Steve,
That's what I always thought, that the device was supposed to be in contact
with a flat surface (flat as much as possible, or machined flat). Once anything
is introduced between the device and the metal, the heat transfer starts to be
interupted. For RF power transistors, I would think one would want the very
minimum amount of compound between the device and the heatsink. Anodizing
is about the same process that's done to the foil on an electrolytic capacitor
except a dye is introduced. The anodizing itself is what creates the insulative
layer, not the dye per say. It may add to it, I don't know. The way I was
always
taught was black absorbs heat, and anything bright like white or silver
reflects it.
I would think the mass or thickness directly under the device is what will help
keep it cool, and the find transfer the heat to cool it off by the air
circulating past
them. I can't see where black would help this at all, unless they say the black
helps with the heat transfer to the air. If so, what does the heatsink to with
the
extra heat it obtains from absorbing infrared radiation if black? I surely
don't see
how it could help heat transfer from the device to the heatsink itself.
A test for anyone that would want to try it is take two pieces of identical
heatsink,
one painted black, one bright (not painted), and place them outside in the
sun. After about an hour, check the temperature of each. I believe I'd want the
coolest one in contact with the transistors myself.
Best,
Will
*********** REPLY SEPARATOR ***********
On 8/16/06 at 9:03 AM Steve Thompson wrote:
>Will Matney wrote:
>> Peter,
>>
>> I agree on the math. The problem I see on using a black heatsink is that
>even
>> though all heat is infrared, and there's a concentrated heat transfer
>from the device
>> to the heatsink, using black supposedly helps this transfer.
>Getting heat from the device into the heatsink is all about contact
>transfer. As far as I'm aware, any surface treatment on aluminium
>increases thermal resistance so makes the transfer worse. The best
>surface for contact with the device is plain metal, machined to a
>flatness of .001"/inch or better, with a very low surface roughness -
>the closer to polished the better.
>
>Thermal compound has a thermal resistance only slightly better than air.
>Ideally there should be a tiny amount of it, just enough to fill
>microscopic crevices. It just doesn't work to take out gaps between the
>device and the heatsink.
>
>Steve
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|