Original poster: Steve Conner <mailto:steve@xxxxxxxxxxxx><steve@xxxxxxxxxxxx>
> In order to keep it a bit nearer to that potential,
> I've included a 1 uF capacitor from that point to
> mains-common;
Ken, this sounds wrong to me. You will be developing
something like 5,000v across the primary coil and tank
capacitor. The ends of the tank circuit that connect
to the H-bridge can't move more than +/-350V, so that
midpoint must be jumping around by thousands of volts,
not zero volts like you say.
I would suggest deleting that capacitor and making
sure the junction between tank cap and primary coil is
well insulated from anything else. Also moving your
feedback coil so it's directly on one of the H-bridge
outputs, then it only gets +/-350V of electrostatic
interference rather than +/-5kV.
Steve Conner