Heatsink material

[yazkaz]

Which material would be most suitable for making custom heatsink canisters? Copper (especially OFC), brass, or aluminum? And why?

I understand that OFC yields the best performance, but I cannot find OFC material in my area. Will brass be a good substitute, or should I stick to aluminum for better performance?

 

[precisionworks]

Copper (especially OFC), brass, or aluminum? And why?

Of those three metals, pure copper (aka OFC or C10100) has the highest thermal conductivity (k) at 223. Pure aluminum is second at 118, and brass is way down the chart.

The most common copper alloy is ETP (C11000, or Alloy 110), which has a k rating only 1% less than OFC. Most electrical & electronic suppliers carry ETP in rounds, flats, and square bars.

http://www.cut2sizemetals.com/copper/rectangular-bar/kbr/?gclid=CLPPto2--54CFRQeDQodWF0vBg

http://www.smallparts.com/s?ie=UTF8...bility_class&gclid=CN_Dno-_-54CFQUeDQodcHi8Aw

http://www.onlinemetals.com/merchant.cfm?id=87&step=2&top_cat=87

 

[StrikerDown]

Barry, you didn't mention gold! :twothumbs

 

[darkzero]

As Fred mentioned, in another thread, Silver! :naughty: "Silver, pure 235"

I'd love to see Fred make a pure silver flashlight! *hint, hint

 

[stinky]

isn't diamond actually the king of thermal conductivity? but seriously, there are many copper alloys that are suitable. you don't have to use ofc. supposedly ofc is difficult to work with compared to some other alloys.

Copper No. Thermal Conductivity, 68°F (Btu/Ft2/Ft/Hr/°F)
Coppers
C10100 226
C10200 226
C11000 226
C12200 196
C14300 218
C14500 205
Copper Alloys
C17200 62-75
C17300 62-75
C17510 144
C18000 125
A945 80
C18200 187 (68-212°F)
C18700 / C99 218
C97 140
Brasses
C21000 135
C22000 109
C23000 92
C26000 70
C28000 71
Leaded Brasses
C33000 67
C35300 67
C36000 67
C36500 71
C38500 71
Tin Brasses
C46400 67
C48500 67
Phosphor Bronze
C51000 40
Leaded Phosphor Bronze
C54400 50
Aluminum Silicon Bronze
C64200 26
Silicon Bronze
C65500 21
Copper-Zinc
C67300
Copper-Nickel
C71500 17
Nickel-Silver
C75200 19

extracted from here: http://www.copperandbrass.com/cnb/cms/groups/intranet/@inter-cabs/documents/web_content/d_004663.pdf

 

[greg_in_canada]

Speaking of silver, here is a cool (PC) heatsink made using a silver coin: http://www.overclockers.com/heres-a-little-teapot-short-and-stout/

Greg

 

[precisionworks]

... you didn't mention gold! ...As Fred mentioned in another thread, Silver!

LOL

Silver did jump out as the leader in thermal conductivity, but the OP asked only about copper, brass or aluminum. In a rare moment, I tried to stay on topic and focus on only those three metals

 

[Justin Case]

Which material would be most suitable for making custom heatsink canisters? Copper (especially OFC), brass, or aluminum? And why?

I understand that OFC yields the best performance, but I cannot find OFC material in my area. Will brass be a good substitute, or should I stick to aluminum for better performance?

What are you selection criteria? Just thermal conductivity? How about thermal mass? Material cost? Machinability? Corrosion resistance? Looks? Anodizability? Hardness? Tensile strength? Etc?

 

[PhotonFanatic]

As Fred mentioned, in another thread, Silver! :naughty: "Silver, pure 235"

I'd love to see Fred make a pure silver flashlight! *hint, hint

Not pure silver--that's too soft, but sterling silver, now that's another story. And non-tarnishing sterling silver--yes, that's on the drawing board.

And, for sure, there's no reason not to use it for the heatsink as well. :devil:

 

[fppf]

Conduction is not the only criteria to look at for a heat sink material.
Most heat sinks that need to conduct heat use copper because of its very high conduction rates, but then when convection needs to happen switch to aluminum.

If you look at most CPU heat sinks there is a copper slug that contacts the chip and then that is pressed into an aluminum fin set to convect the heat away. This gives the best of both materials.

 

[PhotonFanatic]

Since the normal fluid doing the thermal transfer for convective cooling is air, I fail to see how the material makes any difference, other than one being more expensive than the other. Am I missing something?

