Feedback amplifiers and their properties

Loop gain and desensitivity, the four feedback topologies and their effect on impedances, bandwidth, distortion, and stability.

Drafted with Aria, reviewed by the AiCanCode.org team. Spotted an error? Use Give Feedback at the bottom of the page.

Why it matters

Transistor gain varies from device to device and drifts with temperature, yet an instrument amplifier must have a gain you can trust to 0.1 %. Negative feedback is the trick that makes this possible: it trades surplus raw gain for a closed-loop gain set by precision resistors, and at the same time widens bandwidth, reduces distortion and lets you choose whether the amplifier looks like a high- or low-impedance source and load. Every op-amp circuit in the following topics is a feedback amplifier.

Key ideas

The basic loop. A basic amplifier of gain A drives the output x_o. A feedback network samples the output and returns x_f = β·x_o, which is subtracted from the input x_s at a mixing point, so the amplifier sees the error x_i = x_s − x_f. Solving gives the closed-loop gain A_f = A/(1 + Aβ). The quantity Aβ is the loop gain and (1 + Aβ) is the desensitivity factor D.

When Aβ ≫ 1 the closed-loop gain becomes A_f ≈ 1/β: it depends only on the feedback network (usually resistors), not on the transistors.

Benefits of negative feedback (each by the factor D):

  • Gain desensitivity: dA_f/A_f = (1/D)·dA/A.
  • Bandwidth: for a single-pole amplifier the upper cut-off rises to f_H·D and the lower cut-off falls to f_L/D, so gain × bandwidth is unchanged.
  • Distortion: nonlinear distortion generated inside the loop falls by D (for the same output level).
  • Noise or interference injected inside the loop (after the first stage) is reduced relative to the signal; noise entering at the input with the signal is not.
  • Impedances change by D in a direction set by the topology.

Four topologies. Name them by what is sampled at the output and how the feedback is mixed at the input.

  • Voltage-series (series-shunt): samples output voltage, mixes in series with the input voltage. Voltage amplifier. R_in rises by D, R_out falls by D. Example: non-inverting op-amp amplifier, emitter follower.
  • Current-series (series-series): samples output current, mixes a voltage in series. Transconductance amplifier. R_in rises, R_out rises. Example: CE stage with unbypassed R_E.
  • Voltage-shunt (shunt-shunt): samples output voltage, mixes a current at the input node. Transresistance amplifier. R_in falls, R_out falls. Example: inverting op-amp amplifier, collector-to-base bias resistor.
  • Current-shunt (shunt-series): samples output current, mixes a current. Current amplifier. R_in falls, R_out rises. Rule of thumb: series mixing raises input resistance, shunt mixing lowers it; voltage sampling lowers output resistance, current sampling raises it.

Positive feedback (1 + Aβ < 1) increases gain; when Aβ = −1, i.e. loop gain of magnitude 1 with the right phase, the circuit oscillates (Barkhausen criterion). Comparators with hysteresis and oscillators use positive feedback deliberately.

Stability. In a real amplifier A has several poles, so its phase shifts with frequency. If the loop phase reaches −180° while |Aβ| is still above 1, negative feedback turns positive and the amplifier oscillates. Phase margin (how far the phase is from −180° where |Aβ| = 1) of about 45°–60° is a common design target; more feedback (larger β) makes this harder, which is why op-amps are frequency-compensated.

Formulas

A_f = A / (1 + A·β) — closed-loop gain; A open-loop gain, β feedback factor (both dimensionless for a voltage amplifier).

A_f ≈ 1/β when A·β ≫ 1.

dA_f/A_f = (1/(1 + A·β))·(dA/A) — gain sensitivity.

f_Hf = f_H·(1 + A·β), f_Lf = f_L / (1 + A·β) — single-pole amplifier.

R_if = R_i·(1 + A·β) (series mixing), R_if = R_i / (1 + A·β) (shunt mixing).

