Pulsed operation and anode current

[David Pye]

Hi everyone and thanks for letting me join the group!

I am trying to build a multiplexed clock using IN-14 nixies, and have run into a problem.

In order to get acceptable brightness, in pulsed mode, the current has to be much higher to make up for the fact that it is only on for short pulses.

From the data sheet https://tubehobby.com/datasheets/in14.pdf  :

For 'direct drive':

Firing voltage (no more than) 170V 

Current for digits (no more than) 2.5mA 

Multiplex mode :

Power supply voltage 190V 

Average current for digits 0.7 - 1.5mA 

Pulse current for digits 7 – 13mA 

Pulse width (no less than) 70µS 

Period 1 – 1.8 kHz 

However, I cannot achieve those currents.

When I tried calculating the anode resistor value I wanted, (using http://www.csgnetwork.com/anoderescalc.html) - using supply 170, sustain ~130V, current 13mA, it suggests an anode resistor of around 2.6K, but when trying this value, I found that the current that flowed was more like 6mA, and the voltage across the resistor was 10V, and 160V across the nixie.

For further testing, I briefly tried WITHOUT *ANY* anode resistor.    It was extremely bright, with an anode current of around 8mA - and not more!      I checked the PSU voltage while doing this, and it was still at 170V, so the voltage wasn't sagging.

Obviously this would kill the tube if left on for any length of time, but for testing purposes it illustrates my issue.

I've tried briefly with a couple of IN14s and even an IN12, and the above is roughly consistent.

So this leads me to a couple of questions:

1)   Seeing as the data sheet seems to take care to specify a higher voltage for multiplex mode, am I to assume that for some weird reason, the nixies don't behave as I'd expect when you are trying to flow larger currents?  ie they don't actually try to pull the supply voltage down all the way to Vsustain?  Do I need to increase Vsupply to 190V to achieve higher currents e.g 13mA?

2)   If I am happy with the current of 8-9mA at 170V without an anode resistor while multiplexing, does it mean that I can (in this specific application) get away without an anode resistor then?

Suggestions welcome!

Thanks

David

[Bill van Dijk]

How are you measuring the pulse current? This can only be done reliable with a scope measuring the voltage over the anode resistor during the pulse, and calculate the current. You can make a fair approximation by subtracting the tube “on” voltage from your supply voltage, dividing that by your resistor. Any attempt to use an amp meter will NOT give an accurate result.

 

Bill

--
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/7f8e1fd5-d144-452b-8805-35847d8204f4%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[gregebert]

Welcome to the group; I assume no responsibility if you get addicted to nixie tubes, and all sorts of vintage display devices. 

Be very careful about how much current you pass thru a nixie; too much will cause premature wearout and too-little can result in cathode poisoning.

Also, be aware that nixie tubes have non-linear characteristics, especially in the region where ionization occurs.

Initially, the current in the nixie is zero as you increase the voltage to just below the point where ionization occurs. As the voltage is slightly increased you can see very faint glowing, and the current is usually below 100uA. When the voltage is bumped-up slightly, the voltage drop across the tube reduces and the current increases. This region is unstable and exhibits negative resistance. If you limit the current with a resistor, you will see a steady glow and the voltage across the tube will be significantly less than the ionization voltage. The graph below is data I collected for a b7971 tube. Visible glowing occurs above 135V; ionization started as high as 180V in some cases . Once the cathode has ionized, though, it has reasonably linear behavior: More voltage across the tube results in more current, hence higher brightness.

Now, getting to your question, you definitely want an anode resistor to limit the current; without it your design will be unstable and the nixie current will be determined by the tube's raw characteristics, which vary from tube-to-tube, as well as over time.

I suggest you learn first with a direct-drive design, then work on multiplexing.

[David Pye]

Hi,

I'm measuring the current very simply here.  For a brief test (and accepting it may harm the nixie), I simply connected it directly across the 170V supply.  I had made a number of tests with increasingly lower values of current limiting resistor prior to doing this.

What I saw was an anode current of around 8-9mA.  The 170V of the PSU didn't fall so at this current level the nixie's own internal 'resistance' appears to be limiting the current to around 9mA, and not the 13mA that I want to achieve.

