This is an example of a close loop servo system -- DC Motor equipped with encoder wheel to receive feedback about speed.
I looks pretty good.
I have some minor comments, depending on how exact you want to be, and how much you want to describe this particular servo vs servos in general...
- Your picture of the servo signal is "not to scale", which is getting to be a particular annoyance of mine, since it makes it more difficult for people to understand how a microcontroller can operate multiple servos. The maximum length of the signal pulse maxes out at about 1/10th of the total period, whereas your picture shows about 1/5th. See servo-radio | Now that the servos pulses happen one at a tim… | Flickr
- It might be useful to separate the "servo contrller/amp" from the "motor driver."
- The "Feedback device" need not be a potentiometer; optical encoders and hall effect sensors are pretty common (both of which are contact-free, and thus last better.)
- I'm not sure that the control signal is always converted to a voltages, vs converting the feedback signal to a pulse; I'm pretty sure I've seen the latter.
Thank you very much for the good words and the constructive criticisms. In fact, I have been keeping the text description pending in respect of receiving comments on the pictures/images presented in post #40.
I have searched Net and have looked for diagrams which bear (at least at conceptual level) mecanical components (Fig-1(a)) and electronics controller (Fig-1(b) that correspond to actual SG90 Servo System (Fig-(d) to (f)) and Fig-2.
Figure-1:
Figure-2:
I have edited the source diagrams to bring them close to the spirit of SG90 as much as possible. The components/signals of Fig-1(a) and (b) are labelled to allowo readers to make one-to-one correspondance with Fig-1(e).
Fig-1(d) and (e) are due to my efforts of carefully dismantling the SG90 and then putting them back in full working condition.
1. I have worked on single SG90 Servo System. I have now edited the rough scales of the pulse width of PWM signal so that the two signals of Fig-1(b) and Fig-1(c) agree.
2. Motor Driver block is added in Fig-1(b). Now, it agrees with the Q1 and Q2 Motor Driver transistors of Fig-2.
3 As I am referring to SG90 which uses potentiometer as a feedback network, I have chosen a diagram from Net that contains potentionmeter in the feedback path. Of course, other devices like encoder wheel (post #42) etc. are the viable options.
4. This is a very interesting and informative point you have raised. It is much more accurate to process/compare two digital level PWM signals (reference command PWM and KC8801 generated PWM based DC volt of potentionmeter) compare to handling analog signals. Accordingly, I have added a PWM block inside Fig-1(b) which KC8801 must have cotained though I don't have the details of the inside world of KC8801 Controller/Driver.
Moreover, the PWM concept agrees with the L298N Motor Driver that asserts the PWM signal coming from UNO directly on the power driver circuit; the necessary filtering is done by the Motor Coil and other stray capacitances.
I have also added a signal conditioning block inside Fig-1(b) to remove any noise or jitter present on input PWM signal.
//============================================================
I will do some more measurements to record the wiper voltage of potentiometer at diffrent angular positions of the shaft/horn and then see how it goes with input position command PWM.
I have practically observed that the motor terminals are not disconnected (power cut off) from the motor driver. Instead, the terminals remain connected with an almost 0% duty cycle of the resultant PWM when the shaft arrives at the desired position; as a result, the Motor is at completely standstill condition. This is evident from the fact that any external disturbance imposed on the horn (such as a slight touch to change the position manually) causes the motor to jitter, turning back and forth to retrieve the original set point.
//===========================================================