For some time now, there has been some mis-information that folks are a bit confused on.

It has been stated that the various coatings on flashlights have no effect on the temperature of the flashlights, same applies to heatsinks.

So, I ordered some 0.750" diameter Aluminum 6063 rod stock to prepare an actual test to demonstrate the difference.

One of the minor reasons, is contrary to what I mentioned in one of my old posts, where I publically calculated the differences, and how much more cooling a high emissivity coating has as compared to a low emissivity polished surface- some members still claimed I was wrong. Of course I can understand how manufacturers would feel that way, it saves them a few dollars, leading to higher profit margins.

To do a real world test, in order to show the reality of things, I matched up two blue LEDs that are matched to within 0.04V of each other.

I took the rod, and cut two pieces, each to 1.735" long.

I bored holes into the side of each piece, 0.375" deep, for K-type 42 guage thermocouples.

Both pieces were polished up a little, but not nearly anywhere near a perfect mirror finish.

One was coated with black paint, the other was left bare.

The contenders:

Each were handled to get fingerprints and some finger print oil on the surface, much like a flashlight in the real world. If you look at the polished one, please don't lift my fingerprints!

The LEDs were wired in series, each is getting 1000mA on the nose.

The bored holes were filled with thermal paste, to get a more accurate reading.

The LED on the Black one is 40mV higher than the other, so the temperature would technically be slightly higher, if everything was equal- due to the higher power in the LED.

Both LEDs were directly mounted with Artic Alumina thermally conductive epoxy, cured under pressure, over night.

After 20 minutes, when things stabilized, the temperatures for the bare one:

And then I transferred the same thermocouple and measured and photographed the black one:

As you can see, the temperatures are obviously 17.4 degrees C cooler for the black one. Which also works out to a 31.32 degrees F difference.(correction pointed out by Curious-character)

No surprises here.

.

I then set each down on the surface, to help reduce the added cooling of chimney effect, let things cool and repeated the same test.

First the polished one:

Next the black one:

As you can obviously see, there is a 20C degree difference between the two now. This works out to a 36 degree F difference. (correction pointed out by Curious-character)

After 1 hour, the temperatures still read the same, so they had thermally stabilized.

Ambient temperatures in both cases remained at 20.6C.

.

So, in short, using a coating with a higher emissivity has a pretty drastic effect, reducing temperatures by 65 degrees Fahrenheit in the example shown.

This reduction of temperature would cause the die to run cooler, making it more efficient, and producing more light out of the flashlight.

If I had fully polished the bare Aluminum to a mirror finish, the bare one would have performed even worse in comparision.

If the surface area was greater (larger, such as is more common flashlights) than what is found on this 0.750" diameter 1.735" long rod , the temperature differential would be even greater, since the semi-polished aluminum has a very low thermal emissivity ~5%, as compared to the ~95% emissivity of black paint, and emissivity depends heavily on surface area.

Regular anodized Aluminum has an emissivity of ~77.6%

Hard Anodize, Type III, has an emissivity of 83.5% to 85.6% depending on the color, for typical specimens, and this will vary a bit for the quality and thickness of the coating.

Keep in mind, that if the light was tiny, like the Fenix P1, or smaller yet like the CR2 Ion, the cooling provided by a hand would be dominant, but it would still heat up much the same when set down on a surface.

The effects in a Flashlight would be even worse, since the whole flashlight is not a solid core of aluminum, and there are thermal resistances in the path, which would cause the area near the LED, and especially the LED die to run much hotter.

If the flashlight had heavy deep knurling with raised diamond points, the surface area would also be greater, which would benefit the black painted one even further than the polished one. I've seen the effect in a thermal camera many times before.

I hope you had as much fun reading over this as I did doing the demonstration.

For some, this doesn't make sense- the emissivity (the ability to radiate thermal heat) of a high solids white is within a percent as high solids black. However, if you run the test, or look up the data, one will find black or white paint has little difference, they both radiate very well in the thermal spectrum. One unique thing about white paint though, is that it reflects in the thermal spectrum, unlike black paint. However, polishing is a very, very bad thing for emissivity. Here I didn't polish to a mirror finish on purpose, since even Titanium gets small scratches on it on a keychain.

For this test, I used flat black paint, Rustoleum 7778 that I purchased at Lowes. Krylon branded high heat flat black paint is even better yet, but I could not find it at that store.

If we go back and subtract the ambient temperature from the numbers:
Bare 95.0C - 20.6C = 74.6 C rise
Black 77.6C - 20.6C = 57 C rise

1-(57/77.4)*100= 23.4% cooler for the Black vs. Polished

Laying down:
Bare 101.6C - 20.6C = 81.0 C rise
Black 81.6C - 20.6C = 61 C rise

1-(61/81)*100= 24.7% cooler for the Black vs. Polished

I think the number predicted was 24% back in that way old calculation thread- so not too bad.