I'm assuming therefore, that in order to get the current I want, (and have 'room' to include a current limiting resistor of sorts) I'll need to go up to 190V.

David

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/007c01d4b1d6%24c4df4030%244e9dc090%24%40gmail.com.

[David Pye]

Hi Greg,

Thanks!

I've made a number of direct drive designs before, and when trying to pass a current of a couple of mA through one of the small tubes, I've used an anode voltage of 170V, an anode current limiting resistor as calculated, and had no problems achieving the current flow I want.

The problem has occurred here when I've wanted a very much larger current (13mA in this case), and found that I cannot achieve this on 170V even without a limiting resistor!

I've concluded from this, and from the graph you've kindly shared that if I want to achieve 13mA anode current I will have to have a higher voltage across the tube, especially if I want 'room' for want of a better word, to include a current limiting resistor.  I say that because at 170V i am already failing to achieve my desired current without one, so hopefully at 190V (which is the voltage the datasheet specifies for multiplexing) I'll be able to include one.  This means I can, for example, get a 20V drop across the limiting resistor, have a 170V drop across the tube and have a 'limited current' of around 8mA, which I'd be happy enough with here.

The thing that is surprising me here is that I expected the voltage across the tube to fall to Vsustain once conduction started, but when trying to pass a larger current, it isn't doing that.  For example, this tube is specified to have a Vsustain at around 135V.  However, when I had a 2k current limiting resistor, I was seeing a 10V drop across it, and a 160V drop across the tube.

Presumably, once you hare above the 'unstable region' you describe, it demonstrates positive resistance again, and current flow is proportional to voltage across it?

I'm assuming this behaviour is why it specifies 170V as anode voltage for direct drive, and 190V for multiplexing (presumably to achieve the higher pulse currents)?

David

--
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/513eb93d-3672-42c6-a6b9-ded554048f7c%40googlegroups.com.

[Paul Andrews]

What is your pulse width and period? The brightness will be greatly affected by the on time of the pulse as it takes a while for the tube to turn on, meaning it isn’t on for the whole of the pulse on-time.

You can decrease how long it takes the tube to light up by holding the unlit cathodes at between 70V and 100V.

Like Bill said, you need to use an oscilloscope to see what is actually happening to anode and cathode voltages. A multimeter is just going to give you averages.

[Paul Andrews]

I should add that using a higher voltage also reduces the time it takes for a tube to light up,, which is probably why the datasheet recommends it.

> --
> You received this message because you are subscribed to a topic in the Google Groups "neonixie-l" group.
> To unsubscribe from this topic, visit https://groups.google.com/d/topic/neonixie-l/2idq-HQuKMs/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to neonixie-l+...@googlegroups.com.
> To post to this group, send an email to neoni...@googlegroups.com.
> To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/877447b2-0ae8-4a0d-82e6-de75d0d299ea%40googlegroups.com.

[David Pye]

That makes sense to me.

I have been testing, as I say, with continuous current purely to get the anode resistor value set 

Even with the full 170VDC across the tube, the current is around 8mA.  I might decide this is acceptable enough from a brightness I want, but it leaves no room for anode resistor....

It is interesting to me, as most of the calculations rely on the tube pulling the voltage down to Vsustain once it strikes but it seems here that that doesn't happen when passing these kinds of current.

David

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 an email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/3A457756-DC1B-4DD7-AA66-39238516C6B0%40gmail.com.

[gregebert]

Assuming the 170VDC supply is not adjustable, you can easily put another DC supply in-series with it. There are a variety of small DCDC converters available with isolated outputs; just connect the (-) output of the converter to +170VDC, and power your nixies from the (+) output. A +24VDC output will get you above the 190V recommended per the datasheet excerpt.

If the +170 supply is a switchmode converter, you might be able to change the feedback resistor to get higher voltage, but you need to realize that will cause it to run hotter so it's risky.

I usually run my tubes from +210VDC (boosting the line voltage) or +340VDC (diode-capacitor voltage doubler from AC line). I aslo use simple current regulators (NPN or NMOS on the cathode-side ; PMOS on the anode side) for most of my designs so the current is constant regardless of line voltage or tube condition.

