Painting your heatsink black. Ideas, Experiences, comments.

Ever since my dad started telling me about it, and since the idea again popped up here in this thread, i hereby open a thread in which anyone can post their ideas, experiences and comments on the painting of a heatsink.


The idea

Physics tells us that a black surface will radiate its energy (heat) away faster then any other coloured one. Thereby, the cooling capacity of any surface should be improved if it'd be painted black. I've only heard about the ideas, i havent yet performed / seen it myself. If anyone already did it, or is willing to try it, please let me know.


What needed?

An ugly heatsink. One that is allowed to get dirty. No tricking, doing the black painting will destroy its visual value, which is the major reason why i haven't yet done it to my Zalman (siggie)

Black paint is needed. Any paint with a lot of plastic additives (eg. latex or whatever), is useless since plastic only isolates heat. Paint with a 'true-colour additive' (carbon) will work better.


To get the people going: some quotes from the other thread:

sjaak said:

Yes, my dad's profession helped with my overclocking. He already told me 10 times to paint my heatsink black ...black surfaces that are hot will radiate heat much faster then other colours. Just that i dont want to ruin the great copper look of my Zalman.

If anyone would try the black-painting of a heatsink, please let me know the results. My dad calculated to me that it, at best, could produce 30% more radiation, away from the HS....

Click to expand...

theflyingrat said:

You know, is there an 'anodizing' process for copper? As I understand it, anodization can help aluminum with heat dissipation. So would anodizing a heat sink black help it even a bit more? Hmm... maybe that's why the aluminum top of my Alpha 8045 is black anodized.

Click to expand...

sjaak said:

He was talking about just taking the heatsink, dropping it into a bucket of black paint and have it dry....

anodizing, afaik, is used to prevent corrosion, and to eliminate reactive properties of certain metals.

Ill try the painting with some old heatsinks of me in the holiday (couple of weeks). Want to see howmuch it could really help, and what other people can make of it.

Click to expand...

sjaak said:

As said, the black paint could improve the radiation by 30% (much more if it would be some stuff like a layer of carbon you could add to the heatsink).

Radiation from copper / aluminium to air is, with good airflow, about 30% of the total amount of energy being dissipated. (so says my living science library)

0.3 * 0.3 = 0,03

Shows that in the worst case (bare black paint - no additives) the improvement could be 9%

Definatly something to keep an eye on, try if possible.

Click to expand...

Again, i would really like to know if people already did it / are planning to do it. Send me the results, and i will post them here. In the upcoming weeks, i will try to perform it on some of my own heatsinks, and see what results i get. If anyone has ideas or comments, please post them.

Tired now - i will add more later this week.

wait hold on, but wouldnt the paint hurt the heat transferring process, since the paint is another layer of material, and isnt paint a bad heat conductor??

lol
"Physics tells us that a black surface will radiate its energy (heat) away faster then any other coloured one. Thereby, the cooling capacity of any surface should be improved if it'd be painted black."

sounds good, but it simply isnt true. color of a surface is only affected by its absorbtion or reflection of LIGHT, not heat(though the absorbs light is converted to heat). a black surface in a freezer will get no colder or hotter than a white one. unless you introduce LIGHT as the heat source. such as outside, where the heatsource is the sun. put white and black in the shade, and it wont matter.

paint is a bad conductor and will negatively impact your thermal transfer from metal to air no matter what color that paint might be. and further more, black or white paint, they will affect the heat equally.

ares350 said:

lol
"Physics tells us that a black surface will radiate its energy (heat) away faster then any other coloured one. Thereby, the cooling capacity of any surface should be improved if it'd be painted black."

sounds good, but it simply isnt true. color of a surface is only affected by its absorbtion or reflection of LIGHT, not heat(though the absorbs light is converted to heat). a black surface in a freezer will get no colder or hotter than a white one. unless you introduce LIGHT as the heat source. such as outside, where the heatsource is the sun. put white and black in the shade, and it wont matter.

paint is a bad conductor and will negatively impact your thermal transfer from metal to air no matter what color that paint might be. and further more, black or white paint, they will affect the heat equally.

