annodomini2 wrote: If the pulse rate can be efficiently varied, it may even be able to load balance at the generation point.

it may be somewhat variable, but there are constraints: firstly, on waste heat dissipation, which tends to reduce the pulse rate; and on efficiency, since the gases in the chamber will remain ionized only for short intervals (at most 1 ms), and re-heating them costs energy, so this tends to increase the pulse rate.

lots of our back-of-the-envelope calculations assumed a rate that was divisible by both 50 and 60 Hz, to accommodate both north american and European systems;
consequently you will see the number 3000 pulses/second tossed around in the forums. but any number > 1000 that’s divisible by 300 would meet all these criteria.
so, 1200, 1500, 1800, 2100, … etc with upper bound fixed by heat dissipation.