Comparison

Positive vs Negative Feedback

An LM741 wired as an inverting amplifier uses negative feedback to set gain at exactly −10 regardless of the op-amp's open-loop gain of 200,000. Wire the output to the non-inverting input instead — positive feedback — and the same LM741 becomes a Schmitt trigger that snaps between +V_sat and −V_sat and never amplifies linearly again. These are not two versions of the same thing; they produce fundamentally different circuit behaviors.

EEE, ECE, EI

Side-by-side comparison

ParameterPositiveNegative Feedback
Feedback polarityOutput fed back to reduce the error signalOutput fed back to reinforce the error signal
Effect on gainReduces closed-loop gain; A_CL = A/(1+Aβ)Increases effective gain; can reach infinity (oscillation)
Effect on stabilityIncreases stability; Barkhausen criteria not metCan cause instability or sustained oscillation
Effect on bandwidthIncreases bandwidth (gain-bandwidth product conserved)Narrows bandwidth or leads to oscillation
Effect on distortionReduces THD by factor (1 + Aβ)May increase distortion; linearization is lost
Effect on input/output impedanceModifies Z_in and Z_out depending on topologyCan cause impedance extremes (near zero or infinity)
Typical applicationInverting/non-inverting amplifiers, regulators, PLL phase detectorsSchmitt trigger, Wien bridge oscillator, 555 timer SR latch
Oscillation conditionMust not satisfy Barkhausen: |Aβ| < 1 at phase = 0°Barkhausen criteria: |Aβ| = 1, ∠Aβ = 0° (360°)
Example (negative)LM741 inverting amp: A_CL = −R_f/R_inLM741 Schmitt trigger: hysteresis set by R1/R2 divider
Example (positive)Not applicableRC Wien bridge oscillator at f = 1/(2πRC), e.g., 1 kHz with R=15.9 kΩ, C=10 nF

Key differences

Negative feedback reduces gain from 200,000 (open-loop LM741) to a stable, predictable value set by passive resistors — an inverting amplifier with R_f = 100 kΩ and R_in = 10 kΩ gives A_CL = −10 regardless of transistor variation or temperature. This same feedback loop reduces THD by the factor (1 + Aβ) ≈ 20,000 and widens bandwidth from the LM741's 10 Hz open-loop to ~100 kHz closed-loop. Positive feedback does the opposite: a Schmitt trigger wired with 10 kΩ/1 kΩ from output to non-inverting input creates hysteresis of ±0.5 V — the output snaps hard between rails and never lingers in the linear region. A Wien bridge oscillator uses positive feedback at exactly one frequency (f = 1/2πRC) to sustain oscillation; AGC keeps |Aβ| = 1. These two feedback types produce circuits that cannot be designed by the same analysis methods.

When to use Positive

Use negative feedback when you need stable, predictable gain, low distortion, controlled bandwidth, or regulated output. An LM741 with R_f = 100 kΩ and R_in = 10 kΩ gives a gain of −10 that stays within 1% from −40°C to +85°C regardless of op-amp aging or supply variation.

When to use Negative Feedback

Use positive feedback when you need a bistable comparator, a Schmitt trigger with hysteresis, or a sinusoidal oscillator. A 555 timer's internal SR latch uses positive feedback to produce stable HIGH and LOW states; a Wien bridge oscillator with an LM741 and R = 15.9 kΩ, C = 10 nF oscillates at exactly 1 kHz with stable amplitude when gain is trimmed to just above unity.

Recommendation

Choose negative feedback for any amplifier, filter, or regulator where gain accuracy and stability are the goal — it is the foundation of nearly every practical analog circuit. Use positive feedback deliberately and carefully only when bistable switching or sustained oscillation is the intended behavior; applied accidentally, it destroys amplifier stability instantly.

Exam tip: Examiners frequently ask for the Barkhausen stability criterion — state it precisely: for sustained oscillation, the loop gain magnitude must equal 1 (|Aβ| = 1) and the total phase shift around the loop must be 0° (or 360°); knowing both conditions is required for full marks.

Interview tip: Interviewers at analog design firms ask candidates to explain why negative feedback reduces distortion — the correct explanation uses the closed-loop gain expression and notes that any nonlinearity in A appears divided by (1 + Aβ), which for |Aβ| >> 1 makes it negligible; a vague answer about "correction" is not sufficient.

More Analog Electronics comparisons