VFD heater question

Hello Ty,
I assume you mean the Adafruit Ice Tube Clock, don't you? I posses this kit myself but received it through an eBay seller who flashed an alternative firmware. The design has a lot of flaws, one of them is that the filaments are driven with DC voltage which causes a visible brightness drop over the length of the tube. According to Dieter Wächter's IV-18 datasheet translation the filaments need 5V/11mA. The 22R resistor causes a voltage drop of ~0,24V (in the Ice Tube Clock the tube's filaments are connected to the regulated 5V supply through a transistor and this resistor). I assume they used the resistor to limit the peak current when the cold filament is powered up.

Ian,
the resistance of the hot IV-18 filaments should be ~450R. And regarding your second post: Not only the positive voltage does matter - you're right, due to the high voltage there most likely will be no visible effect. But you also will drive the segments slightly below the cathode (=filament) potential, and here every slight deviation could matter.

If you turn off the tubes completely off in some power save mode a constant current source may be helpful, but I drive my tube filaments simply with constant voltage. A simple current source can be built up of two transistors and two resistors. I used such a circuitry in the first incarnation of my VFD clock as the "fresh" tubes had extreme variations in segments brightness and segment currents. I had to learn that this was unnecessary, but I can send you the schematics page of this simple driver.

There are some nice articles about the VFD operation principles, one is at the Noritake page. I remember some other article with the urgent recommendation to to implement some hardware protection in case of multiplexing (like a retriggerable monoflop that will turn off the drivers when multiplexing stops) as when your multiplexer stops the tube will die very quick due to the high anode currents. This can hurt you when debugging your code. I remember darkly that I found that article at Futaba but I can't find it any more.

Regards,
Torsten

Hi Ian,
I just found the application note: www.noritake-itron.jp/eng/cs/appnote/apf140_bdvfd/03-2.html.

Regards,
Torsten

Ian wrote: ... Getting it to work well is tricky, but I'm taking care of that in software by doing all of the non-display processing (e.g. handling the web server requests in the period between digits to allow the capactive residual voltage on the segment bus to decay sufficiently so that there is no bleed over. I'm using only high side drivers, so this is a real problem. Luckily there are lots of other things going on, so I have enough useful work to do between digits.
...

Hi Ian, if you aren't already, you should take a look at github.com/me-no-dev/ESPAsyncTCP. I wrote an async HTTP client, also at github.com/judge2005/ESPAsyncHttpClient and an async WiFi manager: github.com/judge2005/ESPAsyncWiFiManager. Going with an async stack will give you way more time to do other things.

To keep the realtime stuff ticking along, I user a timer0 interrupt handler, then you don't have to try manually interleaving stuff. The downside is that everything called from the interrupt handler has to tagged with the ICACHE_RAM_ATTR attribute, or you will get resets. You can see my drivers at github.com/judge2005/NixieDriver. You can see everything pulled together in github.com/judge2005/SingleTubeClock