3 Axis Servo Bot Experiment

by Trenton Henry

11/10/21


I was going through my collection of treasure/trash andfound enough parts to build to experimental gadgetry. I was thinking about making a pan/tilt mechanism for my GridEye thermal camera using a pair or servos, and ended up geting interested in trying to make a 3 axis gimbal head mechanism. Its silly, but it is an interesting challenge. I went through several different configurations of servos, strving to make something minimalist but reasonably rigid. The fine tuning is ongoing.


The first one is just a "can I hack something together that sort of works?" thing. The servos are just running full speed with a delay loop between position upates to reduce the speed. Working on an improved servo driver...



The second one is an attempt to get the pitch and yaw axes aligned at the approximate center, and the roll as close to that as possible. I.e., to get as close as I can to everything rotating about a fixed point in space. And also to see how minimal I could make the scaffolding.

Even with an improved driver algorithm this is still very jerky and wombly. The obvious issues are the lack of counterweights to balance the load, the backlash in the gears, and the very amateur bracketing to connect the servos at the pivot points.



The third one is a simplification. This is only two axes with no added weight yet. This may be a better setup for mounting a thermal camera than the three axis version. It is simpler, smaller, has better balance, and obviously slightly less complex.



One more attempt at 3-axis motion with an improved servo driver. This version version performs simple acceleration/deceleration that does not attempt to compete with the servo's internal PID controller. The result isn't perfect, but it is much improved. The off-balance weight issues and gear backlash of the inexpensive microservos still contribute to some observable wombliness.

End