MarkT:
...just check the dissipation in that resistor...
Oops: The dissipation in R2 in the circuit that I showed is over a Watt when the Nixie is turned on. We might increase the value of R3 somewhat (will probably have to increase R3 also) to make the dissipation less. If this is going to be used to multiplex, say, 10 digits, we might get away with something less than a 2 Watt resistor, but remember that software is driving this stuff, and a software bug (or hardware glitch just might let some of the smoke out. It is definitely not acceptable (to me) to dissipate a whole Watt of wasted heat in anything like this. See Footnote. On the other hand, when dealing with 200 Volt power supplies and components that need them, sometimes our way of thinking about system design, not to mention debugging, has to change also.
Maybe someone else will try a more detailed analysis with different values...
I won't go back through the numbers, but if I were doing it for a project of mine, I might start with 330K for R2 and 3K3 for R3. Something like that will reduce the dissipation in R2 and still allow Q2 to saturate, I'm thinking.
Anyhow, thanks for pointing it out.
Regards,
Dave
Footnote:
For people who prefer simulation to back-of-napkin analysis, note that Spice (or SWCad or whatever) will sometimes give you nice looking output values for results obtained from trial-and-error twiddling and fiddling (or as some folks call it "heuristic design"), but will not necessarily point out impractical component values. You still have to remember to check power dissipation and maximum voltage and current values before actually hooking stuff up. Otherwise you may let all of the smoke out of something important.