GATE Electrical: Control Systems (block diagrams, time response, stability), previous-year questions solved
Real Control Systems questions from the last five GATE Electrical papers, part 1: block-diagram reduction, damping ratio and natural frequency, settling and rise time, dominant-pole approximation, steady state with a PI controller, Routh stability ranges and root-locus rules. Each one has a step-by-step solution that matches the official answer key, the traps that cost marks, and where the idea shows up in real controller work. Bode, Nyquist, compensators and state space are on the second Control Systems page (linked on this page).
Questions from official GATE EE papers (IIT Kharagpur 2022, IIT Kanpur 2023, IISc Bengaluru 2024, IIT Roorkee 2025, IIT Guwahati 2026); solutions by Skill-Lync. · Updated 16 Nov 2026 · 12 questions
Two feed-forward gains and a positive feedback loop round 1/s: what is Y(s)/R(s)?
- (A)
- (B)
- (C)
- (D)
Answer: (B) Official key: B
- Name the integrator output . The first junction adds and the fed-back (both plus signs), so .
- The second junction adds and : .
- Substitute: , so , i.e. .
- : option (B).
Trap: the feedback sign is plus, which is why the pole lands at (an unstable system). Assuming the usual minus sign gives , option (D). We checked the algebra by solving the two junction equations at a dozen random complex values of : only (B) matches every time.
Feed-forward paths like the 2 and 3 gains here are how real controllers react to a known input before the error builds up: motor drives feed the speed reference forward to the current loop, and EV traction controllers feed forward the torque demand. Simulink models are checked against hand-derived transfer functions exactly like this.
Outer negative loop, inner junction with a minus on the forward signal and a plus on feedback: what is C(s)/R(s)?
- (A)
- (B)
- (C)
- (D)
Answer: (D) Official key: D
- First junction: .
- Second junction: the forward signal enters with a minus sign and is fed back with a plus sign, so .
- , so .
- : option (D).
The signs decide everything here. Both loops feed back the same , so they add up to , and the minus on the forward path flips the overall sign. Checked by substituting random complex values for : only (D) satisfies both junction equations.
Sign errors at summing junctions are one of the most common bugs in real control code: a positive-feedback loop where negative was meant makes a drive or a temperature loop run away instead of settling. Model reviews in automotive and aerospace (MIL testing in Simulink) check each junction sign against the hand-derived closed-loop transfer function.
Two 10/s blocks with an inner and an outer unity loop: what are ζ and ωn?
- (A) and rad/s
- (B) and rad/s
- (C) and rad/s
- (D) and rad/s
Answer: (A) Official key: A
- Inner loop: .
- Forward path of the outer loop: .
- Closed loop with unity feedback: .
- Compare with : rad/s, and . Option (A).
The inner loop is rate (velocity) feedback: it turns the first integrator into a first-order lag and adds the damping. Without it the system would be , an undamped oscillator with .
This is the structure of a cascaded position controller: an inner velocity loop (tachometer or encoder speed feedback) and an outer position loop. Servo drives in CNC machines and robot joints are tuned exactly this way, inner loop first, because the inner loop sets the damping the outer loop sees.
System 1: ωn = 3, θ = 60°; System 2: ωn = 1, θ = 70°. Which one settles slower?
- (A)Settling time of System 1 is more than that of System 2.
- (B)Settling time of System 2 is more than that of System 1.
- (C)Settling times of both the systems are the same.
- (D)Settling time cannot be computed from the given information.
Answer: (B) Official key: B
- The angle is measured from the negative real axis, so and the real part of the poles is .
- Settling time depends only on that real part: (2% band).
- System 1: , s. System 2: , s.
- System 2 settles more slowly: option (B).
Exact 2% settling times from the step responses (computed on a fine time grid) are s and s, so the approximation gives the same ordering. Rule of thumb: poles closer to the imaginary axis settle more slowly, whatever their angle.
Control engineers place closed-loop poles inside a region of the s-plane (a vertical line for settling time and a wedge for damping) before tuning anything; MATLAB's sisotool draws exactly these boundaries on the root-locus plot. Drive and power-converter specs are written as settling time and overshoot for this reason.
1/(s+1) in cascade with (s+40)/(s+20): what is the first-order approximation with the same steady state?
- (A)
- (B)
- (C)
- (D)
Answer: (B) Official key: B
- The pole at decays 20 times faster than the one at , so the slow pole dominates: the approximation is .
- Match the DC gain (the step response's final value): original , approximation , so .
- : option (B).
Check with partial fractions: the exact step response is , against for the approximation. The largest gap over s is , under 3% of the final value. Option (A) has the right pole but the wrong final value; (C) and (D) keep the fast pole and throw away the slow one.
Dominant-pole reduction is how high-order plant models get turned into the first- or second-order models used for controller tuning (for example the FOPDT model behind Ziegler-Nichols and SIMC PID rules in process plants). Keeping the DC gain right matters more than the fast dynamics, or the loop gain is off.
Step-response samples 0.78, 1.65, 2.18 at 0.6, 1.6, 2.6 s, final value 3: what is the time constant?
| Time (sec) | 0.6 | 1.6 | 2.6 | 10 | |
|---|---|---|---|---|---|
| Output | 0.78 | 1.65 | 2.18 | 2.98 | 3 |
- (A)1
- (B)2
- (C)3
- (D)4
Answer: (B) Official key: B
- A first-order step response is with , so .
