On the rectifier thing, we need four diodes for a full bridge, since there is no certer tap to ground( the grounded CT, and two diodes is basically two out of phase half wave rectifiers ). Two diodes( one at each end ) form the curent path, one with the banded end to the filter, and the other with the un-banded end to ground. We keep the two with the unbanded ends to ground, and put a pair of vacuum diodes in place of the others, with the cathode( the hot part in a valve, banded end for SS ) to the filter. It is PN-junction noise free due to the vacuum diodes.

this made sense while I wrote it...and still when I read it. I have my fingers crossed.
regards,
Douglas

If there are various concepts / variations on the power supply side, please share. How easily are they simulated in PSU Designer? (I think that's the name.) Can we just pick one PS to go through the mathematical work-up? What other parameters of the PS can be calculated and can we discuss why they are important? What should we look for in terms of sourcing parts? I would also like to see theoretical discussion, not entirely by you (way too much effort for one person,) of the signal path. I realize this is old hat for you, but there are new guys (OK, me) out there that really want to learn this stuff. If they are anything like me, hands on is a much easier way to learn.

I realize the Holidays are upon us, but I think we can start some of the "thought experiments" anytime......Colin

The other sticking points to a later iteration don't seem like sticking points anymore to me. Build the simplest one and they seem to evaporate with familiarity and practice.

The linestage sounds good. I have not listened to it with the single MOSFET plate load CCS in a ling time. It is easy to modify it to a cascode circuit later. The single *IS* simpler and easier to deal with at first. I would not have suggested it if I were not confident in the result.
regards,
Douglas