jfitzpatrick — 2013-11-19T16:00:58-05:00 — #1
Originally published at: http://www.howtogeek.com/175433/does-the-color-of-a-heatsink-affect-its-performance/
When it comes to the pursuit of true geekdom, there’s no question too arcane. Today we take a look at whether or not the color of a heat sink matters (and if the performance hit or bonus is even worth considering).
anybodysguess — 2013-11-19T21:32:24-05:00 — #2
I do suspect if the surface contacting with the CPU, GPU or other chip was colored it would adversely affect the cooling ability of the device.
wilsontp — 2013-11-20T12:12:19-05:00 — #3
You're thinking of a surface finish, such as paint... but even then, it wouldn't matter which color the paint is: it could still act as an insulator.
The article is pretty much spot-on: color only affects radiation, with black being the most efficient radiator. Convection and conduction aren't affected by color at all.
vassago — 2013-11-20T19:01:12-05:00 — #4
It is a universally known fact that red cars go faster.
Seriously, the material, heatsink design and the composition of the paint will be the main factors in effectiveness. Color itself is just for show, maybe black might be negligibly more efficient but even then ... I'd like to see some numbers on that.
reholmes — 2013-11-21T00:34:06-05:00 — #5
I would think that heat transfer at the convection boundary layer would resemble conduction and "immediately away" (depending on Reynolds Number) from the boundary layer would depend on laminar vs turbulent flow.
Does anyone have a gut feeling here?
sirraf03 — 2013-11-21T03:18:08-05:00 — #6
I'd think that's where the convection makes the difference with air being such a poor conductor. The heated air at the boundary should be acting as an insulator between the radiator and the next (as yet to be heated) layer of air, then the currents created by convection moves unheated air to replace the heated boundary layer, thereby making the transfer of heat more effective by the replacing of the boundary layer instead of waiting for heat transfer through the "insulating" layer.
If you take a liquid cooling system to replace the air cooling system and look at the way heat is transferred, what you'll see is that the majority of the fluid passing through the heating chamber (not all) has collected heat, and the heat is still expanding as it continues through the system, but the first edge of the fluid to go through the collection chamber is moving with the heated fluid to "dump" it's heat.
Of course, I'm just an old "hands on" type with no back ground science to back that up, it's just what my "gut" says might be going on.
bigaldepr48 — 2013-11-21T12:57:37-05:00 — #7
Radiation heat transfer is dependent on the emissivity of the surface which is to some degree a function of the color. It is also a function of the absolute temperature (K or R) of the hot surface raised to the forth power minus the temperature of the surroundings raised to the forth power and its emissivity. Most of the surface area of a heat sink is surrounded by other heat sink surfaces at the same temperature so there is no heat transfer going on. Only the surfaces that can radiate out into the case transfer heat. So heat sink radiation is very limited.
There are two types of convection, natural and forced. Natural convection occurs as the heated air rises as it becomes less dense and is replaced by cooler air. This provides little flow and thus very poor cooling, especially for small things like CPU heat sinks. Add a fan to create forced convection and the heat transfer goes way up. As mentioned before as the Reynolds number increases the heat transfer rate increases because it increases the turbulence causing more mixing. Another factor is the heat capacity of the fluid. Air has a low heat capacity or specific heat (Cp) while water has a much greater Cp. The Prandl number is the dimensionless number that determines the heat transfer based on the specific heat of the fluid. The Reynolds number and Prandl number combine to form the Nusselt number which determines the forced convection heat transfer rate. Air cooling has a high Reynolds number but a very low Prandl number. While water cooling has a little lower Reynolds number it has a much greater Prandl number resulting in the best heat transfer rate.
brotherboard — 2013-11-21T14:42:33-05:00 — #8
t is a universally known fact that red cars go faster.
It is a scientific fact that red cars discolour faster than any other. In fact red isn't a primary colour at all.
doctordeere — 2013-11-21T18:42:05-05:00 — #9
You wear a pocket protector, don't you?
sirraf03 — 2013-11-22T09:19:02-05:00 — #10
Cool! (in all it's meanings as applied to your reply) Interesting info. Now that I have names for the various factors involved, I'm going to find the time to do a little studying on them! Thanks!
sirraf03 — 2013-11-22T09:21:47-05:00 — #11
That's Why they go faster! Zoom! Zoom! Zoom!
wilsontp — 2013-11-27T12:20:47-05:00 — #12
In fact red isn't a primary colour at all.
Tell that to your eyes, which have photoreceptors that respond to frequencies of EM radiation that we perceive as red, blue, and green light.
Actually, the primary colors we were taught in school are incorrect. When doing additive mixing, the primary colors are indeed red, green and blue. However, when you're mixing paint or ink, you're actually doing subtractive mixing, so the primary colors are the inversions of those three.
red = cyan
green = yellow
blue = magenta
You can reproduce the entire color spectrum from cyan, magenta, and yellow pigments. You cannot do so with red and blue pigments.
So red is indeed a primary color, as are blue, green, cyan, magenta, and yellow.
jfitzpatrick — 2013-11-29T16:01:02-05:00 — #13
This topic was automatically closed after 10 days. New replies are no longer allowed.