PEC EPE Mechatronics common mistakes
PEC EPE Mechatronics · Understand the traps and how to avoid them.
Substitutes numbers directly into formulas such as stress = force/area, power = torque × angular velocity, or efficiency = output/input while mixing mm, MPa, rpm, W and kW, then reports a numerically plausible but dimensionally wrong answer.
Under time pressure the candidate recalls the formula but treats units as labels rather than part of the calculation; mixed unit systems in mechatronics problems make this very easy to miss.
What to do instead
Before calculating, write a quantity table with symbol, value, original unit and coherent SI conversion. Carry units through every step, convert to SI at substitution, and do a final dimensional check on the result.
Calculates AC power, current or rectifier output using peak voltage or average current as though it were RMS, and omits power factor or waveform form factor.
The candidate memorises P = VI from DC and applies it to sinusoidal, phase-controlled or nonlinear loads without identifying the waveform.
What to do instead
First identify the waveform and whether the question asks for RMS, average or instantaneous value. Use RMS definitions or known form factors, include cos φ for sinusoidal displacement power factor, and integrate or use rms values for nonsinusoidal loads.
Connects a 3.3 V microcontroller output directly to a 5 V sensor input or relay coil, or ignores source/sink current, pull-up requirements, debounce and isolation in an embedded interface.
The candidate thinks at block-diagram level and treats logic signals as ideal, forgetting electrical compatibility and protection.
What to do instead
Use an interface checklist: logic threshold compatibility, drive current versus load current, pull-up/pull-down, isolation, flyback protection, grounding and debounce. Draw the equivalent source-load circuit before accepting the design.
Selects a motor, gearbox or bearing using only nominal torque or static load, omitting duty-cycle RMS torque, reflected inertia, service factor, equivalent dynamic bearing load and thermal limit.
Textbook selection formulas are often practised with steady nominal values, while real mechatronic drives have acceleration peaks and varying duty cycles.
What to do instead
Build a duty-cycle table with time, speed, torque and inertia. Calculate reflected inertia, RMS torque over the cycle, thermal margin and bearing L10 life with the correct service factor; compare all against catalogue ratings.
Uses steady-state error constants or the final-value theorem before checking closed-loop stability, and confuses system type number with system order.
The candidate memorises the error table for step, ramp and parabolic inputs but forgets that steady-state error is meaningful only for a stable closed-loop system.
What to do instead
Check stability first using Routh-Hurwitz, root locus or pole locations. Then determine system type from open-loop poles at the origin and apply the correct error constant. State explicitly that final-value results are invalid if the system is unstable.
Tunes a PID controller by guessing gains or copying Ziegler-Nichols values without accounting for actuator saturation, derivative noise, sampling time or plant delay, then calls an oscillatory response 'fast'.
The candidate treats PID as three independent knobs and ignores the physical loop, especially saturation and discrete-time effects.
What to do instead
Start from a plant model or identification. Specify bandwidth and phase margin, add anti-windup, filter the derivative term, check actuator limits and sample time, and validate with step and disturbance tests before declaring the loop tuned.
Treats sensor resolution as accuracy and ignores loading, offset, gain, linearity, hysteresis and uncertainty; for example, connects a high-impedance sensor to an ADC without a buffer.
The candidate looks for one headline specification on a datasheet instead of building a measurement chain and uncertainty budget.
What to do instead
Separate resolution, precision and accuracy. Check source impedance versus ADC/load impedance using a Thevenin equivalent, include offset/gain calibration, linearity and hysteresis in the uncertainty budget, and verify at least at zero, mid-scale and full-scale.
Applies homogeneous transformation matrices in the wrong order or mixes base, world, tool and sensor frame labels, producing inverse kinematics or a Jacobian that is algebraically correct but expressed in the wrong frame.
The candidate uses arrow notation loosely and skips frame bookkeeping when under time pressure.
What to do instead
Draw the frame tree and label every transform with superscript/subscript, such as ^A T_B. Verify by transforming a known origin or unit axis, use consistent pre- or post-multiplication, and check a simple numerical pose before trusting the general result.
Writes PLC ladder logic directly from the process narrative without a state table or scan-cycle analysis, so start/stop latches, first-scan bits, interlocks and fail-safe states are wrong, or the emergency stop is implemented only in software.
The candidate reads the narrative as instantaneous cause and effect and forgets that PLC logic executes sequentially and repeatedly.
What to do instead
Prepare an I/O list and state table before coding. Simulate scan order, define fail-safe de-energised states, keep the emergency stop hardwired and independent, and test each interlock and latch transition explicitly.
Answers a tolerance-stack, manufacturing-process or professional-practice question with a general opinion, omitting the datum scheme, worst-case or statistical stack calculation, applicable standard clause, stakeholder duty and documented assumptions.
Candidates concentrate on numerical subjects and treat professional practice as common sense, so they do not use the structured frameworks the question expects.
What to do instead
Use a fixed answer structure: identify the applicable standard or code clause, state datum and assumptions, perform worst-case and RSS stack calculations where relevant, then address safety, ethical duty, documentation and verification.