Module 19 · UPSC Optional: Electrical Engineering

Electrical Engineering Paper II: Core Topics

Updated 17 Nov 2026 · 4,344 words
Updated 21 Nov 2026
Electrical Engineering Paper II: core syllabus areas, key concepts, standard references and answer-writing practice.
Control Systems · Electronics · Communication Systems · Power Electronics
🎯 Take Mock Test

Learning Objectives

  • Understand the fundamental concepts of Electrical Engineering Paper II: Core Topics
  • Apply key formulas and techniques to solve problems
  • Practice with exam-level questions to build speed and accuracy

Electrical Engineering Paper II — Control Systems, Electrical Machines & Power Systems

Control Systems

  • Control System Components: Open loop vs closed loop control systems, feedback and feedforward control, servo mechanisms, synchros, synchro transmitters and receivers, synchro torque differential systems, synchro capacitor, synchro applications.
  • Block Diagrams & Signal Flow Graphs: Block diagram reduction rules, block diagram algebra, Mason's gain formula, signal flow graphs, Mason's gain formula, Mason's gain formula applications.
  • Time Domain Analysis: Standard test signals (step, step, ramp, parabolic, impulse), transient response of first order and second order systems, rise time, peak time, peak overshoot, settling time, steady state error, steady state error for step, ramp, parabolic inputs, static error constants (Kp, Kv, Kv).
  • Stability Analysis: Routh-Hurwitz criterion, necessary and sufficient conditions for stability, special cases (zero row in Routh array, zero row in auxiliary polynomial), relative stability, root locus (root locus construction rules, angle condition, magnitude condition, angle of departure, angle of arrival, centroid, asymptotes, breakaway points, intersection with imaginary axis), root locus applications.
  • Frequency Response Analysis: Bode plots (magnitude plot, phase plot), corner frequencies, phase margin, gain margin, gain crossover frequency, phase crossover frequency, gain margin, phase margin, Bode plot stability criteria, Nyquist plot (Nyquist stability criterion, Nyquist stability criterion, encirclements of -1 point, Nyquist criterion), Nyquist plot construction.
  • Compensators: Lead compensator, lag compensator, lag-lead compensator, PID controller, PID tuning (Ziegler-Nichols rules), cascade control, feedforward control, cascade control.
  • State Space Analysis: State space representation, state space models (controllable canonical form, observable canonical form, diagonal form, Jordan canonical form), state transition matrix, properties of state transition matrix, solution of state equation, controllability (Kalman's test), observability (duality principle), pole placement (pole placement by state feedback), Ackermann's formula.

Electrical Machines

  • DC Machines: DC machine construction (stator, rotor, armature, field poles, interpoles, compensating windings), EMF equation (E = PNZ/(60A) × φN/60), torque equation (T = (PZ/2πa) × φIa), DC generator characteristics (open circuit characteristic, external characteristics of shunt, series, compound generators), DC motor characteristics (torque-speed characteristics, speed-torque characteristics), speed control of DC motors (armature control, field control, Ward-Leonard system), DC motor starters (3-point starter, 4-point starter).
  • Transformers: Transformer construction (core type, shell type), ideal transformer equations, practical transformer (equivalent circuit, equivalent circuit parameters - R1, X1, R2', X2', R0, X0), equivalent circuit parameters determination (open circuit test, short circuit test), voltage regulation, efficiency, all day efficiency, parallel operation of transformers (conditions for parallel operation, load sharing), autotransformer (advantages, disadvantages, applications), Scott connection (3-phase to 2-phase conversion), Scott connection applications.
  • Induction Motors: 3-phase induction motor construction (stator, rotor - squirrel cage, wound rotor), rotating magnetic field, principle of operation, slip (s = (Ns - N)/Ns), rotor EMF, rotor current, rotor frequency, torque equation (T ∝ φI₂cosφ₂), maximum torque condition (R₂ = X₂), maximum torque at starting (R₂ = X₂), torque-slip characteristics, speed control methods (stator voltage control, rotor resistance control, frequency control, pole changing, cascade control), induction generator, induction generator operation (self-excited induction generator, grid-connected induction generator).
  • Synchronous Machines: Synchronous machine construction (salient pole, cylindrical pole), EMF equation (E₄ = 4.44fNΦK_w), synchronous impedance method (open circuit test, short circuit test), voltage regulation (EMF method, MMF method, ZPF method, Potier method), synchronous motor (starting methods, V-curves, inverted V-curves, power factor control), synchronous condenser, hunting (hunting oscillations, damping windings), synchronous condenser applications.
Technique: Thevenin's Theorem
Question: Find Thevenin equivalent across terminals A-B of a circuit with 12V source, 4Ω resistor in series with parallel combination of 4Ω and 8Ω resistors.
Answer: V_th = open circuit voltage across A-B = voltage across 4Ω resistor = (4/(4+4))×12 = 6V. R_th = 4Ω || 4Ω = 2Ω. Thevenin equivalent: 6V source in series with 2Ω resistor.

