Nixie clock build

[Da_Stier]

I recently came across a box of nixie tubes in my stuff.
I bought them many years ago, back when I was still at school, because you could get them for 1€ a piece, including the driver IC.
I don't really like the idea of building clocks out of them, simply because I think they are so much cooler and more versatile than that.
So I find a clock pretty boring.
However after doing nothing with them for over six years, I decided to build a clock. 

I decided to go with a simple 555 timer boost converter for the power supply, however after finishing it, I came across this website:
https://leap.tardate.com/electronics101/555timer/nixiepowersupply/
I really liked the idea of adding a feedback loop to a 555 timer, so I redid my supply and was very pleased with the outcome.

The rest of the circuit is relatively straight forward, it uses four soviet K155ID1 driver ICs, an arduino nano, a DS3231 RTC and a simple transistor to switch the neon bulb second dot.
(I wanted a stand alone second dot, I could have used the dots provided by the nixies)

For construction I went with magnet wire in a point to point style, left over from my tesla coil.

















If someone is interested in the arduino code, here it is:
code.zip

[TMaxElectronics]

Cool project. Can't ever go wrong with a nixie clock

I do however find it quite interesting how the building blocks of nixie clocks don't really seem to have changed much since they became popular. The drivers are still the same old obsolete K155 style ones, and the boost converter still based on a (almost) discrete setup with a 555 timer. With catalog suppliers having so many better parts in stock (and available for normal people to buy) I really don't understand why these circuits aren't being updated... apart from maybe the arduino

In my nixie clock I used a BD9306 boost IC that has all the regulation and everything built in, which can easily run four IN-12 tubes at full power from just a usb supply without even getting hot.
And for driving the tubes I use a HV5122 250V open drain 32bit shift register, that has an output enable input that can be used to pwm the tubes, which makes them dimmable to almost 0% before the illumination starts to get uneven.
And those things also don't cost all that much more (if at all) than the old parts.

And yes I know the irony behind complaining about nixie clock parts being old

[Da_Stier]

You are totally right, apart from the arduino, the design might as well come straight out of an 80s electronics magazine.
I think there are some more modern designs out there with "real" boost converters and a nice PCB.
I like the sound of your way of doing it, do you have some pictures?

My example was not intended to be a nice and polished version, I just used what I had and what made a simple circuit.
Originally I wanted to use optocouplers as drivers or maybe simple NPN BJTs, however this would have been quite a few more wires to solder, since I would need a transistor on each digit.

I also thought about making a matrix out of all the digits, similar to how you control 7 segment displays.
However at this point I found the box with the soviet driver ICs. 

So I used the old ICs, since I already had them anyway. 

(Mr. Carlson from Mr. Carlsons Lab made a frequency display with nixie tubes a while back and he used shift registers with discrete output transistors, if I remember right)

[Zipdox]

If you don't mind me asking, why did you build that monstrosity with enamel copper wire and solder on the back of a TO-247 package? Seems kinda dodgy to me, especially considering the high voltage.

[klugesmith]

Nice work.  What is the source of time reference frequency, and how do you set the clock time?

If smallness is a virtue, don't we need to count the size of wallplug power supply, or whatever is behind a USB power outlet? 
I once made a Nixie-voltage power supply, as trivial modification to a
salvaged AC-to-low-voltage-DC supply of 10 or 15 watt size, that had simple flyback circuit topology.  Removing the outermost secondary windings left room to add a 200 volt secondary, and keep a regulated logic-voltage output.

[Da_Stier]

@ Zipdox: This technique is kind of my style, I used to build a lot of prototypes like this.
The wire is grade 3 and has a rated voltage of 560V, (no idea if this grading is manufacurer specific or not) so the 160V for the Nixies shouldn't be to big of a deal.
The MOSFET is very overpowered so it does not get warm at all and the soldering to the back was simply because it lined up with the other end of the wire.
I agree, that this is a kind of "special" (and propably dodgy) way of doing it, but as I said, I used what I had and did a one of thing in my own style. 

@klugesmith: The RTC (a DS3231) has a 32kHz clock crystal, propably nothing special but good enough for a general purpose clock.
The time is set over serial by sending "hour.minute.second.day.month.year" in ASCII.
Not the cleanest or simplest solution but it kept the hardware interface to a miniumum (0 infact  )
You can take a look at the code if you want to see, how it is done in detail.
My clock runs on a 12V DC power brick.
I thought about using an AC power brick and simply using another small mains transformer on the clock to stet the voltage back up for the nixies.
I like the idea of adding another secondary winding, as you said, I think I try to remember this one. 

[profdc9]

If you're curious, you could check out this Nixie clock project I did

https://github.com/profdc9/NixieClock

It uses an Arduino and no obsolete electronics (except for the Nixie tubes of course) and uses the 555 timer boost converter to avoid proprietary ICs.  It is based on shift registers and is therefore expandable.