ISRO Scientist/Engineer Electronics 2025 paper: Part A, questions 21 to 40, solved
Questions 21 to 40 of the ISRO Scientist/Engineer 'SC' Electronics written test of 26 Oct 2025 (Set A): 20 questions, each solved by Skill-Lync and checked against ISRO's final answer key.
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
Probability that the gap between two Poisson arrivals at a bank exceeds 10 minutes
- (a)
- (b)
- (c)
- (d)
Answer: (a) Official final key: (a)
For a Poisson process of rate , the time between arrivals is exponentially distributed, .
Work in hours: and , so .
Numerically , so , option (a). The paper prints 0.188, a slightly rounded-down value; it is the only option that matches.
Trap: using 5/6 (option c or d) comes from mixing in a 5 minute gap, or from an arrival rate of 5 per hour.
Queue and traffic models for call centres, network packet arrivals and command uplink scheduling use the exponential inter-arrival time of a Poisson process.
Power received from a 2 W, 17 dBi satellite at 40000 km by a 10 square metre aperture
- (a) dBW
- (b) dBW
- (c) dBW
- (d) dBW
Answer: (c) dBW Official final key: (c)
The power density at range is the EIRP spread over a sphere, and the receiver collects it over its effective area:
In decibels: and , so EIRP is (about 100 W).
Spreading loss for the area: with gives , and divided by that is .
, so , option (c). In watts this is about .
A received power must be below the transmitted power over a long path, which rules out the two positive options (b) and (d) at once.
This is the basic link-budget step for any geostationary-orbit link, in which the spreading loss over about 36000 to 40000 km dominates and the ground antenna's effective area recovers a tiny fraction of the power.
Main distinction between Shannon's source coding theorem and channel coding theorem
- (a)Source coding is only for lossless compression, while channel coding is only for lossy compression
- (b)Source coding applies only to discrete sources, whereas channel coding applies to continuous sources
- (c)Source coding deals with compressing data, while channel coding deals with transmitting data reliably over a noisy channel
- (d)Source coding aims to increase redundancy, while channel coding aims to remove redundancy
Answer: (c) Source coding deals with compressing data, while channel coding deals with transmitting data reliably over a noisy channel Official final key: (c)
The two theorems answer two different questions.
- Source coding theorem: a source of entropy bits per symbol can be compressed to about bits per symbol and no lower. It removes redundancy.
- Channel coding theorem: over a noisy channel of capacity , data can be sent with error probability as small as you like if the rate . It adds structured redundancy so errors can be corrected.
So source coding is about compressing data and channel coding is about reliable transmission over a noisy channel, option (c).
Why the others fail: (a) channel coding is not compression, and source coding also has a lossy form; (b) both theorems exist for discrete and continuous cases; (d) swaps the roles, since source coding removes redundancy and channel coding adds it.
Satellite links use both: a source coder (image or telemetry compression) shrinks the data, then a channel code such as convolutional, Reed-Solomon or LDPC protects it against noise.
Logical expressions for difference D and borrow X of a binary half subtractor
- (a)
- (b)
- (c)
- (d)
Answer: (b) Official final key: (b)
Build the truth table for :
| (borrow) | |||
|---|---|---|---|
| 0 | 0 | 0 | 0 |
| 0 | 1 | 1 | 1 |
| 1 | 0 | 1 | 0 |
| 1 | 1 | 0 | 0 |
The difference is 1 when the inputs differ, which is an XOR: . A borrow is needed only when you take 1 from 0, that is and : .
, , option (b).
Option (a) has an XNOR for . Options (c) and (d) have , which is the borrow of , not . Option (d) also gets wrong.
Subtractors, adders and comparators built from these gates make up the arithmetic units of microcontrollers and FPGA fabric; a full subtractor chains two half subtractors.
Wave impedance for TE modes in a rectangular waveguide
- (a)
- (b)
- (c)
- (d)Constant
Answer: (b) Official final key: (b)
For a TE mode the wave impedance is the ratio of transverse electric to transverse magnetic field, . The guide's phase constant is , so
where is the intrinsic impedance of the filling medium (about 377 ohm for air). This is option (b).
Behaviour: at cutoff and approaches at high frequency; for example at it is in air. Option (c) is the TM mode impedance, , which is below . Option (d) is wrong because the impedance depends on frequency.
Waveguide-to-coax transitions, filters and horn feeds are designed using the mode impedance, which changes with frequency near cutoff and sets the matching behaviour of microwave and ground-station feed hardware.