Memory Tricks

  • Kirchhoff's Laws: KCL - algebraic sum of currents at node = 0 (charge conservation); KVL - algebraic sum of voltages around loop = 0 (energy conservation).
  • Thevenin's Theorem: V_th = V_oc (open circuit voltage), R_th = V_oc/I_sc (short circuit current), R_th = V_oc/I_sc.
  • Maximum Power Transfer: Maximum power transferred when load resistance equals Thevenin resistance (R_L = R_th), maximum power P_max = V_th²/(4R_th).
  • Resonance: Series resonance - impedance minimum, current maximum, power factor = 1; Parallel resonance - impedance maximum, current minimum, power factor = 1.
Mains strategy
Electrical Engineering answers: draw circuit diagrams, show step-by-step derivation of circuit equations, state assumptions clearly, quote IS codes (IS 456 for RCC, IS 800 for steel, IS 1893 for seismic design, IS 10262 for mix design), quote standard values (electrical resistivity of copper 1.72×10⁻⁸ Ω·m, copper density 8.96 g/cm³, electrical resistivity of aluminium 2.83×10⁻⁸ Ω·m, aluminium density 2.7 g/cm³).

Power Electronics: The DC Chopper Circuit Family

The study of power electronics in Paper II shifts focus from linear control (amplifiers, analog circuits) to switching devices that convert energy at high frequencies. The core syllabus area here is the DC Chopper, which acts as a variable DC voltage source. Unlike a transformer which changes AC voltage levels, a chopper chops a fixed DC input into a varying average output using semiconductor switches (BJT, MOSFET, IGBT). The fundamental operation relies on the concept of duty cycle (δ = ton/T), where T is the switching period.

The standard reference for this domain is Bernard K. Bose, Power Electronics, which details the topologies extensively. A student must master four primary chopper configurations: Step-down (Buck), Step-up (Boost), Step-up-Step-down (Buck-Boost), and Cuk converters. Each has distinct voltage transfer ratios and polarity characteristics. For instance, in a Buck converter, the output voltage V_o is given by V_o = δ V_s, where V_s is the source voltage. Crucially, the load current must be continuous for these simple topologies to function as intended; if the load is highly resistive and frequency low, discontinuous conduction mode (DCM) occurs, altering the transfer ratio.

A critical mechanism often tested is the flyback diode (or freewheeling diode). When the switch opens, the inductive load tries to maintain current flow. The voltage across the inductor reverses polarity, forward-biasing the diode. Current then circulates through the load and diode, bypassing the switch. This prevents high-voltage spikes that would destroy the switch (arcing). In modern designs using IGBTs or MOSFETs, this diode is often an intrinsic body diode or a fast recovery Schottky diode to minimize switching losses.

The Inverter section complements this by converting DC to AC. The basic single-phase full-bridge inverter uses four switches (typically IGBTs) arranged in legs. By controlling the firing angles, one can generate square waves, quasi-square waves, or stepped waveforms. The harmonic content of these waveforms is a key analysis topic. A pure sine wave requires complex PWM (Pulse Width Modulation) control. The Stevedore's Inverter or specific industrial inverters often use phase-angle control to regulate output voltage by varying the conduction angle of thyristors, though this introduces significant harmonics.

