GATE Electronics (EC): Control Systems (block diagrams, time response, Routh and root locus, Bode, Nyquist, compensators, state space), previous-year questions solved, part 3

Part 3 of 3 of the Control Systems hub: 9 more official GATE EC questions on frequency response: Bode, Nyquist and compensators and state-space models, each solved step by step and checked against the official answer key. Part 1 covers block diagrams, signal flow graphs and transfer functions, time response and steady-state error and stability: Routh-Hurwitz and root locus.

Questions from official GATE EC papers (IIT Kharagpur 2022, IIT Kanpur 2023, IISc Bengaluru 2024, IIT Roorkee 2025, IIT Guwahati 2026); solutions by Skill-Lync. · Updated 17 Nov 2026 · 9 of 25 questions on this page

GATE 2022 · EC · Q.21MCQ · 1 markFrequency response: Bode, Nyquist and compensators

Unity feedback with K = 2: closed-loop poles in the right half plane from the Nyquist plot of G(s)

Consider a closed-loop control system with unity negative feedback and in the forward path, where the gain . The complete Nyquist plot of the transfer function is shown in the figure. Note that the Nyquist contour has been chosen to have the clockwise sense. Assume has no poles on the closed right-half of the complex plane. The number of poles of the closed-loop transfer function in the closed right-half of the complex plane is ___________.
Nyquist plot of G of s in the complex plane. A big circle of radius 0.8 about the origin is traversed clockwise. Two small loops sit on the negative real axis: a clockwise loop from minus 0.8 to minus 0.4 and a counter-clockwise loop from minus 0.4 to 0. The point minus 1 is marked on the real axisRe G(s)Im G(s)−10.40.8
  1. (A)0
  2. (B)1
  3. (C)2
  4. (D)3

Answer: (C) Official key: C

  1. The closed-loop poles are the roots of , that is . So the critical point for the plot of is , not .
  2. Nyquist criterion: , where is the number of clockwise encirclements of the critical point, is the number of open-loop poles in the right half plane and is the number of closed-loop poles there. Here .
  3. Read the plot. It has a big circle of radius drawn clockwise, a small loop between and drawn clockwise, and a second small loop between and drawn counter-clockwise.
  4. The point lies inside the big circle () and inside the left loop (between and ), but not inside the right loop. That is one clockwise turn from the circle and one from the left loop, so .
  5. , option (C).

Trap: option (A), , comes from using the point and ignoring . The curve never goes around (the leftmost point is ), so a reader who forgets that moves the critical point to finds no encirclements and calls the system stable. Option (B), , comes from counting only the big circle, or only the loop. The code rebuilt the plotted curve as a closed polyline with the arrow directions read off the figure and computed its winding number numerically. Around it gets 2.00 clockwise turns, made of 1.00 from the circle, 1.00 from the left loop and 0.00 from the right loop. Around it gets 0.00 clockwise turns, which is the option (A) answer.

Where this shows up in industry

The Nyquist criterion is how a loop with a measured frequency response, such as a motor drive current loop or a power-supply voltage loop tested with a network analyser, is judged stable without knowing its transfer function. Checking where the plot sits against the critical point for the actual loop gain is the same step as checking gain margin on the bench.

GATE 2023 · EC · Q.25MCQ · 1 markFrequency response: Bode, Nyquist and compensators

Phase cross-over frequency of G(s) = k / (s(1+sT1)(1+sT2)) in a unity negative feedback system

The open loop transfer function of a unity negative feedback system is , where , and are positive constants. The phase cross-over frequency, in rad/s, is
  1. (A)
  2. (B)
  3. (C)
  4. (D)

Answer: (A) Official key: A

  1. The phase cross-over frequency is the frequency at which the phase of equals .
  2. , so .
  3. Set this to : .
  4. Two angles add to when their tangents are reciprocals: , that is .
  5. So and rad/s, option (A).
  6. Units check: and are in seconds, so a frequency in rad/s must scale as . Only (A) does: has units of .

Trap: option (B), , comes from solving and dropping the square root. The code swept the phase over to rad/s on points for four pairs, found the first crossing each time, and compared it with all four options: only (A) matched every pair. For s and s, option (A) gives rad/s with phase , while (B) gives rad/s with phase , (C) rad/s with and (D) rad/s with .

