Jerking Servos

[blackflame2996]

We have constructed a gimbal, but have encountered a strange problem. Whenever we run our final conveyor belt that feeds the shooter, both of the servos, which are on top of the robot, move around sporadically. The problem persists even when we run the belt directly through the CAN bus via TI's BDC com 101. We have isolated the wires, swapped cables, swapped the sidecar, and checked for any current through the robot frame, all to no avail. Does anyone have any suggestions?

[Kevin Sevcik]

Your servo is likely seeing the commutation noise from that last motor. The signal to the servos is a standard PWM signal, so electrical noise can appear as pulse width commands to the servo. I'd recommend keeping your servo cable runs as short as possible and away from the motor power cables. You can reduce the noise from the motor by putting a capacitor across the motor leads, as close to the actual motor as possible. Per R66, this can be a 1 microFarad or smaller capacitor. You can also put a small resistor across the PWM signal leads to the servo to act as a shunt load. This would go between the white and black wires, again preferably close to the servo itself. I'm not sure on the value for this resistor. I think I'd start at 1k and slowly work down till it stopped dancing or stopped moving at all.

This resistor is good at filtering the noise from power wires, which is usually inductively coupled to your signal wire. This means it generates small currents in your signal wire. The signal wire is connected to a high resistance input, so small currents make for big voltage swings. By adding a shunt to the circuit, you're reducing the resistance, which reduces the size of the voltage swings. Your only problem is that eventually, the sidecar wont be able to drive enough current through a small resistance, and your servo will pretty much stop moving at all.

[blackflame2996]

Are these modifications legal?

[remulasce]

As per [R66]:

A noise filter may be wired across motor leads or PWM leads. For the purposes of Inspecttion (sic) and rules compliance, such filters will not be considered custom circuits, and will not be considered a violation of Rule [R49] or Rule [R65]. Acceptable signal filters must be fully insulated and are:
A one microfarad (1 µF) or less non-polarized capacitor may be applied across the power leads of any motor on your Robot (as close to the actual motor leads as reasonably possible).
A resistor may be used as a shunt load for the PWM control signal feeding a servo.

So yes, perfectly legal.

[philso]

Kevin

What value resistor would you recommend putting on the servo wires.

Thanks.

Phil

PS: See you in San Antonio and later the Lone Star Regional.