Author Topic: Need help understanding VFD Filament  (Read 5199 times)

0 Members and 1 Guest are viewing this topic.

Offline romhunterTopic starter

  • Regular Contributor
  • *
  • Posts: 104
  • Country: vn
Need help understanding VFD Filament
« on: September 12, 2019, 04:40:47 pm »
I'm trying to work with some IV-3A VFD Tubes from Russia. In the datasheet, the filament is specified to be driven with 0.85V (?). So I just google around for a solution to use this with a 12V power supply, and come accross those ideas:

1. Drive it with DC since it's a short tube: 0.85V regulation with 6 filaments as load, it's a scary thing to do. I'm thinking about a LM317, set the output to its minimum, and drop the output voltage using a few diodes. I might get close if the temperature of the diode does not fluctuate too much.

2. Drive it with AC voltage: I don't quite understand it right here, so I might need some help to clear my head. Anything will be of use. I'll be refer to the Noritake VFD guide:

The output of this "DC-AC Converter" is a square wave? If that's true, can I drive the filament using a micro and a few opamps? (with an output BJT).

3. What's the filament bias voltage? On the Noritake website, it's said to be applied to the center-tap of the transformer. So how can I do it using the forementioned method (MCU and opamps).
 

Offline janoc

  • Super Contributor
  • ***
  • Posts: 3925
  • Country: de
Re: Need help understanding VFD Filament
« Reply #1 on: September 12, 2019, 07:58:16 pm »
I have used this driver for the filament of a salvaged VFD:

http://www.kerrywong.com/2013/06/05/vfd-filament-driver-using-555/

Worked a treat.

The filament voltage/current varies, you will need to find the correct one for your VFDs in the datasheet. Normally the filament wires should only barely glow, anything more than that will seriously degrade their lifetime.
 

Offline romhunterTopic starter

  • Regular Contributor
  • *
  • Posts: 104
  • Country: vn
Re: Need help understanding VFD Filament
« Reply #2 on: September 13, 2019, 02:54:06 am »
I have used this driver for the filament of a salvaged VFD:

http://www.kerrywong.com/2013/06/05/vfd-filament-driver-using-555/

Worked a treat.

The filament voltage/current varies, you will need to find the correct one for your VFDs in the datasheet. Normally the filament wires should only barely glow, anything more than that will seriously degrade their lifetime.

As far as I understand, the transformer in this circuit "worked like" a low pass filter, therefore filter the square wave and output a near sine wave on the output. I'm trying not to use a center-tapped transformer in the design (lower space constraint, lower BOM), but it might serve the purpose. I just don't understand the needs to bias the center-tap to some voltage.

 

Offline Bud

  • Super Contributor
  • ***
  • Posts: 7276
  • Country: ca
Re: Need help understanding VFD Filament
« Reply #3 on: September 13, 2019, 05:03:49 am »
It is done to prevent inactive segments from glowing.
Facebook-free life and Rigol-free shack.
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: Need help understanding VFD Filament
« Reply #4 on: September 13, 2019, 09:12:12 am »
AC is required to keep the segments at the same brightness. It doesn't need to be a sine wave. A square wave will work perfectly.

A basic inverter can be made using a transformer with a centre tapped primary and a couple of transistors. Obviously higher value resistors (try 2k2) and smaller transistors, than the 2N3055 can be used e.g. BC338, PN2222A.


EDIT:
The filament bias voltage can be generated by connecting one side of the transformer secondary to a potential divider. The bias current is tiny, so fairly high value resistors can be used.
« Last Edit: September 13, 2019, 10:54:47 am by Zero999 »
 

Offline janoc

  • Super Contributor
  • ***
  • Posts: 3925
  • Country: de
Re: Need help understanding VFD Filament
« Reply #5 on: September 13, 2019, 04:34:27 pm »
As far as I understand, the transformer in this circuit "worked like" a low pass filter, therefore filter the square wave and output a near sine wave on the output. I'm trying not to use a center-tapped transformer in the design (lower space constraint, lower BOM), but it might serve the purpose. I just don't understand the needs to bias the center-tap to some voltage.

If you don't center tap, you will have a voltage gradient across the filament (the filament is a cathode of the vacuum tube and the DC used to drive segments will be superimposed on the heating AC voltage) - that will cause the segments to glow at uneven intensities from one edge to the other.

It is well explained in this FUTABA app note (page 9):
http://www.vwlowen.co.uk/arduino/vfd/VFD-Application-Note.pdf

If you don't want to mess with transformers, you can drive the filament with DC as well. But that will cause uneven brightness and you need to be very careful to not exceed the maximum voltages on the filament or you will quickly damage the VFD (the filament coating will sputter off if the filament gets too hot).

 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: Need help understanding VFD Filament
« Reply #6 on: September 13, 2019, 08:30:09 pm »
I don't see the need for a centre tap. The transformer can easily biased using resistors. The bias current is tiny: in the pA range, the the values can be high.
 

