Inside of one of these you’ll find the actual DC motor itself, a set of plastic gears, a potentiometer (pot) connected to the output shaft, and a little circuit board that controls it all. The gear train is used to convert the high-speed low-torque output of the motor into its high-torque low-speed output, and the pot reads the orientation of the output shaft so that it can be controlled (servoed) to the correct position. However for today, we’re primarily interested in the case, and most of what’s inside doesn’t really matter.
The lower part of the case comes off with four very skinny, very long screws.
And then, you can pull out the tiny little circuit board.
The DC motor slips out easily, but the three wires to the pot are soldered in, and need to be clipped.
Incidentally, it’s straightforward to hack servo motors, repurposing the circuit board such that (1) the two outputs to the DC servo motor actually control something else and (2) that the input signal from the pot comes from something else. You can read our article about how to make a one-ton servo motor for a good example.
And then there is the matter of the cable. We don’t actually need that much cord hanging out the end — and it weighs something — so we can clip it shorter.
A dab of hot glue secures the cut-off end of the cable to the bottom of the case.
And the finishing touches: Reassemble the case, add the servo horns and finally the earwires. The final weight of each one is about 6.5 g, and the total weight of the components that we removed (motor, wires, circuit board, and cables) is about 5.6 g. As we have left it, the output shaft feels solidly held in place. Turning the output shaft still turns the gear train, and its motion is limited to the servo motor’s original range of travel; about 2/3 of a turn.
You can make it even lighter — all the way down to 5.4 g per piece — by opening it up further and removing the pot and all of the plastic gears except for the output shaft. It looks mostly the same from the outside (with the exception that the gears are no longer visible), but does not feel nearly as nice: The output shaft is only loosely held on its axis, and now able to turn freely through a full circle.
One might imagine taking it the opposite direction too: Building in a little battery and microcontroller, so that the servo motor would turn on its own while dangling from an ear. That version is left as an exercise for the reader, hopefully one with short hair that won’t get tangled.