NOTE This standard governs the field layer of the HVAC controls system: the sensing elements, transmitters, thermostats, and actuators that gather signals for and execute commands from the building automation system. (1.1)

NOTE It begins at the field controller's terminal strip and extends outward to the point of measurement or actuation. Everything from the controller inward — the controller itself, the network, programming, and the written sequences of operation — belongs to Building Automation System. Drawing the boundary at the controller terminal keeps two different trades and two different submittal reviews from overlapping at the most error-prone seam in a controls installation. (1.2)

NOTE The devices covered are the physical instruments a technician can touch, calibrate, and replace. (1.3)

NOTE This includes space temperature sensors and wall thermostats; duct, mixed-air, immersion, and outdoor temperature and humidity sensors; differential-pressure transmitters for duct static, hydronic, and filter applications; CO2 sensors for demand-controlled ventilation; occupancy sensors tied into the HVAC loop; airflow measurement stations; and electric valve and damper actuators furnished under the controls scope. (1.4)

NOTE ASHRAE Guideline 36 is the best single reference for determining which inputs and outputs a given system type requires; the sequences themselves are owned by the BAS standard, but the device list must satisfy them. (1.8)

NOTE The following work is explicitly excluded from this standard and is governed elsewhere. (1.9)

NOTE The boundaries below are stated as scope exclusions, not obligations, because no one performs them under this section. (1.10)

NOTE A factory calibration certificate is not a substitute for field verification. (4.4)

NOTE Field-installed sensors are frequently shipped uncalibrated or drift during shipping and handling. Requiring a NIST-traceable factory certificate establishes the as-shipped baseline; the field verification below confirms the device still reads true after installation. Both are required for the critical points. (4.5)

NOTE A supply-air temperature transmitter on AHU-1 carried as TT-AHU-1-SA on one drawing set and as something else on another is a guaranteed installation error and a recurring source of controls RFIs. Tag consistency is a quality-assurance requirement, not a drafting nicety. (4.11)

NOTE Enclosure selection is a function of where the device lives, not of the device itself. The same transmitter may be Type 1 in a dry electrical room and Type 4X on a corrosive rooftop. IEC 60529 IP ratings are an acceptable equivalent basis where a manufacturer rates only to the IP code: IP54 is roughly equivalent to NEMA Type 12, and IP65 to NEMA Type 4. (5.2)

NOTE Class 2 power limiting keeps the field wiring inside the energy thresholds that permit plenum-rated cable without conduit in many jurisdictions and bounds the fault energy at the sensing element. The conductors and raceways themselves are governed by Conductors And Cables and Raceways And Conduit; this standard fixes only the device-side voltage and circuit class. (6.2)

NOTE A 4-20 mA loop has a finite voltage budget. If the sum of the device burden, the wiring resistance, and the controller input resistance exceeds what the supply can drive, the transmitter saturates below 20 mA and the reading clips at the top of the range. This is checked on paper at submittal, not discovered at startup. (6.4)

NOTE For most building points a 10k Ohm thermistor gives adequate accuracy at the lowest cost and is the default for space sensing. A 1000 Ohm platinum RTD is more linear and stable over a wide span and is preferred for immersion and critical air-handler points. A 4-20 mA transmitter is used where the run length or electrical noise makes a direct-resistance signal unreliable, and a BACnet MS/TP network sensor is used where the project's network architecture favors device-level addressing. (7.2)

NOTE An accuracy figure with no stated span is meaningless: a device can be ±0.5°F at 70°F and ±3°F at the ends of its range. Stating the operating span over which the accuracy must hold is what makes the specification enforceable. (7.5)

NOTE Poor sensor placement is the single most common cause of comfort complaints that no amount of controller tuning can fix. Sensors on exterior walls read the wall, not the room; sensors under a supply diffuser read the supply air; sensors near a radiant panel or in direct solar gain read the source, not the occupant. (7.8)

