Is /any/ kind of Nixie Multiplex-compatible?

161 views
Skip to first unread message

Per Jensen

unread,
Jun 6, 2015, 8:47:21 PM6/6/15
to neoni...@googlegroups.com
Hi.

Being a Nixie-hoarder for many years, i’m actually guilty of not having a working clock at home. The only clock i have working is sitting in the local hackerspace. It’s an direct driven design and have been working perfectly for about 6 years now.

As i have a million “real” projects that i am working on, ii’m pretty sure that getting the time to design a clock “properly” isn’t just around the corner.

On one of those late-night shopping-sprees i came across this board: http://switchmodedesign.com/collections/arduino-shields/products/open-source-nixie-tube-shield-cut-jump-pcb and ordered 2 pcs.

I “trusted” the website-name (switchmode designs) as the board being designed by one that knew what he was doing, so i didn’t check schematic or board layout before i had the boards in hand …. Maybe i should have done that, because that switcher is the worst design i have /EVER/ seen.

Anyway, apart from that, it works ok-ish.

I wanted to use this board as a “backend” for my own board with tubes. I quickly whipped up a board for 4 IN-4’s: https://www.dropbox.com/s/dsk1hv2xly7e9v2/Nixie.png?dl=0

It’s a milled PCB so that’s why the copper is left around the tracks.

Now i am having this stupid problem - ghosting!

No, it’s not software ghosting, because if i only connect a single tube with two wires (the wide cable isn’t connected) , sometimes the neighboring tube lights up, too.

So somehow the tubes being connected only by the cathodes and the anodes floating, powering one can excite the others.

I even removed the pins for the tubes so the anode pins of the unused tubes is not even touching the fibreglass-material. Only cathodes is thouching.

(anode-pin connected by flying wire to anode resistor and driver) https://www.dropbox.com/s/y0xhc9gazrxa120/nixie3.jpeg?dl=0

Is there anything to do here, or should i scrap the boards and design my own direct-driven circuit?


// Per.

blkadder

unread,
Jun 6, 2015, 8:58:29 PM6/6/15
to neoni...@googlegroups.com
I built one of the Switch Mode cut and jump shields, and I didn't have any issues at all.  I did have to adjust the voltage, but I got it working without incident.  Have you contacted anyone over at SWD?  They are really good about customer service.  Not sure why you are having the ghosting.

Per Jensen

unread,
Jun 6, 2015, 9:03:31 PM6/6/15
to neoni...@googlegroups.com
Well the issue has nothing to do with the board i bought, it’s with my own board/tubes that i’m having problems with. So disregard that i am using that board to essentially boost the voltage to 180V and nothing more right now.

As it is right now, disconnecting the 20-way cable from my board and only connecting a current-limited power supply to one of the cathodes and an anode will make a tube light up, and usually one of the others, too. That’s my problem.

// Per.

Dekatron42

unread,
Jun 7, 2015, 8:11:49 AM6/7/15
to neoni...@googlegroups.com
I think that you should try to lower the value of the 100k resistor across the base-emitter of every MPSA92 to say some 10k-22k, they will not be fully turned off in some instances with the 100k resistor.

Did you use the correct anode pin for the IN-4? Have look here: http://www.tube-tester.com/sites/nixie/data/in-4/in-4-sh2.htm , they can be driven in bi-quinary or normal mode and the anode pins are not in the same place depending on what mode is used.

What resistance do you use for the anodes, maybe you need to change that for the IN-4 to work properly compared to the IN-17?

/Martin

Per Jensen

unread,
Jun 7, 2015, 10:41:56 AM6/7/15
to neoni...@googlegroups.com
Please disregard the shield that i am using.

It doesn’t exist right now.

On my bench i have a 180V power supply, my board, two wires and a resistor.

I tried with 2.5 / 3mA current per tube.

Connecting the power supply to the anode-pin (Pin4) on the tube trough a suitable resistor for 2.5mA current and one of the digits to GND.

The digit lights up, but some of the positive current leaks into the other tubes, making them light up too. The only common connection between the tubes is the cathodes.

As you can see, right now Pin4(Anode), Pin7(Not connected), Pin10(Not connected), Pin13(Not connected)


Am I having these issues because of the tubes being semi-biquinary?


// Per.

Dekatron42

unread,
Jun 7, 2015, 10:58:20 AM6/7/15
to neoni...@googlegroups.com
Can you show a photo of the ghosting?

Do you have more than one tube connected when the ghosting appears or just one?

Have you tried biasing resistors on the cathodes?

/martin

David Forbes

unread,
Jun 7, 2015, 1:08:13 PM6/7/15
to neoni...@googlegroups.com
Per,

Hi. I built many copies of several types of Nixie clocks about 10-12
years ago, using multiplexing. They worked very well, with no ghosting
at all.

