15.5 Sensor Selection, Installation, and Input Point Configuration

15.5 Sensor Selection, Installation, and Input Point Configuration

The Building Automation competency block on Sensor Installation and Applications requires four things: identifying the types of sensors that measure temperature, humidity, pressure, velocity, level, proximity, CO₂, motion, and photoelectric signals; describing the proper application of each; wiring sensors to unit controllers per NEC requirements; addressing points for various input sensor types; and configuring inputs to unit controllers.

A control system is only as good as the numbers it is fed. A drifted or badly located sensor makes a perfectly written sequence behave badly, and the workstation will report the wrong value with total confidence. This section is about getting the number right before blaming the logic.

---

1. Sensing Technologies by Measured Quantity

---

2. Input Signal Types and How the Controller Reads Them

Sensors reach a controller as one of four input classes, and the controller's input must be configured to match.

Scaling is a separate step from wiring, and skipping it is the most common commissioning defect. A 0–10 VDC pressure transmitter with a 0–5 in. w.c. range wired to an input that is still scaled 0–100 will read "53" when duct static is 2.65 in. w.c. Everything is wired correctly; the number is meaningless. Configure the input with the transmitter's range from its label, not from memory or from the last job.

4–20 mA and the live zero. Because the loop is a current loop, series resistance in a long conductor does not change the current, so distance does not shift the reading the way it does on 0–10 VDC. And because the bottom of the range is 4 mA, not 0, a broken conductor produces 0 mA — a value outside the valid range that the controller can flag as an unreliable point. An open 0–10 VDC circuit reads 0 V, which is indistinguishable from a legitimate zero. This is the reason 4–20 mA survives in critical and long-run applications.

---

3. Location: The Rule That Beats Accuracy

A ±0.2°F sensor in the wrong place is worse than a ±1.0°F sensor in the right one, because the accurate device makes a wrong number credible.

Space sensors — where not to mount:

• On an exterior wall, where conduction through the wall biases the reading with outdoor temperature
• In direct sunlight at any hour of the day
• Directly above or beside a heat source — a copier, a monitor, a coffee maker, a light fixture
• In the discharge path of a supply diffuser, where it reads supply air rather than space air
• Behind a door, inside a closet, or in a corner with no air movement
• On a wall cavity that is open to a chase or to unconditioned space — stuffing the box with insulation stops the chimney effect that otherwise pulls attic or crawlspace air across the sensing element

Duct sensors:

• Locate in a well-mixed section, at least several duct diameters downstream of a coil, a mixing box, or a branch takeoff.
• Immediately downstream of a coil or a mixing plenum, temperature stratification of 15–20°F across the cross-section is normal, so a single-point probe reports whatever stratum it happens to sit in.
• Averaging elements — a long serpentine capillary or several sensing elements in series, strung across the duct — are the correct answer there. The task list's mixed-air and economizer sequences depend on them: an economizer controlled from a single-point mixed-air probe will hunt or freeze a coil.
• Seal the penetration and support the probe so it cannot vibrate against sheet metal.

Outdoor-air sensors: north side, shaded, out of direct sun and away from exhaust discharge, relief hoods, and condenser discharge. An outdoor sensor mounted above a rooftop condenser reads the condenser, and every reset and economizer decision downstream is then wrong.

CO₂ sensors: in the breathing zone (roughly 3–6 ft above the floor) or in the return path serving the zone, away from the direct throw of a diffuser and away from where occupants exhale onto the device.

---

4. Wiring per NEC Requirements

Every wiring task in this block ends with "per NEC requirements," and the rules from Section 15.2 apply to sensor wiring specifically:

• Sensor wiring is almost always Class 2, power-limited under Article 725.
• Never share a raceway, cable, or enclosure with line-voltage conductors unless separated by a permanent barrier. Running a thermistor pair alongside a VFD output is the fastest way to build an unstable input.
• Use plenum-rated cable (CMP) in ducts, plenums, and ceiling cavities used as return-air plenums.
• Use shielded, twisted pair for analog inputs in electrically noisy areas, and ground the shield at the controller end only — grounding both ends creates a ground loop that injects the noise you were trying to reject.
• Support the cable on J-hooks or in tray at code intervals. Cable draped on ceiling grid or on piping is a citable violation.
• Observe conductor gauge and length. On a 10 kΩ thermistor, the wire's own resistance adds directly to the sensed resistance; on a long run of small conductor this shows up as a fixed offset the technician then "fixes" with a bias, hiding the real problem.
• Label both ends of every conductor.

Two-wire, three-wire, and four-wire transmitters are not interchangeable. A two-wire (loop-powered) 4–20 mA transmitter draws its power from the loop; a three-wire device needs separate power and a shared common; a four-wire device has isolated power and signal. Wiring a three-wire transmitter as though it were loop-powered gives a dead input, and the sensor is usually blamed.

---

5. Addressing and Configuring Inputs

The task list names "addressing points for various input sensor types" and "configuring inputs to unit controllers" as distinct competencies.

A complete input configuration has six parts:

1. Terminal assignment — which physical input the sensor lands on (AI-1, UI-4, BI-2)
2. Input type — resistive, voltage, current, or dry contact, often set by a jumper or DIP switch and in software; both must agree
3. Sensor curve or range — 10 kΩ Type III, 0–10 V, 4–20 mA
4. Engineering-unit scaling — the transmitter's range mapped to the displayed value and units
5. Point name and address — following the site convention (AHU2_MAT, not AI3), with the network address recorded in the address schema (Section 15.7)
6. Alarm limits and reliability handling — what the controller does when the value goes out of range or the point goes unreliable

Configuring a point on a networked sensor bus adds an address step: the room operator or smart sensor carries its own address, set by DIP switch, rotary switch, or software, and it must be unique on that bus.

---

6. Verifying a Sensor Before Replacing It

The discipline that separates a technician from a parts-swapper:

1. Compare to a calibrated reference. Put a known-good instrument next to the sensor and compare readings at the same moment. A sling psychrometer or calibrated digital psychrometer for space conditions; a calibrated thermometer for duct and immersion.
2. Measure the raw signal at the sensor and again at the controller terminals. If the two disagree, the fault is in the wiring, not the sensor. If they agree but the workstation shows something else, the fault is in the scaling or the curve selection.
3. Check the thermistor against its resistance table. A 10 kΩ Type II NTC thermistor reads 10,000 Ω at 77°F (25°C); resistance rises as temperature falls. Measure resistance with the sensor disconnected from the controller and compare to the manufacturer's temperature–resistance table for that curve.
4. Trend the point. A sensor that reads a plausible value but holds it unchanged for hours is failed. A single spot reading cannot reveal that; a trend reveals it immediately. This is the diagnostic the task list's "trending" competency exists for.
5. Check the mounting before the device. If the reading is plausible but always biased one direction, suspect location — sun, a heat source, an unsealed wall cavity, stratification — before condemning the sensor.
6. Then, and only then, replace — and after replacing, verify the new device against the reference rather than assuming a new part is a correct part.