In circuit adjustment of the g2 droping resistor is not too bad if you know it's going to be required from the start. Getting a good quality screen bypass cap is not too difficult either. It aviods the heater cathode issues of the cascode too.

Plate loads for the 6AU6 of ~30k and ~100k seem to work best. The middle ground seems to be less than optimum, but with an adjustable CCS for the cathode load, the experiments can be carried out easily.

Other front end options include the 6EJ7( $3 each ) the 12BY7( $12 ) and the slightly 'spensive 7788( a $35, 45 mA/V gm-monster ).

With pentodes, gain is ~plate load * transcoductance. Even the 6AU6 with 25k has enough gain. EF86 and its octal cousin the 6SJ7 are also contenders. The metal 6SJ7 offers an external shield for free( or more specifically, the expense of lack of a glass bottle ).
regards,
Douglas

I have built with pentode front ends, but not exactly this way. I built a shunt regulator, and had g2 referenced to ground instead of the cathodes. The cascode experiments showed the importance of referencing the upper grid( the faux screen/g2 ) to the cathode node. I have contemplated a resistance loaded EL34 for this job in one of the high voltage amps.

The EL34 is a bit down the road at this point. The more complex cascode seems to be the better option at this point.
regards,
Douglas

Are you guys doing this with the 1624/1625 power stage, or are you doing this with the 6v6 outputs?

Joe

I think I will assemble the next amp this way with the pentode front. It has a better tolerance to transient overvoltage than the MOSFET voltage source does. It is a bunch of trades, and the voltage may not be such a worry if B+ is under 400V. For 650-700, I am going to have some worries with the MOSFET voltage ref. now that I think about it...
regards,
Douglas

I have all the iron, finally. Most of the resistors, leds ect.

Need to get going on the caps, and mosfets.

Maybe by end of summer.

I'll let everyone know when she's finally finished.

Joe