GATE Instrumentation (IN): Control Systems (steady-state error, Routh, root locus, Bode and Nyquist, state space, PID and compensators), previous-year questions solved, part 3

Part 3 of 3 of the Control Systems hub: 8 more official GATE IN questions on state-space models and time-delay systems and controllers, compensators and actuators, each solved step by step and checked against the official answer key. The other parts cover block diagrams, signal flow graphs and time response and stability: Routh, root locus, Bode and Nyquist.

Questions from official GATE IN papers (IIT Kharagpur 2022, IIT Kanpur 2023, IISc Bengaluru 2024, IIT Roorkee 2025, IIT Guwahati 2026); solutions by Skill-Lync. · Updated 19 Nov 2026 · 8 of 23 questions on this page

GATE 2022 · IN · Q.30NAT · 1 markState-space models and time-delay systems

Transfer function (s+1)(s+3)/((s+5)(s+7)(s+9)): minimum number of state variables

The transfer function of a system is:

In the state-space representation of the system, the minimum number of state variables (in integer) necessary is ________

Answer: state variables Official key: 3 to 3

  1. The minimum number of states in a state-space model is the order of the transfer function after any common factors are cancelled. This is the order of the minimal realisation: one state per energy-storing element, or one per pole.
  2. Denominator: has degree 3, with poles at , and . Numerator zeros are at and .
  3. No zero equals a pole, so nothing cancels (common factors: 0). The numerator degree 2 is below the denominator degree 3, so the transfer function is strictly proper.
  4. Minimum number of state variables: 3.

Trap: counting the 2 zeros as extra states, giving 5, or subtracting the zeros from the poles, giving 1. Zeros shape the output matrix and do not add states. We built the controllable canonical realisation with 3 states: its controllability and observability matrices both have rank 3 and 3 (full rank, so no state can be removed), and it reproduces the transfer function at test points.

Where this shows up in industry

The state count sets the size of the model a state-feedback controller or an observer has to carry, for example in a Kalman filter on a sensor. Order-reduction tools look for cancelled or very weak modes to shrink it, which is exactly what this minimal-realisation test looks for.

GATE 2023 · IN · Q.43MCQ · 2 marksState-space models and time-delay systems

Bode plot with a 34 dB peak at 100 rad/s and phase falling to -720 degrees: matching transfer function

The magnitude and phase plots shown in the figure match with the transfer-function ________.
Bode diagram of the transfer function in Q43. The magnitude plot starts at 0 dB at 10 rad/s, rises to a sharp peak of about 33 dB at 100 rad/s and falls to about minus 9 dB at 200 rad/s. The phase plot falls steadily from about minus 30 degrees at 10 rad/s through minus 360 degrees near 100 rad/s to minus 720 degrees at about 188 rad/s, with a steeper step near 100 rad/s−100102030400−180−360−540−72020406080100120140160180200Bode DiagramMagnitude (dB)Phase (deg)Frequency (rad/s)
  1. (A)
  2. (B)
  3. (C)
  4. (D)

Answer: (B) Official key: B

  1. Read the magnitude plot first. It starts at 0 dB (a DC gain of 1) and has a sharp peak of about 33 dB at . A second-order section has its peak at with height .
  2. gives and , so and a peak of , about . That fits options A, B and C. Option D has , so its peak would be at 10 rad/s, not 100 rad/s: eliminated.
  3. Now the phase plot. A second-order section alone can only add a phase lag that levels off at . The plot goes on falling in a straight line to at about . A phase lag that keeps growing linearly with frequency is the signature of a pure time delay: has magnitude 1 at every frequency and phase .
  4. At the second-order part contributes about , so the delay supplies about . Then , matching . Option (B).
  5. Option C has a delay of . At that is a phase of degrees, invisible on the plot. Option A has no delay, so its phase would stop at about .

Trap: choosing A because the magnitude plot looks like a lightly damped second-order system. The magnitude plot cannot tell a delay from no delay, since the delay has gain 1; only the phase can. We evaluated each option from 10 to 188 rad/s with an unwrapped phase: peak and phase at 188 rad/s come to 34.0 dB and -718 degrees for B, against -179 degrees for A and C and a peak at 10 rad/s for D. Only B matches the plot.

Where this shows up in industry

Dead time from fluid transport, long sampling or communication delays shows up on a measured Bode plot as phase that keeps falling at high frequency while the magnitude does not change. Identifying the delay from the phase slope is a standard step before tuning a loop, because the delay limits how fast a controller can safely be made.

