5.3 Sensor Testing and Commissioning

5.3 Sensor Testing and Commissioning

A transmitter that is "calibrated on the bench" can still lie in the DCS if the range, square-root flag, or failure-current interpretation do not match. Topic 1.L is the checkout chain: cold loop → hot loop → range verification → simulation → diagnostics, including where square-root extraction lives.

Cold loop versus hot loop

Cold-loop checkout is unpowered wiring work. Confirm continuity from plus to plus, polarity of two-wire 4–20 mA loops, shield landed at one point (typically the control-system end), insulation resistance, and that the junction-box landing matches the loop diagram. A swapped pair on a two-wire transmitter may still "work" with reverse-polarity protection — or it may not power up. Either way, cold-loop is the cheap time to find it.

Hot-loop checkout is live. Power the loop, prove that 4 mA, 12 mA, and 20 mA (and often 25/50/75% for a five-point check) arrive as the intended process values in the DCS, including engineering-unit scale, decimal places, and alarm setpoints. Hot-loop is also when you find that the historian tag is still scaled 0–150 inWC from last year's range.

Calibration, range, and interpolation

Calibration compares the device to a traceable standard (deadweight tester, documented calibrator, temperature bath) and adjusts zero/span — or a sensor trim versus an output trim on a smart transmitter. Range verification confirms that the configured LRV/URV match the datasheet and the DCS. Those are different stamps: a device can be accurate on 0–200 inWC and still be wrong for the loop if the DCS still believes 0–100 inWC.

Interpolation on a linear 4–20 mA span:

I = 4 mA + 16 mA × (PV − LRV) / (URV − LRV)

PV = LRV + (I − 4) / 16 × (URV − LRV)

Five-point checks (0, 25, 50, 75, 100% up and back down) catch hysteresis and a swapped square-root setting that a 0/100% check will miss.

Square-root extraction: where it is, and what 25 inWC does

For a differential-pressure flow element, Q / Q_max = √(ΔP / ΔP_max) when density is the design density. Someone must take that square root: the transmitter, the DCS, or a flow computer — exactly one place.

Worked: DP transmitter 0–100 inWC, 4–20 mA.

Square root OFF (linear DP, or a wet-leg level):

I = 4 + 16 × (ΔP / 100)

Square root ON (flow percent on this DP range):

I = 4 + 16 × √(ΔP / 100)

At 25.0 inWC with square root on: √0.25 = 0.50 → I = 4 + 16 × 0.50 = 12.00 mA. The linear current at the same DP is 8.00 mA. That pair is an exam favorite. At 64 inWC, √0.64 = 0.80 → 16.80 mA with square root, versus 14.24 mA linear.

If the transmitter already outputs the square-rooted current and the DCS also takes square root, a true 50% flow (25% DP, 12 mA) is interpreted as 50% of span and then √0.50 = 70.7% on the display — high in the middle, coincidentally correct at 0 and 100%.

Simulation: mA injection versus HART / fieldbus

Milliamp injection (loop calibrator in place of the transmitter, or in series per the tool's method) tests the wiring and the analog input card against an independent current. It does not test the transmitter sensor.

HART loop test / DAC simulation commands the transmitter to drive 4, 12, 20 mA with its own output electronics. That tests transmitter output + wiring + AI card. It does not replace a sensor trim. HART can also write a simulated PV into the digital PV while the analog output follows — know which one the DCS is actually using (analog versus digital PV).

FOUNDATION Fieldbus / PROFIBUS PA loops have no 4–20 mA. Forcing a value is a block simulation (AI / transducer block), not a current injection. Bringing a mA calibrator to a fieldbus segment is the wrong movie.

Diagnostics during simulation should show a "loop test" or "simulated" flag so operations does not treat a forced 20 mA as a real 100% process. Clear the force before you leave the loop in service.

NAMUR NE43 failure currents

Smart pressure and temperature transmitters flag sensor failure, out-of-limits, and configuration errors. On analog loops, many devices follow NAMUR NE43 practice:

The unused gaps (3.6–3.8 mA and 20.5–21.0 mA) keep measurement from colliding with failure. NAMUR NE43 is not a supplied NCEES PE Control Systems design standard. If a question gives 21 mA on a burnout-high RTD transmitter, treat it as this industry convention, not as something you search in ISA-5.1 or IEC 61511.

Configure the DCS so that 3.6 mA and 21 mA do not interpolate as −2.5% and 106% of range and quietly drive a controller. They should shed to a bad-PV status, a last-good hold, or a defined fail-safe — matching the cause-and-effect.

Exam trap: re-range the transmitter, forget the DCS

The loop was commissioned as 4–20 mA = 0–100 inWC, linear, square root off. Operations later asks for 0–200 inWC capability. A technician sets the transmitter URV to 200 inWC and does not touch the DCS.

At an actual 100 inWC, the transmitter now outputs 12.0 mA (50% of 0–200). The DCS still maps 12 mA to 50 inWC. The process is twice what the board shows. Control, interlocks, and historians are all wrong in the same direction.

Fixes: change both ends, or use a digital PV (HART/fieldbus) with a single scale, and repeat the hot-loop 4/12/20 check. The same class of error happens when someone enables square root in the field after the DCS was already taking it.

/practice/pe-control-systemsPractice questions with detailed explanations