Japanese solid state amps

[James]

ESR in big caps tends to drop a bit as they run warmer. Which will take a while from a cold start.

In the case of my M-22, that'll be almost never. The quad of 33,000uF caps are functionally mounted in the open and away from the power transistors or heatsinks. Truly a case of design dictating form.

 

[Arkless Electronics]

Thanks for your views. I have been using the much vaunted class A/B Dynavector HX-1.2mk2 over the last couple of days, just to keep its caps formed, before swapping the M-22 back in. Again, I'm reminded there is a very audible improvement in clarity, listenability and naturalness from the class A amp. I seem to recall that the M-22 has temperature sensors on the heatsinks and is designed to self-adjust as it warms up or cools down. Hopefully, those mechanisms are in good order as the way it sounds suggest.I keep my fans running at their lowest speeds, and they seem not to compromise the sonic goodness.

The vast majority of amps have this!! 90%+
As I said earlier it's really important in class A/B SS amps, that don't necessarily get very hot. A class A amp biased at say 1A won't be bothered at all if it goes to 1.2A or 0.9A, it's still in class A. With a class A/B amp which is likely to be biased to around 0.005 - 0.002A OR 0.05 - 0.08A depending on which of the main topologies have been used, it is critical that the bias is kept accurate and common for the distortion quantity and spectra to change somewhat from minute to minute due to this not being as accurate as we would like. There is a narrow "sweet spot" in a class A/B amp (usually not exactly sweet but much better than if it drifts outside the range!) where just enough quiescent current has been applied for crossover distortion to have fallen to a reasonably low level, but not quite enough for distortion to rise again due to transconductance doubling distortion.
As the temperature sensor (usually just another transistor bolted to the heatsink) can only react "after the fact", due to it taking some time for the heatsink to heat up or cool down, a typical class A/B amp is almost always suffering from too much or too little bias depending on the recent history of the signal.... not ideal!

 

[martin clark]

In the case of my M-22, that'll be almost never. The quad of 33,000uF caps are functionally mounted in the open and away from the power transistors or heatsinks. Truly a case of design dictating form.

It's the internal temp that matters, and yes, that will rise owing to Iripple_rms^2 * esr. But mounting reservoir caps away/outside of hot casework for a class-A amp is a solid measure for cap lifespan & reliability!