Again, if this was a hand-held only light, the human body would act as a liquid cooling scheme, which would cool things down a lot more, unless you had to wear gloves, which is common in much of the US in the winter. Outdoors, where temperatures are even cooler, the delta temperature between the radiator and the ambient is even greater, and the black coated one would benefit even further over polished. The black coated light "sees" a colder "target" to radiate into and the black coated one radiates 17 times better than the polished one.

If you ever get a chance, on two occasions where the outside temperatures are the same, you'll find that a heat source runs cooler on a clear night vs. a cloudy night- especially if it is black. Interesting stuff.

Originally Posted by Curious_character

Thanks very much for taking the time to set up and document the demonstration. I'm sorry I missed the post with the calculation. I've always thought about radiative transfer as dominating only at relatively high temperatures, but I see now that's not true if the convective transfer is bad enough.

I did spot one error in your analysis, the C - to - F conversion. The conversion factor is exactly 9/5 (the ratio of the degree sizes), so

a 17.4 degree C change is a 31.32 degree F change, not 63.32
a 20 degree C change is a 36 degree F change, not 68.

Great job - thanks!

c_c

Tonight, I hooked up a 90 CFM fan (I'm not sure of the airflow at the test subjects, the fan just moves that much) and decided to take some measurements, to see if the black coating would reduce the thermal tranfer in air:

Ambient temperature was 20.5C

It appears as though the coating has no effect when there is a lot of forced air blowing on the lights.

Well, I did some more checking, and found that my airflow was not quite uniform over the area. What I did find, that is even in extremely strong airflow, the black one was always the same or a bit cooler than the semi-polished bare one.

Then it occured to be to move the two a bit closer together, to make the airflow more the same. I also moved the air source further from the test subjects, and in this case, it is off to the left side of the picture. I ran two tests, one with the polished one at the top of the photo, and then switched their positions, and ended up with the same answers. I also tried several different spacings between them, in case the polished one radiated enough to effect the black one, or vice versa. I didn't find enough difference to notice. Anyhow, even in airflow, the black one is actually consistently cooler.

For the pictures, I placed the meter near the test subjects, just during the picture taking, to get a better close-up.

Semi-kind of Polished one:

Black one:

Newbie,
Thanks for this thread and your investigation. I recall when I first joined here, thermal discussions were more prevalent and the comments were that among the three means of thermal relief, conduction, convection and radiation, radiation was the least significant. You have shown radiation to play a part greater than one would assume but I suspect there is also the question of its role being relative to the presence or absence of the other means? If you were to sit both samples down on a nice thick slab of Al or copper, would you see the delta you see with them sitting in free and unmoving air? If you were to elevate the humidity level, would this delta remain as high?

Someone posted back then about a motorcycle head that when highly polished would keep the piston from seizing but a rough cast head would cause the piston to seize. I would assume that this is due to a difference in air flow over the surface and means of thermal relief based on the air cooling. Is forced air cooling in the category of conduction or more like enhanced convection?!?

On the head front with blowing air, it sounds like in the rough cast case, the rough surface was increasing the thickness of the boundary air layer, reducing the effect of air cooling, a very well known effect. Engineers often will create turbulent air flow on purpose, to help scrub the boundary layer. But you have to be careful with turbulent air flow, as it can greatly decrease the amount of air passing thru, and if you go too far, you'll actually reduce the effective cooling. I've even seen laminar airflow work better, and from experience, it seems there is a sweetspot between the two. Another effect that happens, when the air is not forced thru an area, is what is known as "flow bypass", where there is enough resistance or back-pressure to airflow, that the air takes a route around the heatsink. Either one of those effects (which are somewhat related), is likely what was actually going on.

Humidity here is about 58%. The effects of humidity, and all the rest of this stuff is in many beginning mechancial engineering books, and is dealt with deeper in the thermodynamics portion of mechanical engineering. Some people even make a whole life long career in this area. There are a number of software suites that will help predict the outcomes under various scenarios, and one of the more common ones that is specialized for cooling/heating is made by Flomerics:
http://www.flomerics.com/

However, in most environments (7-94% humidity, 15-60C) humidity has a negligble effect on heat sink performance.

These pieces that I tested were large enough to be similar to some flashlights offered on CPF.

I didn't really consider setting them down directly on metal, as it is rare for me to find metal to sit my light down on the forest/mountains.

As the light gets much longer, like a Mag 4D, when it is oriented vertical, the chimney effect helps more.

Higher altitudes also reduce the cooling effect of air, it it thinner. As little as 5,000 ft can cause a forced air heatsink to become 10% less effective.

I also ran the air test to see how that contributes, see my last post.

The optimum mix varies depending on the environment, but it looks pretty obvious that a higher emissivity surface due to a very thin layer doesn't hurt in any case.

However, it looks that the surface coating has a large effect on the flashlight temperature when it is not held in the hand. And as I mentioned before, the forced liquid cooling of the human hand can have a very substantial effect, if you are not wearing gloves.