- : s. : s. : s.
- All three agree: s, option (B).
Shortcut: at the output reaches 63.2% of its final value, i.e. . The output passes 1.90 between 1.6 s and 2.6 s, which already points to 2 s. The 10 s point ( s from the formula) is too close to the final value to be useful: small reading errors there blow up.
Reading a time constant off a step test is the first thing done when commissioning a temperature, level or speed loop: bump the input, log the response, fit the first-order model, then tune the PI controller from it. The same 63.2% rule is used to characterise thermocouples and RC filters.
Step response Y(s) = 100/(s(s+100)): what are the rise time, settling time and time constant?
- (A)
- (B)
- (C)
- (D)
Answer: (A) Official key: A
- , so : a first-order response with s.
- Rise time: at and at , so s.
- Settling time (98%): s.
- , , : option (A).
The memory aids and come straight from these logs. Options (B) to (D) are the same pattern scaled by 10, 100 and 1000: only the time constant decides, and .
Datasheets for sensors, current probes and power-supply control loops quote 10-90% rise time and settling time to a band, and for a single-pole system both follow from one number, the time constant. Oscilloscopes measure rise time with exactly this 10% to 90% definition.
PI controller (Kp = 3, Ki = 1) on the unstable plant 1/(s−1): what are the final controller output and plant output?
- (A)
- (B)1, 0
- (C)1, −1
- (D)−1, 1
Answer: (D) Official key: D
- Controller . Characteristic equation: . Both poles at : the loop is stable, so the final value theorem applies.
- Plant output: , which is 1 at . With the integrator, the output follows the step exactly: .
- Controller output: , which is at .
- Check: at steady state , so gives . Option (D): −1, 1.
The trap is assuming the controller output must be positive to push the output up. The plant's DC gain is negative (its pole is in the right half plane), so holding takes . Option (A) is what you get if you forget the feedback stabilises the plant.
Plants that are unstable on their own are stabilised by feedback every day: magnetic levitation, boost converters in some modes, inverted-pendulum balancing robots and fighter-jet pitch control. The steady-state actuator effort, worked out exactly like this, decides whether the actuator saturates.
Unity feedback around k/(s² + 4s − 5): for what k is the closed loop stable?
- (A)
- (B)
- (C)
- (D)
Answer: (C) Official key: C
- Characteristic equation: .
- A second-order polynomial has both roots in the left half plane exactly when all its coefficients are positive (the Routh table's first column is ).
- : , option (C).
The open-loop plant has a pole at (), so it needs a minimum gain to be stabilised. We checked by computing the closed-loop poles for from to in steps of 0.1: stable exactly for .
Routh's test is still the fastest way to find the gain limits of a loop on paper, before simulating. In practice engineers keep the operating gain well inside the stable range (gain margin of 6 dB or more) because plant gains drift with temperature, load and ageing.
Loop gain K(s−1)/((s+1)(s+2)) with unity feedback: which statement about stability is right?
- (A)stable for all .
- (B)unstable for all .
- (C)unstable for all .
- (D)stable for all .
Answer: (B) Official key: B
- Characteristic equation: .
- Stable when both coefficients are positive: and .
- For every the constant term is negative, so one root is in the right half plane: unstable for all , option (B).
Why not (C): at the polynomial is , which is stable, so 'unstable for all ' is false. Checked by computing the roots for from 0 to 20 in steps of 0.05. The right-half-plane zero at is what limits the gain: as grows, one closed-loop pole is pulled towards it.
Right-half-plane zeros appear in boost converters, aircraft pitch dynamics and some steering systems; they cap the achievable bandwidth no matter how the controller is tuned. That is why boost-converter voltage loops are run much slower than buck-converter loops.
G(s) = Ks(s+2)/((s+5)(s+7)): which of −1, −4, −6, −10 lie on the root locus?
- (A)−1
- (B)−4
- (C)−6
- (D)−10
Answer: (A) and (C) Official key: A;C
- Rule: a real-axis point is on the locus () if the number of real open-loop poles and zeros to its right is odd. Zeros at ; poles at .
- Segments: : one to the right (zero at 0), on the locus. : two, off. : three, on. Left of : four, off.
- So and are on it; and are not. Answer: (A) and (C).
Check with the magnitude condition: the gain that puts a closed-loop pole at is . It is at and () at , both positive; at and it is negative (, ), which would need .
Root-locus sketching is how engineers see at a glance what a gain change will do to a loop's poles. It is still the main view in MATLAB's Control System Designer (sisotool) when tuning lead compensators for motor drives and positioning systems.
G(s) = K(s² − 2s + 2)/(s² + 2s + 5): what is the angle of departure from the pole −1 + j2?
Answer: 7° Official key: 4 to 8
- Poles: . Zeros: roots of , i.e. .
- Angle condition at the pole : .
- From the other pole : the vector is , angle . From the zero : vector , angle . From the zero : vector , angle .
- , i.e. between 0° and 360° (key 4 to 8).
Numerical check: for the closed-loop root next to moves off at . Forgetting to wrap into the 0°-360° range is the usual way to lose this question.
Departure angles tell you whether lightly damped poles (flexible modes in robot arms, drive shafts, aircraft structures) move towards more or less damping as gain rises. Notch filters and lead networks are placed using exactly this angle condition.