ISRO Scientist/Engineer 'SC' Electronics 2025 question paper, solved with the final answer key

ISRO's written test for Scientist/Engineer 'SC' (Electronics), held on 26 October 2025, question by question. Every answer follows ISRO's final key of 15 May 2026, which changed three answers and excluded one more question after the December key. Every solution is worked out in full by Skill-Lync, and the answers ISRO changed are marked.

Question paper and answer keys published by ISRO's Centralised Recruitment Board; solutions by Skill-Lync · Updated 18 Nov 2026 · 20 of 80 questions on this page

How many questions were in the ISRO Scientist/Engineer Electronics 2025 paper?

95 multiple-choice questions in two parts, written on 26 Oct 2025. Part A had 80 electronics engineering questions of 1 mark each, with 0.33 deducted for a wrong answer. Part B had 15 aptitude questions worth 20 marks, with no negative marking; they were the same in every discipline and are solved on the Mechanical paper page. ISRO later excluded Part A Q44 and Q49 (and aptitude Q82) from evaluation and revised 3 answers after objections.

Which subjects carried the most marks?

Counting the 80 Part A questions by subject: analog circuits 17, signals and systems 9, networks 9. Together that is 35 of 80 technical marks. The remaining questions spread over 6 more subjects; the table below links every question.

What was the cut-off for this paper?

In January 2026 ISRO evaluated the Electronics paper out of 98 (79 + 19) under its December key: the last candidate short-listed for interview, other than PwBD, scored 49.73, and every candidate also needed 39.5 in Part A and 9.5 in Part B; 92 were short-listed. ISRO then issued a final key on 15 May 2026 that excluded one more question (out of 97) and re-issued the results on 22 May 2026 with 100 short-listed; it did not publish a new cut-off. The panel of 14 Aug 2026 selected 46 plus 1 PwBD, with no wait list.

Questions by subject, ISRO Electronics 2025 (Set A), counted by Skill-Lync
SubjectQuestionsQuestion numbers
Analog circuits173693339434652555859616367697477
Signals and systems9147172042607175
Networks921534374547657980
Electromagnetics, microwaves and antennas95811192528485154
Electronic devices and VLSI9272930386466687072
Communications81014162223265762
Digital circuits and microprocessors712244041497678
Control systems731364450535673
Engineering mathematics51318213235

ISRO 2025 · Electronics Set A · Q.1Part A · 1 mark · -0.33 if wrongSignals and systems

Poles and zeros of x(t) = delta(t) - (4/3)e^(-t)u(t) + (1/3)e^(2t)u(t)

For the following function, poles and zeros will be :
  1. (a)Poles at and , Zero at
  2. (b)Poles at and , Zero at
  3. (c)Poles at and , Zero at
  4. (d)Poles at and , Zero at

Answer: (b) Poles at and , Zero at Official final key: (b)

Take the Laplace transform term by term: , and .

The denominators give the poles: and . Now combine over . The numerator is

So . The zero is at (a double zero), the poles at and . Check from the three terms directly: .

The term is why the numerator has the same degree as the denominator. The classic slip is to assume the zero sits at or to read it off the coefficients and .

Poles at and , zero at

Where this shows up in industry

Pole-zero placement is how filter and control-loop designers read stability and response from a transfer function, from receiver filters to attitude-control compensators.

ISRO 2025 · Electronics Set A · Q.2Part A · 1 mark · -0.33 if wrongNetworks

L-network matching 50 ohm source to 200 ohm load at 10 MHz with high-pass response

A source with internal resistance is to be matched to a load using a lossless LC network at a frequency of . Which of the following configurations will provide proper impedance matching and high-pass filtering?
  1. (a)A series inductor followed by a parallel capacitor
  2. (b)A parallel inductor followed by a series capacitor
  3. (c)A series capacitor followed by a parallel inductor
  4. (d)A parallel capacitor followed by a series inductor

Answer: (c) A series capacitor followed by a parallel inductor Official final key: (c)

An L-network steps resistance up or down by a factor . The shunt element must sit across the higher resistance and the series element on the lower side. Here the source is and the load , so the order must be: source, series element, then a shunt element across the load. That rules out (b) and (d), which put the shunt element on the source side (that arrangement always looks like or more from the source).

