I let this project sit idle for over a month, with the output imbalance issue gnawing a me the whole time ;-) A couple of days ago I put it all back together and came up with a better way to produce a balanced output signal. This version of the circuit adds two resistors to the output side of the modulation transformer, those combine with the 10K balance pot to provide enough deflection plate bias range to properly null out the carrier. I also experimented with lowering the bias voltage on the accelerator element, which adds a tiny amount of limiting action to the tube. That changed the cathode bias, so I adjusted the cathode resistor accordingly. I changed the cathode and deflector bias points down slightly after looking over the tube data sheet some more. (I just noticed that the 2.5V p-p RF carrier raises the grid above the cathode, which should be avoided, I may change to a 750 ohm cathode bias resistor to increase the cathode bias a small amount). The circuit produces a nice sounding signal at this point, it's still far from being something that I'd put on the air.