I grew up long after vacuum tubes had stopped being used, but I always found technology from that area to be fascinating. Nixie tubes in particular I think are very cool. They are filled with neon gas, and work by accelerating electrons from the digits to the grid in the front. You actually get a gradient of different colored plasma as the electrons gain velocity; the layer nearest the digits is a light orange, while the next layer is blue. I suspect that if you looked at running nixie tubes over a wide-spectrum camera that can view ultraviolet and infrared they would look very cool. But I digress. I have seen plenty of nixie tube clocks around, and I thought it would be cool to make my own. I used six IN-12B nixie tubes from ebay with six vintage russian high-voltage driver ICs (to the best of my knowledge there is no modern equivalent of this IC). To run it all, I used a preprogrammed driver microcontroller from AllSpectrum Electronics. I considered programming an Arduino to run the clock functions, by the microcontroller AllSpectrum sells is far more advanced than anything I could program in a reasonable timeframe. I designed an acrylic case to have a unique aesthetic that I think separates my clock from most others that you can find on the internet. I then used a laser cutter to cut out the case, and glued it all together. The tube orientation I used is a bit weird because I didn't want any tubes upside down. In the picture below, the time is 12:03:00. The lit numbers are much clearer in person, it is difficult to capture the way they look on camera, but I tried my best.
Step 1: Parts
You'll need the following stuff:
Six IN-12A or IN-12B nixie tubes (The only difference is the side the decimal point is on) (eBay)
Six IN-12 nixie tube sockets (eBay)
Six 74141 high voltage nixie tube driver ICs (eBay)
A high-voltage supply circuit (around 170 Volts)
Two 1/4"X12"X12" sheets of plastic (or whatever material you need for your own case design) (Mcmaster-Carr)
Two normally open pushbuttons (eBay)
A 6-digit nixie tube clock controller IC (www.allspectrum.com)
Various resistors, capacitors, and supercapacitors as called for in the datasheet for the controller IC
A temperature-compensated crystal oscillator (TCXO) or a regular crystal (I got a free sample of a TCXO from maxim-ic.com)
Neon bulbs (if you want delineating colons and and am/pm indicator, I skipped them)
A circuit board (I used one with copper pads on the bottom to make it easier to solder on components) (futurlec.com)
An epoxy capable of gluing plastic
Step 2: Electronics Testing
Firstly, I soldered 11 wires to each nixie tube socket: one for each digit, and one for the power lead. The grid is powered, while the lit digit is grounded. This way, electrons flow from the digit to the ground. To test the driver ICs, the nixie tubes, and the power supply, I built the test setup shown below. I plugged leads running from the socket into the breadboard for the purposes of this test. I used an Arduino to provide the BCD signal used to control the driver ICs. By setting the Arduino to cycle through all possible four-bit binary numbers, I could cycle through all 10 numbers the nixie tubes can display. I put my power supply in a project box along with a 9V battery so I wouldn't shock myself. I also soldered wires to the sockets and ran them to a breadboard. I also taped the nixie tubes in a earlier faceplate design so I would be shielded from the high voltage running to them while they were being tested. It turned out that one of the ICs I got was burned out, so it was good that I tested them all before soldering them to the final circuit board. The faceplate below was my original design, but I realized it didn't look nearly as cool as it would if I arranged the tubes in a circle.
Step 3: Final Electronics
I followed the circuit designed for the 6-digit nixie tube clock controller. I omitted the am/pm indicator and the hour/minute/second delineating lights because they didn't really fit with the aesthetic of my design. In hindsight I wish I had put in an am/pm indicator because it acts as a negative/positive display for certain settings of the controller IC. I soldered the components down to the board and used solid-core wire to link the components together. I ran stranded wire from the high voltage driver ICs to the nixie tube sockets. I used stranded wire because it is less likely to break when it is flexed repeatedly.
Step 4: CAD Design of the Case
The final design of the case I decided upon is a cube 5.5" on a side. It is comprised of six panels laser-cut from translucent red plastic. I designed them in SolidWorks to ensure that they would fit correctly. I designed the panels to interlock at the edges to produce a strong, easy to assemble case with a seamless look. I ordered fluorescent red acrylic from McMaster-Carr, but it turned out to be more neon orange, which turned out to be fine. I also intended to order 1/4" acrylic, but I accidentally ordered 6mm, so I had to redesign the interlocking edges so that they would fit precisely.
Step 5: Laser Cutting the Case
I cut the panels out of 12"X12"X.236" fluorescent red plastic. I intended to order 1/4" thick plastic, but I ended up getting 6mm thick plastic instead. I had to redesign the case to compensate for this new dimension. I laser engraved the circuit diagram into the bottom panel just as an homage to vacuum-tube era technology which was user-servicable, and because it looks really cool engraved into the plastic.
Step 6: Socket Spacers
To set the nixie tubes back from the outside of the case I designed a set of spacers a pair of which would be glued to each socket to space them 1/2" from the outside of the case. I designed these in SolidWorks and cut them out of the same plastic I used to make the case using a laser cutter.
Step 7: Gluing in the Tubes and Assembling the Case
To mount the nixie tubes to the case, I glued a pair of spacers to each socket, and glued the outermost spacer to the faceplate. I scuffed up the spacers before gluing them so that they would bond well. Using this mounting system gave the case a more seamless look than screws would have, and meant I didn't have to deal with fitting screw heads near the nixie tubes. Gluing the nixie tubes to the face plate before gluing the face plate to the case meant I had easier access to the nixie tubes during the gluing process. To glue the rest of the case together I just put epoxy on the contacting surfaces of the case pieces and held them together to dry with masking tape. To hold on the back, I just glued it in two places so I could pry it open later if I needed to access the internal components.
Step 8: Settings
Once the clock was up and running, I set the time, the date, and a number of settings in the chip using the set and advance buttons. The chip has something like 100 different settings you can use to adjust the minutia of how the clock runs. One issue I ran into is that the colon separator lights are used as indicators in the setting menu, but I had not attached any. It might be a good idea to at least run some test leads if you're building a clock so you can use this feature. My favorite is one where you can set the brightness to vary at different times. Using this, I set it to turn off at midnight so it doesn't light up my room at night. To access the user manual with all of the settings, go to the following website:
This is the AllSpectrum page for the six-digit nixie tube clock controller.