Originally meant as a birthday gift for my girlfriend (november), then pushed back as a christmas present, then valentines day, then graduation(june), and after a complete rebuild of the circuit, then the box, it’s done.

By far the simplest project I’ve worked on in terms of electronics, but for some reason I had the most problems with this. Mostly due to the fact that when I started, I had no idea what I was doing.

Control:

In order to save nixie tube life (~1-2 years running constantly), the tubes automatically shut off after being on for 10 minutes. Press the button on the front to reset the timer (turn on tubes). The back button changes the temperature scale. Default is Fahrenheit, press again for Celsius, press again for Kelvin, press again for Fahrenheit, C,K,F,C,K….

thermometer : nixie tube from Colin Mann on Vimeo.

—-updated Jan 31, 2010—

There’s been interest expressed in how this thing was designed internally.  So, here it is.  Let me first say, that this is literally the first electronics project I really ever worked on, as well as the first project involving microcontrollers.  As a result, this was actually the second version of the clock, and if I was to do the whole thing over again it would still be a lot different.  The first one sort of worked, but looked like crap.  Here’s the board layout and schematic, also available to download at the end.

The board was based around the use of the AD595 chip, which basically allows for the use of thermocouples without having a cold junction present.  Basically the chip biases the thermocouple to provide the cold junction, or something like that.  Ultimately though, I didn’t end up using this method, for a few reasons.  One thing I learned recently about why I had trouble, is you can’t solder to thermocouples!  Thermocouples are specially selected materials that create a natural voltage when touched together, if one is held at 0°C (cold junction).  Thus, they are not copper or anything like that, and the need to be crimped etc.  And ultimately, in order to connect them to the IC, I think you would want the crimp from a copper wire to be very very small length to the thermocouple.  I’m not sure though, and I never actually got that far.

The main reason I stopped using the thermocouple was because every time I would just gently move the box, the temperature reading would go crazy.  I realize now, this is probably because there was no solid connection between the thermocouple and IC due to some forced soldering that really wasn’t bonded.  So finally, I just used a thermistor voltage divider circuit instead.  I believe it was one 10k resistor in series with a 4.7k thermistor, and if you look at the voltage between the two, you can convert to your temperature reading.

One thing to note on the layout if you do end up using this design.  Everything will actually work just fine, the 74141 chips are setup correctly, so is the microcontroller, but the transistors are missing something!!!  There are no resistors going into the base, or at the emitter.  This will actually work alright, but it will source a lot of current through the circuit, and the 7805 voltage regulator will heat up quite a bit.  Save yourself a lot of trouble, add in some 10k resistors in between those transistor bases and the microcontroller pins.  Also, I would instead of using those long 10 pin 1 row connectors for the tubes, change that to a 5×2, that way you can use ribbon cable and a crimping connector, save yourself about 3 hours.  Oh lastly, on this project, I wasn’t ready to build my own dc-dc converter, so I used this one from allspectrum.com.

Really though, since there is so little required for this circuit, you should be able to fit the converter on there with it, but the other one works great, and is pretty cheap also.

I’ll post the code for the microcontroller here as well, although I’m a little embarrassed to even look at it, remember, first time.  Basically all it does is take an analog voltage reading, sum it for a while, and take the average, which is what is output on the tubes.  I only used two 74141′s an 3 transistors which control the 1,2,and 3 digit on the left most nixie tube.  Since it’s not a clock, all the digits don’t have to work for the hundreds spot, just enough to display up to 399 (kelvin).  Feel free to email me if you have more questions.

eagle files

avr-gcc code for atmega88 (after looking at this again, I’m not sure if this is the final version anymore, but it’s a starting off point at least). found better version.

also, pictures of the inside and making the box can be found here.