Another advanced topic is the Cycloconverter, which converts fixed-frequency AC to variable-frequency AC directly without an intermediate DC link. This is used in large motor drives (like rolling mills). The control logic here involves overlapping firing sequences of thyristors to synthesize a lower frequency output from a higher frequency input. Efficiency drops significantly due to the lack of regenerative capability and high harmonic distortion, limiting its use to low-speed, high-torque applications.

Examiner's trap
Question: In a step-down chopper with continuous load current, if the duty cycle is 0.5 and input voltage is 100V, what is the output voltage? Answer: 50V. Trap: Students often forget that in discontinuous conduction mode (DCM), V_o depends on the load resistance and switching frequency, not just duty cycle. Another trap: Confusing the polarity of Buck-Boost (output opposite to input) with Cuk or SEPIC (output same polarity). Always check the diode orientation in the schematic.
Exam-day strategy
When solving numerical problems on choppers, always state your assumption regarding continuous vs. discontinuous conduction mode first. If the problem does not specify, assume Continuous Conduction Mode (CCM) for standard textbook answers, but mention DCM if the load is purely resistive and frequency is low. Always draw the voltage waveform across the switch and the diode to justify your calculation of average voltage.

Communication Systems: Modulation Indices and Bandwidth

The communication systems syllabus focuses on the transmission of information over a channel. The core concept is Modulation: varying a carrier parameter (amplitude, frequency, or phase) in accordance with the message signal. Amplitude Modulation (AM) varies the envelope amplitude; Frequency Modulation (FM) varies instantaneous frequency; Phase Modulation (PM) varies instantaneous phase.

A critical distinction lies in Analog vs. Digital modulation. Analog systems (AM, FM, PM) are susceptible to noise accumulation. Digital systems (ASK, FSK, PSK, QAM) offer better noise immunity but require higher bandwidth and complex coding. The Bandwidth requirement is a frequent calculation point. For AM, BW = 2f_m (twice the highest modulating frequency). For FM, Carson's rule applies: BW ≈ 2(Δ f + f_m), where Δ f is the peak frequency deviation. This explains why FM radio occupies more spectrum than AM but offers superior audio quality and noise rejection.

The Noise aspect is vital. Thermal noise (Johnson-Nyquist noise) is inherent in all resistive components. Shot noise arises from discrete charge carriers. The signal-to-noise ratio (SNR) determines the fidelity of transmission. In FM, the "capture effect" and "threshold effect" are unique phenomena where a stronger signal suppresses a weaker one on the same frequency, or where SNR degrades rapidly below a certain threshold level.

Digital Communication introduces concepts like Error Control Coding (Hamming codes, Reed-Solomon) and Synchronization. Without synchronization between transmitter and receiver, data bits cannot be aligned. Techniques like Manchester encoding ensure self-clocking in the signal. The Shannon-Hartley theorem (C = B log2(1 + S/N)) defines the theoretical maximum channel capacity. Students must understand that increasing bandwidth or SNR increases capacity, but practical constraints (filtering, non-ideal components) limit real-world performance.

The Satellite Communication section involves uplink/downlink frequencies, polarization (linear vs. circular), and path loss calculations involving free-space attenuation and atmospheric absorption. The concept of FDMA, TDMA, CDMA multiple access techniques is essential for understanding cellular networks. FDMA divides frequency spectrum; TDMA divides time slots; CDMA uses spreading codes to share the same frequency band simultaneously.

Examiner's trap
Question: If a carrier of 10 MHz is modulated by a 5 kHz sine wave, calculate bandwidth. Answer: 10 kHz (2 × 5 kHz). Trap: Students often add the carrier frequency to the bandwidth or confuse f_m with the carrier frequency. Another trap: In FM, calculating deviation ratio (β = Δ f / f_m) and forgetting that modulation index affects sideband amplitude distribution (Bessel functions), not just bandwidth.
Exam-day strategy
In communication questions, always define your variables (f_c, f_m, Δ f) clearly before substituting values. If asked about noise figure, recall that it is defined as the ratio of input SNR to output SNR (F = (SNR)in / (SNR)out). Remember that a lower noise figure is better.

