(6″ x 8″, wood, glass, obsolete electronics)
Jenny’s digits are transmitted to infinity using technology of the time of her birth.
This is an art piece I made as part of a series that all fit inside small wooden boxes that mount on the wall. They have a slightly steampunk look and are built using outdated technology. In this piece I use a dekatron for counting, a nixie tube to display the numbers and a whole set of 3-terminal neon-filled trigger tubes to drive the nixie cathodes. Also counting pulses are generated using a neon relaxation oscillator. The changing neon glow in each component makes for a visually interesting display.
All these parts use very high voltages, and the circuit must step up the mains voltage to 400 volts. Sourcing the parts was a challenge. Ebay is good for this. The dekatron came from Ukraine and the trigger tubes from some guy in Scotland.
It was very challenging to fit everything in the case, particularly as the power supply was quite complex, and involved generating numerous different positive and negative bias voltage from the mains input.
Part of the fun I get from my art is the technical challenge. Below you can see the circuit diagram of this piece.
At the top left is a neon relaxation oscillator. I was actually very pleased with this design which I think is quite novel. The problem is to generate a solid 150V pulse to kick the dekatron round to the next cathode at a regular interval every time the neon fires. The 680nF capacitor charge up from 150V via the 10M resistor and the diode to ground. When the neon firing voltage is reached, the conducting neon shorts the left side of the capacitor to ground. This drives the emitter of the transistor below ground and the capacitor discharges through the transistor base, turning it on for a short time, and thus driving the collector to ground. It gives a nice solid full swing of a few milliseconds duration at the collector. I have never seen this circuit before, I just made it up when something else was not working.
The regular pulses are sent to a pulse shaping network to drive the quadrature phase counting cathodes on the A101 dekatron. These pulse shaping networks are pretty touchy. I basically used the one from the application notes for the A101 as each dekatron type has rather different requirements for the shape of the counting pulses. A bias voltage is derived from the 150V supply by a potential divider.
The A101 takes a 400V anode supply (via a resistor which is not visible on the diagram) and has ten counting cathodes – the neon glow can only sit on one of them at a time and is induced to skip from one to the next during counting pulses. Seven of the cathodes are used in this circuit in order to sequence the 7 numerals of the display and the rest effectively cause a pause in the digit sequence which looks good. When the glow discharge lands on a cathode, it raises the voltage on that cathode by a few volts because of the current passing through the 68K resistors to ground. The rest of the circuit is about getting the nixie tube segments to turn on when the appropriate cathode is active.
The nixie driver is quite subtle in design and a number of different things are going on. The anode of the nixie display tube is driven from 250V half wave rectified AC. This is done so the XC18 trigger tubes are able to turn off because their anode current needs to fall to zero to make this happen. In order for the correct turn on and turn off voltages to be obtained at the gate of the trigger tubes, the cathodes need to be biased to -50V. When one numeral is illuminated on the nixie tube it is driven from the 250V sine wave peak through the 47K anode resistor and through the conduction of the trigger tube which introduces a voltage drop of around 75V.
One significant issue that I found was the need to actually bias the anodes of the trigger tubes for correct operation in series with a nixie tube, otherwise they will not fire reliably. This is the reason for the 470K/1M voltage divider and the 1M anode resistors. I had previously had a lot of problems with getting trigger tubes to work in my trigger tube clock when following applications notes blindly. The biasing arrangement I used was gleaned from a schematic that I found for the counter / display module of the vintage Mark 8 Anita calculator. This ring-counter and display driver has a biasing arrangement and so I started experimenting with such a thing when the original circuit I was using failed to work.
The last part of the puzzle is the power supply: how to derive +150V, +400V, -50V, and pulsed +250V from the mains supply in a tiny space using a single transformer? I used a 115V to 115V-0-115V transformer to give 115VAC and 230VAC feeds. The 230VAC was half wave rectified to drive the nixie anode, it was also separately rectified, smoothed and stabilized at +150V using a zener, and also put through a voltage doubler to get around +400V under load. The center tap was rectified, smoothed and stabilized to give the -50V bias voltage.
This project was a nice challenge to build, looks cool, and is reliable and fun to watch. You can see more of the internals in this video: