I also thought of the idea with reducing the current reference. However, in my case it simply doesn't work, because the short pulse time (i.e. blanking time of 1.8-2us) when the motor is standing still is enough to increase the current to the nominal value for spinning. Current regulation is simply lost in this case, I think some of the posts in this thread also discussed this issue.

I don't know about whether pulling the reset will help - in my case I just hardwired the Reset to Vcc and cannot change it without cutting traces on the PCB below the chip, and finding another spare output from microcontroller (can be challenging).

herdmay1, are you sure about this "feature" of A3986 i.e. pulling both reset and enable? Then I can just implement it in the next revision of the PCB, if I find the spare microcontroller outputs.

Regards,
P.