Where this shows up in industry

The phase cross-over frequency is where a loop would oscillate if its gain were 1 there, so the gain margin of a motor-drive speed loop or a DC-DC converter voltage loop is read at exactly this frequency. A type-1 loop with two lag poles, like this one, is the standard model of a position servo with a motor and a current-loop lag.

GATE 2023 · EC · Q.57NAT · 2 marksFrequency response: Bode, Nyquist and compensators

Asymptotic Bode magnitude plot with slopes -20, 0 and -40 dB/decade: value of a+b+c in k(s+z)^a / (s^b (s+p)^c)

The asymptotic magnitude Bode plot of a minimum phase system is shown in the figure. The transfer function of the system is , where , , , , and are positive constants. The value of is __________ (rounded off to the nearest integer).
Asymptotic magnitude Bode plot of a minimum phase system, magnitude in dB against log omega. The plot falls at minus 20 dB per decade up to omega 1, is flat from omega 1 to omega 2, and then falls at minus 40 dB per decade, crossing the horizontal axis after omega 2ω1ω2log ω−20 dB/decade−40 dB/decade|G(jω)| in dB

Answer: Official key: 4 to 4

  1. Each factor changes the slope of the magnitude plot: a pole at the origin ( in the denominator) gives dB/decade at all frequencies. A zero adds dB/decade after its corner , and a pole adds dB/decade after its corner .
  2. At low frequency the slope is dB/decade. Only the origin pole acts there, so and .
  3. At the slope changes from to , a rise of dB/decade, so there is one zero at : .
  4. At the slope changes from to , a drop of dB/decade, so and there is a repeated pole at : .
  5. and .

Trap: reading only the end slope and the first slope, one writes and and forgets the zero that makes the plot flat between and , giving and a sum of . The code built in dB for and and measured the slope in each region. Over all from to only gives slopes , and dB/decade (measured , and ). Dropping the zero () gives , and , which has no flat part, so it fails the plot.

Where this shows up in industry

Reading a transfer function off a measured Bode magnitude plot is how engineers identify a plant from a frequency sweep: a network analyser injects a small signal into a DC-DC converter or motor drive loop, and the corner frequencies and slopes give the pole and zero locations used to design the compensator.

GATE 2024 · EC · Q.12MCQ · 1 markFrequency response: Bode, Nyquist and compensators

Bode magnitude plot: 40 dB/decade is the same as how many dB/octave?

In the context of Bode magnitude plots, is the same as _______.
  1. (A)12 dB/octave
  2. (B)6 dB/octave
  3. (C)20 dB/octave
  4. (D)10 dB/octave

Answer: (A) Official key: A

  1. A decade is a frequency ratio of . An octave is a frequency ratio of . So one decade holds octaves.
  2. A slope of dB over one decade is therefore spread over octaves: .
  3. Same result with logarithms: over one octave the magnitude changes by .
  4. The nearest option is , option (A). (The exact value is , which is why the option reads .)

Trap: option (B), , is what a single pole or zero gives, and it equals (), not . The code took and and measured the magnitude change across one octave and across one decade: gives 6.02 dB per octave and 20 dB per decade, and gives 12.04 dB per octave. Only the value 12.04 matches an option, and it matches (A) alone.

Where this shows up in industry

Loop-shaping a motor-drive or converter control loop is done with Bode plots, and engineers read the roll-off in dB/decade near the crossover frequency. Converting to dB/octave is how audio and filter designers quote the same steepness, for example a 12 dB/octave second-order crossover filter.

GATE 2025 · EC · Q.16MCQ · 1 markFrequency response: Bode, Nyquist and compensators

Nyquist plot with points P, Q, R, S: which are the gain and phase crossover frequencies?

