Digibury: Project Conway Big Reveal Part 2- the hardware build
- 1. Life Is But a Game by Tinkersoc
- 2. The Hardware Brief • To build a 10 x 10 matrix of lights that can play the Game of Life (GOL). • The matrix must be controllable from an external source (the display only needs to render the GOL, it does not need to run it). • A high quality finish is required.
- 3. Initial Concept 100 ping pong balls + RGB lights = win.
- 4. Sourcing Lights • Due to the size of the matrix we can’t plug in lots of lights into a micro (100+ pins needed!). • Multiplexing is one solution: ▪ This would provide very fast IO to render an image, but adds complication to the wiring of the device.
- 5. Sourcing Lights • We don’t need fast refresh rates (the GOL will be slowed down a lot due to the ‘limited’ size of the screen). • Addressable RGB lights would give us expansion room (replacing or adding pixels would be easier)...
- 6. Enter the WS2801 and the WS2812B
- 7. WS2801 & WS2812B • 2 wires (ws2801B) and 1 wire (ws2812b) required to control a huge string of LEDs. • Very bright RGB and individually addressable. • Although timing is sensitive (would require FPGA or some fancy DMA on very large strings) we aren’t running that many LEDs (100 is relatively small). • 100 of these LEDs would need a fair amount of ‘current’ at 5V, but easy to supply (10A external psu).
- 8. Prototype 1. • 5 x 5 Matrix using a WS2801 string. • Bit of hardboard. • Arduino.
- 9. Designing The Final Hardware • WS2812b tape (very cheap per LED). • IKEA HOL table (A 10 x 10 lattice!). • Wire. • Front one-way-mirror.
- 10. A Lot of Soldering Later….. If YouTube decides to work…. http://www.youtube.com/watch? v=M7ut70dOd7o#t=192
- 11. Version 2 • Soldering by hand is slow and potentially unreliable. • Time to produce some PCBS. ! ! ! ! • Produced in KICAD. ▪ Each strip contains 5 LEDs (will need 10 strips of these). ▪ Surface mounting each led and a capacitor. Fits nicely in the lattice gap.
- 12. Result:
- 13. All in:
- 15. DISASTER! ☹
- 16. Diagnosis – Why is my LED not working? • Power is fine (5V consistently across the power rail). • No shorts. • Can run an individual stick. • Should be working and we have yet to diagnose what is wrong (probably something messing up the signal).
- 17. A FEW DAYS LEFT AND WE HAVE NO DISPLAY!?!
- 18. The Society To The Rescue – Scrapping the design and going Plan B. • KISS approach (This had got lost along the way.) • Many Tinkersoc members = many man/woman hours. • More eyes to check the work. • A fun evening of building (make this an actual Tinkersoc project).
- 19. Day 1 - Case Design • Custom case was designed in autoCAD • Lasercut by architecture department (took over 1 hour to cut all the MDF).
- 20. Day 1 - LEDs • Back to the WS2801s (same as those used in the prototype). ! ! ! ! ! ! • Kiss again: 100 LEDSs prewired purchased and measured to fit the custom case. Arrived day 2.
- 21. Day 1 - Software • Arduino micro setup with GOL firmware. • A PhD student wrote this. Uploaded and tested.
- 22. DAY 2: BUILD.
- 23. Lessons Learnt • Things can be swimming along fine and fall over when you least suspect it (generally right at the end). • KISS is the best approach (as long as you KISS the right thing). • Your society can save your butt. • I love the smell of laser’d MDF in the morning.
- 25. Life Is But a Game by Tinkersoc