Monday, June 18, 2012


Continuing my fascination with surplus soviet electronics, I discovered an interesting type of vacuum tube known as a dekatron. These tubes are a type of decade counter, and were commonly used as an early type of memory or to select among several outputs. They are a type of cold cathode gas discharge tube, and use some rather elegant physics to form high performance counters for the vacuum tube era. Inside the end of a dekatron there are a number of metal electrode pins arranged in circle around a central anode disk. There are ten main cathodes and either two or three guide cathodes between each main cathode, depending on the type of dekatron. Upon being powered, one of the main cathodes will conduct, creating a small amount of plasma to form around the conducting cathode. If the next guide electrode becomes more negative than the conducting cathode, the glow will "jump" from the currently conducting cathode to the next one over. Clever circuitry is connected to the guides so that when a pulse is applied to the counter, it causes the guides to transfer the glow from one main cathode to the next one, and since there are ten main cathodes, this creates a decade counter. "Cathode zero" has an output to create a carry pulse into the next dekatron tube. These devices were able to produce surprisingly high count rates in the megahertz range near the end of their development, but were eventually abandoned. They fell out of use once decade counter integrated circuits became available due to their higher performance, lower cost, and ease of use. However, they are still occasionally used by hobbyists as a type of display due to their interesting looks and retro/vintage feel. These devices are available on eBay, of course, and I purchased a few each of two different types. Modern microcontrollers make these easy to drive; a high voltage BJT transistor is connected to each electrode, and software in the microcontroller makes each electrode the most negative in turn. This enables forward and backward motion, as well as allowing the glow to stop on any of the electrodes, not just the ten main ones. I devised a way to determine when cathode zero is the one conducting using an optocoupler to signal the microcontroller, so that I have a home position to start at when the microcontroller first turns on. My plan for these is to create a module to fit in the 5.25" drive bay of my PC and use two of these tubes to indicate the current CPU and RAM usage. The CPU tube will spin at a faster rate when the CPU is under load, and the RAM tube will show a circular bar graph to represent the percentage of used RAM. This project is not yet finished, but I expect it soon will be as I have started working on it again. So far I have the high voltage supply built and the microcontroller operating one of the tubes, and I have a good amount of the code already written as well. More updates soon.

No comments:

Post a Comment