Heat loss and insulation

[JR.]

Most power devices publish a theta junction to case. So if you know the case temp and watts of power being dissipated, you can calculate the rise above the measured device package temp to determine the junction temp.

I have always avoided exceeding 150'C junction temps while I am pretty sure some designers exceed that transiently. I was taught that plastic packages were not mechanically stable above 150'C while silicon in metal packages may survive up closer to 200'C where silicon itself degrades. Of course some devices get funky at very high temps and/or can have local hot spots that exceed the average temp.

The old spit test can insure that the case is below 100'C if the spit doesn't boil away. If the case is 100'C or less the heat sink will be <90'C which IIRC is max allowable by UL for humans to touch.

JR

[mediatechnology]

I have always avoided exceeding 150'C junction temps...

...If the case is 100'C or less the heat sink will be <90'C which IIRC is max allowable by UL for humans to touch.

OK, two good rules of thumb.
I'm well below 150C junction temp and right at the maximum unguarded heatsink temp.

What I don't know is the thermal resistance of the dry mica insulator.
I need to look that one up.

[JR.]

I have always avoided exceeding 150'C junction temps...

...If the case is 100'C or less the heat sink will be <90'C which IIRC is max allowable by UL for humans to touch.

OK, two good rules of thumb.
I'm well below 150C junction temp and right at the maximum unguarded heatsink temp.

What I don't know is the thermal resistance of the dry mica insulator.
I need to look that one up.

My recollection is less than 1'C but heat sink grease makes a big difference to fill in air gaps.

I recently had to buy some HS compound to put between the power resistors and my aluminum heat spreader plate for my cooker. I wanted
to make sure as much heat got to the cooker as possible to keep temp rise on the resistors less. I can see discoloration on the wiring to the
power resistors so I know they are getting a lot hotter than the hot plate does.

JR

[mediatechnology]

I once had the "opportunity" to destroy a big extruded aluminum-encased power resistor burning in a power supply for a little too long.
These would not have passed the spit test.
When the ceramic body (Beryllium I think) came out of the housing a lot of zinc oxide titanium white heatsink grease came out with it.
So apparently they also use grease to couple the resistor to the case internally.

It's cold here today and I need to fire up the class-A.
When I run it under load I prefer to warm my hands over the heatsinks.
The load resistors aren't very satisfying.
Such is the nature of class-A.

[JR.]

Maybe I shouldn't call my Class A amp class A because my static bias current will not be very high, my trick circuit just prevents the opposite power device from ever turing off. In theory this will prevent the classic crossover distortion related to time required to turn a power device on.

I do not expect much audible improvement over a well executed conventional amp that delivers respectably low distortion. Nor a lot of room heat... So two strikes agains my approach,.

JR