Reduction of Ripple Voltage in a Dynamitron

We determined that a precise neutralization of the RF ripple voltage on the high voltage terminal of a Dynamitron has previously been prevented by a nonnegligible phase shift of RF currents in the two halves of the ~l00 kHz class C oscillator tank circuit, which is actually constituted of two slightly unequal high-Q coupled circuits because it has two ground points: the inescapable center-tap-ground in the capacitive legs and a center-tap-ground lead to the induction coil. The latter is needed to prevent damage by flashover transients; equivalent to its removal was the adjusting of RF ground return current to a null by aid of a current transformer on this lead and the suitable adjusting of trimmer capacitance. While the phase shift was thus held to a null, the actual ripple amplitude on the hv terminal was minimized by adjusting additional trimmer capacitances installed in the terminal of the machine. Then p/p 100 kHz ripple at 2 MV DC output was reduced to about 50V and RMS resolution by (p,¿) resonance threshold data near 1 MV was about 250 V. The limit to resolution has various causes including mechanical vibrations and unbalanced harmonics of the RF.