Eeyore said:
Adequate for the application but I'd never use that little heatsinking on a
commercial amp. Probably not that then.
Commercial amps have to deal with a much wider range of conditions and
making them bomb-proof is a much more difficult job. That is why I was
so surprised when this amplifier, which should have been working under
well defined (relatively easy) conditions, suddenly went 'phut'.
The problem with the heat sinking was lack of space; that was the
biggest heat-sink arrangement which would fit into the available space.
If I had used a more elaborate finned arrangement to give better
transfer to the air, it would have cost a lot more and would have hit
another limit: the amount of air available. (A blower was not an
option).
It is interesting to note that the original valve amplifier, which came
out of that space, had a continuous total dissipation of 250 watts and
was nowhere near its thermal limits. If a certain amount of heat
dissipation is inevitable, a valve take up less space than the
equivalent transistor with all its heat sinks. (I know it's not a fair
comparison in terms of output wattage, but it is an amusing thought.)
Probably the cheapest option for reducing the thermal stress would be to
put two output transistors in parallel on each heat sink. That way the
junction-to-plate thermal resistance would effectively be halved and the
plates could be run safely at over 100 degrees C. The available air
volume would be raised to a higher temperature, so that limitation would
be less of a problem too.