R_of = R_o / (1 + A·β) (voltage sampling), R_of = R_o·(1 + A·β) (current sampling). Here A is the open-loop gain measured with the appropriate load (and the output short-circuit or open-circuit value where textbooks specify).

D_f = D / (1 + A·β) — distortion with feedback.

Worked examples

Example 1 (standard: desensitivity). An amplifier has A = 1000 and β = 0.01. Find A_f. If A falls by 20 % because of ageing, by how much does A_f fall?

  1. A_f = A/(1 + Aβ) = 1000/(1 + 10) = 90.9.
  2. New A = 800: A_f = 800/(1 + 8) = 88.9.
  3. Change = (90.9 − 88.9)/90.9 = 2.2 %. (The differential formula gives 20 %/11 = 1.8 %, accurate only for small changes.)

Answer: A_f = 90.9; a 20 % fall in A gives only about a 2.2 % fall in A_f.

Example 2 (GATE level: voltage-series amplifier). A voltage amplifier has A = 10⁴, R_i = 10 kΩ, R_o = 100 Ω and f_H = 1 kHz (single pole). Voltage-series feedback with β = 0.02 is applied. Find A_f, R_if, R_of and f_Hf.

  1. Desensitivity: D = 1 + Aβ = 1 + 10⁴ × 0.02 = 201.
  2. A_f = A/D = 10⁴/201 = 49.75 (close to 1/β = 50).
  3. Series mixing: R_if = R_i·D = 10 kΩ × 201 = 2.01 MΩ.
  4. Voltage sampling: R_of = R_o/D = 100/201 = 0.498 Ω.
  5. f_Hf = f_H·D = 1 kHz × 201 = 201 kHz.

Answer: A_f ≈ 49.8, R_if ≈ 2.01 MΩ, R_of ≈ 0.50 Ω, f_Hf ≈ 201 kHz. Gain × bandwidth = 49.75 × 201 kHz ≈ 10 MHz, the same as 10⁴ × 1 kHz.

Common mistakes

  • Using A_f = A/(1 − Aβ) with negative-feedback sign conventions mixed up; decide the sign once.
  • Applying R_in × D to a shunt-mixing circuit (it should be divided).
  • Assuming feedback reduces noise that enters with the input signal; it does not.
  • Forgetting that the bandwidth improvement holds exactly only for a single-pole amplifier.
  • Using the differential sensitivity formula for large changes in A.
  • Thinking more feedback always makes an amplifier more stable; with several poles it can cause oscillation.

For GATE IN

Expect numericals on A_f, β from given open- and closed-loop gains, the percentage change in A_f for a change in A, bandwidth extension, and input or output resistance with feedback. Also expect topology identification from a circuit (which quantity is sampled, how it is mixed) and loop-gain or phase-margin reasoning. Practise the four topology rules until they are automatic.

Quick check

  1. A = 200 and A_f = 20. What is β?
  2. A = 10⁵ and β = 0.1. What is A_f to three significant figures?
  3. Which topology lowers both input and output resistance?
  4. A single-pole amplifier has GBW = 2 MHz. With feedback giving A_f = 20, what is f_H?

Answers: 1. 0.045. 2. 10.0. 3. Voltage-shunt (shunt-shunt). 4. 100 kHz.

Try answering each one aloud before you open it.

  1. 1.What is a feedback amplifier?Concept

    A feedback amplifier is an electronic amplifier that uses feedback to control its gain and other properties. Feedback is the process of taking a portion of the output signal and returning it to the input. This can be done in a positive or negative manner, affecting the amplifier's performance in terms of stability, bandwidth, and distortion.

  2. 2.Explain the difference between positive and negative feedback in amplifiers.Concept

    Positive feedback occurs when the feedback signal is in phase with the input signal, which can lead to increased gain and potential instability or oscillation. Negative feedback, on the other hand, involves a feedback signal that is out of phase with the input, reducing gain but improving stability, bandwidth, and linearity. Negative feedback is more commonly used in amplifiers to achieve a more controlled and stable output.

