While going through the firmware release notes I've noticed the following:

•Software feedback loop so set vs actual current tracks better

Do I understand it correctly that you've, at some point, made a switch from the analog, hardware feedback loop to driving the pass transistor purely in software loop based on the current shunt readings for the basic operation? Doesn't it run a risk for the device failing to respond and correct properly to the load / source voltage changes due to some yet-undiscovered firmware bugs and/or regressions?

The hardware feedback loop is still the primary way the Re:load Pro regulates. In 1.5, I added a secondary, much slower software feedback loop that exists to iron out discrepancies between the ADC and DAC measurements. The Re:load Pro's ADC is more accurate and more linear than its DAC, so the software loop takes regular measurements of the actual current, and applies a 'tweak' value to the output of the DAC to get them to line up better than calibration alone can accomplish.

Oh, I see, that's clever. Thanks for explaining that.