We had this already implemented, but never released it because of the feature beeing quite confusing, which could also lead to people believing the Smart Lock is broken. The way it worked was the following:
- The Smart Lock never does a recalibration on it’s own. It’s very scary when it suddenly starts to move without user interaction. Therefore everything that tries to identify the position has to happen as part of a lock command.
- When the Smart Lock receives the command “lock” it locks until max resistance.
- When the Smart Lock receives the command “Open door” (= unlatch) it unlocks until max resistance.
- If the Smart Lock receives the command “unlock” it first locks until max resistance and afterwards moves to the unlocked position.
This ensures that it never unintentionally pulls the latch. But it does on every command drive to max resistance at least once, which produces a higher than usual wear on the mechanics/motor and battery drain and you can still end up with lock states beeing reported wrong (e.g. after using the panic function it’s unlocked, but the smart lock believes it is still locked). Therefore this mode never made it into a production firmware.