L-Profile for output transistors-Thickness???

[zox2003]

My another question:

If you have output transistors mounted on L-Aluminum Profile is it better for temperature transfer to have that profile thicker or thiner?

Regards

[suzyj]

Thicker, clearly.

You can calculate the thermal resistance of a given bracket, using the cross-sectional area and the mean distance between the transistors and the heatsink.

I used 40x40x6mm aluminium L shaped extrusion for my 100W amp.

Cheers,

Suzy

[AndrewT]

Hi,
thicker, to reduce the thermal resistance.

try for at least 6mm but preferably 8mm. The legs on emitter and base resistrict the thicker sizes.

An unequal angle with the To3 mounted on the short leg is better, or saw the To3 leg shorter.

An inverted T bar is better still. the top of the T mounts to the sink and the leg takes the device. Use bolts on both sides of the leg to pull down the top flange flat.

Use thermal compound on the bar to sink interface.

[jackinnj]

Originally posted by AndrewT
Hi,
thicker, to reduce the thermal resistance.

try for at least 6mm but preferably 8mm. The legs on emitter and base resistrict the thicker sizes.

An unequal angle with the To3 mounted on the short leg is better, or saw the To3 leg shorter.

An inverted T bar is better still. the top of the T mounts to the sink and the leg takes the device. Use bolts on both sides of the leg to pull down the top flange flat.

Use thermal compound on the bar to sink interface.

The thickness is needed for rigidity. The mass of the object will absorb energy, but not infinitely so -- after you load it with enough energy it will either radiate or change state.

Thermal impedance is linearly proportionate to thickness...in other words, heat transfer is inversely proportionate to thickness:

[AndrewT]

Hi Jac,
I think I understand what you are inferring. If I have read it correctly then you need to think again.

Please re-phrase so that we can clearly understand your message.

BTW. the posted chart is OK.

[jackinnj]

Which will get rid of more heat, a cube of aluminum 1 metre volume, or that same volume of aluminum as a plane 10 millimeters thick?

[AndrewT]

Hi Jack,
there's two different questions in there and they give conflicting answers, that will only confuse those that don't have a full grasp of the theory.

Back to the first topic.

Is it better to go for a thick mounting or thin?
I have said thick and I stand by that answer.

Where do you stand?
Your response in post4 implies that thinner is better, but I know you did not mean that. That's why I gave you the chance to re-consider.

[anatech]

Hi Andrew,
I'll support reasonably thick. We are attempting to trasfer the heat to the radiator (heatsink).

-Chris

[Wavebourn]

Once I glued by epoxe to the toroidal tranny poles cut from 6mm thick aluminum wire, it looked like a hedgehog. For better convection space between them was 8mm. However, in case of forced air cooling they could be glued closer to each other.

Similarly, you can make your L-shape parts with ribs (bent edges), both for rugidity and for better convection.

[Wavebourn]

Oops... Sorry, I misread the question (transformer vs transistor)

Yes, thicker brackets are better, especially when the square of their contact with the heatsink is bigger. Better to use the single long angle of the whole length of the heatsink made of the same material (aluminum angle with aluminum heatsink, copper angle with copper heatsink), if transistors are mouinted in the one row.