This is a re-launch of my previously posted project “Extremely Accurate Dimmable Nixie Tube Clock”. Obviously, I have overestimated the demand/popularity of the project (whether the goal set too high, lack of wide campaigning, or cost of parts – probably a little bit of everything). Anyway, I want to thank again all people who backed the project and try to restart. This time around it ought to fare better, as I plan to make,realistically, only few clocks.
Changes/Improvements over the previous design:
The Nixie Tubes will be pluggable to pin sockets on main board; IN-8 not IN-8-2 will be used in this clock
Since the Nixies are to be pluggable, I can assembly and test the whole clock PCB, and then remove the Nixie Tubes. Ready-made clock will be offer as an option – just functional, fully debugged PCB +wall wart PS
For those who prefer to obtain their own Nixie Tubes, I will make 2 colon lamps (pluggable) as an option, too.
The PM neon indicator lamp will be added for optional 12-hr mode
The 3V lithium backup battery won’t be soldered like it used to be, but instead it will be in the battery holder for easy replacement
I am changing the 12V 500mA wall wart transformer with a 12VDC switching power supply. Transformer based PS gives around 18VDC when unloaded. Switcher is more efficient and draws much less quiescent current.
The initial 9V regulator then is no longer needed as HV boost DC-DC will be fed with stabile 12V from switcher. I will add a protective diode to prevent wrong polarity Power Supply plug-ins.
The switching dc-dc 5V regulator can be replaced with LDO (low drop out) regulator. As I tested it the whole clock consumes less than 1.5W of energy
Since the enclosure glass tube diameter is increased now to 3.25”OD the PCB inside can be a bit wider. With the sockets for Nixies at 0.1” raster - the spacing between the traces can be increased. All vias will be covered now, so perhaps no conformal coating needed (will check that before the release)
Due to low demand I will no longer offer a shorter 4-tube version (also no 4-tube PCB).
The goal of this project is to revive nearly extinct technology of Nixie Tubes dating from 50-ties all the way to early 90-ties. Despite being called "cold" discharge tubes Nixies cast a magical warm glow that is very different from 7-segment LED displays ubiquitous in our day and age. It is much easier (and cheaper) to build LED display, hence the new technology has prevailed. Then why bother, one could ask. Few reasons: It is truly an educational experience to get involved in project like this. It is also a challenge that requires a lot of knowledge, engineering skills, and perseverance and I'm up to it. It is a venue allowing me to express myself artistically. Finally it is a lot of fun, and the task gives me a thrill of trying something new and unproven, a design with some risk of "uncharted territory".
The RTC (real time clock) IC that my current design uses is a state of art DS3231 Extremely Accurate RTC from Maxim yielding at room temperature accuracy of ±2ppm (parts per million), or about ±1min/year.
Nixie Tubes require the Rail Voltage of about 180 - 230VDC. The Nixies rail voltage is converted by a step-up DC-DC switching regulator of my own design - I have adopted boost regulator MAX1771 with an external N-Mosfet and Schottky Diode (both rated for high voltage). This solution is isolated from mains, so it is much safer. I tested my DC-DC Converter even with Dekatron Tubes that require much higher voltage (450 VDC).
One of the most important aspects of the way of driving Nixie Tubes is how it affects their longevity. Yes, these tubes are burning out over time. Can something be done to extend their lifetime beyond several thousand hours? Well, yes, but it is not a straight forward process. Nixie Tube operation doesn’t depend on thermionic emission despite looking like vacuum tube. The striking voltage has to be high enough for the Nixie to ignite the cold discharge glow. And although Nixie Tube exhibits “negative resistance” (glow remains at anode working voltage 10V to 20V below ignition voltage), dimming Nixies by lowering the Rail Voltage proves to be inadequate due to relatively narrow range of regulation. Typical working voltage of IN-8 is in the range of 170V.
Alternatively, chopping Nixie’s current (continuously ON/OFF) allows the tube to be glowing with average intensity as low as 5% of its nominal intensity (at 100% of current). Persistence of vision causes naked eye to ignore fast intensity changes (fusion flicker rate threshold above 16Hz). In a nutshell: it is acceptable for human eye and it is very beneficial for Nixie Tube to be switched on and off with certain frequency. This phenomenon allows to multiplex tubes (MUX i.e. sequentially switch on one tube at a time). Instead of running all of the 6 tubes all the time some tubes can be on while the others wait for their turn.At first I have reasoned that grouping 6 tubes in 3 sections of 2 tubes will result in 50% of their average intensity loss. Surprisingly, the 3x2 tubes clock appeared as bright as 6 tubes running at 100%. This propelled me to building 2 sections of 3 tubes MUX-ed – still without significant loss of illumination. Finally I have built 6x1 tube MUX-ed clock, where each one Nixie is be turned on 1/6 of the total time. This is only 1/6th of the transistors used in my 1st clock, and the clock works great – see the introductory video. Human perception of light, sound, etc. is really measured in logarithmic scale (one has to increase radio volume 10-fold for us to feel as if it’s twice as loud! -same with light).
Actually, the ON time for the single Nixie during its 1/6th time slot can be much shorter and it depends upon the ambient light level. Yes, I have LDR (light dependent resistor) onboard and microcontroller measures the ambient light level. I use so called “rolling average” (calculate average light level of last 10 seconds) to avoid sudden intensity regulations. If the light level is low (room is dark), then each Nixie is barely pulsed, just enough to maintain its faint glow at night. It not only extends greatly your valuable Nixie Tubes, but has added benefit of reducing the glare projected by the clock.
Lastly – the clock enclosure. On one hand I wanted to encapsulate the clock and protect the fragile glass Nixies. On the other hand I wanted to display the same Nixies like a precious trophy along with modern, efficient electronic design). Hence see-through enclosure made out of ¼-inch thick, cast acrylic tube (not the cheapest stuff, but well worth it considering unmatched optical clarity).
Many thanks to my wife Anna for putting up with me, and for my son Michael for lending his voice in introductory video.
Please, help me execute this project. Provided it is fully funded, we all can enjoy a timeless piece of nearly forgotten technology. I have chosen the Russian IN-8 tubes for their viability and reasonable price (around $12 - $15 a piece + shipment from Russia, Bulgaria, Moldova, Ukraine, wherever I can find them).
Risks and challenges
The supply of IN-8-2 is limited. No one really knows how many are out there. I will leave no stone unturned in pursuit of as many sets as we will need to fulfill the project. Problem is the timely shipping from eastern Europe - it always takes few weeks.
The prototype on the photograph is a little bit tight inside. I plan to purchase a tad bit larger diameter cast acrylic tube i.e 3.25-inch OD (outside diameter) rather than 3.00-inch OD to better accommodate the Nixies. Some of the tubes are a taller than other, so it is not exact science, but my enclosure ought to fit all of them.
The Cast Acrylic tube is nice, but expensive. First time around I got a foot long tube from a company in California for about $60 with shipment, plus $9 for a single cut-to-length! Enclosure sides require CNC milling machine – it also adds to cost as I have to outsource. I have found a local manufacturer in Michigan, but they don't post any pricing on their website. I'll have to make a trip to find out.
I just got the acrylic pieces for colon lamps from England. I will sketch the colons and Nixie interface in SolidWorks - hope to figure out details.Learn about accountability on Kickstarter
- (45 days)