Still checking things out on the starved plate circuit. The stock setup on the ART has a heater supply of 5.7 volts so I shorted the diode in the ground leg of the three terminal regulator causing the control grid current to drop to 3uA from 5uA. So I may try reducing the heater voltage to the tube a bit more to see where the tradeoff between gain and grid current is. The gain of a single stage is looking to be approx 31 with the maximum signal at the control grid before distortion at 40mVolts peak to peak. So if you build your own box you will need to buffer the signal to the tube due to its low impedance and reduce its amplititude if you want a clean sound before going into distortion.

So you don’t need to boil off very many electrons with a plate current of only 60uA I have too look around but thought I saw a curve on heater voltage and electron emission.
If I can reduce the grid current more a higher value grid resistor can be used thus increasing the input impedance of the tube.