The Nyquist plot of a system is given in the figure below. Let , , , and be the positive frequencies at the points , , , and , respectively.
Which one of the following statements is TRUE?
Nyquist plot of a system. The branch for positive frequencies rises from omega equals 0 plus at the bottom, passes S which lies on the unit circle, then R, then crosses the negative real axis at Q close to the origin, and continues to the top where omega equals 0 minus. A large arc at infinity runs clockwise from omega 0 minus to omega 0 plus. A small loop between Q and the origin goes to omega equal to plus and minus infinity at the origin. P lies on a dashed line at 45 degrees above the negative real axis and R on a dashed line at 45 degrees below itReal axisImaginary axis−11j−jPQRSω = +∞ω = −∞ω = 0−ω = 0+45°45°
  1. (A) is the gain crossover frequency and is the phase crossover frequency
  2. (B) is the gain crossover frequency and is the phase crossover frequency
  3. (C) is the gain crossover frequency and is the phase crossover frequency
  4. (D) is the gain crossover frequency and is the phase crossover frequency

Answer: (D) Official key: D

  1. Definitions on a Nyquist plot of : the gain crossover is where the plot cuts the unit circle, . The phase crossover is where the plot cuts the negative real axis, .
  2. The dash-dot circle in the figure is the unit circle (it passes through , , and ). Of the four labelled points on the plot, only lies on it. So is the gain crossover frequency.
  3. is where the positive-frequency branch crosses the real axis, to the left of the origin and well inside the unit circle. It is on the negative real axis, so and is the phase crossover frequency. The gain there is below , so the system has a gain margin.
  4. and sit on the two dashed lines, not on the real axis and not on the unit circle, so neither is a crossover point.
  5. Gain crossover , phase crossover : option (D).

Trap: option (C) swaps the two names ( gain crossover, phase crossover). The code read the four labelled points off the page render (unit circle radius px) and found distances from the origin , , , . Only is at distance , and only has zero imaginary part. The code then tested all four options against these two rules and only (D) passed. Angles of and are and , not . The coordinates are read from a drawing, so the check is on positions, not on a transfer function.

Where this shows up in industry

Gain and phase crossover are what you read off a loop-gain plot when you check phase margin and gain margin of a voltage loop in a DC-DC converter or a motor-drive speed loop. A plot that crosses the negative real axis well inside the unit circle, as here, means there is room to raise the loop gain before the loop goes unstable.

GATE 2023 · EC · Q.43MCQ · 2 marksState-space models

Solution of x'(t) = Ax(t) for a second-order system with distinct eigenvalues and eigenvectors

The state equation of a second order system is , is the initial condition.
Suppose and are two distinct eigenvalues of and and are the corresponding eigenvectors. For constants and , the solution, , of the state equation is
  1. (A)
  2. (B)
  3. (C)
  4. (D)

Answer: (A) Official key: A

  1. The zero-input solution is . Distinct eigenvalues give two independent eigenvectors, so any initial state can be written .
  2. For an eigenvector, , so and .
  3. By linearity, , option (A).
  4. Direct check: differentiating one term gives . With the derivative brings down , which is not .

Trap: options (B), (C) and (D) look similar and only differ in a factor in the exponent, so the risk is to pick one by feel. The exponent must carry exactly because the derivative has to give again. The code took (eigenvalues and ), integrated with RK4 from and compared with the eigen-expansion at s: largest difference . It then plugged random into each option and measured the residual : option (A) gave , options (B), (C), (D) gave , and , so they are not solutions.

Where this shows up in industry

Writing a state response as a sum of modes is how engineers read stability and speed straight from the eigenvalues of : a battery-pack or motor model is stable when every has a negative real part, and the slowest mode sets the settling time. The same modal view is used to design the gain of a state observer so that its error modes die out faster than the plant.

GATE 2024 · EC · Q.53MSQ · 2 marksState-space models

Which state-space models have the same transfer function as A=[[0,-2],[1,-3]], B=[1,0], C=[2,-5]?

Consider a system represented in state space as . Which of the state space representations given below has/have the same transfer function as that of ?
  1. (A)
  2. (B)
  3. (C)
  4. (D)

Answer: (A) and (C) Official key: A;C

  1. For a state model , the transfer function is .
  2. System : , with determinant . Its inverse is .
  3. . Multiply by : .
  4. (A): is in companion form, so and gives . Same.
  5. (B): the same with gives , and gives . Not the same.
  6. (C): diagonal , so the transfer function is a sum of two first-order terms: . Same.
  7. (D): . Not the same.
  8. The representations with the same transfer function are (A) and (C).