State-Space Modeling and Controllability Analysis

State-space representation converts an n‑th order differential equation into a first‑order matrix form ẋ = A x + B u, y = C x + D u. The matrix A captures system dynamics, B maps inputs to states, C selects measured outputs, and D represents direct feed‑through. For a linear time‑invariant (LTI) system, controllability is determined by the rank of the controllability matrix 𝒞 = [B, A B, A² B, …, Aⁿ⁻¹ B]. If rank(𝒞) = n, every state can be driven to any desired value in finite time using an appropriate input. Observability follows the dual test with the observability matrix 𝒪 = [Cᵀ, (Aᵀ) Cᵀ, …, (Aᵀ)ⁿ⁻¹ Cᵀ]ᵀ. A system that is both controllable and observable admits a minimal realization; otherwise pole‑zero cancellations hide internal modes that cannot be influenced or measured.

Canonical forms simplify analysis. The controllable canonical form places the characteristic polynomial coefficients in the last row of A and uses a companion‑type B = [0 … 0 1]ᵀ. The observable canonical form is its transpose. The diagonal (modal) form decouples states when A is diagonalizable, revealing each mode’s eigenvalue directly. However, repeated eigenvalues with defective eigenvectors require Jordan blocks, which affect the structure of B and C and may render certain modes uncontrollable or unobservable even if the overall rank test passes.

Pole placement via state feedback u = –K x + r assigns closed‑loop eigenvalues to desired locations provided the pair (A, B) is controllable. Ackermann’s formula gives K = [0 … 0 1] 𝒞⁻¹ φ(A) where φ(s) is the desired characteristic polynomial. For multi‑input systems, the solution is not unique; parametric approaches (e.g., using the controllability subspace) allow additional design freedom such as minimizing control effort or shaping transient response. Observer design mirrors pole placement: a Luenberger observer ẋ̂ = A x̂ + B u + L (y – C x̂) places observer poles via L using the dual system (Aᵀ, Cᵀ). The separation principle guarantees that controller and observer poles can be chosen independently.

In examinations, a typical twist is to provide a transfer function with a hidden pole‑zero cancellation and ask whether the resulting state‑space model is controllable. The correct approach: construct a minimal realization (e.g., via controllable canonical form after cancelling common factors) and then test rank. Another common trap is to assume that a diagonalizable A guarantees controllability; the input matrix B must have non‑zero entries in every modal direction. Always verify the rank of 𝒞 explicitly.

Digital Control Implementation and Quantization Effects

Digital controllers operate on sampled‑data signals with a zero‑order hold (ZOH) equivalent discrete‑time model x[k+1] = A_d x[k] + B_d u[k], where A_d = eA T_s and B_d = ∫₀T_s eA τ dτ B. The sampling period T_s must satisfy the Nyquist criterion relative to the closed‑loop bandwidth; a common rule of thumb is T_s ≤ 0.1 / ω_c where ω_c is the desired crossover frequency. Faster sampling reduces discretization error but increases computational load and quantization noise.

Quantization arises from finite word‑length arithmetic in the processor. Fixed‑point representation introduces a quantization step Δ = 2–(b–1) for a b‑bit signed integer (assuming full‑scale ±1). The quantization error is often modeled as an additive white noise source with variance σ_q² = Δ² / 12. In a feedback loop, this noise is shaped by the closed‑loop transfer function; high‑gain loops amplify low‑frequency quantization noise, potentially causing limit cycles — sustained oscillations of small amplitude even with zero reference. Dithering (adding a small pseudo‑random signal) or using a higher‑precision accumulator for the integrator term mitigates limit cycles.

