Pandicon ZM1200 Questions

[SWISSNIXIE - Jonathan F.]

I would like to build a clock/counter/timer with the ZM1200 pandicon tube, which contains 14 digits 0-9 that can only driven by multiplexing.

Datasheet:
http://www.tube-tester.com/sites/nixie/dat_arch/ZM1200_philips.pdf

What i wonder:

• Does the Bias Voltage or Screen Voltage draw any current, or can it be limited to nearly zero?
• I would create the different Voltages with Z-Diodes, like in the Schematic, page 6 : http://www.tube-tester.com/sites/nixie/dat_arch/ZM1200_valvo.pdf
• For pulse duration i would go with about 300us, should not be a problem for modern microcontrollers, and the HV513 IC
  
• I don't know much about multiplexing at all, is the idea to switch on one digit for 300us, and then the next, or make a small gap where nothing is driven (all low)?

[David Forbes]

On 5/25/16 10:31 PM, SWISSNIXIE - Jonathan F. wrote:
> I would like to build a clock/counter/timer with the ZM1200 pandicon
> tube, which contains 14 digits 0-9 that can only driven by multiplexing.
>
> Datasheet:
> http://www.tube-tester.com/sites/nixie/dat_arch/ZM1200_philips.pdf
>
> What i wonder:
>

> * Does the Bias Voltage or Screen Voltage draw any current, or can it

> be limited to nearly zero?

It's an AC thing, so it uses no current, but add a bypass capacitor.

> * I would create the different Voltages with Z-Diodes, like in the

> Schematic, page 6 :
> http://www.tube-tester.com/sites/nixie/dat_arch/ZM1200_valvo.pdf

> * For pulse duration i would go with about 300us, should not be a

> problem for modern microcontrollers, and the HV513 IC

That might be slow enough. You want to be sure that each cathode
ionizes. Probably not a problem in a single-tube display like this.


> * I don't know much about multiplexing at all, is the idea to switch

> on one digit for 300us, and then the next, or make a small gap where
> nothing is driven (all low)?
>

To prevent ghosting, turn off the active anode, then wait a while (~50
us), then change the cathode selection, then wait another ~50 us, then
turn on the next anode.



--
David Forbes, Tucson AZ

[SWISSNIXIE - Jonathan F.]

Attached i have my first idea of the drive concept, i take much from the datasheet, but with modern parts.

The top two HV513 are set to polarity "1", what means a logic 1 will source voltage, a 0 will tie the output to GND, to avoid short circuit, i added diodes. When one Anode is off, it will be at about 105V clamped through diodes.

For the bottom one, the two HV51 are set to polarity "0", what means a logic 1 will tie the output to GND, a  0 will source voltage from VPP, which is connected to the 75V Voltage. XA15 will be the screen which is permanently at 75V.


Hope this will work.

[David Forbes]

The only thing I see missing is a series resistor on your Zener diode
chain, between D30 and the HV supply.

[gregebert]

I dont see any anode-resistors. How are you limiting cathode current ? 

Also, the diodes will trap the charge at the anodes; I cant do the math right now (t=CdV/I), but the charge will bleed-off during the 'dead' time when no anodes are driven to VOUTHV. I'm assuming here that +105V is not enough to maintain illumination. If you also turn off the cathodes at the same time you turn off the anodes (or before), the charge has no place to go. So make sure you turn off the anode first, wait several usec, then turn off the cathodes. Given the capacitance is a few pF, and current is roughly 1mA, I think there's more than enough time. But I see a lot of postings on this group about ghosting problems with multiplexed displays. Timing is critical.

On my designs, I try to provide a DC path to circuit ground for safety and ESD reasons. Surface mount resistors are small and cheap.

[SWISSNIXIE - Jonathan F.]

Yes, the current will be limited to adjustable 4-6mA

The original Valvo Circuit also uses about 105V, that's why i choose this value.

[SWISSNIXIE - Jonathan F.]

My idea is to drive as follwing

1. turn on cathode of digit 1
   
2. turn on anode of  digit 1
3. wait 200uS
4. turn off anode of digit 1
5. wait 50uS
6. change cathode to digit 2
   
7. wait 50uS
8. turn on anode of digit 2

This will end up in roughly 300us for one cycling and around 4ms for a complete display cycle. I hope this will be fast enough to don't see a visible flicker...

What value of resistor would be required between D30 and the HV Supply?

And where would you mount resistors in the DC path? you mean from anodes to ground or from the hv supply to ground?

I'm actually now evaluating if i should use a single Atmega328, or a 328 and a Attiny. The tiny would only take care about the shifting and stuff, while the mega328 is doing all the other handling. I'm thinking about this solution, because i worry about unknown delays with libraries and stuff.

[David Forbes]

On 5/28/16 4:18 AM, SWISSNIXIE - Jonathan F. wrote:
> My idea is to drive as follwing
>
>

> 1. turn on cathode of digit 1
> 2. turn on anode of digit 1
> 3. wait 200uS
> 4. turn off anode of digit 1
> 5. wait 50uS
> 6. change cathode to digit 2
> 7. wait 50uS
> 8. turn on anode of digit 2

>
> This will end up in roughly 300us for one cycling and around 4ms for a
> complete display cycle. I hope this will be fast enough to don't see a
> visible flicker...
>

Flicker is not a problem if you can complete a full cycle of all tubes
in 10 milliseconds or less. You may want to increase the 'on' time to
500 milliseconds for a brighter, more uniform display. Experiment with it.

>
> What value of resistor would be required between D30 and the HV Supply?
>

Ohm's law. (HV voltage - diode string voltage) divided by Zener current
(>2mA for good regulation)

>
> And where would you mount resistors in the DC path? you mean from anodes
> to ground or from the hv supply to ground?
>
>
>
> I'm actually now evaluating if i should use a single Atmega328, or a 328
> and a Attiny. The tiny would only take care about the shifting and
> stuff, while the mega328 is doing all the other handling. I'm thinking
> about this solution, because i worry about unknown delays with libraries
> and stuff.
>

The concern is that an undesired interrupt will cause a display glitch
by extending the 'on' time of a digit. It's not much of a concern, but
you can prevent it by disabling interrupts during the display time, that
300 us when the anodes is turned on. Glitches during the 'off' time are
not nearly as visible, because they affect all digits equally.

[SWISSNIXIE - Jonathan F.]

I have edited the first schematic and added a 1Meg resistor and some bypass caps.

I also placed a 5.5k anode resistor, that one i have read out from a Philips datasheet. I think Valvo and Philips ZM1200 are the same!

[Dekatron42]

I found this design yesterday: http://www.radiomuseum.org/forum/zm1206_zm1206.html where they use Solid State Relays to drive the anodes and cathodes, easy and simple. They also show just three trim potentiometers for the different bias voltages and a set of diodes to apply those bias voltages to the anodes, cathodes and screen.

/Martin