> I was told that as usual switching amps beyond class C were conceived of in pre-sand days. Don't know anyone who's been fool enough
Switchers are linears with excess gain.
Tubes don't have enough gain to be killer switchers.
Sure, it could be done to prove a point. And vacuum-tube digital computers were well developed (though when you get down to details, they were awful close to their linear roots).
Sure, the power gain of a vacuum tube is close to infinity... at DC. Gain-bandwidth for the common types is 10MHz, which means they don't have a lot of gain at 1MHz broadband. Since the input impedance is largely reactive, tuned narrow-band systems can work faster, but that cripples many switcher techniques.
> duty cycle of semiconductors (or vacuum tubes) as they alternate between high conduction and off states."
Vacuum tubes don't have a high conduction state. The juiciest vacuum is high resistance compared to the skinniest wire. Since most loads are wound wires, tubes don't conduct well into loads.
> Why one would use a filiament-device to make a hi-efficiency amp is beyond me
To avoid the indispensable output transformer. Just as your PC power supply replaces the 60Hz power transformer with a much smaller and more efficient (and cheaper) 25KHz transformer, it might be possible to transform audio at supersonic rates and avoid the big iron. If you have a 500VDC power rail handy, and just want a DC output, it is easy (though I have never seen it done). Switch-mode audio power transformation is harder: you tend to need a many-KHz transformer with response to 20Hz, which is harder than a plain linear transformer.
Switchers are linears with excess gain.
Tubes don't have enough gain to be killer switchers.
Sure, it could be done to prove a point. And vacuum-tube digital computers were well developed (though when you get down to details, they were awful close to their linear roots).
Sure, the power gain of a vacuum tube is close to infinity... at DC. Gain-bandwidth for the common types is 10MHz, which means they don't have a lot of gain at 1MHz broadband. Since the input impedance is largely reactive, tuned narrow-band systems can work faster, but that cripples many switcher techniques.
> duty cycle of semiconductors (or vacuum tubes) as they alternate between high conduction and off states."
Vacuum tubes don't have a high conduction state. The juiciest vacuum is high resistance compared to the skinniest wire. Since most loads are wound wires, tubes don't conduct well into loads.
> Why one would use a filiament-device to make a hi-efficiency amp is beyond me
To avoid the indispensable output transformer. Just as your PC power supply replaces the 60Hz power transformer with a much smaller and more efficient (and cheaper) 25KHz transformer, it might be possible to transform audio at supersonic rates and avoid the big iron. If you have a 500VDC power rail handy, and just want a DC output, it is easy (though I have never seen it done). Switch-mode audio power transformation is harder: you tend to need a many-KHz transformer with response to 20Hz, which is harder than a plain linear transformer.