Click to expand...

The idea is not to use a plastic basef paint but a paint that is thermally conductive. Some of the stuff used for painting engines is good for that. Why do they paint the radiators on cars black then? What about the alpha heatsinks? I am no physicist but I want to see what people say.

but wouldnt black absorb more energy (from light of course)??? instead of radiate it away? Just go outside in the summer wearing black, then change and wear white. You will notice the difference. ares350 is absolutly correct though, it only matters when you talk about light.

why dont you try it and tell us.

my guess is the results would be near impossible to measure in an air cooled system.

put a latex or almost any paint will restrict heatflow GREATLY. Anodizing is nothing like it, and yet it still limits heatflow. it may increase thermal capacity, but this doesn't do anything for your cpu temp. this is why everyone will recommend that u use a thin layer of thermal goop for example. using too much, gets u bad results. Paint will not go on thin. A layer of anodizing is like microns thick while paint will probably be like from 1/2 mm to 1/8mm thick

heatsinks don't corrode. more importantly, copper doesn't corrode even with water (aluminum is what corrodes). copper oxidizes, and discolors. this is more or less NOT important because the heat contact area is filled with thermal goop which won't allow that to happen. oxidizing won't have a great affect on heat disspansion either.

Now the reasons companies oxidize heatsinks isn't to improve there heat transfer or capacity, but to reduce conductivity. aluminum/copper are very conductive and the last thing u want in a computer is something conductive and has a potential of falling off and killing things. Now a days, its not much of a risk and with cpus give off as much as 100watts and more, everything has to be done to get most out of a heatsink.

EDIT:: got beat out by 4 posts by the time i wrote this, lol

radiators are black for appearance purposes. intercoolers are silver for appearance purposes(often chromed silver, though that would add to reflecting light, but I dont think they really consider this). a thin coat wont kill heat transfer(depending what the coat is made of) but in computers your looking for every degree, in cars, they have a nice buffer, little inefficiency is well worth it for cosmetic benefits.

color simply doesnt matter to heat capacity or transfer. only what light waves it absorbs or reflects. NOTHING MORE. light is energy, energy can be heat. thus in the presence of light, black will absorb more light, thus energy, thus heat. but this in in the source of light. enough that it would have an effect.

edit: when I say enough that itd have an effect, I mean ALOT. like a spot light or the sun, white shirt vs black shirt inside under normal lighting, you dont get hotter in black do you?

ares350 said:

lol
"Physics tells us that a black surface will radiate its energy (heat) away faster then any other coloured one. Thereby, the cooling capacity of any surface should be improved if it'd be painted black."

sounds good, but it simply isnt true. color of a surface is only affected by its absorbtion or reflection of LIGHT, not heat(though the absorbs light is converted to heat). a black surface in a freezer will get no colder or hotter than a white one. unless you introduce LIGHT as the heat source. such as outside, where the heatsource is the sun. put white and black in the shade, and it wont matter.

paint is a bad conductor and will negatively impact your thermal transfer from metal to air no matter what color that paint might be. and further more, black or white paint, they will affect the heat equally.

Click to expand...

You took the words right out of my mouth .

I think this guy has the black-body relationship of physics messed up. I'll explain it later when I can better figure out what it means....but its like a description of how a perfectly black object would absorb heat, but turned around, so it's releasing heat.

If someone that has a VERY good background in physics and some quantum mechanics wants to look at it, google "black-body relationship"

PS, god I'm such a fricken nerd....

This has been mulled over many times ever since I first joined... There is always an extensive discussion, which ends up in numerous people trying to get numerous other people to understand that the color of their heatsink makes no difference in its heat transfer capability.

The surface color of the heatsink on a CPU makes no difference.

Car radiators are not trying to get maximum cooling capacity, they are trying to get a certain amount of cooling (its a balance thing), and the paint has little, if anything, to do with it.

Color of heatsinks is marketing, the alpha performance is not improved by its color.

This has been discussed before, please use the search...

http://www.ocforums.com/showthread.php?t=293914
http://www.ocforums.com/showthread.php?t=277742
http://www.ocforums.com/showthread.php?t=203997

In reply to Ares350:

Light is a form of radiation, just as any other beams we know.