GATE 2024 · IN · Q.17MCQ · 1 markState-space models and time-delay systems

State-space model: which system matrix A gives a pole at the origin

Consider the state-space representation of a system where is the state vector, is the input, is the system matrix and is the input matrix. Choose the matrix from the following options such that the system has a pole at the origin.
  1. (A)
  2. (B)
  3. (C)
  4. (D)

Answer: (B) Official key: B

  1. The poles of a state-space system are the eigenvalues of , the roots of .
  2. A pole at means , that is . A singular always has a zero eigenvalue.
  3. (A): . Not zero. (C): . Not zero. (D): . Not zero.
  4. (B): . A zero determinant, so a pole at the origin. Its characteristic equation is (trace 4, determinant 0), giving poles at and . Answer: (B).

Trap: choosing (A) because it has a zero entry. The entry is not the pole: (A) has poles at and and is stable. We computed the eigenvalues of all four matrices numerically: only (B) has an eigenvalue at 0 (the other eigenvalue of (B) is , so the system also has an unstable pole).

Where this shows up in industry

A pole at the origin is an integrator: a tank level, a motor position or any process that accumulates its input. Seeing it in the state matrix tells a control engineer the loop will have zero steady-state error to a step with plain feedback, and that open-loop it will drift.

GATE 2026 · IN · Q.44MCQ · 2 marksState-space models and time-delay systems

Poles of an LTI system in state-space form with A = [0 1; -10 -7], b = [1; 1], c = [1 1]

A linear time invariant (LTI) system is described by:The poles of the system are ____.
  1. (A)
  2. (B)
  3. (C)
  4. (D)

Answer: (A) Official key: A

  1. For , , the transfer function is and its poles are the roots of , the eigenvalues of .
  2. , so .
  3. gives and (check: sum , product ).
  4. No pole is cancelled by a zero. The transfer function is , whose zero at is nowhere near either pole. The pair is controllable and observable, so both eigenvalues are poles of the input-output model. Poles: , option (A).

Trap: taking the characteristic polynomial as , a sign slip, which gives the roots 2 and 5 of option (B). Both poles lie in the left half-plane, so the system is stable. We computed the eigenvalues from the quadratic formula, and confirmed that the controllability and observability matrices both have non-zero determinants: only passes.

Where this shows up in industry

State-space models are what model-based tools such as Kalman filters and multivariable controllers use. The first check on any identified plant model is the eigenvalues of the A matrix: a pole in the right half-plane tells you the plant is unstable before you design anything.

GATE 2022 · IN · Q.53NAT · 2 marksControllers, compensators and actuators

Plant 0.001/(s(2s+1)(0.01s+1)) with lag compensator (s+10)/(s+0.1): asymptotic Bode slope at 3 rad/s

In a unity-gain feedback control system, the plant

is controlled by a lag compensator

The slope (in dB/decade) of the asymptotic Bode magnitude plot of the loop gain at rad/s is ________ (in integer)

Answer: Official key: -60 to -60

  1. On an asymptotic Bode magnitude plot the slope changes by dB/decade at each pole corner frequency and by dB/decade at each zero corner. A pole at the origin gives a slope of dB/decade from the start.
  2. The loop gain is . List the corner frequencies: the integrator at the origin (), poles at rad/s (from ), rad/s (from ) and rad/s (from ), and a zero at rad/s (from ).
  3. At rad/s, count what is already below: the integrator, the poles at 0.1 and 0.5 rad/s. The zero at 10 and the pole at 100 are still above.
  4. Slope dB/decade.

Trap: the lag compensator's zero at 10 rad/s is above 3 rad/s, so it has not yet bent the plot back up. Stopping at the plant's own corner at 0.5 rad/s gives dB/decade, which forgets the compensator's pole at 0.1 rad/s. We counted the corners below 3 rad/s in code and got -60 dB/decade. The exact (smooth) magnitude slope at 3 rad/s is about -58 dB/decade, a little above the asymptote because the nearest corners (0.5 and 10 rad/s) are still close; the question asks for the asymptotic value.

Where this shows up in industry

Reading the slope of a Bode plot at the crossover region is the quickest check on loop shape: a slope near -20 dB/decade through crossover means a healthy phase margin, and -60 dB/decade means trouble. A lag compensator is used to raise low-frequency gain, and its pole and zero must sit well below the crossover so as not to cost phase there.

