Tricky Instrumentation & Control interview questions

75 questions with answers (ICE-426 to ICE-500). Try each one aloud before reading the answer.

Measurement Fundamentals & Errors

ICE-426 · A digital meter shows five decimal places, so it must be very accurate. True?

Trap: Confusing resolution with accuracy. Display digits only show resolution. Accuracy depends on calibration, drift and errors, so a five-digit display can still be wrong by several percent.

ICE-427 · If repeated readings agree closely, the instrument is accurate. Right?

Trap: Equating precision with accuracy. Tight scatter shows precision. A systematic offset can make every reading consistently wrong, so accuracy needs comparison with a traceable standard.

ICE-428 · Averaging many readings removes all measurement error. Correct?

Trap: Assuming all error is random. Averaging reduces random error by √n but does nothing to systematic errors such as zero offset or calibration bias.

ICE-429 · A gauge rated ±1% accuracy is within 1% of the reading anywhere on its scale. True?

Trap: Misreading a full-scale spec as percent of reading. ±1% of full scale on 0–100 bar is ±1 bar everywhere, which is ±10% at a 10 bar reading.

ICE-430 · Once calibrated, an instrument has zero error. True?

Trap: Thinking calibration eliminates error. Calibration leaves residual uncertainty from the standard and procedure, and drift begins immediately, which is why calibration intervals exist.

ICE-431 · Higher sensitivity is always better. Correct?

Trap: Ignoring range and noise. A very sensitive instrument may have a narrow range, saturate easily and pick up more noise; the right sensitivity matches the required range and resolution.

Sensors & Transducers

ICE-432 · All transducers need an external power supply. True?

Trap: Assuming every transducer is passive. Active (self-generating) transducers such as thermocouples, piezoelectric crystals and tachogenerators produce output from the measured energy itself.

ICE-433 · A piezoelectric load cell is ideal for weighing a silo. Correct?

Trap: Ignoring charge leakage. Piezoelectric output decays for static loads, so it suits dynamic force only. Use strain-gauge load cells for static weighing.

ICE-434 · An inductive proximity sensor can detect plastic bottles on a conveyor. True?

Trap: Thinking proximity sensors detect any object. Inductive sensors detect only metals through eddy currents. Use capacitive or photoelectric sensors for plastic.

ICE-435 · A strain gauge measures stress directly. Right?

Trap: Confusing strain with stress. It measures strain through resistance change; stress is inferred using the material's modulus (σ = Eε) within the elastic range.

ICE-436 · An LVDT output is exactly zero at the centre position. True?

Trap: Assuming an ideal null. Harmonics and stray capacitance leave a small residual null voltage, and phase-sensitive demodulation is needed to resolve direction around null.

ICE-437 · Raising bridge excitation always improves strain-gauge signal quality. Correct?

Trap: Ignoring self-heating. Output rises with excitation, but I²R heating causes drift and zero shift, so excitation is limited by gauge size and the substrate's heat sinking.

Temperature & Pressure Measurement

ICE-438 · A thermocouple's voltage is generated at the hot junction. True?

Trap: Locating the EMF at the junction. The Seebeck EMF develops along the wires wherever a temperature gradient exists. The junctions only define the end temperatures, so wire inhomogeneity in gradient zones causes error.

ICE-439 · Copper wire is fine for extending a thermocouple because copper conducts well. Right?

Trap: Treating the thermocouple as a simple conductor. Copper–thermocouple junctions create new junctions at the terminal head, moving the reference point. Use matching extension or compensating cable.

ICE-440 · An RTD is always better than a thermocouple because it is more accurate. True?

Trap: Judging only by accuracy. Thermocouples cover much higher temperatures, are more rugged under vibration, respond faster and are smaller, so they win in furnaces and fast-response applications.

ICE-441 · A thermistor's resistance always rises with temperature. Correct?

Trap: Generalizing from metals. The common NTC thermistor's resistance falls steeply with temperature; only PTC types rise.

ICE-442 · A sealed vessel's gauge pressure reading stays the same when moved to high altitude. True?

Trap: Forgetting gauge pressure is relative to local atmosphere. With the same absolute pressure inside, lower ambient pressure at altitude makes the gauge read higher.

ICE-443 · A gauge mounted 10 m below a water line reads the line pressure. Right?

Trap: Ignoring static head. It reads line pressure plus ρgh ≈ 98 kPa (about 0.98 bar) from the water column, so apply a head correction.

