6x Multiplexing Nixie Current

[scotth]

First of all, thanks for accepting me to the group!

I've built a functioning nixie clock that uses a nixie driver to multiplex all six numbers.  Due to my inexperience, I just figured out the resistor that I needed to drive a single nixie tube at the correct voltage an current, which landed me at 22kOhms.  However, as I'm sure you all know, my numbers were much dimmer once i got the multiplexed clock working with the same 22k resistors.  Because of this, I decided to check the voltage drop and current of my nixies and got some numbers that can't be correct, as it is below the operating specs of the tube.  Is it some sort of average or something?

I've seen another thread on here where this is mentioned and I learned that I could get a more accurate reading of voltage and current using a scope (which I don't have). What I'm wondering is, is there a way to calculate what the current should be without measuring it?  I'm not horribly dissapointed that the numbers are dimmer than driving a single tube, but I'd like to make sure I'm operating at the nominal current in order to maximize my brightness without significantly decreasing the expected life of the tubes.  Any help is greatly appreciated!

Below is a picture of my anode driving transistor circuit (?) not sure what the best way to word that is.

[Tomasz Kowalczyk]

With most DC voltmeters you can safely assume that it measures the average. So having average voltage drop on anode resistor and resistance you can calculate average current, and then multiply it by 6 (if there is a dimming/anti-ghosting dead time between tubes, you might want to take it into account) to get the pulse current. 

Nixie tubes can operate in multiplexed mode with pulses of current higher than nominal current - as long as you aren't pushing average current over the nominal current, it should be OK.

In IN-18 datasheet I found information, that the average current in pulse operation should be a bit lower than during DC operation.

Anyway, if your tube is specified for 2mA and you multiplex 6 tubes, I think you can safely use 5-6mA pulses (~1mA average current) without damaging the tube. But it would be less guessing and more solid information if you would specify which tubes are you using.

[Paul Andrews]

Or you could put a big variable resistor in parallel and gradually reduce it until you are happy with the result. Bearing in mind you don't want to go too low (too much current). The data sheet for the tube might tell you an acceptable multiplex current. If it doesn't, I would look at some data sheets for other tubes that do specify this and see if there is some kind of average current multiplier that is being used. Then drop it some and use that. The regular resistance formula will still give you the instantaneous current.

[scotth]

The tubes I'm using for two different clocks are IN-4s and IN-8s. I found the pulse specs you referenced on the IN-8 datasheet, but not the IN-4. On the IN-8 datasheet, it lists two different modes, A and B. For mode A, which is a pulse duration of 0.1ms, it gives a pulse Voltage of 200V and current of 10mA. It also lists a PMW of 10%, what does that mean?

For mode B, which has a pulse duration of 1-2ms, it recommends 200V at 5mA and the same PMW.  Which mode should I use?

Is the pulse Voltage the same as my input voltage before my transistors?

I will try to measure what current I am at later today.

Thanks!

[scotth]

Whoat, sorry about that last post.  I made the mistake of trying to post from mobile and apparently failed on the formatting front.  Strange considering I didn't change the formatting at all.

[Tomasz Kowalczyk]

To be honest, I don't understand the timing values from datasheets either. I would just assume that if the DC current of IN-4 is 2,5-3mA, and IN-8 - 2,5-3,5mA, then in 6 digit multiplex 5mA in pulse on each of them wouldn't hurt them at all - that is still 1/3 of nominal power dissipation of a DC driven tube at minimum specified current. I bet increasing this to 7,5mA in pulse (1/2 power dissipation) would be still safe, but the glow might start changing colour a bit - I've noticed more blueish glow on overdriven tubes while testing my LC-531s, which due to poor quality have higher striking voltages, so I was doodling with them alot.

[GastonP]

OK you don't need to shout to be heard in this list. :)

1 - You can safely assume that "Pulse Voltage" is the one supplied by your Power Supply. It is higher than the one needed for non-multiplexed operation to guarantee a faster strike.

