The long discussion about designing one's own versus buying a kit determined me 
to write about my first experience in the matter.

It started with a Nixie display module bought from a hamfest (Friedrichshafen), 
in 2008, a module that included the Nixie tubes and 74141 drivers. No HV 
supply, no documentation of any kind. The module included 10 tubes, 2 of witch 
were letters and symbols (M,m,G,u;Hz,V,A,Ohms), 8 with numerals.
At that time it just seemed an unusual type of display and I thought it would 
be an interesting challenge to try and use it for some purpose. A clock was the 
obvious choice, since it seemed quite simple to implement. At the time I had 
almost nonexistent electronics knowledge (besides the basics; e.g. I did know 
what a transistor did but did not know how to connect it, drive it, etc). Since 
I write software for a living I thought the simplest way would be to use a 
microcontroller, but it wasn't something I was familiar with. However, one of 
the books I found at the ARRL booth was "PIC Basics" (by Cliff Powlesland, 
G8CQZ) and I bought it thinking it would provide some useful pointers to 
microcontroller programming. It was a good choice, as it had some electronics 
instructions along the way.
I believe that trying to create something from scratch is a better learning 
experience then simply copying something that already exists so I decided to 
create my own design.

This was made easier by the fact that the "display" module already had the 
drivers connected to the nixies; after some searches on the internet I have 
found the datasheet for the 74141 drivers (thanks Dieter - I think your site 
was one of the first to provide such information back in 2008), so by following 
traces I could figure out the module pinout. I also found the datasheet to the 
tubes and figured out the needed voltage.

I used a 220-110v transformer (I live in a 230v country - close enough) and 
found the resulting 160v dc to be enough to light up the tubes.
The first test (I did not know if the module would work or not) was to simply 
power it up with HV and see what happens - all digits lit. I figured that I 
should pull the inputs to ground so I did that, connected the remaining input 
to ground and indeed one digit was lit. This test was done with all inputs in 
parallel, so in fact it was one digit for each tube.
I started experimenting with the uC. I have bought the kit recommended in the 
book (Velleman K8048) and the 2nd example provided with the kit was enough for 
what I wanted to do next.

Since there were 8 tubes, I connected them in 2 groups of four, to drive them 
from the 8 bit port of the uC. This seemed to work, as well.
Along with the kit, I bought a few TTLs, including some 74LS373 registers, to 
experiment with. Well not much in the way of experiment - their use seemed 
pretty straightforward.
The approach was as follows - I used 3pcs 74LS373 registers to maintain the 
outputs. All PORTB pins from the PIC were used to connected to the input of the 
registers (8 bits) while 3 PORTA pins were used to SELECT each register. The 
datasheets (along with some ideas picked up from the book) were enough for me 
to manage to get a different digit lit up on each tube. After that it was just 
a matter of programming to get a clock running. I added two buttons for setting 
hour and minute and everything was pretty much working. By the way - I used the 
4MHz crystal as timebase as well, setting up a timer in the PIC that would 
"tick" every 100ms (IIRC; I should check the sources) and this proved 
surprisingly precise. I suspect this has to do with minimizing the crystal 
error - n ppm error at 4MHz means a smaller part of a second than n ppm at 
32768Hz.

So now I had a nice nixie clock but with a big bulky transformer from 220v to 
supply the tubes. This part took a little over a week of vacation (but it was 
more than full time work - from wake-up until going to sleep I was working on 
this). However this also sparked an interest in electronics that has not 
subsided yet.

In the end. the design is obviously primitive and lacking all of the features 
of other clocks (most annoying being lack of battery backup, since we tend to 
get power failures around here). I have planned for the past 3 years to design 
and build another, more advanced one, but there are always some other things 
that get in front. By the way, besides the power supply (which by itself took 
some months in all) the case proved most difficult for me. In the end it's only 
two black pieces of plywood with holes for the tubes - just as primitive as the 
rest. The 2 rectangular buttons for Hours and Minutes (what was in my mind to 
pick rectangular ones? I think it was the only thing I found at the time) are 
dangling from a small wire. But it's the first physical thing that I created 
(software seems so ephemeral). And for 3 years it has worked flawlessly in our 
bedroom, it stays on at all times (except when we are on vacation) and still 
works.

Pictures took during various development stages are at 
http://hawk.ro/ceas/index.html

-- 
You received this message because you are subscribed to the Google Groups 
"neonixie-l" group.
To post to this group, send an email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit https://groups.google.com/groups/opt_out.


Reply via email to