Coefficient quantization affects pole locations. A pole at z = r ejθ implemented with quantized coefficients may shift outside the unit circle, causing instability. Sensitivity analysis uses the derivative ∂z/∂a_i where a_i are denominator coefficients. Poles near the unit circle or with high multiplicity are most sensitive. A robust design practice is to keep closed‑loop poles well inside the unit circle (e.g., |z| ≤ 0.8) and to verify stability after coefficient rounding using the Jury test or root locus on the quantized polynomial.

Implementation structures matter. Direct Form I and II realize the difference equation directly but suffer from coefficient sensitivity. Cascade of second‑order sections (biquads) reduces sensitivity because each section handles a conjugate pole pair. Parallel form further isolates quantization effects. In exams, you may be asked to compare the signal‑to‑quantization‑noise ratio (SQNR) of a 12‑bit fixed‑point PID versus a 32‑bit floating‑point implementation; the answer hinges on the dynamic range of the integrator state and the scaling strategy (e.g., block floating‑point). Always state the assumed scaling and overflow handling.

Worked Example: Quantization‑Induced Limit Cycle
A digital PI controller with transfer function C(z) = K_p + K_i T_s / (1 – z⁻¹) is implemented in 12‑bit fixed‑point arithmetic (Δ = 2⁻¹¹). The plant is G(z) = 0.1 / (1 – 0.9 z⁻¹). The closed‑loop is stable with poles at 0.6 ± 0.2j. Estimate whether a limit cycle can appear and suggest a remedy.
The integrator state accumulates the error multiplied by K_i T_s. With 12‑bit resolution, the smallest representable increment is Δ. If the steady‑state error is less than Δ / (K_i T_s), the integrator cannot update and the output stalls, producing a limit cycle of amplitude ≈ Δ / (K_i T_s). For typical gains K_p = 2, K_i = 5, T_s = 0.001 s, the increment is 5×0.001 = 0.005. Δ = 4.88×10⁻⁴, so the minimum error to move the integrator is 0.097. Since the plant gain is 0.1, a steady‑state error of 0.1 yields an output change of 0.01, well below the quantization step. Hence a limit cycle is likely. Remedy: increase word length to 16 bits (Δ ≈ 1.5×10⁻⁵) or use a 32‑bit accumulator for the integrator while keeping 12‑bit storage for other variables. Dithering with a ±Δ/2 pseudo‑random signal also breaks the dead‑zone.
Exam‑Day Strategy for Digital Control
When a problem asks for the discrete‑time equivalent of a continuous plant, write the ZOH formulas explicitly and compute A_d, B_d using the matrix exponential or the Laplace‑domain step‑invariant method. For quantization questions, state the word length, compute Δ, and model quantization noise as additive white noise with variance Δ²/12. Always check pole locations after coefficient rounding; a quick Jury test on the rounded denominator saves marks. If a limit‑cycle analysis is required, compare the integrator step size with the quantization step and mention accumulator widening or dithering as standard fixes.

Resonant and Soft‑Switching Power Converter Topologies

Hard‑switching converters (buck, boost, buck‑boost) force semiconductor devices to turn on and off under full voltage and current, incurring switching losses proportional to frequency. Resonant converters introduce an LC tank that shapes voltage and current waveforms into sinusoids, enabling zero‑voltage switching (ZVS) or zero‑current switching (ZCS). In a series resonant converter (SRC), the resonant inductor L_r and capacitor C_r are placed in series with the load; the switch operates at a frequency f_s near the resonant frequency f_r = 1 / (2π √(L_r C_r)). When f_s > f_r, the tank current lags the voltage, providing ZVS for MOSFETs. The voltage gain is M = V_o / V_in = (1 / (1 – (f_s/f_r)²)) for a half‑bridge SRC, allowing both step‑up and step‑down operation.

