Why not copper?

[bdl666]

Why is it that every one uses aluminum on their heat sink?
Why not copper? Copper is been used by overclockers on computer cpu's for a while now. Why not on a Luxeon? Maybe someone can make a sink with a copper core and the rest on aluminum, like the new Pentium 4 sinks.

 

[chalo]

[ QUOTE ]
bdl666 said:
Why is it that every one uses aluminum on their heat sink?
Why not copper? Copper is been used by overclockers on computer cpu's for a while now. Why not on a Luxeon? Maybe someone can make a sink with a copper core and the rest on aluminum, like the new Pentium 4 sinks.

[/ QUOTE ]

Many modders have been doing just that; namely, using a copper slug underneath the LED and coupling it thermally to an aluminum flashlight body.

Copper is used where heat transfer rate is most important, because its heat conductivity is very high. However, aluminum has a much higher heat capacity per unit weight than copper. So copper is better at shifting heat quickly from a sensitive component to where it does no harm, but aluminum is better at soaking up a lot of heat.

High strength aluminum alloys have thermal characteristics reasonably close to pure aluminum, whereas high-strength copper alloys (bronzes and brasses) usually have dramatically less heat conductivity than pure copper.

So if you made a pure copper flashlight body for its heat transfer capabilities, you'd have to live with its wear-, nick-, and dent-prone softness as a corollary quality.

Chalo Colina

 

[bdl666]

I din't mean the entire flashlight. Just the pedestalwere the emmiter sits.Maybe like the hotlips but with a copper slug .

 

[brickbat]

Copper would have a lower thermal drop across it than aluminum. But at the relatively low power levels of Luxeon's (1-5W) I doubt the difference would amount to more than a degree or two.

 

[Floating Spots]

Plus most flashlights are aluminum and mixing metals in the long term is not favorable.

 

[Jarhead]

[ QUOTE ]
chalo said:

Many modders have been doing just that; namely, using a copper slug underneath the LED and coupling it thermally to an aluminum flashlight body.

Copper is used where heat transfer rate is most important, because its heat conductivity is very high. However, aluminum has a much higher heat capacity per unit weight than copper. So copper is better at shifting heat quickly from a sensitive component to where it does no harm, but aluminum is better at soaking up a lot of heat.

Chalo Colina



[/ QUOTE ]

I'll agree that Aluminum will soak up more heat per weight,
but copper still beats it for size/volume, so per volume, copper will soak up more heat than aluminum, and copper will also have 2-3 times less thermal resistance than alumimum.

Thermal conductivity (W/m*K):

Aluminum Pure ... 247 (rarely ever used in pure form)
Aluminum (6061).. 171
Aluminum (6063).. 193
Aluminum (7075-T6)130
Copper .......... 398
Brass (70Cu-30Zn) 115
Magnesium ....... 170
Diamond ......... 2500
Magnesium alloy ZK60A 117
White thermal compound (AOS52022) 0.7

Specific Heat, J/gm K:
Aluminum 0.900
Copper 0.386

A ratio of 2.33 (Al):1 (Cu) (by mass)

But the density (mass per volume) is greater for copper:

Metal or alloy kg/cu.m
aluminium - melted 2560 - 2640
aluminium bronze (3-10% Al) 7700 - 8700
aluminium foil 2700 -2750
beryllium copper 8100 - 8250
copper 8930

3.31 (Cu) : 1 (Al)

Pure Aluminum's electrical conductivity is only 62% that of copper, but:
Copper 100%
Aluminum 6063 57 - 65%
Aluminum 6061 47 - 56%
Aluminum 7075 44 - 47.5%
Aluminum 7075 T6 series 30 - 35%
In almost all cases, aluminum with a T after it will have lower conductivity than the base alloy.

So, for volume:
- Copper will soak up 142% more heat than Aluminum
- it will have half the thermal resistance at
the same time.
- average 40% better electrical conductivity per size (or 40% less electrical resistance.)

And if you like the feel of heavy things, the copper will be 3 times as heavy as aluminum, for the same machined object.

 

[asdalton]

Heat capacity matters when the heatsink is absorbing transient heat loads, such as during a cold startup or in periods of brief overdriving. However, it is completely irrelevant to the heat transfer characteristics during steady use. Unfortunately, the term "heatsink" is misleading because it implies that the short-time absorption of heat is important instead of what matters for most applications: steady-state heat transfer.