As long as the current is limited, running a nixie tube at a higher voltage is not harmful, though it wastes more energy.

[chuckrr]

Sounds like you are going to wreck something!

I have a complete (16) tube text-scanning array built with B-7971 tubes.
It's direct drive.

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/CAOQ6x0HQqoMQgwXS5WaKgip3hkppKmCNm5Zg66ahTdkhPdem1A%40mail.gmail.com.

[Bill van Dijk]

You can see why the specs recommend 190V as the recommended for multiplexed mode. Below 190 to 200V you will not get it to work properly.

 

Bill

 

From: neoni...@googlegroups.com [mailto:neoni...@googlegroups.com] On Behalf Of David Pye

Sent: Monday, January 21, 2019 4:35 PM
To: neonixie-l <neoni...@googlegroups.com>

--

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/7f8e1fd5-d144-452b-8805-35847d8204f4%40googlegroups.com.

[David Pye]

But interestingly it is working within specified current but without the anode resistor.

I am not sure what advantage I would get increasing the anode voltage to 190v and having a resistor drop 20v to leave my 170v and 8ma current flow through the tube.

Presumably the only advantage would perhaps be quicker ignition?

David 

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/00fe01d4b287%24d1ddcc50%24759964f0%24%40gmail.com.

[Bill van Dijk]

Indeed better “ignition”, as well as better current control. It also provides the headroom to control the current with a resistor. Running the tube without a resistor to limit (and control) the current, you’re playing Russian roulette with your tubes.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/CAOQ6x0GTLY1bwVwvqyYSsVc3s%2BvWEAkFXukayouoB31ecdKzsg%40mail.gmail.com.

[gregebert]

The advantage and importance of the series anode resistor is that it makes your anode current more predictable. Without the resistor, your anode current will be determined primarily by the tube's characteristics, which vary over time and tube-to-tube, ie unpredictable. I think that tube current can start to increase exponentially at a certain point and once that happens it will rapidly heat-up and get destroyed. I took this to an extreme with a 0A2 gas regulator tube, and it got so hot the glass melted. It was hilarious.

You can prove mathematically that the larger your anode resistor is (which of course requires a higher anode supply voltage), the less-dependent the anode current is on the tube's unpredictable characteristics. The drawback is wasted energy.

[Jon]

On Tuesday, January 22, 2019 at 9:05:31 PM UTC, gregebert wrote:

The advantage and importance of the series anode resistor is that it makes your anode current more predictable. Without the resistor, your anode current will be determined primarily by the tube's characteristics, which vary over time and tube-to-tube, ie unpredictable.

And that's important in a multiple tube display (which after all is the motivation to entertain the design idea of multiplexing), because you don't want different currents and therefore potentially visibly different glow intensities across the array. Quicker/more reproducible ignition is also important in the multiplexed setting also because variations in the actual duration of glow during the multiplex time slice are going to be perceived as intensity differences too.

I can't explain exactly what's happening in your current experiments, but suggest that you follow the theoretical & datasheet advice: 190V and a calculated anode resistor that puts the anticipated current in the indicated range. And then adjust the anode resistor experimentally to give you acceptable brightness when driven in the multiplex setting. You may be surprised by how little additional current is needed above the direct drive steady illumination range to generate a perfectly acceptable brightness while multiplexing. That depends of course on the multiplexing ratio that you are looking to use: a 1 x 6 multiplex is going to need to need the peak current pushed harder than using the same 6 tubes in a 2 x 3 multiplex (which was my preferred arrangement on my IN12 42 tube clock).

Jon.

[Dekatron42]

You can look at the Burroughs datasheet: https://frank.pocnet.net/other/Burroughs/B5750B585X.pdf for some details on the multiplexing of Nixies, there are some notes and also diagrams that explain the necessary voltages and currents in relation to the multiplexing timing there. I also know that the German manufacturer RFT had datasheets for multiplexed driving, unfortunately I don’t have those datasheets, I’ve only read about it in German books on display technology.

