To test v5.0’s adaptive brightness using its little photocell, I hooked up the neopixel ring to my arduino’s 3.3v, GND, and D6 pins, with the photocell connected as per this schematic:
The neopixel ring and the photocell were both connected to 3.3v to simulate the 4.2-3v of the rechargeable battery I’ll use for the watch, and when running any code at all, the LED ring would glitch and flash all over the place with random data. To solve this problem, I switched over the circuit to my Sparkcore from Sparkdevices, a little CC3000 based arduino compatible beast, which has a 3.3v logic level instead of the arduino’s 5v, to more closely match the watch’s environment. All my LED shoes that run off of lipos and all the other low power neopixel projects out there work, so there had to be a simple solution. This simple change worked. I copied over the code to the Sparkcore web IDE, and flashed it over the internet (!) It worked perfectly, and dimmed the LEDs in the dark and lit them back up again in the light. Problem solved and project resumed!