Driver heat sink

[80stech]

I recently had to do some work on my Rotel Rb-870 and noticed that the driver transistors where originally leaned up against the output transistors with heatsink compound between the two and have now pulled away. My question is: was this done on purpose to make both driver and output run at the similar temps? Should I re-establish the thermal connection or just put separate heat sinks on the drivers?

Cheers,
Rainer S.

AKA 80stech

[ostripper]

This is something I have both encountered on my study of OEM amps and experimenting on my DIY's. I tried running the drivers separate with a standard one device Vbe thermal compensation. I did not have disaster or thermal runaway , but the compensation was affected much more by the environment (5-8mv OPS Re change) . Running with both the outputs and drivers thermally coupled (same heatsink) reduces this to 1-2mv Re. looking at a OEM originally designed to have separate drivers (JVC and Nikko) I saw a PTC thermister in the Vbe to compensate for ambient . That's how they do it !!

In conclusion , if your amp was designed thermally to compensate for the whole OPS (drivers/outputs) as a single thermal coefficient , your Vbe is specifically engineered to compensate for THAT configuration. Changing the thermal "scheme" would also require redesigning the Vbe.

OS

[80stech]

Thanks ostripper, excellent response!! I'll go out of my way to maintain that thermal coupling. I didn't think it would be like Rotel to cheap out by trying to save on some heatsinks! I'll check my bias voltages as well. BTW the bias pots do look very cheap, any reason to upgrade those?

Thanks again for the great response,
Cheers,
Rainer S.