It’s been a long time since I wrote any updates about my IN-3 matrix project. Whilst I haven’t had much chance to work on it it has come a long long way from the original 4×4 tester.
I built an oak frame, machined an mdf board and attached 128 bulbs. Then I started the long process of wiring up each bulb to the PWM driver boards. In hindsight, I should have designed some kind of PCB strip to connect the bulbs to. This would have saved many hours of labour and been a lot neater. As it is each row of 16 bulbs takes about 1 hour to wire. Very boring so next time I know what to do!
I opted to buy a high(ish) voltage supply from eBay, the display can draw a lot of power when fully lit so despite having literally hundreds of high voltage driver boards for my IN-9 kit I figured to play it safe!
I am still only running some test code to make sure the bulbs light up and in the right order. I’m using an Arduino and the Adafruit Adafruit_PWMServoDriver library to communicate with the PCA9685 PWM drivers. Once I find a couple of hours to finish off the last two rows I’ll focus on what to use as a controller (probably the Teensy I’ve had laying about for 3 years) and actually write some proper firmware.
I’ve also now added a Github repo with all the hardware design files (DesignSpark PCB) for the driver boards and arduino sketches. I’ll continue to add to it as I progress.
It was a pretty straightforward task to assemble the bits into a functioning 4×4 nixie matrix. The tubes are mounted on spray painted black MDF with the PWM driver mounted on the back. They are each held in place with a small square of double sided foam tape. Wires soldered on the back to the tubes create a small nest and will be more neatly managed of the bigger version. I programmed an Arduino with the Adafruit pwmtest example and it all works fine. In fact, it works much better than I expected and I think I’m ready to go full scale, with no real modifications. All seems too easy.
I did a measurement of power consumption and the 4×4 uses a power of 1.8W, so will need about 15W for going full scale. Not too bad really for 128 tubes.
I recently bought a few (200) IN-3 nixie bulbs. Unlike the familiar numeric tubes or the bargraphs I’m fond of, these are just little bulbs. They have commonly been used as dots between digits in clocks and alike. Not content to make just another run of the mill Nixie project I decided to build a Nixie matrix, much like an LED matrix but with the lovely orange glow. My eventual aim is an 8×16 matrix, enough to display basic numbers and graphics. Though I’m starting with a simpler 4×4. This does provide the first challenge, how to drive so many tubes? The common way is multiplexing, however, since this is a one-off build and additional electronics isn’t too important (I’m not trying to optimise a BOM and a bit of extra work is fine) I have opted to directly drive each IN-3 tube. Not only drive each tube on and off but with a PWM signal. This should allow nice brightness variations and fading effects across the matrix. Pretty sexy. So now I just need 128 PWM channels… For ease I’ve gone for a 16 channel LED PWM driver (PCA9685). There’s even a nice Adafruit library for quick Arduino testing. It’s I2C so I intend the chain 8 of them together, these are driving mosfets that switch the high voltage tubes. (I’m also putting together an IN-9 driver kit (watch out Kickstarter…) so I have an abundance of high voltage supplies.)
So far I’ve fully soldered up 4 boards and tested one with 16 tubes. All working nicely. Today I’ve cut out a small 4×4 MDF board to make up a little test board to get everything right before I go full scale. So far, it’s all be smooth sailing. Wonder how long that’ll last.
It’s been a long time since I started the rework of my nixie bargraph clock. Probably about a year, but I can finally sit back and say it’s finished. I plan to make a few, one to repair the original instructable clock and this one which I mounted on a piece oak.
The numbers have been laser etched with the tubes held by copped pipe straps. The board is mounted on the front with a protective acrylic screen. Exposed PCBs aren’t to everyone’s taste but they are mine. There’s a little RGB LED which indicates which mode the clock is in: set time, calibrate hours tube, calibrate minutes tube. Knowing that the tubes have a limited life span I added some code so they turn off between 1am and 6am. All pretty straight forward stuff. Now it hangs on the wall looking ace.
The new boards arrived yesterday, so last night and this evening I populated a board and programmed it. Worked first time, sigh. The old board has been running continuously for over a month and hasn’t lost a minute. So looks like it’s time to finish up the firmware and make some clocks!
The new boards have arrived and it looks like we have a winner. Now time to write some firmware….
I decided to make people Christmas presents last year, and thought I’d make my Dad a weather station. That’s the kind of thing parents like. I wanted to make something a bit nicer looking than an ugly circuit board so had the idea of attaching the wires to a piece of wood. In the end I had to make it in about a day, just before Christmas, and so it didn’t end up as neat as I would have liked. However, I think the result is still pretty nice/interesting. Something I will definitely try again, perhaps with more time and planning.
The heart of the actual weather station is an Atmega328 with an arduino uno bootloader. For the sensing I use an MPL115A2 pressure and temperature sensor, conveniently broken out by Adafruit. For the humidity I use an DHT22 temperature and humidity sensor. Though since I already have a temp sensor I ignore this one.
It is pretty straight forward to get the arduino and sensors talking and displaying the humidity, temperature and pressure on screen. To display it on the weather station I decided on 3 nixie IN-9 bargraphs, as previously used on my clock project. These need a 140v supply which I bought as a small board (visible on the bottom right). To control the height of the plasma the current through each tube needs to be controlled. To achieve this I map the humidity, temp and pressure values to 3 PWM outputs and pass through a low pass filter, with a time constant of about 1 second. This gives a pretty good DC voltage. This is then fed into the base of a high voltage transistor. This then controls the current through the tube. Much simpler than it might sound. Check out my clock instructable for how to control the nixie tubes with an arduino.
So, why is this all 10 months late. Well I did all of the above in about a day. I wired the whole thing up with copper wire and glued onto a wood backing. An old piece of laminate flooring! However, I never actually got round to making scales for the different tubes. As any young scientist knows, without labelling the axes it means nothing! The problem was I could never think of a nice way of doing it, until now. Playing around at the Hackspace I decided to laser cut them. I tried a few materials, wood being the obvious choice, but opted in the end for black acrylic. After a little tweaking, I cut the scales and stuck them on. Now it really is finished and looking pretty good. An instructable may follow…