New SNR of a 10-bit ADC (62 dB) when resolution is increased by 2 bits
- (a)62 dB
- (b)74 dB
- (c)50 dB
- (d)31 dB
Answer: (b) 74 dB Official final key: (b)
For a full-scale sine input, quantization SNR grows with word length :
Each extra bit adds 6.02 dB (it halves the quantization step, so noise power drops by a factor of 4). Two more bits add .
, so , option (b). The formula gives 61.96 dB for 10 bits, which is the quoted 62 dB.
Choosing ADC resolution for a receiver or sensor chain is a direct trade of SNR (6 dB per bit) against power, speed and cost, for example in satellite payload digitisers and automotive sensing.
Which statement is not true about SRAM based FPGAs
- (a)They are reprogrammable
- (b)They require external configuration memory
- (c)Configuration is held in SRAM
- (d)They have less power consumption compared to Flash FPGAs
Answer: (d) They have less power consumption compared to Flash FPGAs Official final key: (d)
In an SRAM FPGA the logic and routing are set by configuration bits stored in SRAM cells. That gives these facts:
- (a) True: the SRAM can be rewritten any number of times, so the device is reprogrammable, even in the field.
- (b) True: SRAM is volatile and forgets the bitstream at power-off, so the configuration is reloaded at every power-up from an external flash or PROM (or a processor).
- (c) True: by definition, the configuration is held in SRAM.
- (d) Not true: SRAM FPGAs generally draw more power than flash FPGAs. The SRAM configuration array adds static power and there is a power-up surge, whereas flash cells are non-volatile and have lower static power.
The statement that is not true is (d).
Low-power, non-volatile flash FPGAs are often preferred where power or instant-on matters, while SRAM FPGAs give the highest density and speed; radiation upsets in the configuration SRAM are a design concern for space use.
What the received power is directly proportional to in the Friis transmission equation
- (a)The square of the distance between antennas
- (b)The square of wavelength
- (c)The square of frequency
- (d)The cube of the transmitted power
Answer: (b) The square of wavelength Official final key: (b)
Friis transmission equation for antennas in free space:
Reading it off, , and (to the first power). Doubling the wavelength multiplies by 4.
So is directly proportional to the square of the wavelength, option (b).
Why not the others: (a) is inverse, . (c) is inverse too, since gives at fixed antenna gains. (d) the dependence on is linear, not cubic.
Friis is the starting point of every radio link budget, from satellite downlinks to Wi-Fi, and it explains why higher frequencies need larger or higher-gain antennas to hold the same received power.
Temperature dependence of carrier mobility in a non-degenerate n-type semiconductor
- (a) increases with due to increased phonon scattering
- (b) decreases with due to decreased lattice scattering
- (c) decreases with due to increased lattice scattering
- (d) is independent of temperature
Answer: (c) decreases with due to increased lattice scattering Official final key: (c)
Mobility is set by how often carriers are scattered. Two mechanisms matter: lattice (phonon) scattering and ionised-impurity scattering.
At ordinary operating temperatures (room temperature and above) lattice scattering dominates. As rises the atoms vibrate more, there are more phonons, and carriers are scattered more often. The lattice-limited mobility falls roughly as
So mobility decreases with because lattice scattering increases, option (c).
Option (a) has the right cause but the wrong direction. Option (b) has the right direction but a false cause: lattice scattering does not decrease with temperature. Option (d) is false. Only at low temperature does impurity scattering () make mobility rise with .
This is why device speed, on-resistance of power MOSFETs and amplifier gain drift as a chip heats up, and why thermal design matters in satellite electronics.
Which power semiconductor device is not a self-commutating device
- (a)Thyristor
- (b)MOSFET
- (c)IGBT
- (d)GTO
Answer: (a) Thyristor Official final key: (a)
A self-commutating (self-turn-off) device can be switched off by its own control terminal. A thyristor cannot.
- MOSFET: off when the gate voltage is removed.
- IGBT: off when the gate voltage is removed.
- GTO: turned off by a negative gate current pulse.
- Thyristor (SCR): once latched on, the gate loses control. It turns off only when the anode current falls below the holding current, by line commutation or an added forced-commutation circuit.
The device that is not self-commutating is the thyristor, option (a).
Converters on the AC mains use thyristors (line-commutated) for their ruggedness, while spacecraft power supplies and motor drives use MOSFETs and IGBTs switched at high frequency.