Online Ian.M

  • Super Contributor
  • ***
  • Posts: 13216
Re: Need help understanding VFD Filament
« Reply #7 on: September 13, 2019, 09:48:27 pm »
I don't see the need for a centre tap. The transformer can easily biased using resistors. The bias current is tiny: in the pA range, the the values can be high.
Errrr . . . .*NO*

A VFD filament is a directly heated cathode, so its bias network carries the full anode and grid current of the active digit.  With all segments illuminated, that can be in the mA to tens of mA range depending on the VFD.  Its possible to use a resistive divider across the filament supply to provide an effective center tap, but that wastes a lot of power if you use low value resistors and introduces a significant variation in cathode voltage with the number of segments illuminated if you use higher resistor values, which can result in an unwelcome brightness variation between '1' and '8', obvious if seen side by side.

Although you *MAY* be able to get away with biassing the filament at one end for single digit tubes, it increases the risk of ghosting unless you increase the bias voltage, which demands an increase in HT voltage to maintain brightness.

Another approach, only applicable to symmetrical filament drive waveforms at a relatively low crest factor, is to use a pair of Zeners, one at each end of the filament, possibly with a pair of high value  pullups to HT or a lower voltage rail to keep them biassed.  The effective bias becomes Vz -Vf_pk2pk/2.

TLDR: Not having a filament supply secondary center tap is a P.I.T.A.
« Last Edit: September 15, 2019, 06:00:13 pm by Ian.M »
 

Offline janoc

  • Super Contributor
  • ***
  • Posts: 3925
  • Country: de
Re: Need help understanding VFD Filament
« Reply #8 on: September 14, 2019, 10:40:49 am »
TLDR: Not having a filament supply secondary center tap is a P.I.T.A.

Even worse - there is really no excuse because winding your own transformer on a toroid for a high frequency AC drive takes maybe 30 minutes tops. You don't need a lot of turns (the voltage is low) and the high switching frequency keeps the size of the device small.

The link I have posted has 40 turns primary and about 20 turns secondary.
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: Need help understanding VFD Filament
« Reply #9 on: September 14, 2019, 01:20:10 pm »
I don't see the need for a centre tap. The transformer can easily biased using resistors. The bias current is tiny: in the pA range, the the values can be high.
Errrr . . . .*NO
Yes, I stand corrected. I was thinking about amplifying thermionic tubes.
 

Online Ian.M

  • Super Contributor
  • ***
  • Posts: 13216
Re: Need help understanding VFD Filament
« Reply #10 on: September 14, 2019, 02:09:08 pm »
It sounds like you were thinking of grid bias in a small signal amplifying stage, which in most cases were run with the grid negative with respect to the cathode so they would draw negligible grid current.   However, even there one could choose to get the negative grid bias by a DC grounded grid and a resistor in series with the cathode so that the cathode current through it would develop a positive bias on the cathode.  One also usually had the advantage of an indirectly heated cathode with an insulating oxide layer between it and the filament, and as long as the cathode circuit impedance was low compared to the filament-cathode leakage current, that gave you a lot more flexibility in biassing. 

Also it should be noted that the active digit of a VFD runs at a significant positive grid voltage, usually equal to the anode voltage, which provides much of the acceleration needed to get enough K.E. into the electrons to effectively excite the phosphor, but as a result the active grid draws a significant current.
 

Offline janoc

  • Super Contributor
  • ***
  • Posts: 3925
  • Country: de
Re: Need help understanding VFD Filament
« Reply #11 on: September 14, 2019, 08:20:06 pm »
Also it should be noted that the active digit of a VFD runs at a significant positive grid voltage, usually equal to the anode voltage, which provides much of the acceleration needed to get enough K.E. into the electrons to effectively excite the phosphor, but as a result the active grid draws a significant current.

Yup! Going to toot my own horn a bit:
https://janoc.rd-h.com/archives/433

At the bottom are pictures of one of the VFDs I have, that thing takes 600mA easily and that's not even at full intensity, running multiplexed (i.e. only 2 characters are on at the time - each grid is common for a column - two characters stacked on top of each other). If that thing is fully on, it is over 1A of current!

VFDs are pretty power hungry beasts.
« Last Edit: September 14, 2019, 08:27:08 pm by janoc »
 

Offline dom0

  • Super Contributor
  • ***
  • Posts: 1483
  • Country: 00
Re: Need help understanding VFD Filament
« Reply #12 on: September 15, 2019, 12:12:53 pm »
You can just drive the filament with an H bridge (3ish V filaments are common and go well with the voltage losses of a bipolar transistor H bridge connected to a 5 V supply) and simply switch the polarity after each multiplexing iteration. Doesn't need a transformer, doesn't flicker and has no brightness drop.
,
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: Need help understanding VFD Filament
« Reply #13 on: September 15, 2019, 12:31:43 pm »
0.85V is a very low voltage.

