When does my class Atoll IN200 A/B integrated switch from class A to class B?

Thanks for the reality check. And the suggestion. 

I wrote to Atoll and conveyed the skepticism posted here about the Class A operation of the Atoll. They replied:

From Atoll engineers:
"Hello, Thank for for this information.
This is one of the most current mistakes people make about our amplifiers; they compare the size of our heat-sink to a single one; this is nonsense regarding to thermal dissipation because when you use a single, big heat-sink you have to put in isolation components that increase (a lot) the thermal resistor. For that reason, and on that point they are correct -- in order to have the same dissipation you would require an almost 10 times bigger heat sink for each channel; but with our trick, the transistors are transferring about 100% of their heat to the heat-sink with no waste of energy. You'll find attached a measurement that shows the distortion when power is rising; as you can see (in the picture), the distortion is under 0,002% up to 50W on both channel. Best regards. ATOLL Team"

I don't think I can post pictures here, but anyone who'd like to see it can P.M. me.

Quick follow up. I discovered this comment about their more advanced amp and I sent it to Atoll:

What I found said,

“The high bias allows the IN400SE to operate in a Class A for the first ten watts before going into Class AB operation."
http://www.theabsolutesound.com/articles/atoll-electronique-in400se-integrated-amplifier/

And I asked Atoll if they were claiming the my IN200SE operates to 40w in class A while a later and more expensive model only operates in Class A for the first 10 watts? If that is true, why did you not employ this technology in the later amp?""

They replied: "The IN400se use almost the same technology as the IN200se so both work in class A up to the third of the maximum power."

So, they're in contradiction. Is it Class A up to 10w or to 40w? They seem to say both but insist (to me) that it's 40w in Class A. Despite the lack of heat, fins, etc.

They're just jerking me around at this point, right?

I think they are jerking you around. Eliminating the electrical insulating material between the transistor and heatsink will reduce the thermal impedance, allowing a bit more of the transistors heat to be transferred to the heatsink, but this doesn't change the laws of thermodynamics. The heat still has to go somewhere. 

This would allow the heatsinks to run a little hotter without risk of damaging the transistors - say 60 or 65 degrees C. But even so, heatsinks this small would not allow nearly enough heat to be dissipated for 40W class A operation. And your case would be so hot that you wouldn't be able to touch it for more than a second or two.

Take a look at the Pass Int-60. This is rated at only 60w/ch into 8 ohm (although that's pretty conservative) and has rail voltages considerably less than your amp, which means the transistors need to dissipate less power at the same idle current. The Int-60 will do class A up to 30w/ch. Now take a look at the size of the heatsinks on this amp. 

Yes.  10W sounds exactly right for the '400 model.

Coda is an interesting company in that it has models with different flavors, i.e. with different levels of class A operation.  Unfortunately this isn't a switch on the face plate; you choose which version when ordering.