Feedback control uses the difference between a desired output (setpoint) and the actual output (error signal) to drive a system toward its goal. Open-loop systems apply input without measuring output, while closed-loop systems continuously correct based on feedback. The block diagram of a closed-loop system includes the plant (process being controlled), the controller, sensors, and the feedback path. Key performance goals include stability, accuracy, and speed of response.
Start by analyzing simple thermostats or cruise control systems as physical intuitions for feedback before formalizing with math. Draw block diagrams of everyday control systems and compare open-loop vs. closed-loop responses to understand why feedback matters.
Imagine a thermostat: you set a desired temperature (the setpoint), the thermostat measures the actual temperature, computes how far off it is, and turns the heater on or off accordingly. This loop — measure, compare, correct — is the essence of feedback control. Every closed-loop system has the same fundamental structure: a reference input, a sensor measuring actual output, a summing junction computing the error, a controller deciding what action to take, and a plant (the physical system being controlled) that responds.
The key insight separating open-loop from closed-loop control is whether the system knows what it actually produced. An open-loop system applies a predetermined input and hopes for the best — like setting a microwave timer without checking if food is cooked. A closed-loop system continuously checks and corrects. This makes closed-loop systems robust to disturbances and modeling errors, which is why they dominate real engineering applications. However, closed-loop control requires reliable sensors and introduces the possibility of instability.
Gain is the amplification applied to the error signal before it reaches the plant. Intuitively, higher gain means stronger corrections, which reduces steady-state error and speeds response. But there is a catch: if gain is too high, the system overcorrects, then overcorrects its overcorrection, and so on — producing oscillations or instability. This tradeoff is one of the central challenges of control design and motivates the rigorous tools you will encounter next (Laplace transforms, transfer functions, root locus, Bode plots).
The block diagram is the standard language for describing control systems. Each block represents a component characterized by how it transforms an input signal into an output signal. The feedback path takes the plant output, feeds it back through the sensor, and subtracts it from the reference at the summing junction. The signal flowing into the controller is always the error — not the reference, and not the output directly. Understanding this distinction is critical when analyzing how disturbances enter the system and where they can be rejected.