IV-17 tube filament voltage difference (DC drive)

[Chaos Hydra]

Hello guys, I made an IV -17 display panel. I made the filaments five in series and put them under 12VDC to get a 2.4VDC filament voltage.
The problem is, there is a voltage difference between each tube's filament as the one closer to ground goes up to 3.2VDC and the one close to 12V gets to 1.6 VDC.
As you can see in the video, the ones on the outside are much brighter (close to ground).  youtube video here.

So is there anything I can do to fix the problem here? I made PCB already, so there is no way I can switch to AC now. Thanks for the help!

[Terry S]

Zeners across the filaments?

[Tony Adams]

If you can increase the anode voltages enough to get the dimmest displays up to an acceptable brightness you could cheat and PWM the brighter displays to balance the brightness in software. Depends if you can get an acceptably high voltage on the centre tube anodes/grid to counteract the 12v you have on the filament.

[Kejian Lin]

Thanks, but please elaborate, I am new to this stuff. So I put a 2.4V breakdown zenner in parallel with each filament?

--
You received this message because you are subscribed to a topic in the Google Groups "neonixie-l" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/neonixie-l/1FZRkfNeuGY/unsubscribe.
To unsubscribe from this group and all its topics, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/2101696b-dd2c-43b6-8bb2-5a86701782e3%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

Cheers

Kejian Lin "Robert"

[Kejian Lin]

Well Tony, my anode is fixed 24VDC but it can lit everything up. The brightness is now a pure filament matter.

--

You received this message because you are subscribed to a topic in the Google Groups "neonixie-l" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/neonixie-l/1FZRkfNeuGY/unsubscribe.
To unsubscribe from this group and all its topics, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/44592a3c-c8b8-4a89-be4d-3520a5e8cc90%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[gregebert]

Is it possible to cut 2 traces on your PCB to isolate the filament chain ? If so, you could inject an AC supply onto the filaments.

As far as I can tell, the reason your tubes have varying brightness is due to the cathode being tied internally to the filament (I did not see a separate cathode pin in the datasheet, but there *is* a cathode in the tube). The voltage-difference between cathode and anode varies due to the progressive voltage drop across the filament chain, and consequently the current, hence brightness, will vary accordingly.

My best guess is that if you can isolate the filament chain, you would want to power it from a 12.6V RMS center-tapped transformer (or two isolated 6.3v windings). The center-tap would connect to the middle tube AND to your circuit ground. The 2 remaining windings would go to each end of the filament chain. The idea here is that the average voltage on any filament will appear to be 0 (gnd). If the brightness vs voltage of the tube is reasonably linear, you should see more uniform brightness.

I dont think an isolated filament supply would solve your problem; in fact, I think it would result in all tubes being dark because there would not be a complete path thru the anode.

I admit I've done very little experimenting with VFD tubes, and that was in the 1970's when Radio Shack sold them. But I did get mine to work.

The only other option I see is to use PWM as Tony suggested. This will allow you to make your brighter tubes dimmer to match the others.

Don't despair or give up! your display looks really neat.

[Dekatron42]

First of all I think you should read this guide from Noritake on VFDs : https://www.noritake-elec.com/display/vfd_operation.html it explains why an AC filament is best to use.

Then I think that you should see if you can do as "gregebert" says in lifting the heater chain from your current circuit and supply it with an AC voltage from a center-tapped transformer winding with the correct voltage/current and with the correct bias voltage. You'll have to put each tube in parallel for this to work and use a transformer with enough current capability to drive all heaters in parallel - even if there is "no way" you can do this. The voltage is quite low so you can easily wind two extra windings on a toroidal transformer to get the necessary center-tapped winding, just don't forget to tape the extra windings in place so that it won't hum or scratch the other windings. You can experiment quite easily with a toroidal transformer to get the voltage/current you need. Since the voltage is quite low you won't have to worry about the mains voltage variation as that doesn't affect the low heater voltage that much, but you should make sure that you use the correct voltage so you might have to try different toroidal transformer until you get the correct volts/turn ratio.

Apart from this there is a slight variation of the luminance between each of these tubes so you might have to match them before using them like you do with a lot of them side by side as it will show more when you can compare them to each other, I've seen quite a big variation myself on batches even though all of them have the quality stamp.