Size it: . The shunt reactance is and the series reactance is . Check: in parallel with gives at the source side, and a series (a capacitor) cancels the imaginary part, leaving exactly . The two reactances must have opposite signs, which can be done two ways:

  • Series L plus shunt C (option a): a low-pass, because the inductor passes DC and the capacitor shorts high frequencies.
  • Series C plus shunt L (option c): a high-pass, because the series capacitor blocks DC and the shunt inductor shorts low frequencies. Values at 10 MHz: and .

(c) a series capacitor followed by a parallel inductor

Where this shows up in industry

Matching networks like this sit at the input and output of RF amplifiers and antenna feeds; the high-pass form also keeps DC and low-frequency noise out of the stage.

ISRO 2025 · Electronics Set A · Q.3Part A · 1 mark · -0.33 if wrongAnalog circuits

Miller equivalent input capacitance for a capacitor across an amplifier of gain A

In Miller's theorem, if a capacitor is connected between the input and output of an amplifier with gain , the equivalent input capacitance is :
  1. (a)
  2. (b)
  3. (c)
  4. (d)

Answer: (c) Official final key: (c) · provisional key was (b)

Let the input voltage be and the output . The current the input drives into the capacitor is

Seen from the input, that is a capacitor of value . This is Miller's theorem for the input side; the output side sees .

For an inverting amplifier is negative (for example ), so : the capacitor looks much bigger. The textbook form is the same result with the sign of the gain taken out, which is why in option (b) is the trap: it is only right if is defined as a positive magnitude.

ISRO's provisional key (Oct 2025) and its December 2025 key both gave (b); the final key of 15 May 2026 changed it to (c).

Where this shows up in industry

Miller multiplication of the base-collector or gate-drain capacitance is the main bandwidth limiter in common-emitter and common-source stages, and cascode stages are used to avoid it.

ISRO 2025 · Electronics Set A · Q.4Part A · 1 mark · -0.33 if wrongSignals and systems

Which of x1(t), x2(n), x3(t) is periodic

Which of the three signals given below is periodic? , ,
  1. (a)
  2. (b)
  3. (c)
  4. (d)None of the above

Answer: (d) None of the above Official final key: (d) · provisional key was (c)

: a complex exponential is periodic, but multiplying by makes it zero for and non-zero after, so no shift can reproduce it. Not periodic.

: equals 1 for every and 2 at , so . The single spike never repeats. Not periodic.

: the sum places at and at for every integer . Shifting by 4 gives the same sequence, so it repeats with period : one cycle is . No shift smaller than 4 reproduces it, since the at is not matched at , 2 or 3.

So, as printed, the technically correct answer is (c). The official final key says (d), none of the above. The paper does not show ISRO's reasoning. One possible reading is that is named (a continuous-time name) but written with the discrete index , which makes it ill-defined; if so, none of the three would count as periodic. We keep (d) as the official answer.

Official key: (d). Our own working gives (c): has period 4.

ISRO's provisional key (Oct 2025) and its December 2025 key both gave (c); the final key of 15 May 2026 changed it to (d).

Note on the key As printed, x3 is a sum of unit samples at n = 4k and negative unit samples at n = 4k+1, so it repeats every 4 samples and is periodic (answer c); the key (d, none) may treat the mismatch of x3(t) with the discrete index n as making x3 invalid.

Where this shows up in industry

Spotting periodicity decides whether a signal has a line spectrum (Fourier series) or a continuous one, which sets how you filter and sample it in any DSP chain.

