Reviewed the design today and the issue is that the way I am testing easily unleash spikes on several hundreds volt since I disconnect a coil with energy in it. The voltage can be extremely high even if it’s close to no energy in it. Two solutions are suggested (1) avoid disconnecting the motor while it is running, or (2) add protection circuitry if I insist on doing it.
What I plan is to use the 3.3V from DRV10983 to drive the hall sensors. This will then be a separate 3.3V from the one driving the MCU. Secondly I need to add at least 4 TVS diodes which is easy since I am using a small, nice array of 4 x 6.1V diodes that would have done the job. Again this should not happen in normal operation.