I think winding a custom transformer is the only option. I don't see how a centre tap is that difficult. In the basic circuit I posted previously, it would need to have 2 turns on the secondary and 56 on the primary. The primary centre tapped and each end is driven alternately, so the effective ratio is 28:2.

Another winding and voltage doubler could be added for the HV DC.
 

Offline janoc

  • Super Contributor
  • ***
  • Posts: 3925
  • Country: de
Re: Need help understanding VFD Filament
« Reply #14 on: September 15, 2019, 12:31:58 pm »
You can just drive the filament with an H bridge (3ish V filaments are common and go well with the voltage losses of a bipolar transistor H bridge connected to a 5 V supply) and simply switch the polarity after each multiplexing iteration. Doesn't need a transformer, doesn't flicker and has no brightness drop.

And you are totally missing the point of why the center tapped transformer is there. It is not because the engineers don't know how to generate AC voltage across the filament by other means. The transformer is there to have a floating neutral point where to apply a positive bias voltage to the filament so that you don't get segment ghosting from segments that are not completely off.

VFD is a directly heated triode, if the grid is at ground potential you still have some current flowing between anode and cathode, so even the nominally off segments will be faintly glowing. You need to make the grid negative with regards to the cathode to prevent the current from flowing (think depletion mode FET). One way to do that is simply to raise the cathode (filament) voltage above the ground, so when the grid is grounded by the microcontroller, it is negative relative to the cathode, thus repelling the electrons being emitted and blocking the current from flowing.

How do you do that with an H-bridge that isn't floating with regards to the rest of the drive circuit?

And another thing is that a VFD needs a high anode voltage, typically around 40-60V. If you have a transformer there already, that is a matter of adding an extra winding, problem solved. No need to build a separate inverter for that.
« Last Edit: September 15, 2019, 12:46:49 pm by janoc »
 

Offline dom0

  • Super Contributor
  • ***
  • Posts: 1483
  • Country: 00
Re: Need help understanding VFD Filament
« Reply #15 on: September 15, 2019, 01:16:21 pm »
How do you do that with an H-bridge that isn't floating with regards to the rest of the drive circuit?

It has been a couple years since I've done this, but IIRC the ~1.5 V drop on the low side of the filament driver was sufficient to avoid ghosting. (These were relatively low voltage VFDs witha nominal anode voltage of I believe around 24 V).
,
 

Online Ian.M

  • Super Contributor
  • ***
  • Posts: 13216
Re: Need help understanding VFD Filament
« Reply #16 on: September 15, 2019, 03:52:10 pm »
How do you do that with an H-bridge that isn't floating with regards to the rest of the drive circuit?
It has been a couple years since I've done this, but IIRC the ~1.5 V drop on the low side of the filament driver was sufficient to avoid ghosting. (These were relatively low voltage VFDs witha nominal anode voltage of I believe around 24 V).
Alternatively, capacitively couple the H-bridge to the ends of the filament, and provide a DC path for the cathode current via a pair of clamping diodes from the filament ends to your logic supply, or toa pre-biassed (via a pullup to HT) Zener to provide the desired cathode bias voltage.   You can also use a resistor divider across the filament to get a neutral point, but as I mentioned above that's a trade-off between wasting power and 1's being noticeably brighter than 8's.
 

Offline janoc

  • Super Contributor
  • ***
  • Posts: 3925
  • Country: de
Re: Need help understanding VFD Filament
« Reply #17 on: September 15, 2019, 05:02:17 pm »
How do you do that with an H-bridge that isn't floating with regards to the rest of the drive circuit?
It has been a couple years since I've done this, but IIRC the ~1.5 V drop on the low side of the filament driver was sufficient to avoid ghosting. (These were relatively low voltage VFDs witha nominal anode voltage of I believe around 24 V).
Alternatively, capacitively couple the H-bridge to the ends of the filament, and provide a DC path for the cathode current via a pair of clamping diodes from the filament ends to your logic supply, or toa pre-biassed (via a pullup to HT) Zener to provide the desired cathode bias voltage.   You can also use a resistor divider across the filament to get a neutral point, but as I mentioned above that's a trade-off between wasting power and 1's being noticeably brighter than 8's.

All that is a lot more complicated (how big would need those caps to be to pass the heating current of the filament?) than a small transformer that you can wind in a few minutes yourself. Or even buy, these are fairly standard parts.
 

Online Ian.M

  • Super Contributor
  • ***
  • Posts: 13216
Re: Need help understanding VFD Filament
« Reply #18 on: September 15, 2019, 05:47:34 pm »
Yes. Way up thread I said "TLDR: Not having a filament supply secondary center tap is a P.I.T.A.".   Assuming you dont have a negative rail available to let you take inactive anodes and grids right down to a good cutoff voltage, you'd have to have a dammed good reason to try to do the filament supply without magnetics.
 
The following users thanked this post: janoc


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf