A feedback system is a control loop where a portion of the output signal is routed back to the input to compare against a reference, automatically adjusting the system's behavior to maintain a desired state. In electronics, this concept is the dividing line between a "dumb" open-loop circuit that blindly pushes current and a "smart" closed-loop system that actively fights off disturbances. What feedback fundamentally changes in a real circuit is disturbance rejection. Without it, a power supply's output voltage would sag the moment you connected a heavy load. With it, the circuit senses the sag and instantly opens its internal valves wider to compensate, holding the voltage steady.

Think of it like driving a car: an open-loop system is locking the gas pedal in one position and hoping you maintain 60 mph regardless of hills or wind. A closed-loop feedback system is cruise control, constantly measuring actual wheel speed and adjusting the throttle to maintain exactly 60 mph.

The Math on the Bench: A Worked Numeric Example

To understand the definition of a feedback system in hardware, we have to look at negative feedback in operational amplifiers. An op-amp like the ubiquitous Texas Instruments LM358 has an internal open-loop gain that is massive but highly unpredictable—typically around 100,000 V/V, but it drifts wildly with temperature and manufacturing tolerances. If you tried to use it open-loop to amplify a 1.5V sensor signal, the output would instantly slam against the positive power rail (saturation).

We tame this by adding a negative feedback network using two precision resistors in a non-inverting configuration. Let's calculate the exact outcome on the bench:

  • Input Voltage (Vin): 1.5V (from a precision sensor)
  • Feedback Resistor (Rf): 10,000 Ω (10kΩ, 1% tolerance)
  • Ground Resistor (Ri): 2,200 Ω (2.2kΩ, 1% tolerance)

The closed-loop voltage gain (Av) formula for a non-inverting amplifier is:

Av = 1 + (Rf / Ri)

Av = 1 + (10000 / 2200) = 1 + 4.545 = 5.545

Therefore, the output voltage is:

Vout = Vin × Av = 1.5V × 5.545 = 8.317V

Because of the negative feedback loop, the op-amp's internal transistors automatically adjust their conduction to force the voltage at the inverting input to exactly match the non-inverting input (1.5V). The circuit's gain is now entirely dependent on the 1% tolerance of your passive resistors, completely ignoring the op-amp's internal 100,000 V/V variance. As detailed in All About Circuits' semiconductor guide, this is the magic of feedback: it trades raw, unusable gain for precision and stability.

Where You Meet Feedback in Practice

You will encounter feedback loops in almost every modern electronic device. Here is where they do the heavy lifting on the jobsite and the bench:

Switching Power Supplies (Buck/Boost Converters): A controller chip like the UC3842 uses a feedback pin (COMP) to monitor the output voltage via a resistor divider. If the 12V output sags to 11.8V under load, the error amplifier increases the PWM duty cycle to the MOSFET, pushing more energy through the inductor to restore the 12V target.
  • Audio Amplifiers: Class AB and Class D amplifiers use global negative feedback to reduce Total Harmonic Distortion (THD). The feedback loop compares the output audio waveform to the input and cancels out the nonlinearities introduced by the output transistors.
  • Phase-Locked Loops (PLLs): Used in radio receivers and microcontrollers to generate precise clock frequencies. A phase detector compares a divided-down version of the output clock to a stable crystal reference, feeding an error voltage to a VCO (Voltage-Controlled Oscillator) to lock the frequency perfectly.
  • Thermal Management: PC motherboard fan controllers use tachometer feedback from the fan's third wire to adjust the PWM drive signal, ensuring the fan spins at the exact RPM requested despite dust buildup or bearing friction.

Real-World Scenario Walkthrough: When Closed-Loop Goes Wrong

Understanding the theory is one thing; tuning a feedback system in the real world is where engineers earn their keep. Here is a classic failure mode involving a closed-loop DC motor speed controller.

The Setup

We are building a constant-speed conveyor belt drive. The hardware consists of a 12V brushed DC motor, an Arduino Uno running a PID (Proportional-Integral-Derivative) control loop, an L298N H-bridge motor driver, and a 360 PPR (pulses per revolution) optical encoder for speed feedback. The target speed is exactly 1500 RPM.

