I remember a coworker telling me about a project that had mysterious battery drain after shipping.
It turns out the microcontroller was doing read-modify-write on a port that had an LED connected, and in bright sunlight the LED input would be high, and then stay high after the write. It was fixed with a firmware update that set the LED bit based on whatever logic determined if it should be on on every write.
It turns out the microcontroller was doing read-modify-write on a port that had an LED connected, and in bright sunlight the LED input would be high, and then stay high after the write. It was fixed with a firmware update that set the LED bit based on whatever logic determined if it should be on on every write.