I think the Berning/Milbert tube amp "impedance converter" is fascinating, but mysterious. So I'm trying to understand it better, with the end goal of building a working model (but puny - maybe 1W or less).
I simplified the schematic from David Berning's patent significantly, and tried to get the results to work in Falstad's circuit simulator here.
I substituted an NPN transistor because the triode didn't seem to work correctly in the sim, and the two DPDT analog switches would have to be replaced by a real MOSFET H-bridge in a working unit, but for the sim I think that's an acceptable substitution. Finally, I have shown the simplistic non-bridged load; for a speaker load obviously you'd need to block DC or use a BTL configuration.
Choosing the right transformer is critical, of course. I don't really want to wind my own so I want to figure out what specs to look for in an off-the-shelf unit, then tailor the rest of the circuit to fit.