ICE-444 · Pressure can never be negative. True?

Trap: Mixing pressure references. Absolute pressure cannot be negative, but gauge pressure is negative under vacuum, and differential pressure is negative when the low side exceeds the high side.

Flow & Level Measurement

ICE-445 · If an orifice DP doubles, the flow doubles. Correct?

Trap: Assuming a linear DP-flow relationship. Q ∝ √ΔP, so doubling DP increases flow by √2 ≈ 1.41.

ICE-446 · A magnetic flow meter works on any liquid. True?

Trap: Ignoring Faraday's law requirements. It needs a conductive liquid; oils, hydrocarbons and very pure water may be below the minimum conductivity.

ICE-447 · A Coriolis meter measures volumetric flow directly. Right?

Trap: Mixing up mass and volume measurement. It measures mass flow directly, and volumetric flow is calculated from mass flow ÷ measured density.

ICE-448 · DP level measurement does not depend on liquid density. True?

Trap: Thinking DP measures height. It measures hydrostatic head ρgh, so a density change reads as a level change unless compensated.

ICE-449 · Non-contact radar is unaffected by process conditions. Correct?

Trap: Assuming radar is universal. Low dielectric constant liquids, heavy foam, antenna build-up, internal obstructions and turbulence weaken or confuse echoes.

ICE-450 · A rotameter can be mounted in any orientation and used for any fluid. True?

Trap: Forgetting gravity and calibration. It must be vertical with upward flow, and its scale is valid only for the calibration fluid's density and viscosity unless corrected.

ICE-451 · A larger flow meter is safer because it has spare capacity. Right?

Trap: Ignoring velocity. Oversized meters run at low velocity and low Reynolds number, giving poor accuracy, vortex dropout or low-flow cutoff. Size for the actual flow range.

Signal Conditioning & Data Acquisition

ICE-452 · Why does a 4–20 mA loop start at 4 mA rather than 0 mA?

Trap: Thinking it is arbitrary. A live zero lets 0 mA signal a fault such as a broken wire, and the 4 mA minimum also powers two-wire loop-powered transmitters.

ICE-453 · Sampling at exactly twice the highest frequency is enough. True?

Trap: Misreading Nyquist. The sampling rate must exceed 2fmax. At exactly 2f, samples can fall on zero crossings and lose the signal. In practice, sample several times faster with an anti-aliasing filter.

ICE-454 · Grounding a cable shield at both ends gives the best noise protection. Correct?

Trap: More grounding is not always better. For low-frequency instrument signals, two-end grounding creates ground loops; the usual practice is single-point grounding at the control-room end.

ICE-455 · A 24-bit ADC gives 24-bit accuracy. True?

Trap: Confusing resolution with accuracy. Noise, non-linearity, offset and reference errors reduce effective bits, so ENOB and absolute accuracy are often far lower.

ICE-456 · Sampling faster always improves a control loop. Right?

Trap: Faster is not automatically better. Very fast sampling increases noise in derivative terms, raises data and CPU load and can cause numerical problems; it only needs to be fast compared with loop dynamics.

ICE-457 · A 0–10 V signal is just as good as 4–20 mA over long cables. True?

Trap: Ignoring line drop and noise. Voltage signals lose accuracy through cable resistance and pick up noise; current signals are unaffected by line resistance within compliance and provide a live zero.

Control System Theory

ICE-458 · Is positive feedback always bad?

Trap: Assuming positive feedback only causes instability. It is used deliberately in oscillators, Schmitt triggers, latches and seal-in circuits.

ICE-459 · Feedback control removes the effect of all disturbances. Correct?

Trap: Overestimating feedback. It only corrects after an error appears, cannot reject disturbances faster than its bandwidth, and passes sensor noise and sensor errors straight to the process.

ICE-460 · A higher-order system always responds more slowly. True?

Trap: Equating order with speed. Speed depends on dominant pole locations; a high-order system with fast poles can outrun a slow first-order system.

ICE-461 · A type 1 system has zero steady-state error for any input. Right?

Trap: Over-generalizing system type. Type 1 gives zero error for a step, a finite error 1/Kv for a ramp and infinite error for a parabola.

ICE-462 · The final value theorem always gives the steady-state output. True?

Trap: Applying it to unstable systems. It is valid only if sF(s) has all poles in the left half-plane. For 1/(s(s − 1)), it gives −1, but the actual output diverges.

ICE-463 · An overdamped response is best because it has no overshoot. Correct?