This formula clearly does not involve the material of the heatsink:

The formula for Rate of Convective Heat Transfer:[5]
q = hA(Ts − Tb) A is the surface area of heat transfer. Ts is the surface temperature and while Tb is the temperature of the fluid at bulk temperature. However Tb varies with each situation and is the temperature of the fluid "far" away from the surface. The h is the constant heat transfer coefficient which depends upon physical properties of the fluid such as temperature and the physical situation in which convection occurs. Therefore, the heat transfer coefficient must be derived or found experimentally for every system analyzed. Formulae and correlations are available in many references to calculate heat transfer coefficients for typical configurations and fluids. For laminar flows the heat transfer coefficient is rather low compared to the turbulent flows, this is due to turbulent flows having a thinner stagnant fluid film layer on heat transfer surface.[6]

 

[fppf]

To clarify

The convection to the fluid (air in this case) is normally the limiting factor of a heat sink.
The reason they use copper for the core/base is because it has such a high conduction rate. They switch to aluminum for the fins because conduction is typically not the limiting factor for the heat to move.

Aluminum can be extruded and is cheaper then copper.

So having a copper core will move the heat from the source out into the field of fins. Once it gets to the fins it must move into the air, which has a much higher resistance then then copper, and even aluminum. This is unless you have REALLY high flow rates.

Here are some links for reading.
http://powerelectronics.com/mag/power_tests_compare_forced/

http://www.thermshield.com/ThermshieldPages/Copper_vs_Aluminum.pdf

 

[MorePower]

Since the normal fluid doing the thermal transfer for convective cooling is air, I fail to see how the material makes any difference, other than one being more expensive than the other. Am I missing something?

I don't think you're missing anything. Cu slugs with Al fins are often used for PC heatsinks for reasons of both cost, ease of manufacture, and weight.

Early motherboards with high wattage CPUs didn't have a good method to attach heatsinks securely. All copper heatsinks weighed enough that simply setting the case down too hard could rip the processor socket clean off the board. To mitigate this problem, the hybrid copper-aluminum heatsinks were developed.

 

[KowShak]

I don't think you're missing anything. Cu slugs with Al fins are often used for PC heatsinks for reasons of both cost, ease of manufacture, and weight.

Early motherboards with high wattage CPUs didn't have a good method to attach heatsinks securely. All copper heatsinks weighed enough that simply setting the case down too hard could rip the processor socket clean off the board. To mitigate this problem, the hybrid copper-aluminum heatsinks were developed.

The hybrid copper-aluminium CPU heatsinks brought about their own set of problems, to begin with the aluminium fins weren't attached to the base using a method that conducted very well, the result was that a solid aluminium heatsink could outperform the first copper/alu hybrid designs. These problems weren't limited to hybrid heatsinks, often pure copper heatsinks would have the same attachment issues.

There was quite a bit of interesting manufacturing went into making heatsinks, one of the heatsinks I bought for my own PC had skived fins, they actually shaved slices of copper off the base of the heatsink, leaving them attached at one end and standing them up. If you look at many aluminium heatsinks now, you find that they have a copper slug pressed into their base.

Surface roughness was a one issue that people trying to get maximum performance out of their computers looked at, people were lapping the base of their heatsinks and their CPUs to get better conduction, does anybody do that with their LEDs yet?

At the time, I wasn't trying to get maximum performance from a PC, I just wanted mine to run quiet, run cool and not cost me an arm and a leg.

 

[mudman cj]

Surface roughness was a one issue that people trying to get maximum performance out of their computers looked at, people were lapping the base of their heatsinks and their CPUs to get better conduction, does anybody do that with their LEDs yet?

I sure do. I hand lap heatsinks and LED slugs down to a 5 micron finish with 3M diamond lapping film.

 

[yazkaz]

OK folks, can someone tell me the difference between C11000 and C1100?
Couldn't find C10100 or C11000 in my area, only C1100.
Is there a problem with C1100?

 

[PhotonFanatic]

The C10100 is 99.9% Cu, the CU11000 is 99.4% copper; I think they both machine about the same--they might be a little gummy compared to other alloys. Since that's what available to you, no harm it using it.

Complete list of Cxxxxx alloys here.

 

[darkzero]

Not pure silver--that's too soft, but sterling silver, now that's another story. And non-tarnishing sterling silver--yes, that's on the drawing board.

And, for sure, there's no reason not to use it for the heatsink as well. :devil:

:twothumbs :naughty: