GATE Instrumentation (IN): Control Systems (steady-state error, Routh, root locus, Bode and Nyquist, state space, PID and compensators), previous-year questions solved, part 2
Part 2 of 3 of the Control Systems hub: 10 more official GATE IN questions on stability: Routh, root locus, Bode and Nyquist, each solved step by step and checked against the official answer key. The other parts cover block diagrams, signal flow graphs and time response, state-space models and time-delay systems and controllers, compensators and actuators.
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 · 10 of 23 questions on this page
Unity negative feedback with loop gain 6/(s(s-5)): is the closed loop causal and stable
The closed-loop system is ________
- (A)Causal and stable
- (B)Causal and unstable
- (C)Non-causal and stable
- (D)Non-causal and unstable
Answer: (B) Official key: B
- With unity negative feedback the closed-loop transfer function is , and its poles are the roots of the characteristic equation .
- gives , that is , so .
- Causal: a rational transfer function is causal when the degree of the numerator is not greater than the degree of the denominator. Here it is 0 against 2, so the system is causal.
- Stable: both poles, and , are in the right half plane, so the system is unstable. The Routh first column of is , with 2 sign changes, which also means 2 right-half-plane roots.
- Causal and unstable: option (B).
Trap: expecting feedback to cure the unstable open-loop pole at . A gain of 6 is not enough: the term is negative, so no positive gain can make every coefficient positive. We computed the roots of the characteristic polynomial and built the Routh array in code, and tested all four options: only (causal, unstable) matches.
An unstable plant, such as an inverted-pendulum robot or a magnetically levitated rotor, can only be held by feedback of the right strength and sign. The first check on any proposed loop is to compute its closed-loop poles and look for any in the right half plane.
Nyquist plot of a stable open-loop system with points P, Q, R given: phase margin and gain margin
- (A)PM and GM
- (B)PM and GM
- (C)PM and GM
- (D)PM and GM
Answer: (A) Official key: A
- On a Nyquist plot the phase margin is read where the locus crosses the unit circle (, the gain crossover), and the gain margin where it crosses the negative real axis (phase , the phase crossover).
- Point has , so it is the gain crossover. Its angle is , in the third quadrant.
- Phase margin . It is positive because sits below the negative real axis, so the phase has not yet reached at gain crossover.
- Point is on the negative real axis, so it is the phase crossover, where . Gain margin (about 13.6 dB). The gain can rise by 4.76 times before the locus reaches the point .
- Both margins are positive: PM and GM . Option (A).
Trap: reading the gain margin as (the DC gain gives , option B). is only the starting point at . The relevant point is where the locus meets the negative real axis. We computed and from the given coordinates and , and checked all four options: only agrees.
Phase and gain margin tell a commissioning engineer how much extra delay or gain a loop can take before it oscillates. A stable plant with a margin of 40 degrees and a factor of 4.8 in gain is a healthy loop; margins below about 30 degrees and 2 are a warning that valve wear or a sensor lag could trip it.
Number of zeros of P(s) = s^3 + 2s^2 + 5s + 80 in the right-half plane
Answer: Official key: 2 to 2
- The Routh array counts roots in the right half plane without solving the polynomial: the number of sign changes in the first column equals the number of roots with positive real part.
- Row : coefficients and . Row : and .
- Row : . Row : .
- First column: . The sign changes from to between and , and from to between and : 2 changes, so 2 zeros in the right half plane (and one in the left half plane).
Trap: stopping at the first negative entry and answering 1. Count every sign change down the column. We built the Routh array in code (first column 1, 2, -35, 80) and cross-checked with the roots: and , so the complex pair is in the right half plane.
The Routh test is the quick stability screen for a loop's characteristic polynomial before any simulation: with a gain that is too high, the sign pattern in the first column flips and tells you the loop will oscillate. Instrument engineers use it when tuning a loop gain on a slow process such as a heater or a level controller.
Times the Nyquist plot of G(s)H(s) = (1/2)(s-1)(s-2)/((s+1)(s+2)) encircles the origin
Answer: Official key: 2 to 2
- The Nyquist plot is the image of as runs along the imaginary axis from to and closes through a large semicircle in the right half plane. By the argument principle, the number of clockwise encirclements of the origin equals , the number of zeros minus the number of poles of inside the contour.
