Improving heat-sink performance.

[Bazukaz]

Hi,
Has anybody tried to improve heat sinks by using hydrochloric acid ?
I tried , and the surface of heat sink became rugged.I have not made any mesurements yet , but i suppose that rugged surface should dissipate heat better than a flat one.

The heat sink i used was not anodised , so it is worth to try

Regards,
Lukas.

[lndm]

Interesting. I guess you could argue that the rough surface combined with the air flow produces micro-turbulence that aids in the air accessing the heatsink. You might also say the turbulence hinders the free flow of air? A fine point, I know.

If you have a h/s with a fin arrangement that ensures a good transfer of heat as is, and if the air is approaching saturation from the heatsink's point of view (approaching the same temp as the h/s) then surely its presence is unwanted, and good flow is wanted.

I used to read car magazines and there was a misconception once that if you put a flow restrictor in place of the thermostat, the water will linger longer in the radiator and would get colder. Of course, the best heat transfer is made when the difference in temperature between the water in the radiator and the outside air, is greatest, and that means flow.

I may have gotten carried away here (thanks, it was fun).

[AndrewT]

Hi,
I'm inclined to think your

micro-turbulence

outweighs all other characteristics.

The added roughness also has the secondary benefit of increasing the air contact surface area. This, I think, promotes increased heat transfer. I realise that conduction is not at play here but I stand by that supposition.

But do not add roughness to the interface with the active devices that need the cooling.

[lndm]

OTOH, there is a slight removal of material which reduces the heatsink's persistance.

[mzzj]

I would bet that improvement is very small if not neglible.

Painting your heatsinks can improve emission coefficient from something like 0.2 of bare aluminium up to 0.9
(effect on temperatures is much smaller because convection cooling is dominating effect in these temperatures. )



BTW: It doesnt matter if they are painted black or white, visible color has very little to do with emission coefficient in far-IR range.

[Stocker]

The best thing you can do to increase your heatsink's performance is to improve the thermal junction. If you are losing most of the heat transfer in the junction, a 0.01% increase in dissipation is not going to help much. (edit: be careful the type and thickness of paint... heavy house paint is probably not your friend! )

A thick mica washer is going to kill the heat transfer. Silicon goop grease is better unless you need the electrical insulation. A phase-change material is expensive, retail, but you get some ridiculously good thermal coupling. I am still trying to figure out how to unload (hopefully sell) a few thousand pats of Powerstrate and until I do, I am using it. My PC processors are in cool-city now. two down, four thousand to go. Before that, I used the goopy stuff. If I have to use mica, I always try to shave it a little thinner than I found it... The amp in my car, I shaved some of the washers to 1/3 their original thickness and still had plenty!

Another idea is being worked on by the engineers at Innovative Fluidics, with their fan-replacement technology, coming soon to some BIG electronics name brands near you. The details are top-secret but very cool (pun intended).

[LineSource]

Several heat sinks are available with serrated (wavy) fins designed to provide greater surface area and increased air friction to improve heat transfer. "That's Hot!"

[Tweeker]

BeO is a breathtakingly effective washer material. Its more thermally conductive than the aluminum.

[jackinnj]

two things which improve heatsink performance -- mount the fins so that they are aligned vertically -- this improves air flow -- second -- you can use a small fan -- purloined from a PC power supply and run at a low voltage -- even a small increase in air flow will have a dramatic impact upon the heat dissipated. A small fan run at 70% of its stated voltage is un-noticeable.

For my 2x120W bridged LM4780 amp I arranged 2 heatsinks as a wind tunnel and drove the chamber with a 5" fan -- I isolated/insulated the heatsinks instead of the chips to save eliminate the 0.4 C/W thermal impedance which a mica washer would present to the device. The heatsinks are kept away from prying fingers by their enclosure in a steel chassis. The whole schmegege makes a lot less noise than the hard drive in the TIVO !!!

[nuppe]

A big fan will give more airflow at lower (percieved) noise levels than a small one, since you can run it at a lower rpm for the same flow.

Also the chimney effect might help get the airflow up without fans.

/Andreas