Kevin,
Thanks for your input. I don't know if you saw this updated schematic, its really not too different except for slightly different operating points for the 300B tubes and only "one" current source shared for the 300B's instead of separate ones (really four cascode CCS's in parallel to relieve thermal stress like you said).
Its still a compromised design using 300B's as output tubes. If I were to use 45's or 2A3's, it might work better, but I need a good 12 watts of power. A beefy preamp will be required, but most I've built can swing a lot of volts.
I just don't know what the 6N6P sounds like at 150V plate, 5V cathode and at 20mA current.
I should built it with a bunch of cheap parts first, and then massage the circuit.
Another option is 160V, 24mA and 5V. At that point I would change the plate choke to a Lundahl 1668 amorphous core. Very close to the dissipation limits of the 6N6P though.
Even another option is to use the other half of the 6N6P in parallel so my plate resistance is halved with the same mu, but the need for twice the current (and a different plate choke)! This may overcome the drive impedance imposed by the phase splitting choke.
Anand.
