If you remember my last VFD clock, I attempted to design an AC filament driver for the display. The reason for this was to avoid using a center-tap transformer. I wanted to keep the clock as small as possible, so jamming a huge transfer into a clock case just wasn't happening. I toyed with the idea of using an 9V AC power adapter, but I'd need to rectify that to DC as well as knock it down to a lower voltage for the logic and the filament. It seemed like a less than optimal solution. In the end I just decided to power the filament off DC. Since the display was multiplexed, it seemed to mitigate the effects commonly seen when one side of the filament is grounded. (i.e., one side of the display being brighter then at the other.)

I searched online for a solution and discovered the LM9022 VFD Filament Driver. Like the name suggests, it's specifically design to supply an AC signal to VFD filament from a DC source. It seemed too good to be true.

It was. The chip has long since been made obsolete. Luckily, I was able to score some on eBay for a few bucks.

The LM9022 has two operating modes: external oscillation and internal oscillation.

Using external oscillation, you have to supply the chip with a signal. Of course, this is very easy to do with an Arduino, but that would mean pissing away an I/O pin. I could have dedicated a cheap ATtiny for the job, but since I was already using an ATtiny to control all the audio stuff, dedicating yet another chip (in addition to the ATmega328 running the whole show) seemed like total overkill. Three micro controllers in a clock? That's just silly.



Luckily, the LM9022 has another mode, when connected using the following schematic, will self-oscillate. Perfect.



I hastily breadboarded the circuit and wasn't the least bit surprised when it didn't work.



That's just horrible. Being frustrated and impatient, and just wanting to push on with the development of the clock, I canned the idea and just powered the filament off DC.



Fast forward 3 months.

I've started developing a new VFD clock using some really large ILC1-1/8L 7-segment displays. I'm really going all-out on this clock, so figured I'd give this chip one more try. I jumped into EAGLE and replicated the schematic called out in the datasheet.



I made the board as small as possible and tried to keep the components as close together as possible. It's a little smaller than a postage stamp.



About two weeks later, I had my boards. I was in no rush to put it together since I wasn't expecting it to work. Besides, I was pretty busy experimenting with the MAX6921 driver chip that I plan to use to drive the displays.



I figured out how to work the chip, had a functioning clock on my breadboard and starting plugging away at some code. All that remains is to nail down how I want to do all the audio stuff, so it seemed like a good point to take a break and with gears while my brain does some thinking in the background.

One night with a few hours to kill, I put together a board and reflowed it in my oven. It looked pretty good.



I connected it to my scope and attached a 3V battery pack.



Boom! Nailed it!



I then plugged it into my board and connected it to 5V and to the filaments of one of my displays.



It's connect to the display on the right. It's not as bright as the other one that's connect directly to 5V and ground. I do have some resistors on all the segment to limit the current, which can be lowered considerably. I put 4.7Ks on there just because they were handy. 100 ohms will do, so I can make them a lot brighter if I wanted to.



In any case, the board works. The only question at this point is, do I incorporate the design into the final board, or just mount a few on the backs of the displays?

See this project from start to finish:
More VFDs!!!
VFD Breakouts
Large VFD Clock - Part I
VFD AC Filament Driver
Large VFD Clock - Part II
VFD AC Filament Driver V1.1
Large VFD Clock - Part III
Large VFD Clock - Part IV
Large VFD Clock - Part V