- Zeros at and : both in the right half plane, so . Poles at and : both in the left half plane, so . The origin is encircled times clockwise.
- Check on the plot: with , the numerator and the denominator are complex conjugates of each other up to a sign that cancels, so at every frequency. The plot is a circle of radius centred on the origin, starting and ending at and traversed twice.
- Number of encirclements of the origin: 2.
Trap: counting encirclements of , which is the point the Nyquist stability test uses. This circle has radius 0.5 and never reaches , so the count about is 0. We evaluated for from to rad/s (modulus between 0.500 and 0.500) and accumulated its phase: the winding number about the origin is , which is two turns clockwise.
The Nyquist plot answers whether a feedback loop stays stable using only the open-loop frequency response, which can be measured from a plant even when its model is unknown. Counting how often the plot wraps the critical point is the check applied to a measured loop before it is closed.
Negative feedback, G = 1/((s-2)(s-3)), PID C(s), root locus given, |C(j)| = 2: find Kd
Answer: Official key: 0.9 to 1.1
- Write the controller over a common denominator: . It adds a pole at and two zeros to the loop transfer function .
- Read the plot: the three crosses are the open-loop poles at . Only one circle is drawn, at , yet two zeros are needed and every branch of a root locus ends on a zero (or at infinity). Two branches end at (the one from the pole at 0 and one that comes back from the circle) while the third runs to , so the circle marks a double zero at .
- A double zero at means the numerator is , so and .
- Check this against the figure: the loop has characteristic equation . Its breakaway points are the roots of , namely , and ; the locus leaves the real axis at and re-enters at (the figure shows about 2.5 and about ), and the circle rises to (the figure shows about 4.4).
- Now use the gain: . At : , so .
- gives , with and (direct check: ).
Trap: reading the single circle as one zero. The count of zeros comes from the controller (two) and the locus branches ending on , not from the number of circles drawn. We rebuilt the locus in code by tracking the three roots of for from to , found its breakaway points and its height, matched them to the plot, and then evaluated .
A PID controller adds one pole and two zeros to the loop, and placing both zeros close together is a common way to shape the root locus so that an unstable plant becomes stable at a moderate gain. Reading the gain, the zero positions and the shape of the locus from a plot is how a tuning is checked before it is loaded into a process controller.
Phase margin of G(s) = 2(1 - s)/(1 + s)^2 in degrees
Answer: Official key: -2 to 2
- Phase margin , where is the gain crossover frequency, the frequency at which .
- Magnitude: . Setting it to 1 gives , so .
- Phase: the factor is a right-half-plane zero. It raises the magnitude like a normal zero but subtracts phase: . Each factor in the denominator subtracts as well. So .
- At : , so (we got -180.0 in code).
- Phase margin . The loop is exactly on the edge of stability.
- Check with the closed loop: gives , so the closed-loop poles sit at on the imaginary axis, a sustained oscillation at .
Trap: treating as a normal minimum-phase zero, adding instead of subtracting it. That gives at the crossover and a phase margin of , which wrongly says the loop is comfortably stable. We bisected in code, evaluated the phase there and solved the closed-loop characteristic polynomial: and roots .
Process loops with transport delay or inverse response (a boiler drum level that first moves the wrong way is the classic case) behave like right-half-plane zeros, which add phase lag without reducing the gain. A phase margin close to zero on such a loop means the controller gain must be cut back, whatever the magnitude plot suggests.
Unity feedback with open-loop poles at 0, -1 and -4: imaginary-axis crossing s = j alpha of the root locus
Note: The poles are marked with in the figure.
Answer: Official key: 2 to 2
- With poles at , , and no zeros, the open-loop transfer function is with . The closed-loop characteristic equation is , that is .
- The locus crosses the imaginary axis where the system turns marginally stable. Use the Routh array: : 1, 4; : 5, ; : ; : . The row is zero at .
- At the auxiliary equation from the row is , so and . So .
- Check with the angle condition at : the angles of the three poles are , and , which sum to . Net angle of is , so the point is on the locus. The magnitude condition gives .
Trap: quoting instead of . We substituted and into : all three give zero.
