negative feedback

Negative feedback is like a thermostat for an amplifier. A thermostat keeps comparing the room temperature to the dial and keeps nudging the heater until the two match. An amplifier with negative feedback keeps comparing its output to what you asked for and keeps correcting itself, moment by moment, so the output stays where you want it. This one idea is the secret behind almost every well-behaved op-amp circuit.

Here is the mechanism in plain steps. You take a fraction of the output, call it beta, and feed it back to subtract from the input. The amplifier amplifies only the leftover difference (error) by its raw gain A. Write it out: Vout = A times (Vin - beta times Vout). Solve for the output and you get Vout/Vin = A / (1 + A times beta). When the raw gain A is enormous, that product A times beta dwarfs the 1, and the gain collapses to almost exactly 1/beta. Example: a bare op-amp gain of 200,000 with a divider that feeds back one tenth of the output (beta = 1/10) gives a closed-loop gain of very nearly 10, set by the divider, not by the wobbly 200,000.

Why this matters: feedback trades away raw gain you do not need in exchange for things you desperately want. The gain becomes set by precise resistors instead of by the op-amp's wildly variable, temperature-drifting open-loop gain. You also get wider bandwidth, lower distortion, and steadier behaviour. The honest catch: if the signal going around the loop picks up too much phase shift (delay), the correction can arrive inverted and reinforce the disturbance instead of cancelling it. Then the feedback turns positive and the circuit oscillates. Most of this field is about keeping that from happening.