Just sharing my latest project with VFD display

[Rolo]

Hi, would like to share the project I'm working on right now, it's driving a VFD display. I know, these are not the latest technology and if you want to display something it's more easy to get a LCD or LCD TFT panel. But I just like the glowing VFD display's and want to get a grip on how to drive these puppies.  This is the setup I have sofar, a filament driver based on a 555 and a simple high voltage generator, also based on a 555. In this picture I'm driving all anodes and grids static. I'm waiting for parts and PCB's now. The idea is to make panel with this display and shiftregisters so I can drive it with serial data.
To be continued.

[QuantumLogic]

That looks like a fun project. 
I also love VFD displays and their awesome glow but haven't attempted to drive one yet.

[Ian.M]

So you've got an 8 character 5x7 matrix.  That's 35 anodes and 8 grids, so you need 43 HV drivers.  How are you planning to do that?   I've done 12 electrodes with a mix of IC shift registers, and level shifters with discrete high side drivers, but I certainly wouldn't want to extend that to 43 electrodes.  I've got a similar 16 character display waiting for a rainy day that's going to need 51 drivers.

For reference:
* What's the display's make and model no.?
* Do you have a datasheet? (If so, what's the URL?)
* What RMS filament voltage are you using?
* What anode voltage are you using?

[joeqsmith]

I admit, I have always liked the looks of VFDs as well. Looking forward to seeing what you come up with.

[Rolo]

Some more info :

The display is a BOE VFD 20-0814F, these are for sale on ebay by several sellers.
They are new old stock, I do have a datasheet but it's not the exact same type (I think), the last letters are different.
When I apply the filament voltage as in the PDF (2.7V AC) the filament glows very visible, that can't be good.
I now have 2V RMS at 100kHz (measured with my scope) and that works fine so far in my test setup.
For the anode voltage I now have about 18V, that just for testing in static mode. When driving it dynamicly I will have to bias the filament voltage to prevent ghosting. This bias voltage will add to the high voltage.
The datasheet shows 31V max. So I need at least 31V + 5V = 36V. 
I'm planning to modify a common step-up convertor to go up to about 45 volt. I want all to be powered by 5V DC.

For driving I want to use three HV5812 (or the A6812, that's an obsolete compatible part, can still be found).
The anodes and grids are tied to the outputs of the HV5812's, they are both driven with the same high voltage.
So I shift in all the bits for a specific digit and grid, strobe the latches, wait for some time and do the next digit.
Dimming will be done by PWM on the blanking pin of the HV5812.
That's the theory. Will have to see if it works, waiting for the parts and PCB now.

I also want to write the software to show nice scrolling text, but first concentrate on driving the display.
On the long run I would like to make a sort of remote display for my domotics system,
scrolling some sensor value's, time, date and other information.

b.t.w. I do like these rolling numbers :

[Ian.M]

I doubt you'll need the full voltage.  Set up your test jig to bias the filament at +5V,  PWM all the grids together with a 1/8 duty cycle, and see what voltage you need for adequate brightness.  Excessive anode voltage degrades the phosphor.

The filament should just have a visible glow if observed with the display blanked in total darkness.   Underrunning it is nearly as bad as overrunning it.   

[Rolo]

I doubt you'll need the full voltage.  Set up your test jig to bias the filament at +5V,  PWM all the grids together with a 1/8 duty cycle, and see what voltage you need for adequate brightness.  Excessive anode voltage degrades the phosphor.

The filament should just have a visible glow if observed with the display blanked in total darkness.   Underrunning it is nearly as bad as overrunning it.

Thanks for the info, I wil build a test setup using a 74HC238 mux and TD62783A 50V driver to drive the grids, my PSU goes to 30V, so minus the 5V bias I can test with 25V effective high voltage. I have these parts laying arround somewhere. Will post on how that turns out. Do you mean powering all the anodes and multiplexing the grids ?
The filament is just visible in complete darkness, I have tried to capture it with my phone camera, I think I hit the right value there, I can adjust it by changing a resistor in the primary coil of the transformer.

[Ian.M]

Yes, all anodes tied to +HT. Although you could multiplex the grids, for brightness testing it would be much simpler to just tie them together and PWM them.  If you cant find the TD62783A, here's a discrete HV driver that will do the job.

I posted about it [here].  You only need one of them + a 555 to drive all the grids for brightness testing.

I can see the breadboard wiring which means the room isn't totally dark. Therefore IMHO those filaments are possibly a little on the bright side.

[Rolo]

This is when all anodes are tied to +32V (minus 5V for bias gives 27V effective). The grids are driven by a TD62783A driver on 1/8 PWM from an arduino UNO.
Alle grids have there own driver, tied together on the logic side to the PWM pin.  My PSU won't go higher, I think it's reasonably bright, have to check in daylight. This is under my desk light conditions.

[Ian.M]

That looks pretty good.  Its hard to judge brightness from photos, but if it turns out to be acceptable in daylight without a filter, that's plenty.

[Rolo]

Update : This is the filament PSU I'm planning to use. Not my design, I found it here http://magictale.com/2603/revisiting-vfd-psu-part-ii/.
I did add the virtual center tap to bias the filament to 4.7V. The theory is that this will avoid ghosting. The PSU works well, supply at 5V and no need for a transformer.
A nice extra is the enable input, with this I can switch off the filament from the controller. Introducing some sort off sleep mode. (Do I have to switch off the HV as well ?)

[Ian.M]

Thanks for that.   
No. you don't have to switch the HT off.

[Rolo]

Update on project : I have designed a PCB to mount the VFD display on, it has 2mm pitch so I can't used a breadboard. On this PCB are three A6812's,  20 output high voltage latched shifregisters. All anodes and grids are tied to an individual output. It also has a blanking pin to control brightness with PWM.
Here are some pictures of the display in action, the software runs on an arduino Nano, for the the highvoltage (30V) i use a XL6009 step-up module.