Multiplex medium size tubes 3+3

89 views
Skip to first unread message

Jonathan F.

unread,
Feb 20, 2016, 9:16:03 AM2/20/16
to neonixie-l
Hello

I'm trying to make a new board for medium size nixie tubes like the IN-8 and smiliar ones.
Since the PCB cant be very big in size i'm thinking about multiplexing the tubes.

But not the standart-multiplexing with an 7414 IC. I'v used the HV5122/5222 32-bit HV shift registers in my last projects (direct drive) and i like to use just one of these in this project.

That would result in two position multiplexing because the IC can handle only 3 tubes at a time.

To load the HV5122 it takes 5us.
And i would leave the digits on for 15us for example.
This would result in 20us cycles at a 50/50 duty rate.

Has anybody experience with multiplexing?

Is 20us too long, just right, too short?
I would like to see no flicker, if possible

Thanks for your help.

John Rehwinkel

unread,
Feb 20, 2016, 9:22:26 AM2/20/16
to neoni...@googlegroups.com
> I'm trying to make a new board for medium size nixie tubes like the IN-8 and smiliar ones.

I just did that for a divergence meter.

> Since the PCB cant be very big in size i'm thinking about multiplexing the tubes.
>
> But not the standart-multiplexing with an 7414 IC. I'v used the HV5122/5222 32-bit HV shift registers in my last projects (direct drive) and i like to use just one of these in this project.
>
> That would result in two position multiplexing because the IC can handle only 3 tubes at a time.
>
> To load the HV5122 it takes 5us.
> And i would leave the digits on for 15us for example.
> This would result in 20us cycles at a 50/50 duty rate.
>
> Has anybody experience with multiplexing?
>
> Is 20us too long, just right, too short?
> I would like to see no flicker, if possible

I'd say 20µs is WAY too short. Due to stray capacitance and anode switching time, you're likely to get severe ghosting.

What you usually need to do is turn one on, then off, have some dead time, then turn the other one on, then off, then more dead time. 100Hz (10,000µs) is fast enough
to avoid visible flicker, but you can go faster if you want to. Note that the faster you go, the higher anode voltage you need so the tubes will ionize quickly enough.

- John

threeneurons

unread,
Feb 20, 2016, 11:14:29 AM2/20/16
to neonixie-l
Like John said, 20uS is way too short. Most nixies won't even respond. It can take close to 50uS for a nixie tube just to ionize. Here's an old scope, photo I took, of some signals going into a NL-905 nixie tube:

I stepped the anode voltage On for 4mS, then kept it Off for 12mS, and repeated, for a 16mS total period, or ~64Hz.


The Anode is turned On at the beginning of the trace. Notice it takes 30mS for the tube to even start conducting. Full conducting at ~50mS. It varies some with different model tubes, and quite a bit with anode voltage. I did some speed tests, some 10 years ago, but lost the file. I used a few types of nixies, and varied the anode supply from 180 to 200V. Got the best results with a ZM1000, at 200V, and using its primer (-100V 10M ohm) electrode. I got it to strike in 7mS. The presents of light also speeds up ionization.


When I multiplex nixies, I nominally use ~200uS interdigit "blanking". Blanking is when no digits are On. Nothing set in stone, about 200uS. That's just close to the minimum, I suggest. Any time from ~150uS to 1mS is good.


100Hz (10mS) total frame rate (cycle thru all digits) was mentioned. So if its 3 x 3, then each set of 3 digits is one for roughly 3.3mS, less the blanking interval. 3.3+3.3+3.3 roughly equals 10 (9.9 is close to 10 for our purposes). Personally, I can't see any flickering, even as slow as 40Hz (25mS), but I'm old. You can always vary the frame rate, and test it on your self. If you're not going to do anything fancy like "cross fading", then that 100Hz is a good figure to use.

threeneurons

unread,
Feb 20, 2016, 11:19:16 AM2/20/16
to neonixie-l
Oops, the second paragraph should be uS not mS is several spots.


On Saturday, February 20, 2016 at 8:14:29 AM UTC-8, threeneurons wrote:
Like John said, 20uS is way too short. Most nixies won't even respond. It can take close to 50uS for a nixie tube just to ionize. Here's an old scope, photo I took, of some signals going into a NL-905 nixie tube:

I stepped the anode voltage On for 4mS, then kept it Off for 12mS, and repeated, for a 16mS total period, or ~64Hz.


The Anode is turned On at the beginning of the trace. Notice it takes ! 30mS (30uS) for the tube to even start conducting. Full conducting at ~ ! 50mS (50uS). It varies some with different model tubes, and quite a bit with anode voltage. I did some speed tests, some 10 years ago, but lost the file. I used a few types of nixies, and varied the anode supply from 180 to 200V. Got the best results with a ZM1000, at 200V, and using its primer (-100V 10M ohm) electrode. I got it to strike in  ! 7mS (7uS). The presents of light also speeds up ionization.



Jonathan F.

unread,
Feb 20, 2016, 2:12:45 PM2/20/16
to neonixie-l
Thanks for quick response, im happy i dont have to care about the speed of optocouplers :)

Now i will just try and see what will happen :)

Sadly, i personally see flickering better that other people...  :(


One question, why does it need a blanking-time?

gregebert

unread,
Feb 20, 2016, 4:01:30 PM2/20/16
to neonixie-l
Any particular reason why you dont want to use direct-drive ? There's a slight additional cost, but you eliminate all of the problems with multiplexing (flicker, ghosting, audible noise, less brightness).

If PCB size is your concern, then use an HV5530 driver. Even though it's surface-mount, it's pretty easy to solder it and 1 device will drive 3 tubes. My latest clock has six IN-18  tubes, plus all the connectors, drivers, level-translator, and current-limiter for each tube on a 8.5" x 3.25" PCB. Half the PCB area is the IN-18 footprint.

 

Jonathan F.

unread,
Feb 20, 2016, 4:04:31 PM2/20/16
to neonixie-l
Hi, HV5530 is very smiliar to HV5122  ans is also a 44pin QPF.... so the sizes are the same :)
Reply all
Reply to author
Forward
Message has been deleted
0 new messages