I have used the 3 diode, power supply (variac) and load resistor setup on various rigs, the first one was the pair of 813's modulated by a pair of 4x150 (4cx250b) tubes in AB1. The circuit worked very well on the mod monitor and scope, and I could get ungodly amounts of positive modulation. Like 2500 to 3000 watts pep out of a 600 watt carrier, without any audible distortion....
I suspect the high voltage brick diodes I use have some sort of problem, switching speed? that caused very bad splatter when I would exceed 100% neg modulation. I saw no signs of it on the scope or mod monitor, but it showed up on the spectrum analyzer. I also put the circuit in the push pull 812a rig, and the pair of 4d32 rf deck, with the same results. Unless you are running a mod transformer within an inch of its life, I don't see the need to protect it that way. I insulate the mod iron from ground, and never had problems. I once had a qso with someone (cant remember who) who told me the circuit was no good, but would work with the addition of another diode...I think that is what he said... I suppose the solid state 14kv 1 amp bricks just can not react fast enough and let a real short pulse through? In the old days, they used tube type rectifiers, and they may have worked better, just a guess... I thought about trying to add a (very high voltage) capacitor someplace in the circuit to damp out any spikes, but never looked into it. I now tend to think limiting the audio in the low level stages and not exceeding 90% negative modulation is the cleanest way to go. Brett N2DTS