  3. 3.What are the main advantages of using negative feedback in amplifiers?Concept

    Negative feedback in amplifiers offers several advantages: it stabilizes the gain, reduces distortion, increases bandwidth, and improves the linearity of the amplifier. By feeding a portion of the output back to the input in an out-of-phase manner, negative feedback helps to counteract variations in the amplifier's performance due to component tolerances and temperature changes.

  4. 4.Why is negative feedback preferred over positive feedback in most amplifier applications?Application

    Negative feedback is preferred because it enhances the stability and performance of the amplifier. It reduces distortion and noise, increases bandwidth, and makes the amplifier's gain less sensitive to component variations. Positive feedback, while useful in certain applications like oscillators, can lead to instability and is generally avoided in amplifiers where a stable and linear output is desired.

  5. 5.What happens to the bandwidth of an amplifier when negative feedback is applied?Application

    For a single-pole amplifier, negative feedback multiplies the upper cut-off frequency by the desensitivity factor (1 + Aβ) and divides the gain by the same factor, so the gain-bandwidth product stays constant. The lower cut-off frequency of an AC-coupled amplifier is divided by (1 + Aβ). For example, A = 10⁴ with f_H = 1 kHz and β = 0.02 gives A_f ≈ 50 and f_H ≈ 201 kHz. With multiple poles the exchange is not exact and too much feedback can cause peaking.

  6. 6.How does negative feedback affect the input and output impedance of an amplifier?Application

    It depends on the topology, and each change is by the factor (1 + Aβ). Series mixing at the input raises the input impedance; shunt mixing lowers it. Voltage sampling at the output lowers the output impedance; current sampling raises it. So a voltage-series amplifier (non-inverting op-amp stage) has very high input and very low output impedance, while a voltage-shunt amplifier (inverting stage) has low input impedance at the summing node.

  7. 7.In a feedback amplifier, what is the effect of the feedback factor on stability?Application

    A larger feedback factor β gives a lower and better-defined closed-loop gain (≈ 1/β) and less sensitivity to device variations. But dynamic stability depends on the loop gain Aβ versus frequency: if the loop phase reaches −180° while |Aβ| is still above 1, the feedback becomes positive and the amplifier oscillates. Increasing β raises the loop gain at every frequency and reduces the phase margin, so unity-gain followers are the hardest case and op-amps are internally compensated to remain stable there.

  8. 8.Calculate the closed-loop gain of an amplifier with an open-loop gain of 100 and a feedback factor of 0.1.Numerical

    The closed-loop gain (A_cl) of a feedback amplifier can be calculated using the formula: A_cl = A_ol / (1 + A_ol * β), where A_ol is the open-loop gain and β is the feedback factor. Substituting the given values: A_cl = 100 / (1 + 100 * 0.1) = 100 / 11 ≈ 9.09.

  9. 9.If an amplifier has a gain of 50 without feedback and a gain of 10 with feedback, what is the feedback factor?Numerical

    The feedback factor (β) can be calculated using the formula: A_cl = A_ol / (1 + A_ol * β), where A_cl is the closed-loop gain and A_ol is the open-loop gain. Rearranging the formula to solve for β gives: β = (A_ol - A_cl) / (A_cl * A_ol). Substituting the given values: β = (50 - 10) / (10 * 50) = 40 / 500 = 0.08.

  10. 10.What is the impact of feedback on the distortion levels in an amplifier?Application

    Feedback, particularly negative feedback, reduces distortion levels in an amplifier. By feeding a portion of the output back to the input in an out-of-phase manner, negative feedback counteracts nonlinearities in the amplifier's operation. This results in a more linear output signal with reduced harmonic distortion, improving the overall fidelity of the amplified signal.

Finished this topic? Mark it so your progress, study plan and readiness keep up.

Stuck on something here?