long, the effective thermal resistance will indeed be 1/4 that of the 3" long piece.

( a little worse actually, depending on the orientation + power dissipated due to temperature differential between top + bottom of the sink, but not much worse)

[UrSv]

Vinay,

Excuse me if I am assuming something here but it seems that your numbers suggest that you use the width rather than the length. The 0.9 K/W/3" is for a profile with a height (which then as it is an extrusion is called length) and the numbers when increasing the length/height are not linear. The profile cross-section (image of the surface where it was cut) stays the same in all dimensions. So to clarify even more the same profile with the fins oriented vertically in a chassis with a HEIGHT of 3" may end up at 0.5 K/W and the same chassi depth but 6" height may end up at 0.35 K/W.

/UrSv

[vpharris]

Those replies are all good points.

At the Aavid site, click on "technical info" at the top of the page, and you can find your way to this page with info about temperature and length correction factors.

http://www.aavidthermalloy.com/technical/correct.shtml

In part they say "Because the heat load is assumed to be at a point rather than uniformly distributed, thermal resistance does not change linearly with length. (The ends of a very long extrusion would be cooler than the center and therefore the transfer of heat to the surrounding air is little, if any.) It is therefore necessary to apply a correction factor to published data for extrusion lengths shorter or longer than three inches. "

There are a number of other technical articles there that ought to help you become an expert too.

[vinay]

Guys,

Thanks for the responses. The point source now makes sense why the thermal resistance does not drop linearly with length of heatsink.

Vinay