Differences

This shows you the differences between two versions of the page.

--- en:projects:nixie-clock:start [2011/02/09 12:27]
alex created
+++ en:projects:nixie-clock:start [2011/02/19 13:51]
alex
@@ Line 1: / Line 1: @@
 ====== Nixie Clock ======
-工事中
+I designed and built this nixie tube clock in my senior year of high school.  It's a basic 4 tube design controlled by a Microchip PIC microcontroller.
+{{:nixieclock-1.jpg?600&direct|}}
+===== Introduction =====
+In my senior year of high school, the [[http://www.nutsvolts.com/index.php?/magazine/issue/2006/10|October 2006]] issue of [[http://www.nutsvolts.com/|Nuts and Volts]] featured an article about a Nixie tube clock.  The [[http://www.nutsvolts.com/index.php?/magazine/issue/2006/09|previous issue]] had discussed a high voltage power supply built around a PIC microcontroller.  I had been experimenting with PIC microcontrollers for quite some time before and I was intrigued by the prospect of using a PIC microcontroller to generate a rather high voltage from a logic-level supply.  I didn't have much use for it, though.  After seeing the Nixie clock article, I now had the perfect application for it.  So, I bought several tubes off of ebay, direct from the Ukraine, and got to work.
+The most important part of the clock is the nixie tube.  I wanted to make a clock with a low part count and a small physical size, so I settled on a four tube clock.  I did some research online to determine the most optimal drive configuration for the least complexity hardware-wise.  Many designs biased the anodes directly and then used either ten transistors and a demultiplexer chip or a special high-voltage demux chip per tube.  I definitely liked the hv demux chip idea because it means I need one part instead of 21 (demux + 10 transistors + 10 bias resistors).  Another site that I looked at added anode-side switching and then used only one driver chip.  I really liked this idea, primarily because the driver chips are expensive and, just like the tubes, I had to order them directly from the Ukraine. Replacing three driver chips with eight transistors and eight bias resistors seemed like a fair trade.
+Then I had to figure out how to do the other thing the clock will have to do well -- keep track of time.  I settled on a Maxim/Dallas part, the [[http://www.maxim-ic.com/quick_view2.cfm/qv_pk/3914|DS1340C]], with an I²C interface and a built-in crystal, further reducing the parts count.
+In researching how I would build my clock, I looked at many pictures online of other clocks.  Some were completely over-the-top, some were too simple.  I wanted a sleek, modern looking clock that isn't too difficult to build.  One of the ideas I liked about one of the completely over-the-top clocks is blue lights underneath the tubes.  The little hole in the bottom of the spacer on the bottom of the tubes that I bought happens to be just the right size for a 3mm LED.  So I bought a bunch of 3mm blue LEDs off of ebay.
+For the case, I decided on a simple design based on acrylic.  The top and bottom would be black acrylic and the sides would be made of clear acrylic.
+I decided to write the firmware for the clock in C in [[http://www.mikroe.com/|mikroElektronika]]'s [[http://www.mikroe.com/en/compilers/mikroc/pro/pic/|mikroC]] development environment; I had bought it along with an EasyPIC4 development board from them some time before.
+The result of a pile of work spread out over several months is shown above.