 

[AilSnail]

Is it true that aluminium ratiates or convects better to the air?

Also does anodyzing help with either transfer method?

 

[Jarhead]

Well, according to the thermal gurus in the thermal design magazines I get, they say that as far as radiation and convection of thermal energy from a heatsink, there is no difference from copper vs. aluminum, and it is an old wives tale.

Yes, Anodizing and Alodyne processes can both increase and decrease the transfer of heat to the air, depending on the specific process.

Regular flat black paint actually acts as an insulator, but I know that Krylon high heat barbecue paint works nearly as well as black body paint. How do I know this? One of the production technicians painted the black body radiator targets with regular Krylon flat-black, which cause mis-calibration and lots of odd looking images. Then we had some paint that cost 350.00 an oz. show nearly the same results as the flat black Krylon high heat barbecue paint.
(oh, forgot to mention, with the thermal cycling of the black body radiators, the paint would flake off, which then would mess up the calibration for the InSb thermal detector arrays)

 

[illumiGeek]

Yes, copper absorbs heat better than aluminum, but it sheds it slower. Remember, the heat has to go somewhere.

In order for copper to be an effective heat sink it needs a large surface area (fins/pins) to radiate the heat away. Much larger than aluminum at a given heat load. Aluminum makes a good heat pump because heat moves through it very well. Copper tends to hold on to the heat.

In theory a copper core with an aluminum sink to get rid of the heat should work well since copper will transfer heat to aluminum better than air.

But in practice it doesn't really make much of a difference. This has been tried on computer CPUs and they dissipate way more power than a LUX LED, and there are all copper and all aluminum sinks that work just as well as the CO/AL combo.

At the power levels we are dealing with I really don't think the material is too critical.

Aloha, iG

 

[Mark_Paulus]

mm for mm (or inch for inch), Aluminum is going to be lighter than copper. And it's easier and cheaper to get aluminum slugs and bar stock than copper, so a home-modder is more likely to have a chunk of aluminum laying around to work with than an equivalent sized chunk of copper.

 

[Slick]

.

 

[NewBie]

[ QUOTE ]
illumiGeek said:
Yes, copper absorbs heat better than aluminum, but it sheds it slower. Remember, the heat has to go somewhere.

Aloha, iG

[/ QUOTE ]


I'm not sure I can agree with you on that point. It will depend alot on the surface characteristics and the finishes that are applied.

Thermal Emissivities from one source (higher is better):
Aluminum foil 0.04
Aluminum disc (rough grooved finish) 0.18
Aluminum household (flat) 0.01
Aluminum polished 0.04-0.06
Aluminum highly polished 0.035
Aluminum (rough plate) 0.06
Aluminum (oxidized) 0.11-0.19
Aluminum (anodized) 0.55
Aluminum (weathered to point of pitting) 0.57
Aluminum carbon coated 0.84-0.98

Brass (polished with machining lines left in) 0.04
Brass (highly polished) 0.03
Brass (quite tarnished) 0.22

Copper (machined surface) 0.18
Copper carbon coated 0.84-0.98
Copper (plate heavily oxidized) 0.78
Copper (polished) 0.02
Copper (oxidized black) 0.89

Nickel (polished) 0.072
Nickel (polished electroplate) 0.05
Nickel (oxidized) 0.37
Nickel oxide 0.59-0.86

Silver (pure, polished) 0.0.020-0.032

Stainless Steel (lightly buffed) 0.14
Stainless Steel (sand blasted) 0.46

3M Black Velvet 9560 1.0
Thin coating of Krylon Ultra-Flat Black 0.97
Krylon High Heat Barbaque paint, after heating 0.99

As you can see, polishing greatly reduces the thermal transfer into the air, which would cause the light to get significantly warmer.

Here is another page, which includes the emission wavelengths (thermal emissivity changes depending on the temperature of the material):
thermal emissivities

 

[turbodog]

someone should look up the data on thermal expansion for these metals

galvanic corrosion between dissimilar metals should also be looked at (where's that table again?)

copper is somewhat harder to machine, or at least it requires more knowledge

 

[turbodog]

[ QUOTE ]
NewBie said:
[ QUOTE ]
illumiGeek said:
Yes, copper absorbs heat better than aluminum, but it sheds it slower. Remember, the heat has to go somewhere.

Aloha, iG

[/ QUOTE ]


I'm not sure I can agree with you on that point. It will depend alot on the surface characteristics and the finishes that are applied.

