a pre-conditioner (the 1-st of 2) for 9V linear reg. amazingly falls to that 10% of circuits that GO as they simulate !
an error-trial fast compile of the 9V from 12.8V - even more amazingly the circuit falls to that 2% of circuits that actually GO better than they simulate -- this 1 has 2...3mV output change on load variation -- donno what happens if we vary the input . . . likely the bias increases decreases along the V.in resulting unspec. change for V.out . . . not tested -- the russian KT815 mounted on about (both sides tot.) 1²' (1side tot.) ½²' heatsink (that seems to be sufficient for <avg.100mA <peak.240mA I.out)