Range of K for stability of s^4 + 2s^3 + 3s^2 + 4s + K = 0 by Routh-Hurwitz
- (a)
- (b)
- (c)
- (d)
Answer: (d) Official final key: (d)
Build the Routh array. The first column must have no sign change.
| 1 | 3 | ||
| 2 | 4 | 0 | |
| 0 | |||
All first-column entries must be positive: gives , and .
, option (d). At a pair of roots sits on the imaginary axis (marginal stability).
Gain limits found this way set the safe range for loop gain in attitude-control, power-converter and PLL loops, before they start to oscillate.
Which property holds for similar matrices A and B (determinant, trace, eigenvalues)
- (a)
- (b)
- (c) have the same eigen values, same characteristic polynomial
- (d)All of the above
Answer: (d) All of the above Official final key: (d)
Similar means for some invertible . Then the characteristic polynomials are equal:
Equal characteristic polynomials mean the same eigenvalues (option c). The determinant is the product of eigenvalues, so (option a). The trace is their sum, so (option b).
All three statements are true, so the answer is (d) All of the above.
Similarity transformations (change of state-space coordinates, diagonalisation) leave eigenvalues, and hence stability and natural frequencies, unchanged, which is the basis of modal analysis and state-space control design.
Single-phase fully controlled bridge, inductive load, firing angle 30 to 120 degrees
- (a)Output voltage increases, power factor improves
- (b)Output voltage decreases, power factor improves
- (c)Output voltage decreases, power factor worsens
- (d)Output voltage increases, power factor worsens
Answer: (c) Output voltage decreases, power factor worsens Official final key: (c)
A highly inductive load keeps the load current constant, so each thyristor pair conducts for a full half cycle starting at the firing angle . The mean output voltage is
At : . At : . The output voltage decreases, and past it goes negative: the converter now works as a line-commutated inverter.
The source current is a square wave of amplitude shifted by , so its displacement factor is and the power factor is . This is at and has magnitude at . The larger the firing delay, the more the current lags the supply voltage, so the power factor worsens.
Output voltage decreases, power factor worsens, option (c). A delayed firing angle trades voltage and power factor for control.
Phase-controlled rectifiers like this drive DC motors and battery chargers and regulate DC supplies; the poor power factor at large firing angles is why they need power factor correction, and the inverting region is used for regenerative braking.
Voltage V across the 20 ohm resistor in a circuit with a 75 V source and a 6 A source
- (a)50 V
- (b)36 V
- (c)38 V
- (d)12 V
Answer: (c) 38 V Official final key: (c)
Use node voltages. The top of the 75 V source is fixed at 75 V. Let be the voltage across the 20 ohm resistor (the junction of 5 ohm and 8 ohm), which is , and the top of the 12 ohm resistor. The 6 A source takes 6 A out of the node and delivers it to the 75 V node.
KCL at the 12 ohm node: the 8 ohm resistor brings in , which must equal .
KCL at the node: .
From the first, . Multiplying the second by 40 gives , that is . Substituting, , so . Then
, option (c).
Nodal analysis with mixed voltage and current sources is the everyday method for solving bias networks and power distribution paths in electronics, and it is how circuit simulators set up their equations.
Solution of the heat equation u_t = 4 u_xx with u(0,t) = u(pi,t) = 0 and u(x,0) = sin 3x
- (a)
- (b)
- (c)
- (d)
Answer: (a) Official final key: (a)
The initial profile is already one Fourier sine mode, so try . It meets because .
Substitute in : , so and . The initial condition gives .
Option (a). Check the others quickly. (b) gives , wrong sign. (d) has but , so it does not satisfy the PDE. (c) fails both the PDE and the initial condition. The exponent is : higher modes decay faster.
The heat equation with fixed-temperature ends models thermal soak of structures and electronic heat spreaders, where the highest spatial modes die out first and only the smooth profile remains.
General form of a lead or lag compensator in the Laplace domain
- (a)
- (b)
- (c)
- (d)
Answer: (c) Official final key: (c)
A first-order lead or lag network has one zero and one pole, both on the negative real axis, and a finite non-zero DC gain :
If the zero is closer to the origin than the pole and the network gives phase lead (lead compensator). If it gives phase lag (lag compensator). One form covers both, so option (c).
- (a) Two zeros and no pole: its phase is always positive and its gain grows without limit, so it cannot be a lag network.
- (b) Has a pole at the origin (): that is a PI-type controller with an extra zero, not a lead or lag network.
- (d) Has a zero at the origin and rewrites to : it blocks DC.
Lead networks add phase margin to speed up a loop, lag networks raise low-frequency gain to cut steady-state error; attitude-control and servo loops in satellites and antenna drives are tuned with both.
