> Mike Harris>>>
>>> HV reed relays may be worth a serious look.
>>> Will be a whole lot easier than a semiconductor solution.
>>
>> I have lots of experience with various 10kV-rated relays,
>> reed-type and otherwise. They're not quite fast enough.
>
> How about a number of lower-rated reed switches in series,
> in the same coil..?
Would that be fast enough for the OP's 500Hz requirement?
Anyway, the problem with reed switches isn't the switch gap, which is in vacuum and typically can handle very high voltages, but with the insulation between the reed's leads and the low-voltage magnetic coil. I've used may HV reed switches made with special potting compounds, only to have them eventually fail with a carbon breakdown path forming inside the potting material between one lead and the coil.
10kV insulation is a considerable challenge over say 5kV.