to multiplex or not

[linux-works]

hi, new member here.

I'm trying to decide if I have to mux the nixies or if I can run them constantly on?

what are the design trade-offs or issues to worry about?

is it a dc-dc converter power issue?  will it overheat if you drive more than 1 digit at a time?  how many can you drive, safely, at a time?

is it a lifetime issue for the tubes?

my application (one of them) will be as a display of volume level inside an audio system.  if I don't have to mux the tubes, I'd just assume not.  it will be much less activity for the arduino and less noise inside my audio chassis.

I can add extra 'pins' via port expanders and I already have 2 shields that I can use for the 4 digits that I need.  I can mount one (or both) shields on a board by themselves and then wire them to the arduino along with extra port expanders to drive the 2nd shield.  the 2nd shield could have its own dc-dc 555 circuit on it to power its pair of tubes, if one 555 is not enough.

looking for some advice on this.

thanks

[Risen]

Here's the best discussion I've come across:
https://groups.google.com/forum/#!topic/neonixie-l/gqA6OpDbdXM

[linux-works]

thanks.

it seems that if I choose to go non-mux, I can eliminate the driver transistors and basically just tie all anodes to +170v and simply send bcd nibble changes to each 74141 chip and latch them there until the digit/nibble changes.

I measured about 2ma of current on my nixies and so with 4 nixies on at the same time, its only 16ma of current.  I would think that the 555 circuit can drive this, but I guess I'll find out

if I use the arduinix shield, the transistors are there already and I guess I would either not map them to arduino pins or map them to a single pin and make that pin pwm-able so I can dim the lights.

I've used PCF i2c port expander chips before and will likely bring one in for the extra set of 8 bits I need.

I'll post the progress as I get further.

thanks.

[Risen]

it seems that if I choose to go non-mux, I can eliminate the driver transistors and basically just tie all anodes to +170v and simply send bcd nibble changes to each 74141 chip and latch them there until the digit/nibble changes.

You might not want to do this if you plan to have blank digits. While many 74141 chips use the high values (sending any hex digit A-F) as "blank" on the tubes, I've read that some don't, so that could cause an problem. At any rate, I found my chips to leak a fair amount of current when "off", so killing the anode line appears to be more efficient than sending a "blank" digit.

If you use I2C, make sure you add filter caps on the data lines. I ran an MCP23008 on mine to control the separators. The ArduiNix is noisy, though, and kept interrupting the I2C communication until I filtered it.