Sensors and feedback signals let engineered systems measure what is actually happening instead of only assuming a command was followed. A motor, heater, robot arm, or pump can drift away from its target because of friction, load changes, noise, or changing environmental conditions. By sensing variables such as position, temperature, pressure, or speed, a controller can compare reality with a desired setpoint.
This is the foundation of closed-loop control in modern machines, vehicles, factories, and medical devices.
A typical feedback system starts with a sensor that converts a physical variable into an electrical signal such as voltage, current, pulses, or resistance. Signal conditioning then filters noise, amplifies weak signals, and converts the signal into a form the controller can read. The controller subtracts the measured value from the setpoint to produce an error signal, then calculates a correction for the actuator.
The loop closes when the actuator changes the process and the sensor measures the new result.
Understanding Engineering: Sensors and Feedback Signals
A sensor is not a perfect observer. It has range, resolution, accuracy, repeatability, and response time. Range is the span it can measure.
Resolution is the smallest change it can notice. Accuracy tells how close its reading is to the true value. Repeatability tells whether it gives nearly the same reading when the same condition occurs again.
A bathroom scale may show changes of one tenth of a kilogram, yet still be offset from the true mass. In engineering, an inaccurate offset can make a system settle at the wrong temperature, speed, or position.
Calibration connects a sensor reading to a known physical value. Engineers compare the sensor with a trusted reference at one or more points. They then correct the reading in hardware or software.
Many sensors are not perfectly linear. A change of one unit in the physical quantity may not produce an equal change in the electrical output across the full range. Temperature probes, pressure sensors, and light sensors can all show this behavior.
A controller may use a calibration table to turn raw readings into useful measurements. Calibration must sometimes be repeated because age, vibration, contamination, and temperature changes can alter a sensor.
Real signals contain noise. Electrical interference can come from motors, power cables, radio transmitters, or switching circuits. Mechanical vibration can shake an accelerometer or position sensor.
If a controller reacts to every tiny fluctuation, an actuator may chatter, wear out, or waste energy. Filtering smooths rapid unwanted changes, but strong filtering creates delay. That delay matters because the controller is acting on older information.
A thermostat can tolerate a slow reading, while a drone balancing in the air needs measurements that arrive very quickly. Engineers choose filtering by considering both noise and the speed of the process.
Feedback can fail when the measurement is misleading. A temperature sensor placed too far from a heater may report a cool value after the object is already overheating. A wheel encoder can count motion even when a vehicle is slipping on ice.
A blocked pressure tube can make a pressure reading respond slowly. Safety critical systems often use more than one sensor or compare different kinds of measurements. A car may compare wheel speeds with acceleration data to detect slipping.
Students should pay attention to units, sensor location, delays, and possible failure modes. These details often explain why a well written control rule behaves poorly on a real machine.
Key Facts
- Error signal: e(t) = r(t) - y(t), where r(t) is the setpoint and y(t) is the measured output.
- Sensor sensitivity: S = Δoutput / Δinput, such as volts per degree Celsius or volts per pascal.
- A thermocouple produces a small voltage related to temperature difference, often in millivolts.
- An encoder measures position or speed by counting pulses: angle = pulse count × degrees per pulse.
- Signal conditioning can include amplification, filtering, isolation, linearization, and analog-to-digital conversion.
- Negative feedback reduces error by applying a correction that drives the output toward the setpoint.
Vocabulary
- Sensor
- A device that detects a physical quantity such as temperature, pressure, position, or speed and converts it into a usable signal.
- Setpoint
- The desired target value that a control system tries to maintain or reach.
- Feedback signal
- A measured signal returned to the controller so it can compare actual output with the desired value.
- Signal conditioning
- The process of improving or modifying a sensor signal so it can be accurately read by a controller.
- Actuator
- A device such as a motor, valve, heater, or piston that changes the physical system in response to a control signal.
Common Mistakes to Avoid
- Confusing the sensor signal with the control signal is wrong because the sensor reports what happened, while the control signal commands the actuator to make a change.
- Ignoring units in sensor sensitivity is wrong because a value like 0.02 V/°C only makes sense when input and output units are tracked.
- Assuming all feedback improves stability is wrong because incorrect sign, delay, or excessive gain can make a system oscillate or become unstable.
- Forgetting signal conditioning is wrong because raw sensor outputs are often noisy, weak, nonlinear, or outside the safe input range of a controller.
Practice Questions
- 1 A temperature sensor has a sensitivity of 10 mV/°C. What voltage change occurs when the temperature increases by 35°C?
- 2 A position encoder has 1000 pulses per revolution. If the controller counts 250 pulses, what angle has the shaft turned in degrees?
- 3 A motor speed controller has a setpoint of 1200 rpm, but the feedback sensor measures 1100 rpm. Explain the sign and purpose of the error signal if the controller should speed up the motor.