Controlling a Servo like a Stepper Motor

[JackSparrow]

Hi Guys,

I have a problem with my servo motor. I'm using a PIC18F452 to control my motor and I want it to rotate in 1 or 2 degree steps, like a stepper motor. But varying the duty cycle of a normal PWM just doesn't work for me. It jumps great angles some times but I'd like it to rotate step by step, spending the same amount of time at each step.

Do you have any useful codes or advise or anything helpful?

Thanks in advance...

[bungee-]

If you have RC servo then you have to drive it in a certain way. Typically RC servo needs PWM that is 50Hz and pulse width from 1ms to 2ms, where 1.5ms means center position.

Try to search on the forum for servo. If your servo is some other, please ignore my post.

[RLScott]

An industrial servo motor is normally coupled with an optical position encoder for feedback. Then some controller electronics varies the motor drive in a closed-loop fashion to achieve a certain position. Are you providing that closed loop?

A stepper motor, even without position feedback, has an inherent sense of position, and so single steps make sense. There are controller devices that can simulate the stepper motor interface for a servo motor, but that still involves a lot of electronics. Do you have such a device? Or are you trying to make such a device? If so, then what are you using to read the optical encoder?

Robert Scott
Real-Time Specialties
Embedded Systems Consulting

[Ttelmah]

I presume you are actually talking about a radio control 'servo'?. These are of 'varying' accuracy. Basically, they have a closed loop, formed by a potentiometer, feeding back into the control chip, which is used to control an analog multivibrator, which is compared with the value being generated from the incoming pulse. How well this is done, varies with the make of the servo. The overall accuracy of the potentiometer, is also 'variable'. Many types, will exhibit significant 'jitter' (where they will hunt backwards and forwards around the requested position), and also linearity is not fabulous in many cases....
Most units are based on variants of the old NE544, or more modern M51660L control IC. These have a number of controlable parameters, set by resistors/capacitors, which alter how well they behave.
There is a 'deadband' control, which if set to low, leads to the jitter behaviour, while if it is set too high, small 'step' movements, may well be ignored completely.
If you take units from a number of manufacturers, and feed them with identical pulse trains, you will find a massive variation in how well they move.
Generally, makes like Futaba, and Yamaha, tend to produce units that perform comparitively well. Also, many manufacturers offer 'high precision' servo models, which use improved potentiometers, and more accurately setup controller IC's, and these are more likely to work for an application like this, than 'general purpose' servos.
With a good RC servo, a properly setup PWM, should happily allow reasonably repeatable movements of only a very few degrees. If the unit is not behaving properly, you may need to look again at how your timings are set. However wth some makes, it is pure 'luck' how well they will perform, and this may be the problem...
An industrial servo, is a different 'kettle of fish' here.

Best Wishes