/Martin

[David Pye]

Thanks for that, Martin,

This graph is really helpful, and is extracted from the Burroughs data:

This is for a different tube, but it does suggest that a current of around 8mA (which is what I am getting) with 170V across the tube is not entirely unexpected.

I will report back when a few new supplies I ordered (which can deliver a higher current) have arrived, and I can then include an appropriate anode resistor.

It looks like I will have to figure out the anode resistor value experimentally, because at these currents there is no such thing as a 'Vsustain' voltage to make it simple to calculate the anode resistor value, because in pulse mode, the graph suggests that the current in this operating range increases linearly with voltage, and the inverse resistance behaviour doesnt occur here....

David

On Wed, 23 Jan 2019 at 10:25, Dekatron42 <martin....@gmail.com> wrote:

You can look at the Burroughs datasheet: https://frank.pocnet.net/other/Burroughs/B5750B585X.pdf for some details on the multiplexing of Nixies, there are some notes and also diagrams that explain the necessary voltages and currents in relation to the multiplexing timing there. I also know that the German manufacturer RFT had datasheets for multiplexed driving, unfortunately I don’t have those datasheets, I’ve only read about it in German books on display technology.

/Martin

--
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 an email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/f9b96415-4581-4573-bc69-029b70ddb8d6%40googlegroups.com.

[Dekatron42]

If you want more information on how Nixies behave in pulsed mode you should read the book ”Winfried Müller
Elektronische Anzeigebauelemente Electronica 171”, a book from the 70’s from the East German military publisher (DDR). It is written in German but scanning it with OCR and then using Google Translate makes it into easy reading. There are curves for the RFT Nixies there and also proper explanations on how to use pulsed mode. There are some errors in the figure numbering and some other small errors but in large it is a very nice read.

There are a few on eBay almost always and sometimes those sales can be found on Abesbooks or Amazon.

/Martin

[David Pye]

That's brilliant, thank you.

I have put it on order, with an intent to translate it, and hopefully share some of its' nuggets for other people who might be trying to follow in my footsteps :-)

On the positive side, I've managed to work out how to adjust my miniature nixie DC-DC converter board to get it to spit out ~190V, so I don't need to wait for other ones to arrive.

https://hackaday.io/project/163516-multiplexed-esp8266-nixie-clock-project/log/158689-modifying-the-high-voltage-dc-dc-converter

in case anybody wants to know how to do it!

David

--
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 an email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/98767f19-a8f3-432f-85db-ae4c8a2e4998%40googlegroups.com.

[gregebert]

Make sure you test the DCDC converter under full-load. Every one of these I've ever designed & built always worked beautifully under low load, and as the load increases all sorts of gremlins start creeping out of the jungle.

Some of them are easy to spot, like overheating. Others require you to go poking around with a scope to find excess voltage or current.

[David Pye]

Still testing,and indeed, finding gremlins :-/

Even the copy of Elektronische Anzeigebauelemente Electronica 171 has a gremlin in it, being largely misprinted with huge sections missing, including conveniently, the section on pulsed nixie operation.

On a slightly different note, when anode-multiplexing, I can't see why one anode resistor doesn't suffice, as only one tube is ever on at once.

I was pondering this kind of layout:

HV supply->Anode resistor->Multiplex drive circuitry->TUBES

Any reason why that would be ill-advised?

David

On Fri, 25 Jan 2019 at 20:53, gregebert <greg...@hotmail.com> wrote:

Make sure you test the DCDC converter under full-load. Every one of these I've ever designed & built always worked beautifully under low load, and as the load increases all sorts of gremlins start creeping out of the jungle.

Some of them are easy to spot, like overheating. Others require you to go poking around with a scope to find excess voltage or current.

--
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/11d50a9e-1533-450d-b34b-7d0c63aabe6b%40googlegroups.com.

[gregebert]

Are you seeing residual glow on digits/segments, aka ghosting ? If so, you will want to bias the inactive anodes to a voltage below the sustaining voltage.

There are several threads in this forum about ghosting with multiplexed displays

[David Forbes]

Hi, I have some experience building multiplexed Nixie clocks that work correctly. Firstly, do you have an oscilloscope? It's difficult to understand what is going on without one. 