NOTE Mixed air stratifies: the cold outdoor stream and the warm return stream do not blend completely within the mixing box. A short single-point sensor placed in either stream reports a temperature that does not exist anywhere the air is actually delivered, which corrupts every economizer and freeze-protection decision downstream. A serpentine averaging element that crosses the full duct face is the only reliable way to read true mixed-air temperature. (7.16)

NOTE A thermowell lets a sensor be pulled and replaced without draining the system or scalding the technician, and it isolates the element from the pressure boundary. Thermowell length and insertion depth must match the pipe diameter and the sensor element length so the tip sits in the flowing stream, not in a stagnant boundary layer at the wall. Coordinate the immersion taps with Hydronic Piping. (7.19)

NOTE A bare outdoor sensor in sunlight reads the solar load on its own housing, not the air, and can run many degrees high on a clear day. A ventilated radiation shield is what makes an outdoor reading usable for economizer and reset logic. (7.21)

NOTE Capacitive thin-film elements are the building-controls standard: they recover from saturation, resist contamination better than bulk-polymer elements, and hold accuracy over a wide range. Bulk-polymer elements are cheaper but drift faster and are reserved for non-critical monitoring. (8.2)

NOTE Air leaving a cooling coil at part load is at or near saturation, and condensate carried onto a capacitive element destroys it. Locating the sensor well downstream gives the air room to mix and warm above its dew point before it reaches the element. (8.6)

NOTE Piezoresistive transmitters are stable, have no moving linkage to wear, and provide a clean analog or network output. Mechanical pressure switches and gauges are reserved for simple local indication and high/low limit safeties, not for the modulating control and reset inputs this section covers. (9.2)

NOTE A transmitter operating at the bottom of its range wastes resolution and amplifies sensor noise into the control loop; one operating near full scale clips during transients. Placing the design setpoint near mid-scale gives headroom in both directions and the best signal-to-noise for reset control. (9.4)

NOTE ASHRAE Guideline 36 requires the static pressure sensor be located remote from the fan. Sensing at the discharge makes the static-pressure-reset logic travel a much wider band than necessary and starves the terminal boxes at the end of the run, the exact opposite of what duct static reset is meant to achieve. (9.7)

NOTE NDIR is the only technology accurate and stable enough for ventilation control. Dual-beam NDIR, which uses a reference wavelength to cancel lamp aging and window fouling, is preferred over single-beam for long-term stability without constant recalibration. (10.2)

NOTE ASHRAE 90.1-2022 Section 6.4.3 mandates demand-controlled ventilation for zones with a design occupant density of 25 people or more per 1000 ft². (10.3)

NOTE Where the energy code triggers DCV, CO2 sensing is the input that drives it, so the sensor selection below is code-driven, not optional, for those zones. ASHRAE 62.1-2022 governs the ventilation calculation the DCV logic executes. (10.4)

NOTE ABC logic assumes the space empties and falls to outdoor CO2 concentration at least once every seven days, then resets its zero to that low point. A 24/7 facility never reaches that baseline, so ABC drifts its zero upward and reports falsely low CO2, which starves the space of ventilation exactly when it is occupied. In those spaces ABC must be turned off and the sensor calibrated by hand on a schedule. (10.10)

NOTE ASHRAE 62.1 DCV math assumes a 400 ppm outdoor baseline. Actual outdoor CO2 at grade in dense urban locations runs 450 to 600 ppm, so a fixed 400 ppm assumption systematically under-ventilates. A measured outdoor reference removes that error in high-occupancy buildings; smaller projects in clean-air locations may omit it. (10.13)

NOTE Passive infrared (PIR) alone detects motion and is adequate where occupants move regularly. Dual-technology sensors add ultrasonic or microwave detection and hold occupancy through low-motion activity such as seated desk work, preventing the HVAC from prematurely dropping a still-occupied zone to setback. This sensor is the HVAC loop's occupancy input only; sensors used solely for lighting are governed by Lighting Controls. (11.2)

NOTE A networked thermostat exposes its setpoint, mode, and space temperature to the BAS for scheduling, reset, and trending. A standalone thermostat is an island the operator cannot see or override remotely, acceptable only where the served equipment is genuinely outside the BAS scope. (12.3)