ISRO 2025 · Electronics Set A · Q.5Part A · 1 mark · -0.33 if wrongElectromagnetics, microwaves and antennas

EIRP of a quarter-wave monopole radiating 20 W with 3 dBi gain

A monopole antenna radiates with gain of . What is its Effective Isotropic Radiated Power?
  1. (a)
  2. (b)
  3. (c)
  4. (d)

Answer: (c) Official final key: (c)

EIRP is the power an ideal isotropic antenna would need to give the same peak field in the main direction: radiated power times gain over isotropic, .

Convert the gain: (3 dBi is almost exactly a factor of 2). Then

Adding 3 dB to 20 W is the same as doubling it. Option (a) wrongly halves the power, and (d) treats 3 dBi as a factor of 3 ().

Where this shows up in industry

EIRP is the regulatory and link-budget figure for every transmitter, from satellite beacons to ground-station uplinks.

ISRO 2025 · Electronics Set A · Q.6Part A · 1 mark · -0.33 if wrongAnalog circuits

Key advantage of Class AB amplifiers over Class B

What is a key advantage of Class AB amplifiers over Class B?
  1. (a)Lower bias current
  2. (b)Higher crossover distortion
  3. (c)No need for heat sinks
  4. (d)Reduced crossover distortion

Answer: (d) Reduced crossover distortion Official final key: (d)

A Class B push-pull stage biases both transistors exactly at cut-off. Each conducts for one half cycle, but near zero volts neither conducts until the input passes about 0.6 V, which leaves a flat spot at the zero crossing: crossover distortion.

Class AB adds a small bias, just enough to keep both transistors slightly on near zero, so one takes over smoothly from the other. This removes most of the crossover distortion.

Why not the others? (a) is wrong: Class AB has a higher quiescent current than Class B, which is the price paid. (b) is the opposite of the truth. (c) is wrong: output power devices still need heat sinks.

Reduced crossover distortion

Where this shows up in industry

Class AB output stages are the usual choice in audio amplifiers and linear regulators where low distortion matters and the small idle current is acceptable.

ISRO 2025 · Electronics Set A · Q.7Part A · 1 mark · -0.33 if wrongSignals and systems

Is x(t) = sin(100 pi t)/(pi t) time limited or band limited

Signal is
  1. (a)Time limited and band limited
  2. (b)Time limited but not band limited
  3. (c)Not time limited but band limited
  4. (d)Neither time limited nor band limited

Answer: (c) Not time limited but band limited Official final key: (c)

This is a sinc pulse. The Fourier pair is , equal to 1 for and 0 outside. Here , so the spectrum is flat up to and exactly zero beyond. The signal is band limited.

A signal cannot be both time limited and band limited. This spectrum has a sharp edge, so in time the signal keeps ringing with decay for ever: it is non-zero between its zero crossings at , for arbitrarily large . So it is not time limited.

Not time limited but band limited

Where this shows up in industry

Ideal low-pass reconstruction in sampling theory uses exactly this sinc, which is why real converters need anti-alias filters and finite approximations of it.

ISRO 2025 · Electronics Set A · Q.8Part A · 1 mark · -0.33 if wrongElectromagnetics, microwaves and antennas

Device used for low noise amplification in microwave receivers

Which device is typically used for low noise amplification in microwave receivers?
  1. (a)Klystron
  2. (b)Travelling Wave Tube Amplifier
  3. (c)Tunnel Diode
  4. (d)HEMT

Answer: (d) HEMT Official final key: (d)

A low-noise amplifier (LNA) sits at the very front of a receiver, where its noise figure sets the noise of the whole chain. The standard device for that job at microwave frequencies is the HEMT (high-electron-mobility transistor). Electrons move in an undoped channel, which gives high gain and a noise figure well under 1 dB at several GHz.

The klystron (a) and the TWTA (b) are vacuum-tube power amplifiers used in transmitters: high power, higher noise. A tunnel diode (c) can amplify using its negative resistance, but such amplifiers are narrow-band, hard to stabilise and no longer the normal choice.

