Improve dead battery behavior

[manuelkasper]

Test what happens if the AVR is unpowered because the battery is completely dead (or disconnected). Do we need Rd pull-down resistors?

[manuelkasper]

Starts up just fine with only USB connected (no battery). Check how the charger IC behaves if the battery shows 0 V because its protection circuit has disconnected it, and needs an external "zap" to wake up.

[manuelkasper]

I did a test with an old KX2 battery (original). Discharged without mercy until its BMS cut it off at 7.5 V. At that point it read 0 V across its terminals. Then connected it to the KXUSBC2. Had some troubles getting it to accept power from a USB-C PD adapter when it was otherwise unpowered. Seems like the dead battery startup from USB-C PD is not reliable. What did work was a non-PD supply – the KXUSBC2 started up and applied some trickle charge to the battery, which sprang back to life. However, the battery's BMS then seemed to be stuck in some kind of high impedance mode, not really accepting a charging current, leading to OVP events reported by the charger IC (really seems like the BMS's fault). Disconnecting the battery for a moment and reconnecting it fixed that, and the battery was then charged normally.

Scope trace of the battery voltage, with the KXUSBC2 attempting to charge the battery with its 7.81 V open circuit voltage; apparently the BMS is in high impedance mode, so the charging voltage overshoots and the charger IC gives up, and then the voltage drops to zero.

So to sum it up: charging a dead battery seems to work OK (if its BMS doesn't act up), but starting from USB PD only is not reliable at the moment. The FUSB302B has dead battery resistors, but that doesn't seem to work for some reason.