Sensor selection, location and calibration

A building automation system can only be as good as the information it receives. A supply air temperature sensor that reads 1.5 °C low, a CO2 sensor that has drifted 200 ppm, or an outdoor air temperature sensor in direct sun will quietly cause wasted energy, poor comfort and misleading data for years. Sensors are among the least expensive components in an HVAC system, yet they decide whether sophisticated sequences such as resets and demand-controlled ventilation work at all.

This guide covers sensor types, typical accuracy, installation location and calibration practice for the most common HVAC measurements.

Why sensor accuracy matters

Small errors compound. Some examples:

• An economizer comparing outdoor and return temperatures can lock out free cooling, or bring in hot humid air, if either sensor is off by a degree or two.
• A chilled water delta T calculated from two sensors each accurate to ±0.5 °C can be wrong by up to 1 °C, a large fraction of a 5 or 6 °C design delta T.
• A drifted CO2 sensor in a demand-controlled ventilation system either over-ventilates, wasting energy, or under-ventilates, harming air quality.
• Static pressure reset depends on accurate VAV airflow readings at low flow, where velocity pressure signals are very small.

Accuracy requirements depend on the purpose. Sensors used for control, energy calculation or fault detection generally need better accuracy than those used only for monitoring.

Typical sensor types and accuracy

ASHRAE Guideline 36 and energy codes reference minimum accuracy for some sensors used in specific sequences. Specify accuracy across the expected operating range, not just at one point, and select transducer ranges that match the expected values: a 0 to 2,500 Pa transducer measuring 25 Pa will be inaccurate.

Location: getting a representative reading

Space temperature and humidity

• Mount on interior walls at about 1.2 to 1.5 m (4 to 5 ft) above the floor, considering accessibility requirements
• Keep away from supply diffusers, direct sunlight, exterior walls, doors, copiers and other heat sources
• Seal the conduit or wall opening behind the sensor so air from the wall cavity does not affect the reading

Duct temperature

• Use averaging sensors across the full face of coils and mixing plenums where stratification is likely, such as mixed air and coil discharge
• Place discharge sensors far enough downstream of coils and fans to allow mixing
• Single-point probes are acceptable in well-mixed ducts after fans

Outdoor air temperature and humidity

• Install on a north-facing wall or in a shielded enclosure, away from exhausts, roofs and reflected heat
• Avoid locations near intake louvres where building air or exhaust can influence the reading
• Consider using one high-quality outdoor sensor shared across systems to avoid conflicting values

Static pressure

• Locate duct static sensors about two-thirds to three-quarters of the way down the main duct, or at the location identified by the designer, away from fittings and turbulence
• Use a protected outdoor reference for building pressure, sheltered from wind

CO2

• Mount in the breathing zone, about 0.9 to 1.8 m (3 to 6 ft) above the floor, away from doors, windows, supply outlets and where people stand close to the sensor
• Return duct CO2 sensors average a zone but can miss poorly ventilated rooms. See demand-controlled ventilation with CO2 sensors.

Water temperature and flow

• Use thermowells with thermal paste, inserted to the correct depth
• Install flow meters with the straight pipe lengths required by the manufacturer, typically several diameters upstream and downstream

Calibration and verification

Sensors drift. A calibration program keeps them trustworthy.

1. Verify at installation. Check every sensor against a calibrated reference during start-up and commissioning, and correct offsets.
2. Set intervals by type. Temperature sensors are generally stable; humidity and CO2 sensors drift faster; gas detection sensors require scheduled bump tests and calibration per manufacturer.
3. Use traceable references. Reference instruments should be calibrated to traceable standards, with their own calibration current.
4. Check in place where practical. Compare sensors against a reference at two or more points across the operating range when possible.
5. Use data to find drift. Compare redundant or related sensors in BAS trends: return air and space temperatures, CO2 readings at unoccupied times (which should approach outdoor levels), and supply and return water temperatures at zero load.
6. Document results with as-found and as-left values.
7. Replace sensors that cannot be brought within tolerance or have reached their rated life.

Fault detection software can flag many sensor problems automatically, such as values stuck at a constant, readings outside physical limits, or sensors that disagree. See fault detection and diagnostics.

Common problems found in the field

• Space sensors on exterior walls or in direct sun
• Single-point mixed air sensors reading stratified air
• Outdoor sensors on roofs or near exhaust fans
• Static pressure sensors right after a fan or fitting
• CO2 sensors never calibrated since installation
• Software offsets added to hide a failed sensor
• Transducer ranges far larger than the values measured

Our control system verification work routinely includes point-to-point checks and sensor calibration verification.

Key takeaways

• Sensor accuracy directly affects energy use, comfort and the reliability of advanced sequences.
• Match sensor type and range to the measurement, and specify accuracy across the operating range.
• Location is as important as accuracy: avoid heat sources, stratification and turbulence.
• Verify sensors at commissioning, calibrate periodically with traceable references, and use BAS data to detect drift.
• Humidity, CO2 and gas sensors need more frequent attention than temperature sensors.