The ADC sometimes responds with negative voltages read on the ADC input due to noise. This can lead to very high illuminance readings at night (4095 << 4 to be exact due to two's complement :) ).
This commit caps ADC readings at 0.
The nRF52833 has only 512 kB of flash. The .dts file was incorrect, as a
product of having it forked from the nRF52840 .dts files, which has 1 MB
of flash.
The new, correct partitions are closely related to the
[nrf52833dk](ec52722861/boards/arm/nrf52833dk_nrf52833/nrf52833dk_nrf52833.dts (L215)).
Beside having a `CONFIG_BOARD_REVISION` Kconfig string that represents
our semantic versioning for board revisions, the ZigBee sample requires
an integer representation of it. Instead of conditioning the integer
version on the string version only for that sample and having to touch
the ZigBee sample whenever we add a revision, I decided to have a
board-wide config, so that both string and int versions live close to
each other.
Fixes#138.
This will enable us to transparently use different calibration for
different board revisions. The 2.0.0 revision changes the sensing
circuitry and will need its own coeffs.
In this commit, both models are run in parallel for debugging.
The new SDK requires 802.15.4 to be enabled in the board DTS file.
Also disable IPv6 explicitly to avoid compile errors related to its
network stack, since it is not used currently.
Closes#94