Almost all of the old Russian schematics that I have looked at have tied all of the heaters in parallel, there have only been very few designs which have used four VFDs or less which have put them in series and they have then said that they don't care about the luminance difference as they have used a filter in front of the VFDs to even it out.

/Martin

[Chaos Hydra]

MY Thanks Dekatron and Greg! The good thing about my PCB is the filament supply come in from one thick trace and I can easily make a clean cut there. I found an old 9VAC power supply which is essentially a transformer in jack. When driving low load, the voltage goes up to about 10.6VAC so it comes to about 2.1VAC to each tube so worked out nicely. I am planning to put one end to the ground then the other to the filament supply in-feed line. Sounds plausible?

[gregebert]

It's worth trying 1 cut between the 12VDC supply and the filament chain, and applying AC between GND and the cut-trace. It may not give you the best results, but it could produce less variation in brightness.

[threeneurons]

Read the Noritake guide, that was previously listed. All VFDs are directly heated. No separate cathode. The filament is in between your eyes and the glowing anode/segments all the way in the back. Your problem is due to the anode cathode voltage difference not being the same for each tube. I think that has already been mentioned. If you're going to hook the filaments is series, the filament "heating" current needs to be AC, so its addition to the anode-cathode potential averages out to zero. You can use an old audio amp chip to generate that AC:

[Chaos Hydra]

I made a clean cut on the trace and patched in a 10VAC transformer. That solved the overheating filament problem and now has a decent brightness overall. The only problem is: the certer tubes' "off segments" are quite bright. It is kind of difficult to read sometimes.

[David Forbes]

You may be able to correct the "off" brightness by controlling the DC
voltage of the center tap. That will change the bias voltage on the
cathodes relative to the grids.

However, I have no experience working with these tubes.

--
David Forbes, Tucson AZ

[Kejian Lin]

The big problem of my PCB is everything is tied up, so it is hard to fix. Well, these lesson can go to my next design.

--
You received this message because you are subscribed to a topic in the Google Groups "neonixie-l" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/neonixie-l/1FZRkfNeuGY/unsubscribe.
To unsubscribe from this group and all its topics, send an email to neonixie-l+...@googlegroups.com.

To post to this group, send an email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/56469E65.5050909%40dakotacom.net.

For more options, visit https://groups.google.com/d/optout.

[gregebert]

Good progress!

Actually, the change @nixiebunny recommended is probably simpler than you imagined.

If you are using a center-tapped transformer, with the center tap tied to ground, just put a variable power supply between the center tap and your GND, then vary a few volts. You may need to reverse polarity. I suspect you will have to trade-off between max & min brightness, basically like the contrast knob on a TV.

If you are just using a non-center-tapped transformer, you can do the same trick because one leg of the transformer has a path to GND.

Whichever route you try, make sure you dont put the variable supply in series with the filaments because that could burn them out, or needlessly stress them. The exercise is to change the offset-voltage between the filaments (I still want to refer to them as cathodes, because they are the source of emitted electrons inside the tube) and the grid.

Just curious: What voltage levels are you driving on the grids ?

[Dekatron42]

I am glad it works better now!

Have a look at paragraph 6.2 in the Noritake link, there they use a zener diode to bias the center tap, you can usually just use a string of ordinary diodes to try out what bias voltage that you need for your design to have a proper cut-off voltage.

Then when you know that cut-off voltage you can either use one zener with the nearest voltage or you can use two (or a string of diodes) and then you can short part of this diode chain to have an effective dimming capability - the Russians (probably others too) used this way of dimming the VFD displays in clocks for night usage.

So using a center-tapped transformer has its good sides.

/Martin

[Chaos Hydra]

The Grid and Anode runs on 24VDC. Datasheet nominal is 25VDC, works well though. I will try your suggestion but maybe later, thanks.

[Chaos Hydra]

yeah, that's what I need to do next, reduce the background illumination of those tubes. I will look into this. Thanks.

[Tristan]

If the transformer lacks a real center-tap then a virtual center-tap can be used instead. Simply use a pair of resistors of a reasonable value, say 470ohm
 as a divider across the secondary of the transformer and use the center of the divider instead to bias the cathodes instead. If you were running an even number of tubes in series you could use the middle filament connection instead.