I think you should buy some (maybe a box of ? ) transistors and a bigger heatsink and some plain board and read up and start from there?
?
When messing with input stages ?you¡¯ll find the input impedance all over the place (hence the pi attenuator input)
?
When messing with output stages the dissipation will increase and the impedance varies with the point in the cycle (so the LPF action is degraded) and remember the 50W amp has never been wideband?
?
Remember the interesting way Hans avoided RF feedback at switching points by boosting the switching diode bias from the RF waveform - that probably needs checking ?too
?
?