Danny Lewin's patent of Oct 22 posted by Triin is hard to get your head around, but my take on it is that one of the ways AP has approached the noise issue was by adjusting the controller to both receive feedback from the arrays in order to achieve optimum control command results, and to also restrain idle arrays. Is that how others read it?:
"A system which controls a group of multiple arrays of actuator elements, each of the actuator elements including a moving element which moves between a first and a second extreme positions; the system including: a controller, configured to generate control commands for at least one array of the multiple arrays in each individual cycle out of a series of sampling cycles, based on obtained information; and (b) an interface configured to transfer the control commands to the at least one array, thereby resulting in releasing from the first extreme position during the intended cycle of at least one moving element that is included in the at least one array; wherein release of all moving elements of at least one restrained array out of the multiple arrays is prevented during the intended cycle."