Tom,
You modulator efficiency is a bit lower due to the slower turn on time of the tubes to get analog compensation. I think you are measuring at the saturation voltage of the switch which is quite good.
Yep, agreed. I believe Patrick was quoting a theoretical number in the post above.
I actually measured 90.1% efficiency BEFORE I added the analog compensation. I did it by shorting the switch tube to ground and then letting it run full on and comparing the two.
Of course, this does not include the losses of screen current, grid current, filaments and the pulse slope loss from the analog comp - which will bring the overall efficiency lower. But all tube modulators have to deal with this anyway - even the slope loss, in the form of class B operation.
T