Things to look for:

1. Ensure that you have a "dead time" after the anode is turned off,
with the same cathode selected, for about a millisecond. The reason is
that if you immediately change the cathode after turning off the anode,
the residual charge on the anode will cause the newly selected cathode
to light in the tube that was just turned on.

2. Any cables connecting the tubes to the drivers will extend this
storage time due to cross-coupling between the wires, so experiment with
longer "off" time with the previous cathode still selected.



On 6/6/15 5:47 PM, Per Jensen wrote:
> Hi.
>
> Being a Nixie-hoarder for many years, i’m actually guilty of not having a working clock at home. The only clock i have working is sitting in the local hackerspace. It’s an direct driven design and have been working perfectly for about 6 years now.
>
> As i have a million “real” projects that i am working on, ii’m pretty sure that getting the time to design a clock “properly” isn’t just around the corner.
>
> On one of those late-night shopping-sprees i came across this board: http://switchmodedesign.com/collections/arduino-shields/products/open-source-nixie-tube-shield-cut-jump-pcb <http://switchmodedesign.com/collections/arduino-shields/products/open-source-nixie-tube-shield-cut-jump-pcb> and ordered 2 pcs.
>
> I “trusted” the website-name (switchmode designs) as the board being designed by one that knew what he was doing, so i didn’t check schematic or board layout before i had the boards in hand …. Maybe i should have done that, because that switcher is the worst design i have /EVER/ seen.
>
> Anyway, apart from that, it works ok-ish.
>
> I wanted to use this board as a “backend” for my own board with tubes. I quickly whipped up a board for 4 IN-4’s: https://www.dropbox.com/s/dsk1hv2xly7e9v2/Nixie.png?dl=0 <https://www.dropbox.com/s/dsk1hv2xly7e9v2/Nixie.png?dl=0>
> https://www.dropbox.com/s/yn42qtdki3vr7bo/Nixie2.jpeg?dl=0 <https://www.dropbox.com/s/yn42qtdki3vr7bo/Nixie2.jpeg?dl=0>
>
> It’s a milled PCB so that’s why the copper is left around the tracks.
>
> Now i am having this stupid problem - ghosting!
>
> No, it’s not software ghosting, because if i only connect a single tube with two wires (the wide cable isn’t connected) , sometimes the neighboring tube lights up, too.
>
> So somehow the tubes being connected only by the cathodes and the anodes floating, powering one can excite the others.
>
> I even removed the pins for the tubes so the anode pins of the unused tubes is not even touching the fibreglass-material. Only cathodes is thouching.
>
> (anode-pin connected by flying wire to anode resistor and driver) https://www.dropbox.com/s/y0xhc9gazrxa120/nixie3.jpeg?dl=0
>
> Is there anything to do here, or should i scrap the boards and design my own direct-driven circuit?
>
>
> // Per.
>


--
David Forbes, Tucson AZ

Per Jensen

unread,
Jun 7, 2015, 2:03:42 PM6/7/15
to neoni...@googlegroups.com
Well, i guess it’s my own fault starting to describe the “driving”.board when having this problem that i put myself into.

What you see on these two pictures is what i am experiencing. No microcontroller, Arduino or shield is used here. Only my homemade board with 4 tubes.


What is off camera is a regulated lab power supply set at 180 Volts.

For the sake of eliminating capacitively coupling from one anode to the others i even removed the anode pins from the PCB. I have soldered a 15K resistor direcly on the anode-pin (Pin4) which is in turn connected to the power supply.

As you see, one of the other tubes is lit, too!

So somehow the anode voltage is capacitively coupling inside the tubes to a nearby cathode and trough that, powering another tube.

I /do/ know about blanking in software, but that’s not the problem here.

I don’t know what i could do about that.

// Per.


--
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/55747A7B.3000108%40dakotacom.net.
For more options, visit https://groups.google.com/d/optout.

John Rehwinkel

unread,
Jun 7, 2015, 2:14:03 PM6/7/15
to neoni...@googlegroups.com
> What is off camera is a regulated lab power supply set at 180 Volts.

So, a DC supply, no multiplexing.

> As you see, one of the other tubes is lit, too!
>
> So somehow the anode voltage is capacitively coupling inside the tubes to a nearby cathode and trough that, powering another tube.

If it's a DC supply with no multiplexing, there's no AC component, and therefore no capacitive coupling.

What you're seeing is leakage, AKA resistive coupling. There's probably some flux residue, dust, fingerprints, or similar on the board, providing
a high resistance leakage path, which is enough to get a partial glow like that.

- John

Dan Harboe Burer

unread,
Jun 7, 2015, 2:29:09 PM6/7/15
to neoni...@googlegroups.com
..or a very thin piece of copper strand left over from the milling.. have
you done a VERY close inspection of the board? (magnifier or a microscope if
available?)