HEMT

Where this shows up in industry

HEMT LNAs, often cooled, are the front end of satellite ground-station and deep-space receivers, where every tenth of a dB of noise figure is link margin.

ISRO 2025 · Electronics Set A · Q.9Part A · 1 mark · -0.33 if wrongAnalog circuits

Condition for the virtual ground assumption in an op-amp circuit

Which of the following conditions must hold true for the virtual ground assumption to be valid in an op-amp circuit?
  1. (a)The op-amp must have infinite input impedance and zero output impedance
  2. (b)The op-amp must be operated in open-loop mode
  3. (c)Negative feedback must be present and the op-amp should be operating in the linear region
  4. (d)The op-amp should have a very low gain

Answer: (c) Negative feedback must be present and the op-amp should be operating in the linear region Official final key: (c)

The virtual short says . It holds because with very large: if the output stays finite, the difference must be almost zero. That works only if the output is not stuck at a supply rail (linear region) and negative feedback pulls the inverting input toward the non-inverting one.

Open loop (b) just drives the output to a rail. Very low gain (d) would break the argument, which needs high gain. High input and low output impedance (a) are useful ideal-op-amp traits but do not by themselves create the virtual ground.

Negative feedback present and the op-amp in its linear region

Where this shows up in industry

Every summing amplifier, transimpedance stage and active filter design starts from this assumption.

ISRO 2025 · Electronics Set A · Q.10Part A · 1 mark · -0.33 if wrongCommunications

Average Huffman codeword length for source probabilities 0.4, 0.3, 0.2, 0.1

Given source symbols A(0.4), B(0.3), C(0.2), D(0.1). What is the average codeword length using Huffman coding?
  1. (a) bits
  2. (b) bits
  3. (c) bits
  4. (d) bits

Answer: (a) bits Official final key: (a)

Huffman coding repeatedly merges the two least likely symbols. Merge D (0.1) and C (0.2) into a node of 0.3. Now there are A (0.4), B (0.3) and the 0.3 node. Merge B and that node into 0.6. Finally merge A (0.4) with the 0.6 node.

The code lengths are: A 1 bit, B 2 bits, C 3 bits, D 3 bits.

Check against the entropy: bits, and Huffman guarantees .

Where this shows up in industry

Huffman-style variable-length coding sits inside image, audio and telemetry compressors, trimming the number of bits a downlink has to carry.

ISRO 2025 · Electronics Set A · Q.11Part A · 1 mark · -0.33 if wrongElectromagnetics, microwaves and antennas

Phase velocity of a lossless line with L = 0.25 uH/m and C = 100 pF/m

A lossless transmission line is 80 cm long and operates at a frequency of . The line parameters are and . Find the phase velocity :
  1. (a)
  2. (b)
  3. (c)
  4. (d)

Answer: (b) Official final key: (b)

For a lossless line the phase velocity depends only on the per-metre inductance and capacitance:

, so and

The 80 cm length and 600 MHz are not needed for the speed. They give a wavelength of , so the line is 2.4 wavelengths long.

Where this shows up in industry

Velocity factor of a cable sets its electrical length, which matters for phase-matched feed networks, delay lines and timing of high-speed interconnects.

ISRO 2025 · Electronics Set A · Q.12Part A · 1 mark · -0.33 if wrongDigital circuits and microprocessors

VHDL delta cycle: value of signal B right after simulation starts

In VHDL simulation, consider the following process :
signal A, B : std_logic := '0';
process begin
  A <= '1';
  B <= A;
  wait;
end process;
What will be the value of signal B immediately after simulation starts?
  1. (a)'1'
  2. (b)'0'
  3. (c)'U' (undefined)
  4. (d)'Z'

Answer: (b) '0' Official final key: (b)

In VHDL a signal assignment with <= does not take effect immediately. It is scheduled for the end of the current delta cycle, and every statement in the process reads the signal values as they were at the start of that delta cycle.

