Alek, I don't think the opto idea would work.? ?All the resistors are in series, and regardless of value..... the current is the same. (current in a series circuit is the same).?
IE: say 7500 vdc supply, and 25 x 100k @ 3 watt mof resistor is series.... then add a 26th resistor (1666 ohms).? Current in the string is just 3 ma.? ?3 ma x 1666 ohms = 5 vdc.

If an opto, whose input side operated on 5 vdc..... the same opto will draw a lot more than 3 ma on it's input side, to function.? ?Most are typ 10-20 ma on the input side.? ?Where is that current going to come from??

I may well be missing something here.?

?

And for K6LPM, opto's are alive and well, used all the time.? ?The ODC-22's I use to key the sped up vac relays operate in 100u secs....and rls in 750u seconds.? They also have 5 kv isolation between the input side and the output side.? ? They came in a choice of? 9-15-24 vdc input sides? (with each of those having a wide variance)? and a choice of either 60 vdc @ 3 amps ccs...or 200 vdc @ 1 amp for the sec side.? ?That's just 1 opto.? There are hundreds more to pick from.?

Switch transistors are a lot faster.? Like 3.6 microseconds, 300V rated at 500ma.? $1.85 from mouser.? ?My issue with switch transistors is.... IF they ever failed shorted, then the entire OCV? (open circuit voltage) is applied to the xcvr. IMO, the huge isolation between input and output of the opto is a huge advantage.? The input and output are totally isolated from each other.? The opto part is just the laser / light firing a beam from one side to the other.? There is no mechanical or electrical connection between the input and output.? ?The output side is essentially just a SPST-NO electronic switch.?

?