Ronenash -- Thanks for providing your adventures with the D115. Working on one is always a memorable experience!
I would offer a couple of comments to the conversation:
1. Your modification to run the unit with balanced inputs no doubt returned superb results. In my own project, I had never had an ARC piece on my bench before, so I was focused on just trying to get it to square one. However, it was rather obvious from the get-go that the goal of the previous owner of the "Parts and Pieces" collection was also in fact to modify the design to provide for a balanced input, since all of the phase inverter components previously installed had been removed. Of course, he never got very far with that effort as the pics showed for the condition it was in when I purchased it.
It was quite interesting to me that other than for compensating for the heater issue, the unit actually operates quite well with a single ended input and the phase inverter tube removed (!) -- because the cross-coupled stages effectively then became the phase inverters. Performance was certainly better with the phase inverter tube in place, but the unit would still deliver nearly 90 watts per channel (both driven) with the phase inverter tube removed from its socket.
2. I have no doubt that eliminating the phase inverter stage did in fact eliminate the audible cannon shot thump at shut down, as then all the remaining amplifier stages operate in a fully symmetrical fashion with their counterpart on the opposite side of the push-pull connection. Any bumps and thumps created on each side then cancel out in the OPT to effectively eliminate any such pulses in the speaker. In terms of saving the speakers connected to these amplifiers then, going to a balanced input will certainly get the job done.
What I would add however is that while that does save the speakers, it still leaves the output tubes subjected to the large pulses created when the regulators powering the small signal stages -- with their high gain error feedback loops -- collapse out of regulation at shut down. In my unit, tests at the time showed that while removing the phase inverter tube did in fact eliminate the speaker pulse, all the output tubes still endured a significant inaudible shutdown pulse that in my experience, proves to be a real output tube killer. It is particularly hard on the tubes in the D115 because of the huge capacitance in the main B+ supply maintaining significant B+ voltage levels even when the power is initially removed, and the number of regulators involved, coupled with the output tube heaters cooling down with the power removed from them. This causes the electron cloud that forms around the cathode which normally supplies all of the electrons needed for normal current flow in the tubes to quickly diminish, allowing any large pulses during this time to draw electrons directly from the cathode, which causes significant damage to it. In my own tests many years ago, brand new GE 6550A tubes subjected to similar conditions were effectively dead after just 250 hours of otherwise normal power cycling and use. For this reason, I chose to resolve the pulse in the manner I did with my own D115 to eliminate the turn off pulse it otherwise creates. As a result, the original Sylvania manufactured 6550A tubes I first installed in the amplifier are still installed, none have failed with normal weekly power cycling, and at last test (@ ~ 2200 hrs of use), all still produce an average of 94% of Average New Power Output levels for vintage American manufactured tubes in my power output tube tester. Most D115 amplifiers have either blown up or burned out numerous sets of output tubes long before this amount of use time has been seen.
Dave
