Pull them up to what? I have experimented with clamping them to 75V using 330K resistors. The context was a multiplexed clock on a breadboard, so trying to avoid ghosting rather than avoiding blue dots in IN-18. My prototype PCB has slots either for pull-up resistors or diodes. I haven’t finished these particular experiments.
I’m not sure whether diodes are really necessary, I haven’t used them yet, but I wanted to compare the effects to using resistors.
It does seem like a lot of extra components for a direct drive clock.
Ian wrote: H Paul, I've been thinking about adding some high value (>=220k) pull-ups to the unlit cathodes. Do you have any opinon on that? It's a bit of a sweat and a lot of extra components, and propbably not at all necessary, but this is a recurring problem, and always causes consternation.
Another solution would be push-pull drivers, I suppose... The HV5622 is open-drain, nad there does not seem to be an equivalent part in push-pull (80V seems to be the limit).
I just came across this old post searching for IN-18 maintenance voltages. Anyway, regarding using diodes instead of pull-up resistors, it is a waste of time.
Regarding using push-pull, the HV9808 works fine if you clamp the voltage to 75V - that is the voltage for the off anodes - so it stays within its spec, plus stops the off anodes floating. I haven't tried it for IN-18.
I recently built a clock that can drive IN-18. It is direct drive and is driving them at 6mA using a HV5523 - so open-drain. It works fine, but maybe I just got lucky with my IN-18 tubes.