GATE 2024 · IN · Q.12MCQ · 1 markControllers, compensators and actuators

Control valve with decreasing sensitivity at constant pressure drop: flow against stem position

The pressure drop across a control valve is constant. The control valve with inherent characteristic has decreasing sensitivity. If represents the fraction of maximum stem position of the control valve, then the function representing the fraction of maximum flow is
  1. (A), where is constant
  2. (B)
  3. (C)
  4. (D)

Answer: (B) Official key: B

  1. The inherent characteristic of a valve is flow against stem position at constant pressure drop. Its sensitivity is the slope : how much extra flow one more per cent of stem travel gives.
  2. Decreasing sensitivity means the slope is largest at small openings and falls as the valve opens. That is the quick-opening characteristic.
  3. Test the options. For the slope is , which falls from very large at to at . Decreasing sensitivity, and , .
  4. (linear) has a constant slope of 1. has slope , which rises. is the equal-percentage valve: its slope grows with , so each step changes flow by a fixed percentage of the flow already passing.
  5. Answer: (B) .

Trap: picking the equal-percentage form because it is the one with a named constant . Its sensitivity increases. We tested all four options for , and a falling slope (with = 0.5, 2, 10 and 50 for option A): only passes. At the valve already passes 50% of full flow.

Where this shows up in industry

Quick-opening valves suit on-off duties and relief lines where full flow is wanted fast. Equal-percentage valves are the usual choice for temperature and pressure loops because the loop gain stays roughly constant as the installed pressure drop changes.

GATE 2025 · IN · Q.47MCQ · 2 marksControllers, compensators and actuators

Plant 1/s^2 with a lead compensator: phase lead needed for poles at -1.5 +/- j sqrt(27)/2

The figure shows a closed-loop system with a plant and a lead compensator . The compensator is designed to place the dominant closed-loop poles at . From the following options, choose the phase lead that the compensator needs to contribute.
A unity negative feedback loop. The reference enters a summing junction (plus on the input, minus on the feedback), then goes through the compensator C of s, then the plant G of s, and the plant output is the loop output and is fed back to the junction+−C(s)G(s)CompensatorPlant
  1. (A)
  2. (B)
  3. (C)
  4. (D)

Answer: (B) Official key: B

  1. Root locus angle condition: a point is a closed-loop pole of a unity feedback loop only if the loop transfer function there has phase (an odd multiple of it): .
  2. The pole is . Its angle from the positive real axis is .
  3. The plant has phase .
  4. The compensator must supply of phase lead. Option (B).

Trap: giving the angle of the pole itself, (option D), or forgetting that the plant has two integrators and so contributes , not . As a check we built a real lead compensator with a zero at , a pole placed for exactly of lead (pole at ) and a gain of from the magnitude condition: the closed-loop polynomial has roots and , so the required poles are reached.

Where this shows up in industry

Phase-lead design is how motion and position loops with a double integrator, such as a pointing platform or a ram, get damping that the plant alone cannot give. The angle condition tells the designer how much lead a compensator must add before choosing its pole and zero.

GATE 2025 · IN · Q.50MSQ · 2 marksControllers, compensators and actuators

Plant a/(s^2 - b^2), a, b > 0, unity feedback: which of P, I, PI, PD controllers cannot stabilise it

The plant in the feedback control system shown in the figure is , where and . The type(s) of controller that CANNOT stabilize the plant is/are
A unity negative feedback loop. The reference enters a summing junction (plus on the input, minus on the feedback), then goes through the controller C of s, then the plant P of s, and the plant output is fed back to the junction+−C(s)P(s)ControllerPlant
  1. (A)proportional (P) controller
  2. (B)integral (I) controller
  3. (C)proportional-integral (PI) controller
  4. (D)proportional-derivative (PD) controller

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

  1. Closed-loop poles are the roots of . A quick test: a polynomial can have all roots in the left half-plane only if all its coefficients exist and have the same sign. A missing power of is a failure.
  2. P controller, : . There is no term, so the roots are either real and opposite (, one in the right half-plane) or purely imaginary: never strictly stable. Cannot stabilise.
  3. I controller, : . The term is missing, and the coefficient is negative. Cannot stabilise.
  4. PI controller, : . The coefficient is still zero. Cannot stabilise.
  5. PD controller, : . For and every coefficient is positive and a second-order polynomial is then stable. So PD can stabilise it.
  6. The controllers that cannot stabilise are (A), (B) and (C).

Trap: assuming a large proportional gain will pull the poles into the left half-plane. For this plant it only moves them onto the imaginary axis. The derivative term is what gives the missing coefficient, which acts as damping. We checked with , by computing the closed-loop roots over a grid of gains from 0.01 to 100: the lowest value reached by the largest real part of the roots was 0.00 (P), 1.74 (I) and 0.00 (PI), never below zero, against -13.53 for PD, which is well below zero.

Where this shows up in industry

An unstable first-principles plant such as an inverted pendulum, a magnetic levitation stage or a balancing robot needs derivative action to supply damping, which is why those controllers are PD or PID, not plain P or I. The same coefficient test is used when a tuning attempt keeps oscillating forever.

From GATE prep to a instrumentation or 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.