Parallel resonant converters (PRC) place the tank in parallel with the load, yielding ZCS for diodes. The gain expression becomes M = (f_s/f_r)² / (1 – (f_s/f_r)²). A widely used hybrid is the LLC resonant converter, which adds a magnetizing inductance L_m in parallel with the transformer primary. The three‑element tank (L_r, C_r, L_m) provides a wide gain range with ZVS across the entire load range, making it a common choice for high‑efficiency offline adapters and server power supplies. The normalized gain curves are plotted against the frequency ratio f_n = f_s / f_r and the quality factor Q = √(L_r/C_r) / R_ac, where R_ac is the reflected load resistance.

Soft‑switching techniques also include active clamp and synchronous rectification. An active clamp forward converter recycles the leakage inductance energy into a clamp capacitor, resetting the transformer and achieving ZVS for the main switch. Synchronous rectifiers replace output diodes with MOSFETs driven by the secondary winding voltage, reducing conduction loss at low output voltages. The control challenge is to maintain ZVS/ZCS under varying load and line conditions; adaptive frequency control or phase‑shift modulation (in phase‑shifted full‑bridge) adjusts the switching instant to keep the resonant current sufficient for commutation.

Examiners often ask for the design of an LLC converter given input voltage range, output power, and desired efficiency. The steps: (1) select resonant frequency f_r (typically 100–200 kHz), (2) choose L_r and C_r to set Q for the required gain range, (3) determine L_m to set the no‑load gain, (4) verify ZVS condition ILr,peak > Imagnetizing at minimum load, (5) calculate component stresses (peak voltage on switches, RMS current in L_r). A common trap is to ignore the effect of transformer parasitic capacitance on the resonant frequency; always include the effective C_r = C_r,external + C_oss,primary + C_oss,secondary reflected.

Orthogonal Frequency Division Multiplexing and MIMO Fundamentals

Orthogonal Frequency Division Multiplexing (OFDM) splits a high‑rate data stream into N parallel low‑rate subcarriers spaced by Δf = 1 / T_u, where T_u is the useful symbol duration. Orthogonality is maintained because each subcarrier completes an integer number of cycles over T_u, making the inner product of different subcarriers zero. A cyclic prefix (CP) of length T_cp ≥ channel delay spread converts linear convolution into circular convolution, eliminating inter‑symbol interference (ISI) and inter‑carrier interference (ICI) provided the channel is time‑invariant over one OFDM symbol. The overhead factor is η = T_u / (T_u + T_cp); typical values are 1/4 or 1/8.

Peak‑to‑Average Power Ratio (PAPR) is a major drawback: the time‑domain OFDM signal is the sum of many independent subcarriers, leading to a high PAPR (≈ 10–12 dB for large N). This forces the power amplifier to operate with large back‑off, reducing efficiency. Techniques to reduce PAPR include clipping and filtering, selected mapping (SLM), partial transmit sequences (PTS), and tone reservation. In exams, you may be asked to compute the complementary cumulative distribution function (CCDF) of PAPR for a given N and compare the reduction achieved by SLM with U phase sequences.

Multiple‑Input Multiple‑Output (MIMO) exploits spatial diversity and multiplexing. For a narrowband flat‑fading channel with N_t transmit and N_r receive antennas, the channel matrix H (size N_r × N_t) has singular values σ_i. The capacity with perfect channel state information at the receiver (CSIR) and equal power allocation is C = Σi=1r log₂ (1 + (ρ / N_t) σ_i²) where r = rank(H) and ρ is the average SNR. With channel state information at the transmitter (CSIT), water‑filling across eigenmodes maximizes capacity. Spatial multiplexing transmits independent streams on the strongest eigenmodes; diversity schemes (e.g., Alamouti space‑time block code for 2×1) achieve full diversity order N_t N_r without CSIT.

In OFDM‑MIMO (e.g., LTE, Wi‑Fi 802.11n/ac), each subcarrier experiences a flat MIMO channel. The receiver performs per‑subcarrier channel estimation using pilot symbols, then applies linear equalizers (Zero‑Forcing, MMSE) or maximum‑likelihood detection. The cyclic prefix must be longer than the maximum excess delay across all transmit‑receive paths. A typical exam question: given a 4×4 MIMO‑OFDM system with 64 subcarriers, CP length 16 samples, and a channel with 8‑tap impulse response, determine whether ISI/ICI is avoided and compute the spectral efficiency if 16‑QAM is used on all spatial streams with coding rate 3/4. The answer: CP ≥ delay spread ⇒ no ISI/ICI; spectral efficiency = (N_data / N_total) × (bits per symbol) × (coding rate) × (spatial streams) = (48/64) × 4 × 0.75 × 4 = 9 bits/s/Hz.