Thermal Emissivities from one source (higher is better):
Aluminum foil 0.04
Aluminum disc (rough grooved finish) 0.18
...


[/ QUOTE ]

I'm not sure I can agree with the relevance of your data. /ubbthreads/images/graemlins/tongue.gif

At these temperatures conduction, not radiation, is the primary method of heat loss from the material.

 

[NewBie]

Depends on if you are holding the light or not. You should recall folks mentioning some lights getting extremely warm when not held...

If you are not holding the light, its going to make a difference. Many of these numbers are in the 10um emission range, which if I recall correctly is somewhere around 20 degrees C. Thats why I provided the link, so folks could look things up.

Lets see...Wien's Law
Peak wavelength = 2.898 x 10^-3 / T (in degrees Kelvin)

25 degrees C = 298.15 degrees Kelvin 9.72e-6 meters or 9.72um

Many flashlights here reach 40 degrees C:

40 degrees C = 313.15 degrees Kelvin 9.25e-6 meters or 9.25 um

Some here tested a few lights, not held, and recorded 156 Fahrenheit. This is 68.89 Celcius.

68.89 degrees C = 342.04 degrees Kelvin 8.47e-6 meters or 8.47 um

Anyhow, it turns out that 10um black body emission wavelength is 16.65 degrees Celcius.


As things warm up, the wavelength of emission gets shorter, until you start seeing a red glow, and eventually end up with wavelengths in the middle of the visible range, such as the tungsten filament in a light bulb.

A little grease and keeping the moisture out (which grease will also help with) will go a very long way towards preventing galvanic corrosion. Anodizing/Chemkote/Alodyne is also commonly used in many industries to prevent/reduce corrosion. One nice thing about Alodyne (when done properly), is it lowers the electrical resistance on aluminum surfaces. It is common place to see heatsinks with copper inserts to enhance their performance these days.

 

[ViReN]

so, If I leave Rough Machining & Let it Oxidize... my heat sinks (to dark tan color) it will Dessipate more heat?

(i am aware of surface area) but this oxide thing is new... Thanks for the information NewBie /ubbthreads/images/graemlins/smile.gif

(perhaps this is the reason, why the Solar Water Heaters have Copper Black in the Solar Panel Pipes)....

 

[NewBie]

I don't think that will make much difference in your heatsinks, since they conduct heat to outer surfaces. If they were the outer surface, it would be a different story.

 

[wasBlinded]

I wonder what the ratio of heat loss via radiation vs. conduction/convection in air is? Assume a black anodized aluminum tube. Just curious, I have no idea myself.

 

[NewBie]

Thermal conductivity of air is 0.026 vs. 193 of 6063 Al alloy.

Energy radiated per second:
H = esAT^4
e = emissivity (0-1)
s = Stefan-Boltzmann constant
= 5.67 x 10-8 J/(s-m^2-K^4)
A = surface area of object in meters
T = Kelvin temperature

Many flashlights, like the LH, have a surface area of 13 sq. in. Standing in free air someone measured it at 156 Fahrenheit, or 342.0389 Kelvin. To keep things simple, lets figure the item is in free space, and is 156 Fahrenheit. Lets assume it is a nice black, lets use 0.9 for the thermal emissivity of the surface. Doing the math (assuming I did it right, and am thinking about everything right, as I just woke up...and feeling rather dunder headed), if it was black means it would radiate

5.86 Joules per second.

A 1 watt source will dissipate 1 joule per second.

So if the light had a 1 Watt LED in it (which turned all the power into heat), and the light was black, due to the radiation of the heat, it would cool down from 156 Fahrenheit.


Now the same light, in polished Aluminum would have a thermal emissivity of roughly 0.02

So if you do the math, the polished Aluminum light would radiate 0.13 Joules per second.

So, if it was polished aluminum, it would heat up quite considerably beyond 156 Fahrenheit.

If it was chromic acid process Anodized Aluminum, natural color, we'd be looking at 3.58 Joules per second, so it would cool down a bit.


If it was Polished Brass (not highly polished, which would be nearly the same as polished aluminum), we'd be looking at 0.65 Joules per second, so it would also heat up.


On a small light such as the LH, sitting on a flat surface, there isn't much heat lost due to convection, even less if it is on it's side.


In reality, the flashlight probably doesn't transfer the heat evenly over it's whole surface, so these numbers would be a bit different.