Dan


-----Oprindelig meddelelse-----
From: John Rehwinkel
Sent: Sunday, June 07, 2015 8:14 PM
To: neoni...@googlegroups.com
Subject: Re: [neonixie-l] Is /any/ kind of Nixie Multiplex-compatible?
--
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/258279AC-5824-4C8F-AACF-6F7689ABDC3A%40mac.com.

Per Jensen

unread,
Jun 7, 2015, 3:50:42 PM6/7/15
to neoni...@googlegroups.com
Well, tried cleaning the board to the extremes, inspecting under microscope, removing the IDC-connector etc. and no change.

I’m feeling that i’m overlooking some simple, so i whipped up a set of wires connecting two nixies only by wires - only the 10 cathodes are connected, not the anodes.

Now this is what happens:


Can we all agree on it’s not the PCB that is at fault - it’s the tubes …

Even if i ground the anode on the offending tube it still lights up.

// Per.


--
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.

David Forbes

unread,
Jun 7, 2015, 4:37:11 PM6/7/15
to neoni...@googlegroups.com
Per,

Very interesting result.

Does this experiment produce the same result with fewer wires connecting
the tubes? I so, what number of wires is needed to produce this result?

Is there any AC component to your HV power?

Does this experiment produce similar results with a different type of
tubes?

Connecting resistors to the unused cathodes to set their quiescent
voltage level, may change the results.

On 6/7/15 12:50 PM, Per Jensen wrote:
> I’m feeling that i’m overlooking some simple, so i whipped up a set of wires connecting two nixies only by wires - only the 10 cathodes are connected, not the anodes.
>
> Now this is what happens:
>
> Can we all agree on it’s not the PCB that is at fault - it’s the tubes …
>
> Even if i ground the anode on the offending tube it still lights up.
>
> // Per.
>


Dekatron42

unread,
Jun 7, 2015, 5:35:43 PM6/7/15
to neoni...@googlegroups.com
This is due to the unused cathodes floating, as they are not supplied with a voltage that ensures that they are turned off - use biasing resistors and they will turn off.

Have a look at this thread: http://www.vintage-radio.net/forum/showthread.php?t=82934 where there is a nice circuit diagram at the beginning, click on the small thumbnail, and you will see a proper biasing network for the cathodes. This is a circuit for a small motorized clock.

/Martin

Per Jensen

unread,
Jun 7, 2015, 5:57:46 PM6/7/15
to neoni...@googlegroups.com
Did some testing now, I pulled wires off until the problem went away. I only need two wires connected to reproduce the issue!

See this:

Cathodes for 2 and 6 is connected, and it works with both 2 and 6 illuminated in the “master” tube.

Very strange :O

My high voltage bench supply is switchmode-based and there is a small amount of noise on the output (approx 11mV, so that’s not the issue)

I did these last two tests with a 100uF/250V cap over the supply too.

I will have a look at trying some Biasing resistors like Martin mentioned!

// Per.

Dekatron42

unread,
Jun 7, 2015, 6:19:05 PM6/7/15
to neoni...@googlegroups.com
Also do some tests where you raise the voltage to 200V with a correct anode resistor as some of the Russian datasheets mentions a minimum anode voltage via an anode resistor of 200V to get proper function, 180V is a bit low.

You can also try to connect all three of the anodes together, pins 4, 10 & 13 to see if it alters the behaviour, or use any of the pins 10 or 13 instead of the pin 4, just use the correct anode resistor and all will be fine.

/Martin

Per Jensen

unread,
Jun 8, 2015, 6:14:46 AM6/8/15
to neoni...@googlegroups.com
I adjusted the voltage to 200V and used 22K as anode resistor. Adding the biasing network (47K for each cathode and then 150K to B+) did the trick. Now i see 52-57V on the different extinguished cathodes :D

I tried using the “No connect"-pins as anodes with varying success. The best readout is obtained by using the pin4 (at least on my tubes)

// Per.


--
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.

Dekatron42

unread,
Jun 8, 2015, 6:57:35 AM6/8/15
to neoni...@googlegroups.com
Nice!

Thanks for reporting what happened when using the other anodes, that is good to add to the common knowledge!

There is a good document written by Burroughs which discusses multiplexing and special considerations for their Nixies here: http://worldpowersystems.com/archives/Burroughs/ the N101 document at the top, you will also find two more documents there that are well worth reading.

Using a biasing network was very often done in early Russian designs if they included some switching of the Nixies but most designs used direct drive, usually by a KD155ID1. The Polyatron A-201 also uses a biasing network to make sure that the digits are off in the Nixies it drives, but also for making sure that the stepping of the Polyatron works smoothly when the current drawn by the Nixie varies between the different digits that are ignited.

/Martin
Reply all
Reply to author
Forward
0 new messages