Questions related to reverse current and diode emulation mode

[heuj]

Hey @fl4p, great project, thanks for sharing all the insights!
I am developing a DIY mppt charge controller (for way too much time by now...).
When the original ASCAS project came along I wondered about the mysterious low side fet burning topic described there, even tough the ASCAS description is extensive, I have the feeling its not completely clear. As you went into the topic as well I wondered why not open a discussion here.

From my experience the problem can be described in relatively simple way:
The control strategy must never try to actively lower the battery voltage. This may e.g. happen if battery gets charged up to some high value and then the target is reduced to float voltage, if not implemented correctly the control loop reduces duty cycle and tries to "pump back charge" to achieve (lower) bat voltage.
As the charger must not move charge in reverse direction the only way is to wait until bat voltage has decayed.

On inductor level, the average inductor current must not get negative, however negative inductor currents during each pwm cycle per se are not harmful as long as the avg inductor current is >=0. Charge will cycle between in/out caps in this case, which is maybe inefficient but not directly harmful.

Furthermore if control loop has some residual error in limiting reverse current the resulting negative effects may depend on the type of input source: If a typical power supply or some type of rectified generator voltage is used which does not sink the reverse current, the voltage will rise quickly in possibly destructive manner. However in case a typical PV module is connected as soon as the input voltage gets boosted above OC voltage of the module, the module will begin sinking current and likely nothing bad happens (if the controller does not keep cranking up the reverse current).

Does this description fit your experience in the project?