This version only made it do the development board phase. Hence the images here are all renders. This version featured the following:

  1. The ring is now charge via ambient light using 4 large photodiodes.
  2. The display is moved to the palm side of the ring, to allow the photodiodes to point outwards.
  3. The ring is activated with a resistive contact “button”. When the user connects the two gold contacts using their thumb, the GPIO is pulled low, activating the display. This method has no moving parts, and looks great!
  4. The PCB still only has parts on one side.
  5. The power storage element is 4 CeraCharge batteries, 1.5v 0.1mAh, with a 2 serial 2 parallel connection.
  6. The ring can be reprogrammed using pins exposed on the final form factor.

These features led to the following issues with the design.

  1. The batteries give 3v with 0.2mAh, which is good, though the maximum current would be ~3 or 4mA. This would be OK for indoor use, but not outdoors.
  2. There was no space for a boost regulator, meaning the brightness of the LEDs would be hard to control, and would sag with the battery voltage, particularly under load.
  3. The total surface area of the photodiodes was limited by the requirement of having the board with parts on one side. Meaning the batteries and other components are using up outward-facing real-estate.
  4. The BQ5504 used for energy harvesting would only be effective in direct sunlight with such a small photodiode area. The rest of the time the small amount of power available do just enough to keep the converter awake (which wastes more power than it can generate!). For indoor use the converter actively wastes power. This chip also uses lots of board area due to the passives that it needs.
  5. The (nearly) floating GPIO ends up using lots of power (~30-100uA) if it is held in the middle of the voltage levels, which happens often if the pullup is weak enough to allow the thumb resistance to pull it down.