Feedback Amplifier Interview Questions

Q1. What is negative feedback in an amplifier and what are its benefits?

Negative feedback takes a fraction of the output signal and subtracts it from the input, reducing the effective input signal and thus the gain, but simultaneously improving gain stability, bandwidth, linearity, and input/output impedance control. An LM741 with open-loop gain of 200,000 and 1% feedback fraction β has closed-loop gain stabilized to 1/β = 100, varying by only 0.0005% even if AOL doubles. The gain is now determined almost entirely by the feedback network (resistors), not the transistor parameters — which is why op-amp circuits like inverting amplifiers deliver precise, stable gain despite wide transistor spread.

Follow-up: By how much does negative feedback reduce harmonic distortion?

Q2. What are the four types of feedback topologies in amplifiers?

The four types are: series-shunt (samples output voltage, feedback in series with input — increases Rin, decreases Rout, stabilizes voltage gain); shunt-shunt (samples Vout, feeds back current — decreases Rin and Rout, stabilizes transresistance); series-series (samples Iout, feedback in series — increases both Rin and Rout, stabilizes transconductance); shunt-series (samples Iout, current feedback — decreases Rin, increases Rout, stabilizes current gain). An LM741 inverting amplifier is a shunt-shunt topology where the feedback resistor Rf samples output voltage and feeds back current to the summing node. Identifying the topology determines how feedback changes input and output impedances.

Follow-up: How do you identify the feedback topology from a circuit schematic?

Q3. What is the gain desensitization property of negative feedback?

Gain desensitization means that percentage changes in closed-loop gain are much smaller than percentage changes in open-loop gain by a factor of (1 + Aβ) — the desensitivity factor. If an LM741 open-loop gain varies ±20% with temperature but the feedback factor is β = 0.01 (closed-loop gain = 100), the closed-loop gain varies only ±20%/(1 + 200,000 × 0.01) = ±0.01%, a 2000× improvement. This is why op-amp gain-setting resistors need to be only 0.1% tolerance, not the 0.001% that would be required if using the transistor gain directly.

Follow-up: What is the desensitivity factor and how is it calculated?

Q4. How does negative feedback affect bandwidth?

Negative feedback extends the -3 dB bandwidth by the desensitivity factor (1 + Aβ), trading gain for bandwidth as required by the gain-bandwidth product conservation. An LM741 with 200,000 open-loop gain and 1 MHz unity-gain bandwidth, when configured for closed-loop gain of 100 (β = 0.01), has closed-loop bandwidth = GBW/Av_CL = 1MHz/100 = 10 kHz — the same GBW product holds. This constant gain-bandwidth product means wider bandwidth always comes at the cost of lower gain, which is why broadband IC designers cascade low-gain, wide-bandwidth stages rather than using a single high-gain stage.

Follow-up: What is the gain-bandwidth product of an op-amp and why is it constant?

Q5. What is loop gain and how does it determine stability?

Loop gain (Aβ) is the product of the open-loop amplifier gain and the feedback fraction, and the system is stable if the loop gain magnitude is less than 1 (0 dB) when its phase shift equals -180°. For an LM741 with AOL = 200,000 at DC and dominant-pole compensation, the loop gain crosses 0 dB (unity) at 1 MHz with approximately -135° phase (45° phase margin), ensuring stability. If stray capacitances add more phase shift to push the phase to -180° before gain crosses unity, the amplifier oscillates — which is why PCB layout must minimize parasitic capacitance at op-amp inputs.

Follow-up: What happens to the output if the phase margin is exactly zero degrees?

Q6. What is Bode plot and how do you use it to assess stability?

A Bode plot shows loop gain magnitude (dB) and phase (degrees) versus log frequency; stability is assessed by finding gain crossover frequency (where |Aβ| = 0 dB) and reading the phase at that frequency — phase margin = 180° + phase at crossover. For an LM741 with gain crossover at 1 MHz and phase of -135°, phase margin = 180° - 135° = 45°, giving a stable but somewhat resonant response. A closed-loop overshoot below 20% requires phase margin > 45°, which is the standard design target in control loop compensation for switching power supplies and motor drives.

Follow-up: How do you draw the approximate Bode plot from the open-loop gain specification?

Q7. What is gain margin and how does it complement phase margin?

Gain margin is the additional gain in dB that can be added before the system goes unstable, measured as the negative of the loop gain at the phase crossover frequency (where phase = -180°). The LM741 with 20 dB gain margin means the open-loop gain can increase 20 dB (10×) before oscillation — a comfortable margin for component variation. Both gain margin and phase margin are needed: a system can have adequate phase margin but poor gain margin if the gain rolls off too slowly near the -180° phase point, causing instability at high gains.

Follow-up: What is the relationship between gain margin and the distance from the Nyquist plot to the (-1, 0) point?

Q8. How does the emitter degeneration resistor in a CE amplifier implement negative feedback?

The emitter resistor RE provides series-series negative feedback by sampling the emitter current (≈ collector current) and converting it to a voltage that opposes the input base-emitter voltage, reducing the effective transconductance from gm to gm/(1 + gm×RE). For a BC547 with gm = 38.5 mA/V at 1 mA and RE = 470 Ω, the effective gm_eff = 38.5/(1 + 38.5×0.47) = 2 mA/V, and gain = -RC × gm_eff = -4.7kΩ × 2mA/V = -9.4, stable against transistor variation. Removing RE (with bypass capacitor) removes this stabilizing feedback and gain jumps to -181, now sensitive to temperature and beta variations.

