Nixie Ionisation and Maintaining Voltage

[Alaba Baju]

Hi people!,

Thanks to whoever approved my group joining request :)

I own a few tubes, the IN-1 and the IN-9. Never powered a nixie device before, I'm building a switch mode boost converter and I have a few questions please! I understand the IN-1 needs about 170v to light up and about 133v to stay on. One thing I've noticed though is that most people only seem to build switching supplies for the higher voltage. I don't understand! On all the threads no one seems to talk about how the power supply lights it up at 170v and then dials back to maintaining voltage! I'm imagining I'll need to provide both the 170v and 133v levels in my circuit. I'm building a clock/stopwatch with 4 tubes and I'll want to sometimes switch it from HH:MM to MM:SS . The digits on hours wouldn't change much, so I plan to light them up on the 170v line and then switch them to the 133v line. The digits on seconds though I imagine would not spend enough time lit to necessitate switching to a maintaining voltage.

Is my reasoning wrong? It's just that I don't seem to see any threads talk about them this way! Also the SMPS for nixies I see mostly just have a 180v line.

If you knowledgeable lot do see it reasonable to provide two voltage lines for my nixies, then could people suggest methods they'll take to do this? Is there a more energy efficient way of doing this than simply a resistor based voltage divider on the 170v output? The annoying thing about doing two SMPS is that I'll need two separate power sources to my circuit so I don't create a short circuit. Not pleased.  

I was considering using a transformer for isolation and use one power source for two SMPS, but using transformers seem to have a lot of drama, or maybe I'm just lazy?

I'll love to hear thoughts. Thanks!

.... and Google keeps thinking I'm trying to spell 'pixies' when I type 'nixies' ! :D :D

[gregebert]

I dont recommend the IN-1; it's a low-quality tube. See my other postings about it.

You should think of nixie tubes as current-controlled devices; it requires a certain amount of current to produce the desired brightness. As with any device, you dont want to exceed the maximum recommended current, otherwise the device's lifetime will reduce.

That said, it takes a certain voltage to initiate the glow (discharge), which is the ionization voltage. Before the onset of ionization, the current is very low. As the tube ionizes, the number of ionized gas atoms increases exponentially, and so does the current. If you were to allow the current to continue unabated, it would get very large and rapidly destroy the tube. An extreme example is a lightning-strike. By adding a properly sized series-resistor, as the current through the tube increases (a direct result of ionization), the voltage-drop across the resistor will increase, and the voltage across the tube will decrease (this assumes a constant supply voltage), which tends to reduce the current. An equilibrium is reached where the voltage across the tube is sufficient to keep it illuminated.

Be aware the voltage-drop across the nixie tube while it's illuminated will vary from tube-to-tube, and also during it's lifetime. Also be aware that a small change in voltage will result in a large change in current, therefore you really cant rely on a constant-voltage across the tube. Now, you can take the design of a nixie-tube driver one step further, and make a constant-current driver, which is what I did on my big clock. I wont go into details here.

The easiest way to create the roughly 170 volts to ionize a nixie is to rectify+filter the 120V AC line. However, as the tube ages it's ionization voltage might increase above this and your tube will no longer illuminate. So, you can run at a higher voltage; on my first nixie clock I used a voltage-doubler to create +340VDC, then used a larger series-resistor. In this case, I used an 82K resistor to drive a Burroughs 5092 nixie at the optimum 2.2mA. The IN-1 runs about 3mA so you would use a lower resistor around 70K.

Running at 340V does not harm the tube, but it does waste more energy in the form of heat dissipated by the series resistor.

Another option is to use a "boost" transformer; I did this on my big clock to produce +210V. This gives me plenty of margin for the ionization voltage without wasting as much energy as a 340V supply. With 15 tubes, there is a noticeable savings in energy running at 210V vs 340V.

[Arne Rossius]

Hi,

'Alaba Baju' via neonixie-l wrote:
> I understand the IN-1 needs about 170v to light up and about 133v to
> stay on. One thing I've noticed though is that most people only seem
> to build switching supplies for the higher voltage. I don't
> understand!

The higher voltage (ionization voltage) is the *minimum* voltage your
power supply needs to provide. If the output voltage of your power
supply is lower, the tubes may not light up at all. The lower voltage is
the (approximate) voltage at the tube anode when the tube is on, similar
to an LED's forward voltage. You use a resistor between the voltage
supply and the tube's anode to get these two voltages automatically.
When the tube is off, (almost) no current is flowing so the resistor
doesn't drop any voltage, and the full power supply voltage is available
at the anode. Once the gas ionizes and the tube lights up, it starts
drawing current, thus dropping voltage across the resistor. In effect,
the difference between the two voltages is dropped across the resistor,
and you don't have to do any switching.