Initial voltage, final voltage and time constant of a capacitor after a switch from a to b
- (a)30 V, 125 V, 25 ms
- (b)30 V, 25 V, 25 ms
- (c)−30 V, 125 V, 2.5 ms
- (d)−30 V, 25 V, 2.5 ms
Answer: (d) −30 V, 25 V, 2.5 ms Official final key: (d)
Initial voltage. With the switch at 'a' for a long time the capacitor is open, and it sits at the midpoint of the 15 k and 10 k divider. The source has its minus terminal at the top and plus at the bottom (ground side), so the top node is :
After the switch moves to 'b' the divider is disconnected from the capacitor. The capacitor now sees a 20 k resistor in series with the 5 k resistor, which is in parallel with the 5 mA current source (pointing up into the 20 k side).
Final value (capacitor open, no current in the 20 k): .
Time constant: the current source is open when switched off, so and
, option (d). Options (a) and (b) lose the sign of the initial voltage, and they also use 25 ms, which would need a 250 k resistor. Option (c) has the right initial voltage and time constant but the wrong final voltage, 125 V.
Switched RC networks with a Thevenin equivalent are how engineers size settling times of sample-and-hold stages, power-supply soft starts and debounce circuits.
Effect of increasing channel length modulation parameter on a common source NMOS amplifier
- (a)Output resistance increases
- (b)Gain increases
- (c)Output resistance decreases
- (d)There is no change in output characteristics
Answer: (c) Output resistance decreases Official final key: (c)
Channel length modulation makes the drain current rise slightly with in saturation: . The transistor's own output resistance is
A larger means a smaller . For example with , going from to drops from to .
The gain also gets smaller, so (b) is wrong. The output resistance decreases, option (c).
Short-channel transistors in modern CMOS have large , which is why analog designers use longer channels or cascodes when they need high gain or good current-source behaviour.
Fundamental amplitude of a bipolar PWM inverter when modulation index goes from 0.8 to 1.2
- (a)It increases linearly with Ma even beyond 1
- (b)It increases with Ma even beyond 1 and then saturates
- (c)It saturates and remains constant beyond
- (d)It decreases beyond
Answer: (b) It increases with Ma even beyond 1 and then saturates Official final key: (b)
In the linear range the fundamental amplitude is , so Ma from 0 to 1 gives a straight line.
Beyond the sine reference exceeds the triangular carrier peak for part of each cycle, so the pulses disappear there and the output is stuck at (overmodulation). The waveform tends to a square wave, whose fundamental is .
For an ideally clipped sine reference (high carrier ratio), the fundamental in units of is: at , at , at , at and at .
So from 0.8 to 1.2 it is still rising, but no longer linearly (1.2 would be ), and it flattens out at for large Ma. It increases beyond 1 and then saturates, option (b). Option (c) puts the saturation at , which is where the linear range ends, not where the amplitude stops rising.
Overmodulation is used deliberately in motor drives and grid inverters to squeeze out extra output voltage from a fixed DC bus, at the price of low-order harmonics.
Minimum number of states of a Mealy detector for two or more 0s then two or more 1s
- (a)2
- (b)3
- (c)4
- (d)5
Answer: (c) 4 Official final key: (c)
The output goes to 1 on every '1' from the second '1' of a run onward, provided that run was preceded by at least two 0s. The machine only has to remember:
- S: no useful history (start, or inside a run of 1s that was not preceded by 00).
- Z1: one 0 seen.
- Z2: two or more 0s seen.
- O: inside a run of 1s that followed 00 (whether it is the first '1' or a later one).
Transitions (input/output): S: 0 goes to Z1/0, 1 stays in S/0. Z1: 0 goes to Z2/0, 1 goes to S/0. Z2: 0 stays in Z2/0, 1 goes to O/0. O: 0 goes to Z1/0, 1 stays in O/1.
A natural first draft has five states, separating the first '1' after 00 from the later ones. Those two states have identical next states and outputs (on 1: go to O and output 1 from the second one, on 0: go to Z1), so they merge. No other pair merges: S and Z1 differ after another 0, S and O differ on a '1' (output 0 against 1), and Z2 is the only state where a '1' enters O.
The four states are pairwise distinguishable: input 1 separates the start state from the others, and 11 separates the remaining ones, so none can be merged. 4 states, option (c).
Sequence detectors like this appear in framing and sync-word detection in serial links, command parsers and protocol decoders; minimising states cuts flip-flops and logic in the FPGA or ASIC.