While not exactly related to this, I'll post here a message I submitted to ToyKeeper recently (≈2 days ago):
This passiveness could mean lack of interes, although I'll keep listening. :-)
I wanted to say this because among other things ToyKeeper created Andúril and many of these sophisticated firmwares. ;-)
Concerning moonlight modes, I'll share with you a nice idea.
As the technically inclined among you already now, the very low current required for moonlight modes usually causes some trouble with the either the low resolution of MCUs (mainly), sense amplifiers, low or super low sense voltages (which are a must for high power and efficient designs), switching noise… or a combination of any of these.
There's a cheap way to solve this despite using low resolution MCUs, at least for linear or variable load regulated drivers. This is to use an alternative output path, a resistor in parallel with the standard driver output, for a moonlight mode. This way, fully closing the MOSFET gate in regulated linear or variable load drivers would only allow the very low current of the resistor in parallel from input to output to go through. Since a “3V” led has ≈2.5V of Vf at super-low currents, it is easy to calculate a resistor for a regulated linear driver presuming ≈3.6V of input voltage. R = V / I, and so a moonlight mode resistor for, let's say, 100µA moonlight, would be Rmoon = (3.6V - 2.5V) / 0.0001A = 1.1V / 10−4 = 1.1V × 104 = 11kΩ. This moon mode would not be constant current, with moon current varying from 154.5̅4̅µA with 4.2V at the input, the no-load voltage of a fully charged li-ion battery, to 45.4̅5̅µA with 3V at the input, the no-load voltage of a technically depleted li-ion battery. But this is a minor complaint, imho. The moon current would be flowing at all times into the emitter.
Hope this is of service. O:)