How to control single servo with pwm

Check out https://learn.adafruit.com/adafruits-raspberry-pi-lesson-8-using-a-servo-motor/software

Nice project.

But I do not understand why you have the servo so far away from the first rotation point. Why did not remove one step and put the servo on the first finger joint?

Usually servos are simple to program. just a pulse varying from one (1) millisecond to two (2) milliseconds with a repetition rate or 50 or 100 Hz.

Jean-Paul

AC9GH

Thanks for the info, the reason for the servo so far apart is because Im planning to separate the servo and build a complete hand (all five fingers) and the servo has to go on the forearm not the hand, bu its a good idea to have it on the joint, thought about it at first but its wasnt going to work for me

If I'm not mistaken the frequency or duty cycle determines the position of the servo motor. The timings just increase the speed of each turn.

Nope, it is the actual pulse width as correctly stated above, a nominal 1 - 2 mS with about 1.5 being the centre, of course some manufacturers deviate a bit from this to get more range etc., but for basic RC Servos, that's it

the repetition rate is typically about 20mS and many receivers use this as a way to detect a disconnect and move the servos to a default "Safe" value perhaps to put a plane into a slow circle pattern hoping to pick up the transmitter again

If the servo looses its signal, it will typically just sit where it was but stop powering the motor so other forces could move it externally

the "Wiggling" will be due to the fact that the PI is a multi tasking OS (AKA Linux) and therefore you can not guarantee exact timing so from that perspective your kinda stuck with it, you could try sending the pulses long enough to get to position but then stop sending them, this may work but probably wont maintain the "Grip" as the drive system in the servo will probably stop. Adding more servos to the PI will just serve to make this worse.

a better way to drive the servos would be to use an ATMega328 or an MSP430 micro controller which are very cheap but will maintain a very accurate timing thereby reducing the wiggling considerably. In my testing there was NO declarable wiggle

the PI can send the commands for positioning to the controller and it will simply do all the work for you. The pi is then left with the higher level brain functions

you can test this out simply by using an Arduino UNO or something (With a PDIP 328) then once it is working it can be removed and placed into a smaller and dedicated pcb/circuit. you then have the ability to make multiple independent limbs and have the PI orchestrate them.  NICE.

there is plenty of code in the forums and on the Arduino site regarding this. there is even a dedicated Servo library

Regards

peter

Thanks for the reply, that is exactly my plan but I needed to start by testing the servos one by one and then move to a more robust and multi servo platform.

Thanks for the tip on the wiggling now I know the reason for it, thought the servo was faulty.

well for testing, why not test with the platform your going to drive it with, AKA an Arduino UNO or similar. This will also confirm the servo is indeed not faulty

Peter

I'm not sure how the Raspberry pi "servo library" works. Or whatever Omar tried.... But from "pwm" I would think that the raspberry pi has a hardware PWM module (which I know it has) and that this is being used to send out the servo pulses. That should be "not wiggling" even if there is a multitasking OS driving the values for the servo position.

On the other hand, if it's user-space GPIO implementation then things can indeed be bad.

Searching for what you can/should be using, I found: http://pythonhosted.org/RPIO/pwm_py.html, which uses DMA to achieve hardware PWM even on pins that are not PWM according to broadcom....

If you buy the super-cheap servos, yes, they sometimes wiggle. Hmm. IIRC the one that I had that on had a capacitor fall off the PCB inside the servo....  (damaged in transit).

interesting stuff to know, thanks

Peter