This is the progress for today:

I finally transferred the dead-bug build on a more permanent perfboard:

To test it, I make a quick-and-dirty 2S 18650 pack:

I need a datalogger, but, I don't feel like building one, so, I used NI myDAQ to log the voltage curve. Big shoutout to Ashley Roach for sponsoring the myDAQ! :) This is my test setup:

The yellow and the blue traces are opamp's output and half-bridge's output respectively.

Good news: it is running in closed-loop. Bad news: the switching bridge is noisy. :(

The 50% duty cycles indicates that it will make both of the cells to hold the same voltage, i.e. 50% for cell 1, 50% for cell 2.

Result so far:

-Good: Cell 1 and Cell 2 both is approaching the middle voltage.

-Bad: It takes too long.

Note: Cell 2 has more 340kHz ripples because it has more internal impedance.

Next plan are:

-Reduce the switching noise

-Make it balance the voltage faster!