Now that we know how to make measurements from the outside world and how to generate various control signals, lets now look at how to combine them in order to regulate a process. An example of the kind of thing we want to do is controlling the speed of a DC motor. Based upon a previous column ( FD XVIII-2 ) we know that we can set the motor speed with a PWM control signal. If we had perfectly calibrated the system, so that say a 75% duty cycle produced a motor speed of 1000 RPM, we would have an open-loop controller. The downfall of an open loop controller is dealing with all the real world problems like: inaccurate calibration, variations in supplied power, or variations in load on the motor. These complicating factors can only be handled if we actually monitor the motors speed and somehow use the difference between the commanded and actual speed to correct the control signal. This is a closed-loop controller.
How we actually implement a closed-loop controller depends upon what we are trying to control (e.g. motor speed, or motor position), and its relationship to what we are measuring back from the controlled system (e.g. motor RPM or motor angular position). As we will see, the implementation of the controller is also driven by the behaviour we can tolerate when the system is trying to correct itself when it is suddenly perturbed.