Trap: Ignoring speed. Overdamped systems are sluggish; a slightly underdamped response (ζ ≈ 0.6–0.8) often reaches setpoint faster with acceptable overshoot.

ICE-464 · Adding a zero to a transfer function is always harmless. True?

Trap: Ignoring zero location. A left-half-plane zero increases overshoot; a right-half-plane zero causes inverse response and limits achievable bandwidth.

PID Control & Tuning

ICE-465 · Why not just use a very high proportional gain to reduce offset?

Trap: Thinking more gain is always better. High gain reduces offset but reduces stability margins, causes oscillation, amplifies noise and saturates actuators.

ICE-466 · Why not apply derivative action on a noisy flow signal?

Trap: Assuming D always helps. Derivative amplifies high-frequency noise, so it makes the output jitter. Flow loops are fast and noisy, so use PI, or filter and use derivative-on-measurement if D is needed.

ICE-467 · Integral action makes a loop more stable. Correct?

Trap: Confusing offset removal with stability. Integral adds up to 90° of phase lag and reduces stability margins; too much causes oscillation.

ICE-468 · A PI controller keeps PV equal to SP at all times. True?

Trap: Mixing steady-state with transient behaviour. It eliminates offset only at steady state; during setpoint changes and disturbances, errors still occur.

ICE-469 · PID settings can be copied directly between controller brands. Right?

Trap: Assuming one PID standard. Vendors differ in form (ideal, series, parallel) and units (gain or proportional band, minutes or seconds, minutes/repeat or repeats/minute), so settings must be converted.

ICE-470 · Retuning will fix an oscillation caused by a sticky valve. Correct?

Trap: Treating hardware faults as tuning faults. Stiction creates a limit cycle that tuning can only reshape; the valve must be repaired.

Stability Analysis (Root Locus, Bode, Nyquist)

ICE-471 · A stable open-loop system is always stable in closed loop. True?

Trap: Assuming feedback preserves stability. Too much gain or phase lag, especially with dead time, can push closed-loop poles into the right half-plane.

ICE-472 · An open-loop unstable plant cannot be stabilized. Correct?

Trap: Thinking feedback cannot fix instability. Feedback stabilizes many unstable plants, for example K/(s − 1) with K > 1, an inverted pendulum or an exothermic reactor.

ICE-473 · A large gain margin guarantees a robust loop. Right?

Trap: Relying on one margin. A loop can have large GM but tiny PM, or both margins fine while the Nyquist curve passes close to −1 elsewhere. Check PM and peak sensitivity too.

ICE-474 · Bode gain and phase margins always tell you whether a system is stable. True?

Trap: Over-trusting Bode margins. For open-loop unstable or conditionally stable systems, or those with multiple crossovers, Bode margins can mislead; use the Nyquist criterion.

ICE-475 · Adding a pole to the open loop makes the system more stable. Correct?

Trap: Thinking filtering always helps. An extra pole adds phase lag and bends root locus branches toward the right half-plane, reducing margins.

ICE-476 · Higher bandwidth is always better. True?

Trap: Ignoring cost of speed. Higher bandwidth amplifies sensor noise, demands more actuator effort and reduces robustness to unmodelled dynamics and dead time.

PLC Programming

ICE-477 · A physically NC stop button must be programmed with a NC (XIO) contact. Correct?

Trap: Confusing field wiring with program instructions. The NC button keeps the input on when healthy, so the program uses an XIC (examine-if-closed) instruction, which goes false when the button is pressed or the wire breaks.

ICE-478 · When a rung goes true, the physical output changes immediately. True?

Trap: Ignoring the scan cycle. The output image table updates during the scan, but physical outputs are written at the output update stage, after the program is executed, unless immediate output instructions are used.

ICE-479 · Using the same output coil in two rungs acts like an OR. Right?

Trap: Thinking duplicate coils combine. The last rung executed overwrites the bit each scan. Use one coil with parallel branches.

ICE-480 · PLC timers are accurate to the millisecond. True?

Trap: Ignoring scan dependence. Timers are updated when executed, so accuracy depends on scan time and timer base; errors of up to one scan are normal.

ICE-481 · Latched outputs are safe because they remember state after a power failure. Correct?

Trap: Ignoring restart behaviour. Retentive latches can restart motors unexpectedly when power returns; motor starts usually need non-retentive seal-in logic and a deliberate restart.

ICE-482 · An emergency stop can be implemented purely in standard PLC logic. True?