Electromagnetic Compatibility and Filter Design in Power Electronics

Power converters generate conducted and radiated emissions due to fast voltage and current transitions (high dv/dt, di/dt). Conducted emissions on the AC mains are regulated by standards such as CISPR 32 / EN 55032 (superseding CISPR 22), which define limits for quasi‑peak and average detectors in the 150 kHz – 30 MHz band. The primary mitigation is an input EMI filter comprising a common‑mode (CM) choke, differential‑mode (DM) inductors, X‑capacitors (line‑to‑line) and Y‑capacitors (line‑to‑ground). The CM choke presents high impedance to common‑mode currents while presenting low impedance to differential currents; its inductance L_cm is chosen so that the CM corner frequency f_cm = 1 / (2π √(L_cm C_y)) lies well below the lowest regulated frequency (typically < 10 kHz).

Differential‑mode attenuation relies on the DM inductance L_dm and X‑capacitance C_x. The DM corner frequency f_dm = 1 / (2π √(L_dm C_x)) should be placed near the switching frequency to attenuate the dominant switching harmonics. However, excessive L_dm degrades transient response and can cause instability in the control loop due to added phase lag. A practical design uses a two‑stage filter: a first stage with a small L_dm1 and large C_x1 for high‑frequency attenuation, and a second stage with larger L_dm2 for low‑frequency compliance.

Radiated emissions are addressed by PCB layout: minimizing loop areas of high‑di/dt paths, using ground planes, and placing decoupling capacitors close to semiconductor pins. The near‑field magnetic field from a current loop of area A carrying a harmonic current I_h at frequency f is H ≈ (I_h A) / (2π r³) for distance r ≪ λ. Reducing A by routing the return path directly under the forward trace (microstrip) cuts radiated fields dramatically. Shielding cans over the switching node and using spread‑spectrum frequency modulation (dithering the switching clock) spread the spectral energy, lowering peak quasi‑peak readings.

Filter component non‑idealities matter. Real inductors have parasitic capacitance C_p creating a self‑resonant frequency (SRF) above which they become capacitive. Capacitors have equivalent series resistance (ESR) and equivalent series inductance (ESL); the impedance minimum occurs at the series resonance f_sr = 1 / (2π √(C ESL)). In EMI filter simulations, include these parasitics to avoid unexpected resonance peaks that can amplify emissions at certain harmonics. A common exam trap: given a CM choke with L_cm = 10 mH, C_p = 10 pF, and Y‑capacitors C_y = 4.7 nF, compute the CM resonance frequency and verify it is below 150 kHz. The resonance is f_res = 1 / (2π √(L_cm (C_y/2 + C_p))) ≈ 33 kHz, satisfying the requirement.

  • Assuming a state‑space model is controllable just because the transfer function has no pole‑zero cancellation; always test the controllability matrix rank.
  • Neglecting the effect of coefficient quantization on pole locations in digital controllers; verify stability after rounding using the Jury test.
  • Designing an LLC resonant converter without checking the ZVS condition at minimum load; the magnetizing current must be sufficient to discharge the switch output capacitance.
  • Forgetting that OFDM cyclic prefix must exceed the maximum channel delay spread across all MIMO paths; otherwise ISI/ICI reappears.
  • Using only the nominal inductance of an EMI choke and ignoring its parasitic capacitance, leading to an unexpected self‑resonance inside the regulated band.
  • These traps highlight the areas where candidates most often lose marks.

Mastering these advanced topics and their associated pitfalls will significantly improve problem‑solving accuracy in the examination.

Practice Questions

← Electrical Engineering Paper I: Core Topics ← Course Home