Follow-up: What does series-series feedback do to input and output resistance?

Q9. What is conditional stability in a feedback amplifier?

Conditional stability occurs when a feedback amplifier is stable at its nominal loop gain but would become unstable if the gain were reduced — the Bode plot shows the phase passing through -180° twice, with the gain above unity at both crossings. Some high-order control systems exhibit this behavior, where reducing supply voltage (which reduces open-loop gain) paradoxically causes oscillation. This is why amplifiers must be tested for stability across all operating conditions, not just nominal gain — a conditionally stable design will fail during startup when gain builds up from zero to its final value.

Follow-up: How is the Nyquist stability criterion more powerful than Bode analysis for conditionally stable systems?

Q10. What is dominant-pole compensation and why is it used in op-amps?

Dominant-pole compensation adds a large capacitor internally (30 pF in the LM741) to create a very low-frequency pole (~10 Hz) that makes the open-loop gain roll off at -20 dB/decade all the way to unity-gain frequency, ensuring phase margin > 45° at unity gain for any resistive feedback network. Without this compensation, the LM741's three high-frequency poles would make the phase approach -270° before gain reaches unity, causing oscillation in virtually any feedback configuration. The trade-off is severely reduced bandwidth — the LM741's compensated GBW is 1 MHz versus an uncompensated potential of hundreds of MHz.

Follow-up: What is the advantage of external compensation versus internal dominant-pole compensation?

Q11. How does positive feedback differ from negative feedback in amplifier design?

Positive feedback adds the feedback signal to the input (in phase), increasing the effective loop gain and causing the circuit to either latch to one state (comparator with hysteresis) or oscillate (oscillator) rather than amplify stably. A Schmitt trigger uses positive feedback: once output goes high, the feedback pulls the input threshold higher, creating a bistable latch until the input clearly exceeds the new threshold. This is why any parasitic positive feedback path (like capacitive coupling from output to non-inverting input) causes oscillation in an op-amp circuit, even when designed with negative feedback through Rf.

Follow-up: What determines the frequency of oscillation in a positive feedback oscillator?

Q12. What is the Barkhausen criterion for oscillation?

The Barkhausen criterion states that a circuit will sustain oscillation if the loop gain magnitude equals exactly 1 (0 dB) and the total loop phase shift equals 0° (or 360°) at the oscillation frequency. A Wien bridge oscillator using a TL071 op-amp with RC network sets f = 1/(2πRC); with R = 16 kΩ and C = 10 nF, f ≈ 1 kHz. In practice, gain slightly above 1 is used (via lamp or AGC circuit) to start oscillation, with the amplitude-limiting mechanism bringing loop gain back to exactly 1 at steady-state.

Follow-up: Why must the loop gain be slightly greater than 1 to start an oscillator?

Q13. What is Miller's theorem and how does it apply to feedback amplifier analysis?

Miller's theorem states that an impedance Z connecting input and output of an amplifier with gain Av can be replaced by Z/(1+Av) at the input and Z×Av/(Av-1) at the output, eliminating the cross-coupling and simplifying circuit analysis. For a CE amplifier with 4 pF Cμ (Ccb) and Av = -100, Miller capacitance at input = 4pF × 101 = 404 pF — this large capacitor limits bandwidth and is the key result of Miller analysis. Miller's theorem is used both to analyze bandwidth degradation and to motivate design solutions like the cascode topology that prevents Miller multiplication.

Follow-up: How does the cascode configuration circumvent the Miller effect at the CE transistor?

Q14. What is the effect of negative feedback on output distortion?

Negative feedback reduces the harmonic distortion of an amplifier by the desensitivity factor (1 + Aβ): if an open-loop CE stage has 10% THD and is placed in a feedback loop with Aβ = 100, the closed-loop THD is reduced to ~0.1%. The LM1875 audio power IC uses deep negative feedback (loop gain > 60 dB at 1 kHz) to reduce the inherent nonlinearity of bipolar output transistors from percent-level to its specified THD < 0.015%. However, feedback linearization has limits — it reduces distortion but does not eliminate it, and clipping distortion when the amplifier hits its output limits is not reduced by feedback.

Follow-up: Why does clipping distortion not improve with more feedback?

Q15. What is the return difference in feedback amplifier theory?

Return difference F = 1 + Aβ (also called the desensitivity factor) is the factor by which all beneficial effects of negative feedback are multiplied: gain is reduced by F, gain sensitivity to transistor variation is reduced by F, distortion is reduced by F, bandwidth is extended by F, and output impedance is reduced by F for shunt output sampling. For the LM741 at 1 kHz with Aβ = 1000 (60 dB loop gain), F = 1001, meaning gain stability is 1000× better than open-loop, bandwidth 1000× wider, and distortion 1000× lower. Understanding return difference as a single number summarizing all feedback benefits is what separates a circuit designer from someone who just uses formulas.

Follow-up: How is return difference measured experimentally in a real amplifier circuit?