NOTE A pitot-array (flow-cross) station averages velocity pressure across the duct and is robust and inexpensive, but loses accuracy at very low velocities where velocity pressure becomes negligible. A thermal-dispersion array measures mass flow directly and holds accuracy down to very low velocities, at higher cost, and is preferred for outdoor-air and minimum-ventilation measurement where the flow is small and the accuracy matters most. (13.2)

NOTE A flow station reads accurately only in a developed, non-swirling velocity profile. Mounting it too close to an elbow, transition, or damper produces a distorted profile and a systematically wrong reading no field calibration can correct. The required straight-run distances are part of the device specification, not a field option. (13.5)

NOTE This section covers only the actuator, its control signal, and its fail-safe behavior. The damper blade, frame, seals, and leakage class belong to the damper standard. The two must be coordinated so the actuator torque matches the damper the manufacturer actually ships. (14.2)

NOTE The bare-area rule covers the air load, but the torque needed to drive a damper closed also has to overcome blade-seal friction, linkage losses, and the differential pressure across the closed blade. Sizing on area alone is the most common cause of actuators that stall short of full close at design static, which shows up as a TAB failure and a string of RFIs. The submitted torque calculation must account for seal friction and applied safety factor explicitly. (14.4)

NOTE With a 0-10 VDC actuator, a lost or broken signal reads as 0 V, which the actuator interprets as a valid full-open (or full-closed) command rather than a fault. A 2-10 VDC actuator treats 0 V as signal-loss and drives to its spring-return fail-safe position instead. Specifying a 2-10 VDC live-zero signal is what makes signal loss a safe event. (14.9)

NOTE The valve body and trim are part of the hydronic system; this section covers the electric actuator that drives it, its control signal, and its fail-safe direction. Industrial process control valves and positioners are not covered here and are governed by Control Valves And Actuators. (15.2)

NOTE An actuator that can position the valve mid-stroke but cannot drive it fully closed against system differential pressure leaks flow at the closed position, defeating temperature control and wasting energy. Sizing for 1.25 times the close-off pressure gives margin for pump head variation and valve wear. (15.6)

NOTE Fail direction is a life-and-property decision, not a default. A chilled-water coil valve should fail closed so a power loss cannot flood the coil and the space below it. A hot-water or steam coil in a freeze-prone air stream may need to fail open to keep flow through the coil and prevent a freeze burst. The spring-return direction is fixed when the actuator is bought, so the failure mode must be settled before procurement, not at startup. (15.8)

NOTE BACnet MS/TP and BACnet/IP are the dominant building-controls protocols; Modbus RTU appears on some packaged equipment. The network architecture, addressing, and controller-side integration are owned by Building Automation System; this section requires only that the field device speak the protocol the project has standardized on. (16.2)

NOTE Wireless mesh sensors are a retrofit tool for places a wire cannot reasonably reach, not a default. ASHRAE Guideline 13 sets a 5-year minimum battery life so a ceiling full of sensors does not become an annual ladder-and-battery maintenance burden. Mechanical rooms are RF-hostile and may need wired repeaters that must be accounted for in the BAS scope, not discovered at startup. (16.6)

NOTE A pipe temperature tap or flow station buried under continuous insulation is unreadable and unserviceable. Coordinating the insulation scope at submittal keeps the sensing points accessible and the insulation continuous everywhere else. (17.3)

NOTE Point-to-point verification confirms that the device wired to a given controller terminal is the device the drawings say is there and that it reads or strokes correctly. It is a device-level check; integrated functional performance testing of the assembled system is owned by Commissioning, and instrument-calibrated flow measurement is owned by Testing Adjusting And Balancing. (18.2)

NOTE Construction dust coats optical CO2 windows and contaminates humidity elements, shifting calibration before the device ever sees service. Keeping these sensors sealed until the space is clean preserves the factory calibration the project paid for. (19.5)

NOTE A long calibration-stability warranty is the manufacturer's commitment that the device will not drift out of tolerance over its service life, which is the property that actually matters for a DCV sensor. A one-year defects warranty says nothing about drift. (20.3)