At initialisation the process runs once. A <= '1' schedules A to become '1'. B <= A reads A, which is still its initial '0', and schedules B to become '0'. The bare wait; then suspends the process for ever, so it never runs again to pick up the new A.

After the delta cycle A is '1' and B is '0'. B would follow A only if the process ran a second time (for example with a sensitivity list).

B = '0'

Where this shows up in industry

Delta-cycle behaviour of signals versus variables is a daily source of simulation-versus-synthesis surprises in FPGA and ASIC RTL work.

ISRO 2025 · Electronics Set A · Q.13Part A · 1 mark · -0.33 if wrongEngineering mathematics

Inverse Laplace transform of 10/(s^2 + 5s + 4)

Inverse of Laplace transform of is
  1. (a)
  2. (b)
  3. (c)
  4. (d)

Answer: (d) Official final key: (d)

Factor the denominator: . Split into partial fractions:

and . Each term inverts to an exponential:

Check: at it is 0, consistent with the initial-value theorem, since the transform falls as .

Where this shows up in industry

Partial-fraction inversion is how the step and impulse responses of second-order filters and control plants are found by hand.

ISRO 2025 · Electronics Set A · Q.14Part A · 1 mark · -0.33 if wrongCommunications

Metric minimised by a Minimum Mean Square Error (MMSE) equalizer

Which performance metric is minimized in the Minimum Mean Square Error (MMSE) equalizer?
  1. (a)Bit error rate
  2. (b)Mean squared error between transmitted and received signals
  3. (c)Inter-symbol interference
  4. (d)Noise power

Answer: (b) Mean squared error between transmitted and received signals Official final key: (b)

The MMSE equalizer chooses its tap weights to minimise , the average squared difference between the transmitted symbol and the equalizer output.

It is not minimising bit error rate directly (a), which is a consequence rather than the cost function. It does not drive ISI to zero (c): that is the zero-forcing equalizer, which can amplify noise badly. MMSE balances leftover ISI against noise enhancement, so (d) noise power alone is also not what it minimises.

Mean squared error between transmitted and received signals

Where this shows up in industry

MMSE equalization, usually adaptive, is used in modems, OFDM receivers and wireless links to undo channel distortion without blowing up the noise.

ISRO 2025 · Electronics Set A · Q.15Part A · 1 mark · -0.33 if wrongNetworks

Practical utility of Middlebrook's theorem

Which statement best describes the practical utility of Middlebrook's Theorem?
  1. (a)It replaces all needs for simulation
  2. (b)It allows simplification of component count in circuits
  3. (c)It predicts stability in cascaded systems based on impedance interaction
  4. (d)It eliminates switching losses in converters

Answer: (c) It predicts stability in cascaded systems based on impedance interaction Official final key: (c)

Middlebrook's work on cascaded systems, such as an input filter feeding a switching converter, shows that connecting two individually stable blocks can still give an unstable system. The interaction is captured by the ratio of the source output impedance to the load input impedance, (the minor-loop gain). If at all frequencies, the blocks do not interact and the cascade stays stable.

So the theorem is a stability predictor based on impedance interaction. It does not remove the need for simulation (a), reduce component count (b), or remove switching losses (d).

It predicts stability in cascaded systems based on impedance interaction

Where this shows up in industry

Power systems with an input EMI filter feeding point-of-load converters are checked against this impedance criterion, because a regulated converter looks like a negative resistance to its source.

ISRO 2025 · Electronics Set A · Q.16Part A · 1 mark · -0.33 if wrongCommunications

Total power of a 10 kW carrier AM-modulated by two carriers at 30% and 40%

A carrier is sinusoidally modulated by two carriers corresponding to amplitude modulation index of and respectively. The total radiated power is
  1. (a)
  2. (b)
  3. (c)
  4. (d)

Answer: (a) Official final key: (a)

For one tone, total power is . With two tones the sideband powers simply add, so the effective index is :

, so

Multiplying the two single-tone factors, , is the trap: the sideband powers add, they do not multiply.

