See my little blog post about this here: http://coddingtonbear.net/maxwell-bicycle-electrification/.
I have a bicycle dynamo on my bike, a rudimentary understanding of electronics, and a bunch of neopixels, gps, and bluetooth modules to use. This is what happens when those things combine.
Maxwell is a two-part bike computer I've designed to allow me to do a few things:
- Light and control neopixel LEDs wrapped around my bike for added safety while biking at night, and added fun when biking with groups.
- Charge an easily-swappable 18650 Li-On cell for potential use when bike touring.
- Gather and transmit GPS coordinates for display on the internet somewhere.
- Display status information including my current speed, battery voltage, and current consumption, as well as disable or enable bike features including, most importantly, disabling and enabling the LEDs from a small screen mounted at the front of the bike.
- Send interesting data for display on the internet for completely impractical reasons.
The "remote" unit is based around the STM32F103CB microcontroller, and the "base" unit is based around its slightly-more-full-featured brother, the STM32F103RE. Both communicate with one another over a CANBus (routed through two of the conductors in a length of ethernet cable) to share status information and emit/handle commands.
Checked items indicate errata that have been corrected in the schematic.
None as of rev 1.7.
- Incorrect diode footprint for polarity protection. There also might be a power deliver problem through those diodes. Investigate that the selected mosftet is ON at Vgs -3.3.
- I think the gerbers I generated originally may have excluded the rear solder mask, as the resistors and vias were covered.
- Board-to-board interconnect should probably either not use an FPC connector, or use multiple conductors for power/ground. I had to build a workaround connector for transferring power between the two boards.
- Resistor R2 of maxwell-remote-2 isn't really necessary; it could be connected directly to VCC.