Trap: Trusting general-purpose logic for safety. A standard PLC can fail with outputs on, so E-stops must use hardwired circuits or safety-rated controllers and relays.

DCS & SCADA

ICE-483 · More alarms make the plant safer. Correct?

Trap: Equating alarm count with protection. Excess alarms cause floods and nuisance alarms that operators ignore. Rationalized alarms that each need action improve safety.

ICE-484 · SCADA is suitable for fast closed-loop control across a pipeline. True?

Trap: Confusing supervisory and regulatory control. Communication latency and outages make SCADA supervisory; fast control and protection must run locally in PLCs or RTUs.

ICE-485 · A fully redundant DCS can never fail. Right?

Trap: Ignoring common-cause failures. Software bugs, shared power, configuration errors, cyber attacks and environmental events can defeat redundancy.

ICE-486 · An air-gapped control network is immune to cyber attacks. True?

Trap: Trusting physical isolation. Removable media, maintenance laptops, vendor access and supply-chain compromises can bridge the gap, as the Stuxnet attack showed.

ICE-487 · Historian data stored at 1 s scan is a full record of the process. Correct?

Trap: Ignoring compression. Exception and compression deadbands discard points, and fast transients between scans are never captured, so details may be missing in incident investigations.

ICE-488 · A DCS is just a big PLC. True?

Trap: Comparing only hardware. Modern hardware overlaps; the difference is the DCS's integrated database, engineering tools, operator environment, alarm and batch functions designed for plant-wide process control.

Industrial Communication (HART, Modbus, Profibus, Fieldbus)

ICE-489 · HART replaces the 4–20 mA signal with digital data. Correct?

Trap: Thinking HART is fully digital. In point-to-point mode, the analog 4–20 mA remains the primary process variable, and HART adds digital data on top.

ICE-490 · Modbus address 40001 means register number 40001 in the message. True?

Trap: Confusing the reference convention with the protocol address. 4xxxx denotes holding registers; 40001 is sent as offset 0 with function code 03.

ICE-491 · Every RS-485 device should have a terminating resistor. Right?

Trap: Over-terminating. Terminators go only at the two physical ends of the trunk; extra terminators load the line and can cause communication failure.

ICE-492 · Fieldbus is always faster than a 4–20 mA signal. True?

Trap: Assuming digital means faster. FF H1 runs at 31.25 kbit/s with a scheduled macrocycle, so updates are discrete; an analog signal is continuous, though its speed depends on the input card scan.

ICE-493 · Standard Ethernet is deterministic. Correct?

Trap: Assuming switched Ethernet has guaranteed timing. Queuing and traffic make latency variable, so real-time protocols, priority tagging or TSN are needed for deterministic control.

ICE-494 · Profibus DP and PA devices can share the same cable. True?

Trap: Treating them as one physical layer. DP uses RS-485 at up to 12 Mbit/s, while PA uses MBP at 31.25 kbit/s with bus power, so a coupler or link is required between them.

Control Valves & Safety Instrumented Systems

ICE-495 · An oversized control valve is fine because it has spare capacity. Correct?

Trap: Thinking bigger is safer. An oversized valve throttles near its seat, with high gain, poor resolution, seat erosion and cycling; size it to work in the mid-travel range.

ICE-496 · A fail-open valve is always the safe choice. True?

Trap: Assuming one failure position is universal. The safe position depends on the process: cooling water fails open, but fuel gas fails closed.

ICE-497 · Using a SIL 3 certified transmitter makes the loop SIL 3. Right?

Trap: Confusing device capability with function integrity. The SIL of a SIF depends on all elements, architecture, proof test interval, common cause and the final element, which is often the weakest link.

ICE-498 · Proof testing more often always improves safety. True?

Trap: Ignoring test-induced risks. Shorter intervals reduce PFDavg, but each test brings bypass time, human error and wear, so the interval is optimized rather than minimized.

ICE-499 · Adding a second sensor in 1oo2 always makes the SIF much safer. Correct?

Trap: Ignoring common cause and spurious trips. Shared impulse lines, process taps or calibration errors limit the benefit, and 1oo2 roughly doubles the spurious trip rate.

ICE-500 · A control valve at 0% gives tight shutoff, so no isolation valve is needed. True?

Trap: Confusing control with isolation. Control valves have defined leakage classes and may not seat fully at 0% signal; tight isolation requires on-off valves designed for shutoff.

Other levels: Basic · Intermediate · Advanced · All Instrumentation & Control questions