Black does absorb energy from radiation, but thats only if the temperature difference between the black surface and the environment is marginal.


We are talking about a black heatsink being heaten up, and in that case, black is supposed to give more of that heat away using radiation then any other colour. Yes, if the wrong paint is used, it will screw up (read the start post), yes, colour is often used just for looks and marketing purposes, but that doesnt take away the thermal behaviour of black. (BINAS, anyone?)

I will ask more people who are experienced with (car / machinery) cooling, and post what they tell me. All you're replies just made me more anxious to try it

so paint the bottem black an the top white

mabey no??yes?? like its been said use the search this has been done like 4 times (i was right in there with the firts one i saw but turns out that nothing made a diff ven with the good readers u cant tell its just to small of an amount to worry about (look col though black hs on an orange asus mobo (got one)

Thanks to KCT2 for this link: http://www.ocforums.com/showthread.php?t=277742

Where it is said that a black surface would help on passive cooling, but not on active cooling (or at least, thats what i draw from all posts).

A google search yielded both positive and negative experience.

Got 2 variables already: colour, airflow

syberspy@school said:

so paint the bottem black an the top white

mabey no??yes?? like its been said use the search this has been done like 4 times (i was right in there with the firts one i saw but turns out that nothing made a diff ven with the good readers u cant tell its just to small of an amount to worry about (look col though black hs on an orange asus mobo (got one)

Click to expand...

any color will not make a difference. The whole color thing only works with light. If u wanted to put ur case outside, i'd paint it white instead of black so it doesn't get any heat from the light. Adding anything, and i mean anything, thats not as conductive as copper will act like an insolator between the cpu and heatsink and restrict heatflow. When you're dealing with contact, only thermal conductivity properties matter. lightwaves, radiation, convection by air or water all work differently (most of them anyways)
man i talk too much..

Sjaak said:

Thanks to KCT2 for this link: http://www.ocforums.com/showthread.php?t=277742

Where it is said that a black surface would help on passive cooling, but not on active cooling (or at least, thats what i draw from all posts).

A google search yielded both positive and negative experience.

Got 2 variables already: colour, airflow

Click to expand...

i haven't learned natural convection yet, but that sounds right.

ya about the bottem of hs black or copper i know copper will work better but why not try ever think about the thermal pads that are on the bottem of stock intel hs??? there black thermal pads mabey for a reson

but im just thinging i know that in the end the sp94 and my vanteck stealth pull away on top temps/dba rating so im good

Did some research on this after being in a similar thread. Basically you would decrease the performance of the HS, but probably (hopefully) not too much.

1) Radiation is of fantastically small importance compared to convection as far as HS cooling is concerned.
2) Even less so when you consider that if you have 50 fins, all but the outside edges of the outer two are radiating at each other.
3) The total thermal resistance of the heat sink is additive of all the bits. You want resistance down and conduction up. Putting a layer of paint on must therfore always increase the total resistance to the conduction to the outside of the HS (now the outside skin of paint). This is bad, but if done with some very special paint and done very carefully in an incredibly thin layer probably not amount to much.
4) Black radiates the heat away. This is the confusion with "black body radiation". A black body is a physics term describing a theoretical object whose radiated energy is wholly determined by its temperature and not any incident radiation. Basically it's irrelevant to the discussion.
5) Black t-shirts, white t-shirts etc. When we're talking about radiated heat we're talking about a wholly different part of the spectrum. Basically the colour that we perceive an object to be depends on how it absorbs / reflects visible light. With regard to the CPU, it doesn't generate heat in the visible spectrum. If it did, your computer is dead... Hence the colour of a paint that we see is irrelevant and says nothing about whether it absorbs / reflects in the heat part of the spectrum.

Hope that's of some help.

mantrogo said:

[..........]

Hope that's of some help.

Click to expand...

Definatly. Much better then then the people that are just laughing.

Starting to have my doubts, still, i want to try it on a few old heatsinks. I've contacted someone who said that he did it with succes, so i'd like to know how after reading your arguments.