Control system performance is based on feedback. Control of electric motors, however, continues to be a bit mysterious because the common conventions associated with motor control are often driven by cost considerations. The feedback component is often target for elimination in cost constrained systems.
Control systems can be described as “open loop” or “closed loop” depending on the whether or not the system being controlled is well characterized. Many forms of motor control seek to be “open loop”, that is, without the use of a feedback device. However, this notion should be modified to open loop meaning without an explicit feedback device. This is because great effort is expended to “infer” what is going on in the motor through various means. The most common of which is current.
In the world of electric motors, the alternating current motor of Nicola Tesla is well understood, and rarely requires a feedback device. Motor speed is derived from the frequency of the power being supplied minus losses depending on the details of rotor construction and how a specific load affects the motor. The standard ac motor has a small amount of rotor “slip” from 1800 rpm to 1750 rpm which reflects the magnetizing current losses in the motor and magnetic features in the rotor that would be needed to maintain perfect synchronism with the line frequency.
Load variations can be measured by sensing the current in the line going to the motor. So there is a feedback element available from which a great deal of information can be derived. This is where the ambiguity about feedback comes in. The current needed to run the motor with no load is fixed value, so more current read on the motor leads is load, until the motor reaches locked rotor current or stall.
In brushless dc systems a similar approach is used. Detecting the zero crossing point of the phase current establishes precise timing of the rotor speed and is used to regulate timing of current pulses to all three phases of the motor. In this way even the brushless dc motor can be operated without an explicit feedback sensor. However the tradeoff here is very poor low speed regulation of the motor which makes this approach unsuitable for many applications.
From a control system standpoint, feedbacks are the last, slowest loop in the control scheme of the motor. This makes sense in the context of position control as it is normally executed in a PLC or motion controller. However, this makes load regulation more of a challenge since the actual error detection of the control system is being done a level removed from the actual load.
A host of mathematical tools from the signal processing domain have traditionally been employed to characterize the lag created by the control system and the interaction of the controls at varying speeds. All of which works well, but has also lead to “rules of thumb” that are not very clearly understood and which are sometimes misleading.