The Numbers

The Arduino reads the encoder and calculates the PID output every 20 milliseconds (a 50Hz loop rate). We set our initial tuning parameters aggressively to get a fast response:

  • Kp (Proportional): 4.0
  • Ki (Integral): 0.5
  • Kd (Derivative): 0.0 (disabled to keep it simple)

The Outcome

When powered on, the motor doesn't smoothly ramp to 1500 RPM. Instead, it shrieks, violently oscillating between 1100 RPM and 1900 RPM. The conveyor belt shudders, and the L298N driver chip gets dangerously hot from the rapid current reversals.

What Went Wrong (And How to Fix It)

The system suffered from phase lag and integral windup. The 50Hz sampling rate means there is a 20ms delay between the motor changing speed and the Arduino "seeing" it via the encoder. Because the Proportional gain (Kp = 4.0) was too high, the Arduino slammed the PWM to 100% the moment it saw the speed was below target. By the time the 20ms delay passed and the Arduino realized the motor had overshot 1500 RPM, the motor's physical inertia had already carried it to 1900 RPM. The Integral term (Ki) then "wound up," accumulating error over time and making the overshoot even worse on the next cycle.

The Fix: We dropped Kp to 1.2 to soften the initial reaction. We added a derivative term (Kd = 0.8) with a low-pass filter, which allows the Arduino to predict the *slope* of the speed change and back off the throttle before it overshoots. Finally, we increased the PID loop frequency to 200Hz (5ms sampling) to minimize the phase lag. The motor now locks to 1500 RPM within 0.4 seconds with zero oscillation.

Common Confusions: Feedforward and "Positive" Feedback

When studying the definition of a feedback system, hobbyists and students frequently trip over two related concepts.

Feedback vs. Feedforward

Feedback is reactive; it waits for an error to occur before correcting it. Feedforward is proactive. In a sophisticated 12V-to-5V buck converter, the feedback loop monitors the 5V output. But a feedforward path might also monitor the 12V input. If the 12V battery suddenly drops to 11V, the feedforward circuit instantly increases the PWM duty cycle before the 5V output has a chance to sag. Feedforward anticipates the disturbance; feedback cleans up whatever error is left over.

The Danger of "Positive" Feedback

In everyday language, "positive" means good. In circuit theory, positive feedback means the output signal is routed back to the input in phase with the original signal, reinforcing it. This is highly desirable in oscillator circuits (like a 555 timer astable multivibrator or a Wien bridge oscillator) where you want the circuit to continuously toggle or ring.

Safety Warning: Accidental positive feedback in power amplifiers or linear regulators leads to thermal runaway or latch-up. If a power transistor gets hot, its leakage current increases. If that increased current is fed back in a way that drives the base harder, the transistor will conduct more, get hotter, and rapidly destroy itself in a cascade failure. Always verify your feedback polarity with an oscilloscope before applying full load to a custom power circuit.

Frequently Asked Questions

Can a feedback system be purely mechanical?

Yes. The classic example is the centrifugal governor on a steam engine, invented by James Watt. As the engine spins faster, heavy flyballs swing outward due to centrifugal force, mechanically pulling a linkage that closes the steam valve. This is a pure mechanical negative feedback loop requiring zero electricity.

Why does my op-amp feedback circuit ring or oscillate?

Op-amp oscillation is almost always caused by capacitive loading on the output or poor PCB layout introducing stray capacitance in the feedback path. This adds phase shift at high frequencies, turning your designed negative feedback into positive feedback at a specific frequency. According to Electronics Tutorials, placing a small isolation resistor (e.g., 22Ω to 100Ω) between the op-amp output and the capacitive load usually cures this instability.

What is the difference between analog and digital feedback?

Analog feedback (like the LM358 resistor network) operates continuously in real-time with theoretically infinite bandwidth. Digital feedback (like the Arduino PID motor controller) samples the output at discrete time intervals. Digital systems are easier to tune and can implement complex math (like adaptive filtering), but they are fundamentally limited by their sampling rate and the processing delay of the microcontroller.