as it was Gary's 47 amp which led me to the E-Linear connection discovery. I found that the amp ran into headroom troubles when we ran right off of the plate, hence the use of the screen tap for driver voltage. I did not attempt it with pentodes so there is likely not such a problem as with triodes.My amp is a bit simpler than that, but probably similar in performance( IMO, likely better, but I am biased, no? ). It does not have the direct coupling issues( but instead, capacitor ones ). there are a few differences in execution, but for thigs like filament TX choice, Gary has got it nailed. I also think 10k a-a is a better load for U-L connected finals. I am also doing away with the input TX. The cascode differential amp/long-tail-pair phase splitter will handle that better IMO.
regards,
Douglas