I swapped in a 0.9Ω resistor for the old 0.1Ω current sense resistor, giving 9X better current accuracy without reducing the 2A design max discharge current. The original was just a poor choice.
Since that resistor will dissipate 9X more power, I cranked it up to a 4W rating by using two 2W (1.8Ω) resistors in parallel. (Couldn’t find a 5W 0.9Ω.) I mounted them up off the board for good air flow.
I ran current calibration tests at ~100mA and ~2A with two good meters, and tweaked the calibration constant in the code to report the correct current. It’s all relative anyway, since the A/D reference is using a local ~1.88V zener-derived reference voltage rather than the 5V supply.
One more small improvement, and one more small knuckle-rap for being short-sighted in the initial choice of current sense resistor value.