The gain at the imaginary-axis crossing is the ultimate gain of the loop, and its frequency is the ultimate frequency, the two numbers behind Ziegler-Nichols PID tuning. Plant engineers find them by raising the gain until the loop just oscillates.
Transfer function G(s) = 1 + (2s - 1)/(s^3 + 5s^2 + 3s + 22): number of zeros in the left half of the s-plane
- (A)0
- (B)1
- (C)2
- (D)3
Answer: (D) Official key: D
- The zeros of are the roots of its numerator once everything is over one denominator. Do not read them off the fraction: .
- Count left half-plane roots of with a Routh array, which counts sign changes in the first column without solving the cubic. Row : . Row : .
- Row : . Row : .
- The first column is : all positive, no sign change, and no zero entry. So no root is in the right half-plane or on the imaginary axis, and all 3 zeros lie in the left half-plane. Option (D).
Trap: treating as the zero polynomial gives one zero at , in the right half-plane, so none in the left (option A). Or counting the poles instead: the denominator Routh array has entry , two sign changes, so two poles in the right half-plane and only 1 in the left. We built the Routh array in code for the combined numerator and confirmed it against the numerical roots: three left half-plane roots.
Locating zeros matters because right half-plane zeros limit how fast a feedback loop can be made without going unstable, and they show up in real plants such as boiler drum level and some flow processes. Combining a parallel path into a single fraction first is how engineers read the true zeros of a model made of several blocks.
Nyquist contour (counterclockwise) encircles two poles of G(s)H(s): statement for closed-loop stability
- (A)The locus of should encircle the origin twice in the counterclockwise direction
- (B)The locus of should encircle the origin twice in the clockwise direction
- (C)The locus of should encircle the point twice in the counterclockwise direction
- (D)The locus of should encircle the point twice in the clockwise direction
Answer: (B) Official key: B
- The closed-loop poles are the zeros of . Stability means none of them is in the right half-plane, which is the region the contour encloses.
- Argument principle: when goes once round a closed contour, the image of winds round the origin, in the same sense as the contour, a number of times equal to (zeros inside) (poles inside). The poles of are the poles of , so there are 2 inside. Write for the closed-loop poles inside: net winding .
- Stable means , so the net winding in the contour's own direction must be : two encirclements in the opposite direction. The contour here is counterclockwise, so the locus of must go round the origin twice clockwise. Option (B).
- The same statement in terms of the plain loop gain: must go twice clockwise round the point , because the origin of is the point of .
Trap: option (C) is the textbook rule (two counterclockwise encirclements of ), which belongs to the usual clockwise contour. Here the contour direction is reversed, so every encirclement direction reverses too. Options (A) and (D) use the wrong critical point: for the point is , and for it is the origin. We checked this on for , 10 and 30, which has two right half-plane poles and a stable closed loop (characteristic polynomial , roots and , for a gain of 10). Walking the counterclockwise contour gave a net winding of for about the origin in every case, which is two clockwise turns, while option (D) failed at .
Nyquist reasoning is how engineers certify that a feedback loop around an open-loop unstable process, such as an exothermic reactor or an inverted-pendulum style actuator, can be closed safely. It can be applied directly to a measured frequency response, with no model equation needed.
G(s) = K/(s+1)^4 with a gain margin of 20 dB: the value of K
Answer: Official key: 0.39 to 0.41
- The gain margin is how much the loop gain can rise before the closed loop becomes unstable. It is read at the phase crossover frequency , where the phase of is : , or in decibels .
- Each factor contributes a phase of , so gives and .
- Magnitude there: .
- A gain margin of 20 dB is a factor of , so : , which gives .
- Cross-check with the Routh criterion: the closed loop has its roots at , and these reach the imaginary axis when , that is at . The gain can then rise from to , a factor of 10, which is 20 dB.
Trap: setting instead of (the margin is a gain you can still add, so the loop gain at the crossover is below 1), which gives and an unstable loop. We found by bisection on the phase, evaluated the loop gain there, and got , giving a margin of and a critical gain of 4.
Gain margin is the safety factor that a commissioning engineer wants on a loop: 20 dB means the loop gain can rise tenfold before oscillation. Many slow thermal and level processes behave like several equal lags in series, and for these the margin is used to choose how aggressively a controller can be tuned.