I found that 190V was sufficient to run ZM1040 tubes with a 10k anode resistor shared by all tubes. I would have to measure it again to know the current and anode voltage, because it's been about fifteen years since I last made one. 

Ghosting is a problem unless you time the anode and cathode switching properly. I turned off the anode, waited a few milliseconds for the tube to turn off, changed the cathode selection, waited a few milliseconds, then turned on the next anode. I also used PC board cathode wiring instead of discrete wires, to reduce capacitance. 

I remember that the anode current was 5 mA, and the anode voltage was 140V, but I could be off.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/CAOQ6x0ErixCXr_SD2-6DNX_1kiGx2Dwd7KbaAh6BdptNXiXDUQ%40mail.gmail.com.

[Dekatron42]

Where did you buy the book? I’ve never seen an iriginal one that has been missing any chapters. Maybe you got a reprint from one of the many places that sell poor, and sometimes incomplete copies? There were two original books for sale on eBay the other day.

/Martin

[Paul Andrews]

Other things to try:

Pull-down resistor on the anode to make it drop to zero faster.

Pull up resistor on the cathodes to 75V. Cathodes act as additional anodes and floating cathodes can float at a high enough voltage for that to become noticeable.

[David Pye]

Hi Martin,

Via Amazon from a German bookseller (who had also listed it on abebooks).

I'm fairly sure it's genuine (it certainly looks and smells like a 1970s book, complete with the vintage paper feel).

A number (maybe 20) of the early pages are blank, so I suspect it's a quality control issue ;-)

Refund pending, will return.

And will try to buy again.  I've found that there are large numbers listed on ebay.de which I can't search on ebay.co.uk.

Will keep trying!

David

On Fri, 1 Feb 2019 at 07:36, Dekatron42 <martin....@gmail.com> wrote:

Where did you buy the book? I’ve never seen an iriginal one that has been missing any chapters. Maybe you got a reprint from one of the many places that sell poor, and sometimes incomplete copies? There were two original books for sale on eBay the other day.

/Martin

--
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 an email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/b6e1de76-f6d8-49e2-8963-1387034364e2%40googlegroups.com.

[David Pye]

Hi,

The gremlins seem to be largely PSU related - the oscilloscope is showing weird problems like the voltage decaying during the pulse (and not in the way you'd normally expect once the nixie conducts).

I have swapped to a different HV PSU module, which seems to have the capability to generate the necessary current and voltage, so I should hopefully be able to characterise my tubes a little better now.

I should be able to produce voltage/current graphs for the IN12 and IN14 ( or at least, mine!) which will aid anyone else hoping to multiplex and not fall into my various selection of pitfalls...

David

On Fri, 1 Feb 2019 at 00:05, gregebert <greg...@hotmail.com> wrote:

Are you seeing residual glow on digits/segments, aka ghosting ? If so, you will want to bias the inactive anodes to a voltage below the sustaining voltage.

There are several threads in this forum about ghosting with multiplexed displays

--

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/3720cf20-3631-46a4-beba-adee4e6059e1%40googlegroups.com.

[David Pye]

Hi David,

I do indeed have a scope and am starting to get the kind of results i wanted now I have changed out the PSU for a more capable one.

So far, from my testing:

HV 200V.

Anode resistor of 3k9.

Voltage across anode resistor during pulse is between 42 and 48v (depending on the individual tube - some are a little different)

This equates to a pulse current of about 12.3mA, which is within spec for an IN14, which specifies a maximum pulse current of 13mA.  (and no idea about an IN12, as I can't find any reference to multiplex current mode in any of its' data sheets, but as it's a similar size, I'm assuming similar values will be OK.

With two 'test' tubes I didn't seem to have any issues with ghosting as long as a sensible blanking interval is applied, but once I've put the whole assembly together, I will see what happens this time!

Thanks for the advice!

David

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/CAPbqtvcqdakNW30G6DK%3DkTnM8EDPbUc2R%3Du77eUUbHr9zx9SnQ%40mail.gmail.com.