I'm traveling for the next 6 weeks so don't have access to my desktop computer with all the info on it.
The 2A3-40 is biased at Ep = 300V, Ip = 65mA, Eg = -65V; the 6K5 Ep = 155V, Ip = 1.3mA, Eg = -1.3V. Plate load for the 6K5 is cascoded mosfets. Power out was 3.5W at 1% - 2% THD. 2A3 grid current was the determining factor with respect to power out as it occurred at about -60V. Overall power supply B+ is just over 500V. As I said above, the 6K5 is direct coupled to the 2A3 grid by using a large cathode resistor composed of a two resistor voltage divider (bypassed to the B+). The voltage divider uses the upper resistor to set the bias for the 2A3. The lower resistor sets the plate voltage for the 6K5.
For a 300B I think the B+ would have to be ~600V and the driver would have to be changed to a 6F5 and a source follower employed.
The amplifier is extremely clean with highly detailed sound. I am very impressed with the sound as I have always used a high transconductance triode or triode connected pentode in the past to drive a 2A3. These 6K5s have the cylindrical plates that make the 6F8 and Tungsol 6SN7 tubes so popular. Maybe there is something special about this tube construction.