Maker Pro
Maker Pro

Nixie Tube Driver Question (w/ Schematic)

air_gap

Oct 21, 2023
8
Joined
Oct 21, 2023
Messages
8
I hope there are some members with grey hairs here. This question involves 1950s-1970s tech.

I'm exploring ways of driving a Nixie Tube (basically a neon lamp that displays numeric digits). The Nixie (IN-14) I'm trying to drive has a running current of 2mA. Since it's a gas discharge lamp, the resistance is high when off then quickly goes low once lit (struck). For this reason some sort of external ballast current limiting circuitry is necessary.

Most schematics I've found online simply use a fixed anode resistor to regulate the current. While I'm sure that's probably adequate, I was looking for a more precise way of controlling and minimizing the current through the Nixie to maximize it's life (Nixie's cathodes get damaged if overdriven even slightly).

So I am trying to design a active constant current source / current limiter circuit to use with Nixie tubes. Your insight and comments would be greatly appreciated.

SCHEMATIC (Rough Draft):

driver.jpg


DESCRIPTION:

The way I'm hoping this works is when one of the cathodes is turned on by saturating one of the cathode transistors through its CTRL line, the Nixie strikes and current starts flowing through the anode resistor, through the Nixie, then through the saturated transistor and finally through R1.

Once the current through R1 rises above one Vbe drop (~0.6V) Q1 will begin to turn on, shunting current away from the Nixie tube (and associated cathode circuit) until the voltage at R1 is at or below Vbe again. This (I think) will have the effect of establishing a negative feedback loop and thus keeping the voltage across R1 at Vbe, and keeping the current through the cathode circuit at ~2ma.

This is advantageous over a fixed resistor IMO because different cathodes in the Nixie can have slightly different electrical properties and this circuit tailors each cathode to the specified 2ma instead of the one-size-fits-all approach of a fixed anode resistor.

Or is this overkill and I'm missing something? I've obviously never worked with Nixie tubes before.

All replies welcome, thank you.
 

Delta Prime

Jul 29, 2020
1,793
Joined
Jul 29, 2020
Messages
1,793
Welcome to Maker Pro.
I hope there are some members with grey hairs here. This question involves 1950s-1970s tech
Everyone but me...
That's not gray hair on my chin. It's soup.
I mean, the third chin down. ;)
 

Attachments

  • NixieTransistors.pdf
    53.1 KB · Views: 7

Harald Kapp

Moderator
Moderator
Nov 17, 2011
13,632
Joined
Nov 17, 2011
Messages
13,632
Or is this overkill and I'm missing something?
Imho overkill, though I haven't ever worked with Nixies. It seems, however, that you are not alone with your concerns. This website as well as page 5 in the pdf posted by @Delta Prime show a simple way to add negative feedback i.e. current limiting. Basically it is your circuit without Q1 and its 4.7 k base resistor.
The voltage drop across R1 reduces VBE of the driver transistor for the active cathode, thus reducing the collector current aka cathode current.
 

Bluejets

Oct 5, 2014
6,855
Joined
Oct 5, 2014
Messages
6,855
Last edited:

AnalogKid

Jun 10, 2015
2,860
Joined
Jun 10, 2015
Messages
2,860
I'm not a big fan of shunt regulators, and this one is gonna dissipate a lot of unnecessary heat.

If you don't like the simple 1-resistor current limiter, replace it with a 2-transistor constant current source. Then, all of the segment drivers can be emitter-grounded saturated switches with nice, consistent base drive characteristics.

Here is a typical circuit. In your case, the load is the nixie anode connection. If you are worried about component failure, put a resistor in series between V+ and the external high voltage source. Size that resistor for 2.5 mA without the limit circuit, and select Rsense for a 1.5 mA running current. Or something like that - use values appropriate for your device.

1698087337926.png
ak
 

air_gap

Oct 21, 2023
8
Joined
Oct 21, 2023
Messages
8
After studying many of the circuits provided above (thank you) I came up with this which does all the active current regulating on the anode side, so the cathodes can be controlled with simple saturated transistors like Q3.

hv-reg.jpg

When power is applied, the zener voltage (Vz) minus the Vbe drop of Q1 appears at the base of Q2. The Vbe drop of Q2 raises the voltage back to Vz at Q2's emitter, putting Vz across R1 and thus setting the collector current of Q2, and thus nixie current at ~2mA.

Since Q1 and Q2 are thermally coupled (mounted on the same heatsink or just physically touching) any temperature rise (and corresponding decrease in Vbe) in Q2 will be compensated for by an (approximate) equal but opposite decrease in Vbe in Q1, keeping the voltage across R1 (relatively) stable at Vz.

The 1M resistor is provided to ensure there is high voltage on the collector of Q2 at all times so the nixie can strike (light) and R2 is ballast to turn on Dz and Q1. Q3 is a simple transistor switch.

