Engineer Electronics common mistakes
Engineer Electronics · Understand the traps and how to avoid them.
When finding the DC operating point of a BJT, the candidate assumes V_BE = 0.7 V and immediately writes I_C = β·I_B without first checking whether the transistor is actually in forward active mode, so a circuit that is really saturated gets an active-region answer.
Textbook drills almost always use a fixed 0.7 V drop and a single β, so candidates memorize the active-region shortcut and skip the mode check. They also forget that β varies widely between devices and with temperature, so treating it as a design constant hides the saturation case.
What to do instead
After computing the assumed active-region currents, calculate V_CE. If V_CE falls below about 0.2–0.3 V, re-solve assuming saturation with V_CE(sat) fixed and V_BE ≈ 0.7 V, then recompute I_C = (V_CC − V_CE(sat))/R_C. Do the check on every bias problem before finalizing the answer.
The candidate designs or analyzes an op-amp circuit by applying the ideal gain formula without checking whether the required output voltage or current exceeds the supply rails, and reports a gain that the amplifier cannot deliver.
Ideal op-amp assumptions (infinite gain, virtual short, zero output impedance) are taught first and are easy to apply mechanically. Rail limits feel like secondary practical detail, so they are not checked until the answer is already computed.
What to do instead
After computing V_out, compare it against the positive and negative supply rails, and compare the output current against the op-amp's rated short-circuit current. If the result exceeds either limit, state that the output clips at the rail and re-evaluate the feedback loop under that condition.
The candidate minimizes a Boolean function with a Karnaugh map but leaves the map cells in the wrong order or fails to wrap groups around the map edges, producing an expression that is not minimal or is simply wrong.
Gray-code ordering of rows and columns is easy to mislabel under time pressure, and edge wrapping is a visual operation that is skipped when the candidate is rushing. Static-1 hazards from non-overlapping groups are also missed.
What to do instead
Label every row and column header explicitly before filling cells, and after grouping, re-derive the expression by reading off which variables stay constant across each group. Verify by evaluating the original truth table against the simplified expression for all input combinations.
The candidate computes the Fourier series or transform of a signal and drops the symmetry, scaling, or time-shift properties, so the magnitude spectrum is right but the phase or the DC term is wrong.
Candidates focus on the integration and treat transform properties as optional shortcuts. The 1/(2π) or 1/T scaling convention and the sign of the time-shift phase factor e^(−jωt0) are easy to lose, and DC is often forgotten when a signal has a nonzero mean.
What to do instead
Before integrating, identify the signal's symmetry (even, odd, half-wave) and use it to predict which harmonics must vanish. Then apply the shift and scaling properties explicitly and check that the DC term equals the time average of the signal. Confirm units and Parseval's relation as a final sanity check.
The candidate solves a transmission-line or plane-wave problem using the wrong characteristic impedance or wave impedance, for example using 377 Ω for a line with a different dielectric or forgetting the intrinsic impedance depends on the medium.
Free-space values (η0 = 377 Ω, c = 3×10^8 m/s) are over-learned and get substituted reflexively even when the problem specifies a dielectric with ε_r ≠ 1. The distinction between characteristic impedance of a line and intrinsic impedance of a medium is blurred.
What to do instead
Write down the medium parameters first (ε_r, μ_r) and compute η = η0/√(ε_r) and the propagation velocity v = c/√(ε_r) before touching any formula. For lines, use Z0 from the geometry (coax, microstrip) or the given value, and keep line impedance and medium impedance as separate symbols.
The candidate analyzes a controlled rectifier or chopper assuming ideal switches and continuous conduction, then reports an output voltage that is invalid because the load current actually becomes discontinuous at the given firing angle.
Continuous-conduction formulas (V_dc = (2V_m/π)cos α for a full converter) are the ones practiced, and the discontinuous-mode boundary depends on the load inductance and firing angle, which candidates rarely check.
What to do instead
Compute the critical inductance or the extinction angle for the given α and load, and compare with the actual load. If conduction is discontinuous, use the discontinuous-mode relation or solve the conduction interval explicitly. State the assumed conduction mode in the answer.
The candidate determines stability from the open-loop poles alone or reads a Bode plot's gain margin and phase margin at the wrong frequencies, concluding a system is stable when the closed-loop poles are actually in the right half-plane.
The Routh–Hurwitz test and the Nyquist criterion are learned as separate procedures, and the crossover frequencies (gain crossover for phase margin, phase crossover for gain margin) are easily swapped. Candidates also forget that a positive gain margin in dB and phase margin in degrees must both be checked.
What to do instead
Use one consistent method and cross-check with another: apply Routh–Hurwitz to the closed-loop characteristic polynomial, then verify with the Bode margins. Label the gain-crossover frequency (where |G| = 1) and phase-crossover frequency (where ∠G = −180°) explicitly before reading margins.
The candidate writes an interrupt service routine or peripheral initialization sequence that omits clearing the interrupt flag or configuring the correct edge/level trigger, so the program hangs, re-enters the ISR, or never fires.
Register-level details differ between families (8051, AVR, ARM Cortex-M) and are memorized as isolated bits. The flag-clearing requirement and the order of enabling the peripheral before the interrupt are easy to invert.
What to do instead
Follow a fixed initialization order: configure the pin or peripheral, set the trigger condition, clear any pending flag, enable the peripheral interrupt, then enable the global interrupt. In the ISR, clear the flag as the first or last action per the datasheet, and keep the ISR short by deferring work to the main loop.
The candidate applies a numerical method (Newton–Raphson, trapezoidal rule, Euler) without checking convergence or step-size error, and reports a result from a diverging or under-resolved iteration as final.
Hand calculations favor a single pass, and the convergence criterion or error estimate is treated as optional. Newton–Raphson in particular diverges near a zero derivative, which candidates do not test.
What to do instead
State the convergence criterion (residual tolerance or relative change), perform at least one refinement step, and compare results between two step sizes to estimate error. For Newton–Raphson, check that f′(x) is not near zero at the initial guess, and verify the root by substitution.
The candidate computes a Laplace or z-transform and inverts it without checking the region of convergence or the initial/final value, so the time-domain answer has the wrong sign, a missing step, or an unstable term.
Partial-fraction expansion is mechanical, and the ROC is treated as a formality. The initial-value theorem and final-value theorem are forgotten as free checks, and the distinction between causal and anticausal terms is lost.
What to do instead
After inversion, apply the initial-value theorem (lim s→∞ sF(s)) and the final-value theorem (lim s→0 sF(s), valid only if all poles are in the left half-plane) to check the endpoints. Confirm the ROC matches the stated causality, and verify the sign of each residue against a physical expectation.