Fred Bloggs said:
It's coming out exactly right- with the pnp-e at ~0.77 then
Vout=0.5*(5.8+0.77)=3.3V- so the pnp is working at DC anyway. The
5.8V at npn-e means 25.8/2.37=11mA through the emitter resistor
and 18mA through the 140R for 30mA total- this may be a symptom
of open base lead and partial short c-e of some kind- it's shot
for sure.
Nearly, but not quite, it now appears.
In situ, the b-e and c-b looked like diodes, but c-e looked
something like <100 ohms in both directions.
Just now, out on the bench, c-e looked ok in both directions,
and it now works fine as an emitter follower.
So, maybe it was the transistor, maybe it was some resistive
contamination across c-e (that got shifted when the transistor
was changed). The transistor was in a black plastic carrier
with fingers that clamped around the (live) case.
Doesn't matter though, it's done now, and all working ok.
BTW: It was a BSX20, which is an old Mullard TO-18 NPN Si,
Vcbo= 40v, Vceo= 15v, spec'd for 10mA Ic, and 500MHz fT.
I stuck a ZTX653 in there, which is better all round,
except for a 175MHz fT. The 20MHz square wave looks ok
though.