Nixie cathode poison techniques...

[alank2]

My order from www.tayloredge.com arrived, so I can finally generate 170V and have a nixie to play with.  Mr. Taylor was very cool.

It has a dimming feature that is running at 100 Hz and does a duty cycle of 1% to 100%.  You can go down to 50% or even 30% without losing a huge amount of brightness.  I can see how you could use this to do an animation of moving from one digit to another if you wanted to like some of the YouTube videos I've seen of running nixie clocks.

My question has to do with how to prevent cathode poisoning.  I've seen the method where it cycles around to get to the digit, but I'm not feeling that so much.  I know that could be done in the middle of the night, and I like that better, but I have another question:  How long does it take to prevent the buildup on an unused digit?  I think I'd prefer to keep the main digit displayed and rotate through the unused ones very quickly at 1 to 2 milliseconds each.  Or perhaps while transitioning between two numbers, light an unused for 1 to 2 mS.  How much do the unused digits need to be active per day roughly, any guesses?  Any other thoughts?

[woodchips]

My understanding is that a clock showing hours and minutes isn't any sort of problem, the digits change sufficiently frequently. The problem comes if your clock shows date and year, then you will get problems with these digits. I have not come across and reference to the technique of rapidly multiplexing the unchanging digits as being useful.

[jazer]

Simple NE-2s get damaged easily enough if fed DC (flickery in months or less and dead soon after), so I think nixies might have trouble, and not just for years etc. digits.

Any consensus on the best modern chips to drive nixies? Old drivers are hard to come by and lots of discretes aren't appetizing.