I wouldn't build a separate depoisoner; just build-in to your clocks.
I've had very few issues with IN-18 poisoning. If your tubes have been sitting around unpowered for years, they shouldn't be poisoned and should be ready-to-use. My IN-18 clock runs about 16 hours per day, and at night it runs a 1-hour depoisoning routine that cycles thru the unused digits at a rate of once persecond. Tubes that display 0-9 (such as unit seconds) are left off, because they get uniformly used during the day; tubes that cycle 0 thru 5, such as tens-seconds and tens-minutes, are cycled 6-9. The remaining 10 tubes ( hours, month, day, year) are cycled 0-9 because they are essentially static.
You would think the year tubes would have poisoning problems, but they dont, so the 1-hour cycling seems to be enough. I dont run them at elevated current for depoisoning. The most-troublesome tube is for the unit months, and it can take a few days at the beginning of certain months to fully clean-up.
I've never seen a 5092 show signs of poisoning; they are outstanding tubes. They are used in my first nixie clock, run 24/7, and have no depoisoning algorithm.
My 7971 clock runs a 24/7 depoisoning algorithm when the display is idle (based on a PIR sensor), and runs 1 segment of 1 tube for about 0.3 seconds, cycling thru all 15 segments and all 8 tubes repeatedly. Never saw even a hint of poisoning.