Servo Calibration

[dsheller]

I'm curious how you all calibrate your servos to run close to the same speed per pulse... I'm talking about continuous rotation servos.
I bought a couple parallax continuous rotation servos a little ago and just started using them, I noticed they were off through initial testing... so I hooked up a cardboard wheel and marked a position, I then let it run for 10 seconds and count how many rotations I get for that pulse width.

What do you guys do?

[Captain_Tuna]

the should have a trimmer on one side. you need to rotate that to calibrate them, well...that's what i did.

[dsheller]

I'm aware of the trimmer, I set both servos to idle at 1.5 ms pulse width, but the servos -- for whatever reason -- don't respond close enough to drive straight if they were both being given the same pulse.

[Captain_Tuna]

i don't think there's much to do then... i mean, it's the same thing that happens with standard dc motors, there's always one that has one more loop of coil than the other one. To overcome that obstacle i just set them at different pulses (i've also seen some other guys do this). But maybe someone more expert than me has a better solution.

[dsheller]

Yeah that's what I ended up doing, just going through the range of pulses the servo accepts and writing down their revolutions. Then I just use the closest matches in my code, kind of a dirty method, but I'll keep using it until someone suggests something better.

[Admin]

I've never found this to be a problem.

For a robot with closed-loop feedback, sensors will detect and correct for speed imbalances in software.