2 - Both modes A and B are safe from an operational standpoint and are given as indicators of how far you can go with current for the target time and duty cycle (that's how PWM translates in this case).
What it means is that for each 0.1ms a tube is on at 10mA you must wait 0.9ms before turning it on again or you will burn its life away quite quickly. The same is valid for mode B with 1ms and 5mA: you must wait 9ms before turning that tube again on. There is a turn-on delay and also a turn-off delay so not all of this wait-time is really so. As a reference you can read this, from Mike Moorrees "Pile of poo" which is an extremely good guide to multiplexing nixie tubes and quite the opposite to his name choice :)

You need to be careful and stick to what the specifications say regarding to maximum voltage and currents, specially when multiplexing, or as I said before, you risk to burn your tubes lifetime quite quickly.

[scotth]

Thanks everyone for the responses.  I'm working on the IN-8 clock at the moment.  I set the voltage to 200V.  When i had the 22k resistors in, my current on one tube was 0.45mA, so multiplying that by six I only got 2.7mA.  I put a 10k in and am getting about 0.88mA which gives me 5.28mA, which is a little over what the datasheet says, is going over by that much OK?  I don't think the 5mA says maximum, does it?  Hard to translate the Russian.  I like the brightness much better on that than what I get with a 15k and I don't have anything in between.

Thanks, I have had the privilege of reading some of that excellent "Pile of poo" page.  I think my blanking meets the specs that he calls for.  I am a little confused by what you wrote, though.  I believe I have the anodes blanking for 0.2ms and then turning on for 1.8ms; I think this meets threeneurons' criteria of "Typical blanking intervals, for nixies, are between 100 to 300uS." and "It should be ON for a display interval between 1mS to 4mS" What you said seems to be the opposite, right? Off for longer than it is on?  I must be misunderstanding something.

[Tomasz Kowalczyk]

If we assume that IN-8 and IN-8-2 are electrically the same, then I've digged some info. For pulsed operation from 50Hz (one half rectifying? "one-half ??? without filter, sounds like just a diode) the average current should be 1mA. I think that we can extrapolate this data to just "average current in non-dc operation".

Which suggests that you've picked exactly the right resistor (10k)!

BTW IN-8-2 datasheet also states that DC current can be as high as 4,5mA - it just isn't in the operating region which guarantess at least 10000h operating time.

[Tomasz Kowalczyk]

Even more info!

https://tubes-store.com/images/in4_manual.jpg

Same goes for IN-4, there is specified current for half-wave rectified 50Hz voltage. It is 1,5mA for IN-4.

Aaaand we've got it! (As long as we assume that we can extrapolate the half-wave rectified average current to average current in any pulsed operation)

http://www.tube-tester.com/sites/nixie/dat_arch/russian_tubes_02.pdf

Here we can see that:

1. IN-8 and IN-8-2 are thrown onto the same page, so they are electrically the same

2. On the bottom of the page there is again mentioned half-wave rectified supply and the current is 1mA

3. There are A and B regimes mentiones again, but not explained at all

GastonP, where did you find data about what they are? 

[scotth]

Oh, since I am turning on one anode at a time, by the time I get back to the same anode, i have gone through 5 more 2ms steps so it has been 10ms, right?  Is that close enough, or should i adjust it so that it is 18ms, which would make the 1.8ms on time 10%?

On Tuesday, June 6, 2017 at 1:12:52 PM UTC-4, GastonP wrote:

[scotth]

Actually, I'm not sure how I'd meet that 10% and stay within threeneurons' specs.

[scotth]

Thanks for all the research!  Does it matter that, being in the US, my frequency standard is 60Hz?  Also, does the fact that I am using a standard 9v DC power supply make a difference?

[Tomasz Kowalczyk]

If you're using a wall wart and then a DC/DC converter, then it doesn't matter at all.

 All this 50Hz stuff comes from times when DC/DC converters weren't popular or cheap, so to make transformers smaller, the display was connected directly to mains, without any isolation.

Some nixie clocks use it, I believe I've seen one project using biquinary nixies, where one anode was driven by one half of sine wave, and other anode was driven from the other. 

This method is generally not recommended for safety reasons - both user and tube safety, you should always isolate your devices from mains via a transformer or a wall wart. 

[scotth]

I'm using a switch mode boost converter with a 9VDC input, so I think i'm good on that count.  The high voltage is scary enough, I don't think I'd want to run the nixies directly from mains :/.  I think I'll go with the 10k unless anyone thinks I should be more conservative and use a 15k.  If anyone can clear up the point that GastonP made about the duty cycle, that would be awesome.