PID Control, made SIMPLE

The PID control can be broken into its three parts. Intuitively it makes sense to have a control action which is some constant times the error, this should ensure that the system finally reaches the desired state, and at the desired state, the control action would become zero. This makes quite sense but sadly doesn't guarantee that the system will stabilize about the desired set point.

Imagine your system has almost reached the desired point, but the next control action, makes it jump over the desired point. Now, the control action would be negative (as error is now negative). What if this action brings it back to the state just before the desired point (from where we overshot). Now the system will get caught in an endless loop of oscillating about the desired point, but never reaching it. Hence only Proportional action won't do. The Integral action is used to ensure that the system stabilizes about the desired point and the above stated problem doesn't happen.

What if the system takes too long to reach the desired point? The derivative action is then used to ensure it reaches there quickly, but this is bound to cause some over shoot and oscillations. It must be noted, that it is not always required to use all three P, I and D control actions. Most applications use only P and I actions, reason being that the derivative actions amplifies high frequency noise. Thus until unless you really need it, it derivative action should be avoided.