A simple way to approximate the resistor value is this:

R = (V_supply - V_sustain) / I

where V_supply is your power supply voltage (may be higher than 170V),
V_sustain is the tube's sustaining voltage (133V) and I is the desired
current (a typical or maximum value can usually be found in the tube's
datasheet).

The formula isn't quite exact because the sustaining voltage will be
dependent on I, but it is good enough to get an idea for the correct
anode resistor value. If you are unsure, start with a high resistance
and measure the current.

If you use a lower current than the datasheet recommends, the tube life
will be prolonged but the digits will be less bright (obviously) or not
light up completely. Some datasheets also list a minimum current for
this reason.

In my nixie clock with four NL-840 tubes, I used a 180V power supply
with 18k anode resistors (one for each tube). The datasheet recommends
10k at 170V and 47k at 250V. From that, we get I = 2.2 mA and
V_sustain = 148V, so with my 18k resistors the current should be about
1.8 mA, just above the minimum value of 1.5 mA and well below the
maximum value of 3.5 mA.


HTH,
Arne

[gregebert]

I had a brain-meltdown and didn't realize you were making a battery-operated device.

The wristwatch I'm working on right now has a DC-DC converter (flyback converter)  that boosts 3.7V to about 180V to ionize my display, and then throttle-back to around 140 to 160V to save some battery energy. The watch is controlled by a small FPGA, so I use a voltage-divider for the FPGA to monitor the anode supply and regulate it. I'm just about ready to send the board out for fab, so it's going to be a few weeks before I have this debugged. Although I've simulated it extensively in SPICE and Verilog, I dont have a good way to simulate the entire closed-loop system so do expect some surprises.

So, the short answer is yes, you want to modulate your high-voltage supply to save battery energy, but you will still need to burn some extra energy via dropping resistors, or in my case, NPN-constant-current-driver, just because the neon tubes have nonlinear & varying behavior.

[Alaba Baju]

Thanks a LOT Anne! I'm finally understanding that it's sort of like an LED! To test if I understand it properly, would I be correct if I simply connect the nixies to 180v and use a BJT current limiter to limit the current to 2mA?

[Alaba Baju]

Thanks gregebert! :) Just read a post you made on the IN-1 :/ I have them already along with the IN-9 and my  IV-18, those are my only tubes. Once I've built my first project properly, I'll go for other tubes if they don't last long. Good thing I have 6 of them and need only 4 in my design. Thanks for the explanation on the design too. :)

[David Forbes]

On 8/11/2014 2:46 PM, 'Alaba Baju' via neonixie-l wrote:
> Thanks a LOT Anne! I'm finally understanding that it's sort of like an LED!
> To test if I understand it properly, would I be correct if I simply connect
> the nixies to 180v and use a BJT current limiter to limit the current to
> 2mA?
>

It depends on whether you are driving only one tube at a time, or more than one
tube. If you have several tubes on at the same time, then you need anode
resistors to ensure that all tubes get the same current, as the ionizing voltage
will be different for each tube, and different for every cathode in each tube.

If you multiplex the display and drive only one tube at a time, then a current
regulator will work, and there is no anode resistor to waste power.

I do that with my Nixie tube watch. you can see the schematic diagram here:
http://www.cathodecorner.com/nixiewatch/firmware/nwrf-schem.gif
The 150 ohm resistor R8 senses the current through the tube and adjusts the
feedback signal VFB accordingly.

[gregebert]

Thanks a LOT Anne! I'm finally understanding that it's sort of like an LED! To test if I understand it properly, would I be correct if I simply connect the nixies to 180v and use a BJT current limiter to limit the current to 2mA

Yes, I would recommend that. Make sure your NPN is rated for at least 200V, and I would recommend even more. The Icbo rating is important, because if the transistor has high enough leakage, you will have glowing cathodes. I would recommend Icbo of less than 1uA. My watch uses a PMBTA42DS dual NPN device, which is rated at 300V and has 100nA Icbo.

For the base-drive, if you are running from 3.3V, then to obtain 2mA, you would use an emitter-resistor of (3.3-0.7)/0.002 = 1300 ohms. 

[jpeakall]

Only need to worry about a single supply. Lower voltages are if you want to dim. You can just set and foreget, using a current limiting resistor. Or you can pwm the HV to dim/lower the voltage, which is what i do. I have also used a digital pot to control the power supply voltage. So lots of ways to get fancy but no need to if you don't want to.

Jonathan

--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, 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/5ed6136b-a75c-4227-aa1d-8ed98e3a5cdb%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Alaba Baju]

Wow, thanks a lot everyone, I finally understand it now. Such a relief I don't have to build a second power supply line.

Thanks gregebert, Arne Rossius, nixiebunny and Jonathan