GATE EC common mistakes
GATE EC · Understand the traps and how to avoid them.
Spending the first 10–15 minutes re-deriving a formula from scratch (e.g., rebuilding the small-signal hybrid-pi model or the DFT symmetry identities) instead of recalling it, then rushing the remaining numerical questions.
GATE EC has 65 questions in 180 minutes, so roughly 2.7 minutes per question. Candidates who never built a memorized formula sheet feel they must re-derive to be safe, which burns the buffer needed for the 2-mark numericals later in the paper.
What to do instead
Maintain a one-page-per-subject recall sheet (hybrid-pi parameters, Y-parameters, Fourier/Laplace pairs, Maxwell and transmission-line equations, 8085/8051 flag rules, C operator precedence). Do 10 timed 3-minute drills per day where you must start writing the solution within 20 seconds of reading the question.
Leaving Virtual Calculator-dependent numericals (logarithms, exponentials, square roots, dB conversions, trig) until the end, then discovering the on-screen calculator is slow and making arithmetic slips under time pressure.
Many candidates practice with a physical scientific calculator or mental math and only meet the GATE virtual calculator on exam day. It lacks memory keys and history, so multi-step expressions must be re-entered, and a single mistyped digit silently corrupts the answer.
What to do instead
Practice exclusively with the official virtual calculator interface for the last 6–8 weeks. Learn to break expressions into bracketed chunks, write intermediate values on the rough sheet, and pre-compute constants (e.g., 20log10 values, e^1, ln 2) so you only verify rather than explore.
Marking a numerical answer to too many decimal places or in the wrong unit (e.g., entering 0.5 kΩ as 500, or giving 3.1416 when the accepted range is tight), and losing the mark despite correct physics.
GATE numerical answer type (NAT) questions accept a range, but candidates often carry intermediate rounding errors or forget that the answer field expects a plain number in the unit stated in the question. Unit slips are especially common in EM (Ω vs mΩ), power (W vs dBm) and time-constant questions.
What to do instead
Adopt a fixed final-step checklist: (1) re-read the unit asked, (2) convert all quantities to SI before computing, (3) keep 4 significant digits through the working and round only at the end, (4) sanity-check magnitude (is a 1 nF capacitor at 1 MHz really giving ~160 Ω?).
Attempting every 2-mark question in sequence rather than triaging, so a single hard Network Theory or Communications numerical eats 12–15 minutes and forces blind guessing on three easy 1-mark questions.
The paper is not ordered by difficulty, and candidates anchor on question numbers. Sunk-cost thinking keeps them on a problem they have already invested in, while easy marks later go unattempted.
What to do instead
Use a three-pass strategy: Pass 1 (about 60 min) take all 1-mark and clearly routine 2-mark questions; Pass 2 (about 80 min) attempt the remaining 2-mark questions; Pass 3 (about 25 min) return to flagged items and verify. Hard-stop any single question at 4 minutes and flag it.
Applying ideal-op-amp assumptions without checking negative feedback, or ignoring the finite open-loop gain when the question explicitly gives it, and also forgetting input bias current and saturation limits.
Textbook drills condition candidates to write V+ = V− immediately. When the circuit has positive feedback, a comparator configuration, or a specified finite A_OL, the virtual-short assumption is invalid, and the resulting linear answer is wrong.
What to do instead
Before writing any node equation, ask three questions: Is there a feedback path from output to the inverting terminal? Is A_OL stated as finite? Is the output within the supply rails? If A_OL is finite, use V_out = A_OL(V+ − V−) as an extra equation rather than the virtual short.
Forgetting that MOSFET body effect, channel-length modulation, and BJT Early effect change the small-signal parameters, and computing g_m and r_o from the wrong bias point (e.g., using I_D from the wrong region of operation).
Candidates memorize g_m = 2√(K I_D) and r_o = 1/(λI_D) but skip the region check. If the device is in triode, saturation formulas do not apply; if λ or V_A is given, r_o must be included in the gain expression.
What to do instead
Always begin analog problems by determining the region: compute V_GS − V_T and compare with V_DS. State the region explicitly on the rough sheet, then pick the matching current equation. Include r_o in parallel with the load whenever λ or V_A is nonzero and the gain is asked.
Treating sequential and counter circuits as combinational: computing the next state without tracking the present state, or misreading active-low resets, preset inputs, and edge-triggered versus level-triggered behavior.
Under time pressure candidates jump to a truth table for the next state and forget the state register delays the output by one clock. Asynchronous clear/preset pins drawn with bubbles are also frequently read as active-high.
What to do instead
For every sequential problem, draw an explicit state-transition table with columns for present state, inputs, flip-flop inputs (J/K or D/T), and next state. Mark clock edges with an arrow and label any bubble as active-low before writing equations.
Using the wrong Fourier or Laplace property (missing the time-shift phase factor, the scaling factor 1/|a|, or the differentiation-in-frequency −j d/dω), and confusing the DTFT 2π periodicity with the CTFT.
Signals and Systems has the densest set of transform identities, and candidates learn them by pattern rather than by derivation. A sign error in the shift theorem or forgetting the 1/|a| in time scaling produces answers that look plausible.
What to do instead
Build a transform-property table and, for each entry, do one 60-second derivation from the definition. Drill mixed problems where the same signal is transformed under shift, scale and modulation, and verify energy/DC value as a cross-check (e.g., X(0) must equal the integral of x(t)).
In Communications, plugging numbers into the Shannon capacity or BER formula without converting SNR from dB, or mixing up E_b/N_0 with S/N and bit rate with symbol rate.
The formulas are simple, so candidates rush. dB-to-linear conversion, the relation C = B log2(1 + S/N), and the difference between E_b/N_0 = (S/N)(B/R_b) are the usual error points, and Q-function arguments are frequently inverted.
What to do instead
Write the dimensional chain before substituting: convert SNR_dB to linear, confirm B and R_b are in the same units, and check that the Q-argument is dimensionless. Memorize Q(0) = 0.5 and Q(−x) = 1 − Q(x), and sanity-check that higher SNR gives lower BER.
In Electromagnetics, applying the wrong boundary condition or forgetting the direction of wave propagation, so reflection coefficient sign, VSWR, and input impedance of a transmission line come out wrong.
EM questions combine geometry, coordinate conventions, and sign conventions. Candidates memorize Γ = (Z_L − Z_0)/(Z_L + Z_0) but then use it with the wrong reference plane, or forget that a short circuit gives Γ = −1 and VSWR = ∞.
What to do instead
For every transmission-line problem, sketch the line, mark the load end, and write the reference plane explicitly before computing Γ. Verify limiting cases: Z_L = Z_0 gives Γ = 0 and VSWR = 1; open/short give |Γ| = 1. For boundary conditions, state which field component is tangential or normal before applying them.