Brett, thanks for the kind words. Yes, my grandfather was very serious about attention to detail. He liked things balanced and symmetrical, maybe attributable to his other passions: Astronomy, Building reflecting telescopes, and boat building with sheet steel, typical were 36 foot or so cruisers and sailboats. He taught me to solder when I was four, and then at eight he taught me oxy-acetylene and electric arc welding. I spent a great deal of time on the restoration, no small amount was invested in locating and repairing vintage matching meters. Unfortunately, some of that 250-TH rig did not make the trip from NJ to FL when dad retired. We no longer have much of the original power supply, or the exciter/driver deck using the 35-T and 100-TH that was mounted directly below the final deck as seen in the earlier photos. You can see in the recent restoration that the modulator was moved up directly below the final. It used to have metering for the final grid and plate current, plate voltage, and plate current for each of the 810 modulator tubes. There was no overload protection, and no fuses or circuit breakers anywhere in the rig. The final used just grid leak bias, and the modulator a dry battery under the chassis. When I restored it, I added regulated bias supplies, with grid voltage adjustable for each individual modulator tube. Over-current trip relays with indicators were added as well. Adjustable filament voltage with metering was added for mod and final tubes. Variac for plate voltage, and separate rectifier and filter circuits for mod and final. I restored the mod and final decks pretty much as they were originally. Circuit breakers and fuses were added where appropriate. The mod deck has an input transformer with 500 ohm primary, and uses the high-end "CHT" (Chester H Thordarson) iron, very pretty units. A separate speech amplifier was used, and I restored a 1955 version of the UTC MLF (multiple loop feedback) amplifier for this function. The engineers at UTC spent the better part of a year analyzing the Williamson, and several other designs, and determined that large amounts of global feedback actually caused problems such as ringing, overshoot, instability, etc. These problems were apparent when using a step-function input signal, but they were obscured when using a typical continuous sine wave input for testing. They arrived at a combination of a small amount of global feedback, and three local feedback circuits throughout the amplifier. This resulted in superb stability, low distortion, and low output impedance which turns out to be ideal for driving triodes into grid current. This is quite an amazing feat for 5881s (5L6s). There were at least two articles available in periodicals describing the work, and I can make them available if you are interested. I noticed in another thread that you were using local feedback, but you were not a strong proponent of large amounts of global feedback.