Is this a more appropriate way of active current regulation through nixie tubes or am I still in the overkill category? I like the circuit AnalogKid provided above as it provides a similar function but is simpler.

Thank you.
 

air_gap

Oct 21, 2023
8
Joined
Oct 21, 2023
Messages
8
I have an IN-14 nixie coming in the mail this week so I thought I'd cobble something together to determine what the correct anode resistance should be (given I decidedto go that route). What do you all think? Will this work?

test-circuit.jpg

Basically, T1 is a 1:1 isolation transformer providing 120VAC to D and C, which form a half wave rectifier. Since 120 * sqrt(2) =~ 170 then about 170VDC will appear at R. R is connected as a rheostat that will be set to it's maximum 100k value at the start of the test.

Then I will apply power by closing SW and (very) slowly cranking down R until the connected cathode starts to glow and I have ~2mA on the ammeter. Then I will switch off, remove R from the circuit, measure it's set resistance and write that value down.

I will then do the same with the other nine cathodes, writing down the value of R each time, then take the average of all ten and choose the closest standardized value to that.

Will that get me an appropriate value for the anode resistor?

Thank you.
 

AnalogKid

Jun 10, 2015
2,860
Joined
Jun 10, 2015
Messages
2,860
In #9, be sure to put a fixed resistor in series with the pot, so the current through the tube is not fatal when the pot is turned all the way to the left.

The circuit in #7 will regulate much better than the one in #8 because of #7's large amount of negative feedback, versus none in #8

Also, in #7 all of the current needed to run the circuit goes out to the tube. In #8, R2 is a path to GND that has the full power supply voltage across it, wastes energy, and forms a voltage divider with R1 that loosens regulation.

ak
 

air_gap

Oct 21, 2023
8
Joined
Oct 21, 2023
Messages
8
The circuit in #7 will regulate much better than the one in #8 because of #7's large amount of negative feedback, versus none in #8

Yes, I agree your circuit #7 is the overall best HV constant current regulator of any of the multitude of schematics posted in this thread.

My only question, and it may be a dumb one, is will there be voltage present at the bottom of Rsense if there is no load connected (ie open circuit)? Remember a nixie is essentially a open circuit until it lights. A diode, or transistor Vbe junction, needs current flowing through it in order to turn on and have voltage appear on both sides of it.

Before the nixie lights, and it's still open circuit, will there be sufficient HV present at the output of circuit #7 (bottom of Rsense)?

Would a high value resistor like a 750k be needed across the nixie to connect to ground to ensure some small amount of current is flowing through Q1 (mislabeled T1) in circuit #7 so it remains turned on very slightly if/when no cathode is lit?

Thanks again.
 

air_gap

Oct 21, 2023
8
Joined
Oct 21, 2023
Messages
8
I have an IN-14 nixie coming in the mail this week so I thought I'd cobble something together to determine what the correct anode resistance should be (given I decidedto go that route).

^Typo and it won't let me edit.

...(given I decide to go that route).
 

AnalogKid

Jun 10, 2015
2,860
Joined
Jun 10, 2015
Messages
2,860
Would a high value resistor like a 750k be needed across the nixie to connect to ground to ensure some small amount of current is flowing through Q1 (mislabeled T1) in circuit #7 so it remains turned on very slightly if/when no cathode is lit?
No.

There is leakage current in the universe, and through the transistors, and through the nixie tube. Even with nothing but the input impedance of a DMM, around 10 M, the meter will register an output voltage very nearly equal to V+.

And - that image is a lift off of the innergoogle, not my own work. "T" is the least weird of the many opinions I've seen regarding how to refdes a transistor.

ak
 

AnalogKid

Jun 10, 2015
2,860
Joined
Jun 10, 2015
Messages
2,860
Arguably more intuitive than "Q", isn't it?

I think choosing Q was an act of rebellion. After having to adopt V because the Brits got there first, we decided to do something that would irritate *everyone*.

Of course, here in Amurka, we think V stands for vacuum . . .

ak
 

Delta Prime

Jul 29, 2020
1,793
Joined
Jul 29, 2020
Messages
1,793
One could be rebellious and kind at the same time! I just don't understand why? anyone, would go out of their way; to be anything else.
As a complement to post number7
I submit to you an additional resource,which would compensate for the negative temperature coefficient.
In using two NPN bjt's ;even if one were to make an simple improvement by using a complementary pair NPN/ PNP, BJT in the same configuration.
The negative temperature coefficient for example; is approximately−1.6 mV/°C, which causes the current value to vary widely with temperature.
 
Last edited:

Martaine2005

May 12, 2015
4,888
Joined
May 12, 2015
Messages
4,888
Hehe, V = voltage.
E = electro motive force, or something similar.
We Brits shorten the amount of words needed!.
Example: I’m going to powder my nose.
Brits: having a slash.
:)
 
Top