question

What battery and BMS is this?

Some more specifics on the system would be helpful.

So, for the lack of a better option AFAIK, hacking the fudge factor in.

/opt/victronenergy/dbus-systemcalc-py/delegates/dvcc.py, line 426 ff:

                if len(chargers) == 1:
                    # The simple case: Only one charger. Simply assign the
                    # limit to the charger
                    sc = chargers[0]
                    cc = min(ceil(max_charge_current), sc.currentlimit)
                    sc.maxchargecurrent = cc*0.95
                elif max_charge_current > self.capacity * 0.95:
                    # Another simple case, we're asking for more than our
                    # combined capacity (with a 5% margin)
                    self.maximize_charge_current()
                else:
                    # The hard case, we have more than one CC and we want
                    # less than our capacity
                    self._distribute_current(chargers, max_charge_current*0.95)

I added the *0.95 factors (other than the second one which was already present). Problem solved, or rather, circumvented until somebody writes an actual PID loop for this.

Lowering the max system voltage is good for the battery, too. Most of the corrosion and degradation occur at the extremes of charge - and the capacity difference between 3.55 and 3.65V per cell (depending on chemistry) is minimal. Once you back off from the absolute limits set by the BMS (or Cell data), then you can successfully balance a system with battery charge current ~=0. A little overshoot caused by measurement errors and sampling time will not matter as the delta volts will remain within limits. I work out max battery charge current by:

I_chg = Q_nom*4*(3.55-vc_max)

where vc_max is the maximum cell voltage measured in the last scan. Q_nom in the nominal battery capacity in Ah, and 4 is a constant that can be tweaked for stability.