Amplifier feedback is the process of routing a fraction of an amplifier's output signal back to its input to precisely control gain, reduce distortion, and widen bandwidth. When you build an audio preamp, a sensor signal conditioner, or a guitar pedal, relying on an operational amplifier's raw "open-loop" gain is a recipe for severe clipping, thermal drift, and unusable noise. By intentionally sacrificing a massive amount of raw voltage gain, you buy mathematical predictability, rock-solid stability, and a flat frequency response across the audio spectrum.
The Core Mechanism: What Amplifier Feedback Actually Changes
Applying a feedback network fundamentally alters four critical parameters of an amplifier circuit. First, it desensitizes the gain to component tolerances and temperature variations. Second, it extends the bandwidth by enforcing a constant Gain-Bandwidth Product (GBWP). Third, it reduces harmonic distortion by continuously comparing the output to the input and correcting errors. Finally, it modifies input and output impedance, allowing you to match stages without loading down your signal source.
Think of it like a thermostat-controlled water valve. If the municipal water pressure (open-loop gain) fluctuates wildly, the thermostat (feedback network) measures the actual flow rate and adjusts the valve to maintain a steady, predictable output regardless of upstream pressure spikes.
Beginners commonly confuse negative feedback with positive feedback, or mistake feedback-induced oscillation for a "broken" chip. Negative feedback (routing to the inverting input) stabilizes and linearizes the circuit. Positive feedback (routing to the non-inverting input) creates hysteresis for comparators, Schmitt triggers, or intentional oscillators. If your linear audio amp is screaming at 500 kHz, you haven't necessarily blown the IC; you have likely introduced an unintended phase shift that turned your negative feedback positive at high frequencies.
Worked Numeric Example: Calculating Closed-Loop Gain
Let us look at a standard non-inverting amplifier configuration using a ubiquitous NE5532 dual op-amp (currently around $0.50 per unit). We want a voltage gain of 11 V/V to boost a line-level signal.
- Open-Loop Gain ($A_{OL}$): 100,000 V/V (100 dB) at DC.
- Feedback Resistors: $R_f = 10\text{k}\Omega$, $R_i = 1\text{k}\Omega$.
- Feedback Fraction ($\beta$): The voltage divider ratio fed back to the inverting pin. $\beta = R_i / (R_i + R_f) = 1000 / 11000 \approx 0.0909$.
The magic number in feedback theory is the Loop Gain, which is the product of the open-loop gain and the feedback fraction ($A_{OL} \times \beta$). Loop gain dictates how much error-correction the amplifier can apply.
Loop Gain: 9,090 (79.1 dB)
Now, we calculate the actual closed-loop gain ($A_{CL}$) using the exact feedback equation:
$$A_{CL} = \frac{A_{OL}}{1 + (A_{OL} \times \beta)} = \frac{100,000}{1 + 9,090} = 10.9988 \text{ V/V}$$
Notice that the ideal textbook gain ($1 + R_f/R_i = 11$) is nearly identical to the exact calculated gain. The massive loop gain forces the circuit to obey the resistor ratio, rendering the op-amp's internal transistor variations irrelevant.
Distortion Reduction: If the raw NE5532 open-loop Total Harmonic Distortion (THD) is roughly 1% at a given frequency, the closed-loop THD is reduced by the factor of $(1 + A_{OL}\beta)$.
$1\% / 9,091 = 0.00011\%$. This is why negative feedback is the cornerstone of high-fidelity audio design.
Where You Meet Amplifier Feedback in Practice
You will encounter specific feedback topologies depending on the IC and the application. Local feedback occurs around a single gain stage, while global feedback wraps around the entire amplifier chain, including the output power transistors.
| IC / Module | Typical Application | Feedback Topology & Notes |
|---|---|---|
| NE5532 | Audio Preamps / Mixers | External resistors set voltage gain. Requires a 100nF bypass cap near pins 4 and 8 to prevent high-frequency feedback loop instability. |
| LM386 | Low-Voltage Speaker Amps | Internal feedback sets default gain to 20. Bypassing pins 1 and 8 with a 10µF capacitor drops the internal feedback fraction, pushing gain to 200. |
| OPA2134 | High-End DAC I/V Stages | FET-input op-amp used in transimpedance (current-to-voltage) feedback loops. Feedback resistor directly sets gain; parasitic capacitance here is the #1 cause of ringing. |
| TPA3116D2 | Class-D Power Amp Boards | Global negative feedback from the LC output filter back to the error amplifier. Modifies output impedance to maintain damping factor across varying speaker loads. |
In power amplifiers like the TPA3116D2, the feedback network must sample the signal after the output inductor and capacitor filter. If you attempt to modify the feedback resistors on a Class-D board to increase gain without recalculating the compensation network, you will alter the phase margin and likely destroy your tweeters with ultrasonic oscillation. Always consult the manufacturer's application notes, such as the Analog Devices guides on op-amp feedback stability, before altering power stage feedback loops.
Frequently Asked Questions
Why does my amplifier oscillate when I add a feedback capacitor?
Adding a capacitor in parallel with your feedback resistor creates a pole in the feedback network. While this is often done intentionally to limit high-frequency bandwidth (acting as a low-pass filter), it also introduces a phase shift. If the phase shift of the feedback network, combined with the internal phase shifts of the op-amp's gain stages, reaches 180 degrees while the loop gain is still greater than 1 (0 dB), the negative feedback mathematically becomes positive feedback. The circuit turns into an oscillator. To fix this, you must ensure adequate phase margin (ideally >45 degrees) by reducing the capacitor value, adding a small series isolation resistor at the output, or selecting an op-amp with better internal Miller compensation.
What is the difference between negative and positive amplifier feedback?
The difference lies in which input terminal receives the routed signal. Negative feedback routes the output signal back to the inverting (-) input. This subtracts from the input differential, forcing the output to stabilize and linearize the transfer function. It is used in 95% of linear analog circuits (amplifiers, filters, regulators). Positive feedback routes the output to the non-inverting (+) input. This reinforces the input differential, driving the output rapidly to the supply rails. It is used intentionally in non-linear circuits like comparators with hysteresis (Schmitt triggers), active oscillators, and latch circuits.
How does amplifier feedback affect the input and output impedance?
Feedback topology dictates impedance transformation. Series feedback (where the feedback signal is applied in series with the input voltage, like a non-inverting amplifier) multiplies the input impedance by the loop gain factor $(1 + A_{OL}\beta)$, making it ideal for buffering high-impedance sensors. Shunt feedback (where feedback current sums at the inverting node, like an inverting amplifier or transimpedance amplifier) divides the input impedance, creating a virtual ground that is perfect for current-sensing applications. On the output side, voltage feedback (sampling the output voltage) divides the output impedance by the loop gain, yielding a stiff voltage source capable of driving heavy loads without sagging.
