Heatsink Colour ?

Hi all, I don't know if this is the correct forum for this question but I am sure it will be placed somewhere else if necessary.
I am using a four section heatsink tunnel with fan mounted below blowing air out the top. The heatsink sections are "normal" aluminium colour. I read Rod Elliots article on heatsinks and he advises that aluminium should either be anodised black or painted black to achieve maximum dissapation.
I then read another article that clearly stated that if the heatsink is to be fan cooled it should remain un-coloured to achieve maximum cooling but if it is passively cooled it should be dark coloured
Any views on this please
Alan
 
If the only cooling is by forced convection then a surface coating adds extra thermal resistance - unless the coating roughens the surface, creates a little local turbulence and so helps destroy the boundary layer of air.

In most cases some heat will be lost by radiation too so blackening the surface will help. This of course requires that there be somewhere (other than the rest of the amplifier!) where the radiation can be 'lost' to.

To do a full analysis probably need a package like Comsol Multiphysics.
 

Ian Finch

Member
Paid Member
2010-04-11 4:22 am
Coffs Harbour, NSW
Black is beautiful.
As regards not needing black colouring for fan assisted cooling, it is correct, at least as far as typical industry practice goes. This is because the heat transfer is principally by conduction to the air stream - any assistance from radiation is likely insignificant at the maximum dissipation rate. The benefits of a thin oxide layer as exists on natural aluminium to a thicker one with anodizing may not be all that significant but if you are using serious and appropriate tunnel extrusions, anodizing probably won't be a purchase option anyway. If you are just strapping a small fan to a convection heatsink, any difference in surface coating will likely will be unimportant.

Heatsinks without fan cooling work by natural convection at much lower cooling rates per unit area, though still the same transfer process but radiation becomes a more significant contributor to cooling, Radiation is most efficient from black bodies so you get the picture that this is the way to go.

In a nutshell; Black anodizing is best for convection cooling but unnecessary and inferior to natural plain extrusions when a high degree of fan-forced cooling is used.
 
Last edited:

Vostro

Member
2012-05-12 7:20 pm
Ive read somwhere that sometimes there could be a bennifet of having a shinny heatsink.
I agree fully that the main heat dissipation working with convection should be matt black,

But lets say you have a casing with a high internal temperature, and lets say you have the VAS transistors on smaller heatsinks inside the case.
If those heatsinks were black, they would absorb the heat from the ambient aswell as from the transistor.
In this case it Might and I say Might be a case where a shiny heatsink might 'reflect' ambient heat, but still transfer sufficient heat from the transistor that is directly attached to it, and that sufficient convection will still take place.

Regards
 
Fins effectively make black emissivity insignificant.
That sounds odd, why?

The surface of a matt black fin does emit more heat than a polished bright coloured fin.
But the fins face each other. The adjacent fin because of it's inherent emissivity also works well at absorbing heat.
The adjacent fins heat each other.

As far as emitting heat to the external space one needs to look at the effective radiation surface area. This is not the fin area. It is the Black box area that encloses the whole heatsink. Some of that black box area faces into the amplifier internals. The emitting heatsink heats up the internal components and surfaces. The emitting heatsink also heats up the external surroundings and surfaces by radiation due to the area of black box facing to the outside.
The fin area heats the moving air in proportion to the surface area of the fins.

Due to the BIG difference in fin area to black box area, the convection emission dominates.

I have seen a figure stated that just 8% of the external dissipation is due to the matt black radiation.
Further that same paper stated that some non black and non matt surfaces are nearly as good as the matt black radiator. I suspect part of this conundrum lay in the unknown thickness of the "black" or non black insulating layer.
 
Last edited:
It could be argued that black fins help propagate heat to the ends of the fins, so assisting the conduction in the metal, and thus helping convection. However, Andrew's basic point is correct: the useful radiating surface of the heatsink is not the same as the useful convecting surface. Adding fins makes little difference to radiation but a big difference to convection.
 

Ian Finch

Member
Paid Member
2010-04-11 4:22 am
Coffs Harbour, NSW
Interesting:
I had another look at an ESP article on heatsinks quoting >15 times emissivity (energy radiation coefficient) for black anodizing V polished aluminium. That's a staggering difference. There was no figure for unpolished, as in raw aluminium nor for the dull and grimy surface of older equipment in service. From other surface types quoted, I'm interpolating the difference between black anodized and natural, aged aluminium surfaces is probably only ~1.6 times the emissivity.

ESP also quote the the total cooling as 25% or less attributed to radiation. My tests some years ago on 200 x 75mm matt black painted heatsinks of 0.55C/W rating, indicated around 22% of the heat was radiated, based on mean front+back IR temperature sensing. As AndrewT comments, this is subject to dissipating the heat transferred to case air and internal surfaces. As the case temperature stabilses, that heat must eventually be transferred to convected air and secondary radiation from the case.
Ventilation slots are going to help greatly there, whether the fins are exposed or not.
ESP - Heatsink design and transistor mounting
 
Thank you all for your very informative replies, much appreciated.
Why I posed the question was due to an article I read on the internet. For some reason I cannot get a hyperlink to it but if your interested I googled "transistor heatsink colour" and in the list was the article, which was "Themal Design- Homo Ludens", makes interesting reading.
thanks again
Alan
 
Several good points here.
1) For heat radiation, it's the "color" of the heat sink in the mid-infrared that matters, not the visible light color. I routinely use thermal cameras in my line of work for component/heat sink temperature measurement (who wants to try and stick a thermocouple to a teensy surface mount part?). Polished aluminum surfaces have low infrared emissivity, and show up as looking cold using a thermal imager, even if the surface is hot enough to take the skin off your fingers. Painting the aluminum surface with "white out" or putting a piece of polyester film insulating tape (also white) on it drastically alters the infrared emissivity so that the true temperature of the aluminum shows up, even though both the tape and the white-out, look white in the visible spectrum.

2) As Andrew pointed out, increasing the surface area of the heat sink with fins does little or nothing for the radiative properties of the heat sink, though the fins improve the convective cooling tremendously.

3) Improving the emissivity of a surface by painting it can screw up the far more important convective cooling effect by adding a layer of insulation on the heat sink surface. For a densely finned heat sink, it's better just to leave the surface alone, or have it anodized if you have access and/or can afford it.