Where this shows up in industry

Broadcast AM transmitter power ratings are quoted at the carrier, but the supply and cooling must be sized for the higher modulated power.

ISRO 2025 · Electronics Set A · Q.17Part A · 1 mark · -0.33 if wrongSignals and systems

Z-transform of a shifted unit impulse x[n] = delta[n - k]

The -transform of a signal , (where is a positive integer) is :
  1. (a)
  2. (b)
  3. (c)
  4. (d)

Answer: (b) Official final key: (b)

The -transform is .

The shifted impulse is 1 only at and 0 everywhere else, so the sum has a single term:

The transform of is , option (b). This is the time-shift property: a delay of samples multiplies the transform by , and itself transforms to 1.

Option (a) would be an advance, . Option (c) is the transform of the unshifted impulse. Option (d) is not a function of at all.

Where this shows up in industry

Delays written as powers of are how digital filters, FIR taps and sample-rate converters are designed and coded in DSP firmware and FPGA signal chains.

ISRO 2025 · Electronics Set A · Q.18Part A · 1 mark · -0.33 if wrongEngineering mathematics

Which probability distribution is memoryless

Which distribution is memoryless?
  1. (a)Normal
  2. (b)Log-normal
  3. (c)Exponential
  4. (d)None of the above

Answer: (c) Exponential Official final key: (c)

A distribution is memoryless if, having already waited a time without the event, the chance of waiting a further is the same as if you had just started:

For the exponential distribution , so

The identity holds exactly, so the exponential distribution is memoryless, option (c). It has a constant hazard rate .

The normal and log-normal distributions have a hazard rate that changes with age, so the conditional tail differs from the fresh one. The exponential (in continuous time) and the geometric (in discrete time) are the only memoryless distributions.

Where this shows up in industry

Constant-failure-rate models for electronic parts, Poisson arrival of packets or commands, and inter-event times in communication-link and ground-station queue models all rest on the memoryless exponential.

ISRO 2025 · Electronics Set A · Q.19Part A · 1 mark · -0.33 if wrongElectromagnetics, microwaves and antennas

Input impedance of a quarter-wave 100 ohm line loaded by a one-wavelength 50 ohm line

A one wavelength long, 50 ohm line is terminated in ohms. Its input terminals are connected to the load end of a 100 ohm, quarter-wavelength line. What is the impedance seen at the input to the 100 ohm line?
  1. (a)
  2. (b)
  3. (c)
  4. (d)None of the above

Answer: (c) Official final key: (c)

Step 1: the 50 ohm line is exactly one wavelength long. The impedance along a lossless line repeats every half wavelength, so a one-wavelength line shows its load unchanged at the input:

Step 2: this is the load of the 100 ohm quarter-wave line. A quarter-wave line inverts the load:

Multiply top and bottom by the conjugate: .

, option (c). Note the sign flip: the capacitive load becomes inductive after the quarter-wave inversion. Option (a) is the conjugate of the load and would come from a half-wave step done wrongly.

Where this shows up in industry

Quarter-wave sections are the standard impedance-matching transformer in RF feeds, antenna networks and microwave front ends of ground stations and transponders.

ISRO 2025 · Electronics Set A · Q.20Part A · 1 mark · -0.33 if wrongSignals and systems

Output of an LTI system with h[n] = delta[n] + delta[n - 1] for a unit step input

A discrete-time LTI system has impulse response What is its output for an input ?
  1. (a)
  2. (b)
  3. (c)
  4. (d)

Answer: (a) Official final key: (a)

The output of an LTI system is the convolution . Convolving with just delays the input by samples, so

Check the values: (only is on), and for . Option (a), , gives exactly this.

Option (b) would give , which is wrong, because . Option (c) is the impulse response itself, not the step response.

Where this shows up in industry

A two-tap sum like this is a moving-average (low-pass) FIR filter, used for smoothing sensor data and telemetry samples in embedded DSP.

From ISRO prep to a core electronics 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.