Trap: option (D) and option (B) share the poles with (the eigenvalues of every here are and , which the code confirmed with np.linalg.eigvals), and the denominator alone is not enough, because the numerator depends on and . The code computed all five transfer functions symbolically with sympy: gives , (A) gives , (B) gives , (C) gives and (D) gives . It also compared them numerically at six random complex values of (the set that matches is exactly ).

Where this shows up in industry

Many different state-space models (different choices of states) describe the same input-output behaviour, which is why a controller or battery-state estimator can be put into a canonical form or a diagonal form without changing the response. Control engineers convert between them, using companion form for design tools and diagonal form to read the modes directly.

GATE 2025 · EC · Q.53MSQ · 2 marksState-space models

d/dt of three states with A = diag(2, -2, 0) and B = (1, 1, 1): which boundedness statements are true?

Consider a system where , , and are three internal state signals and is the input signal. The differential equations governing the system are given byWhich of the following statements is/are TRUE?
  1. (A)The signals , , and are bounded for all bounded inputs
  2. (B)There exists a bounded input such that at least one of the signals , , and is unbounded
  3. (C)There exists a bounded input such that the signals , , and are unbounded
  4. (D)The signals , , and are unbounded for all bounded inputs

Answer: (B) Official key: B

  1. is diagonal, so the three states are decoupled: , , . The eigenvalues of are , and .
  2. has a pole at (unstable). Even a constant input gives , which grows without limit.
  3. has a pole at (an integrator). With , , also unbounded.
  4. has a pole at (stable). For any bounded input , , so stays bounded for every bounded input.
  5. Statement (A) is false ( is a bounded input with unbounded). Statement (B) is true. Statement (C) is false because is bounded for every bounded input, so the three signals can never all be unbounded. Statement (D) is false because with zero initial state keeps all states at zero.
  6. Only (B) is true: option (B).

Trap: option (C), 'all three unbounded', is tempting after seeing and blow up. The code integrated the state equation with RK4 (step , ) for , , , and . The largest over all of them was , never above the bound , while reached and reached for (the exact value of is , matched by the code). The eigenvalues from np.linalg.eigvals are 2, 0, -2.

Where this shows up in industry

This is the bounded-input, bounded-state test that decides whether a state-space model of a plant is safe to drive open loop. A positive eigenvalue, or a pure integrator such as a motor position or a battery charge counter, means a small constant bias in the input accumulates until the state saturates, which is why such plants are always put inside a feedback loop or given anti-windup limits.

GATE 2026 · EC · Q.41MCQ · 2 marksState-space models

State model A = [-4 -1.5; 4 0], B = [2; 0], C = [1.5 0.625]: Y(s)/U(s)?

The state and output equations for a control system are: , . Which of the following expressions correctly represents the transfer function of the system with zero initial conditions?
  1. (A)
  2. (B)
  3. (C)
  4. (D)

Answer: (C) Official key: C

  1. With zero initial conditions, (here ).
  2. . Determinant: .
  3. Adjugate: . So .
  4. Multiply by : .
  5. , option (C). The poles are the eigenvalues of , so the system is stable.

Trap: option (B) has the right numerator but the denominator , which comes from losing the sign of the cross term, instead of ; sympy confirms the true determinant is and the slip gives . Option (D) keeps from the first state only and drops contributed by the second state; the code confirmed that using only the first row of gives the numerator exactly. The code evaluated numerically at complex values of (np.linalg.solve) and compared with all four options: only (C) matches. The eigenvalues of from np.roots([1, 4, 6]) are .

Where this shows up in industry

Converting a state-space model to a transfer function is how a plant model from a simulation tool is checked against a measured frequency response, and the denominator tells you the pole locations at a glance. Battery state estimators and motor-drive observers are written in state-space form for exactly this reason, and the characteristic polynomial sets the dynamics of the estimate.

From GATE prep to a controls career

Talk to a counsellor.

Talk to a career counsellor

Free 20-minute call. We'll map the right course to where you want to work.

By submitting you agree to our Terms and Privacy Policy.

Talk to a career counsellor

Talk to a career counsellor

Free 20-minute call. We'll map the right course to where you want to work.

By submitting you agree to our Terms and Privacy Policy.