I'm trying to limit power used by servos. The application is such that I don't mind if the servo's position drifts a bit, so I just use an N-MOSFET to shut it off when not in use. A conversation on an Arduino forum brought up the following concern:
Blocking the current flow path on the ground lead may force voltage/current to flow back thru the servo control wire back thru the Arduino to the common ground between the Arduino and servo power supply. If the servo starts to heat up or behave erratically, this might be the reason. I haven't tried it, but a P-MOSFET on the servo power supply might be another choice.
I am seeing some erratic servo behavior, though I think it has more to do with some sort of reset not happening when I bring power back on. Is the logic in the quote sound? Here's a snippet of the current servo control circuit: