Damn that’s another way to see it.
I always thought that the purpose of the dc bias was because of some parasitic property of the tube that glowed even at 0Vak. Your reasoning makes a lot of sense, I understood perfectly without even a sketch.
The only other reason I can think of for having a negative bias, is to have a faster on->off transition for the segments when using multiplexed tubes, that should avoid ghosting to some extent. Maybe I’m plain wrong.
One of my questions is: are you driving in direct or multiplex mode? Both are possible with the 5812.
If you are driving with multiplex mode, you'll usually have to use a higher supply voltage to get the brightness in the segments, say 50V. If you are driving direct (a nicer solution, IMO) you can drop the supply voltage to around 25V.
The voltage gradient across the filament is around 1.3V, and of course in multiplex mode with a supply of 50V, a 1.3V differential is negligible. Even with 25V it doesn't play much of a role. In my clock, I run 3 tube filaments in series, and even the 4.5V or so drop across the 3 tubes can't be seen.
So, summary: IMO don't worry about the AC or the negative biassing, especially as you are using an HV5812, which has a push-pull output. Just use DC. You'll be fine. The only reason to do more is if you want to have something to experiment with and have the time and will to follow the rabbit all the way down the hole.
Looked at the thread... okay, I know Mr. Nixie and if he says the diode is needed I'm willing to take his word for it. The diode, being a plain old silicon type, drops 0.7 volts across it. Thus the "low" end of the filament is sitting 0.7V above ground. Look at the voltage call-outs on his sketch: he's running 1.6V on his filament but instead of +1.6 on one end and gtound on the other, he puts +2.3V on the high end, 0.7V in the low end, difference is 1.6V to the filament but it is sitting at least 0.7V above ground along the entire length f the filament. Or IOW there is a +0.7V "DC bias" to guarantee full turn-off of segments.
Look into it later when the dust is clearing off the crater.
Yep that’s what I was asking for. I knew all of what you said, so my question was: is it necessary? And also, 0.7v drop across an 1N4007 is not realistic, it would be something between 0.7 and 0.8. I wanted a solution with a known really constant voltage to bias the filament so that I can calculate the components for the regulator.
I believe the most of this discussion is due to the fact that I know most of what I need, but you can’t know that since you’re not in my head ofc, so. Still it is informative to have a word with people in this regard
Hmmm Ty, interesting... On my schematic I've got the dropper diodes all on the positive side of the filament. Putting one on the negative side would probably help with any ghosting. Thanks for pointing that out...