Application Guidance: Carbon Dioxide (CO2) Sensors
When selecting a CO2 sensor, consider the following application guidance:
CO2 sensors may be stand-alone or integrated in a thermostat for an HVAC system. They have a variety of data output options, which may include:
- Analog (voltage or current signals)
- Digital (primarily BACnet protocol)
- WiFi to cloud-based data storage
Analog outputs are widely compatible with all building automation systems (BAS) and DCV controllers. BACnet is also generally compatible with most BAS but may require additional hardware and specialized expertise to interface between the sensor and BAS. WiFi to cloud-based data storage systems are growing in popularity and have some benefits and drawbacks that should be considered:
- Benefit: Faster installation with no communication wire
- Benefit: Automatic updates with new feature releases as software is improved over time
- Drawback: Loss of WiFi or internet services can temporarily render devices inoperable
- Drawback: Connectivity to existing BAS hardware may be difficult or impossible. An application programming interface (API) may allow for interfacing with an existing BAS.
- Drawback: Long-term reliance on a vendor for system operation as well as data subscription and storage costs
The CO2 sensor and the associated BAS may have reporting capabilities to alert facilities management that the CO2 level in a space has exceeded a programmed threshold. Alert methods may include email, text, or cell-phone app. To be compatible with California’s CalSHAPE program in schools, CO2 sensors must have an alert method.
The CO2 sensor may display the CO2 value, which is helpful for occupants to know if the space they are in is sufficiently ventilated. The display maybe on the physical sensor or accessible via the web or cell phone app. To be compatible with California’s CalSHAPE program in schools, CO2 sensors must have a display accessible to the teacher.
Most CO2 sensors work by using non-dispersive infrared detection (NDIR) technology. NDIR sensors use a light source at one end of a sample chamber and a detector at the other end of the chamber to measure the intensity of a specific wavelength of light that is absorbed by CO2. The intensity of the light transmitted is correlated to the number of CO2 molecules present, which is then converted to a CO2 concentration based on the volume of air in the chamber. The accuracy of a CO2 sensor can drift over time due to aging of the light source.
Figure 1-Components of an NDIR CO2 sensor 
To maintain accuracy over time, commercially available CO2 sensors generally contain one of two calibration methods: automatic background calibration (ABC) or a dual beam infrared detector approach. Both types of calibration methods generally work well in most commercial building applications. In the ABC approach, the sensor continuously tracks the lowest level of CO2 observed over a period of time (generally 1-7 days). Because most buildings are vacant during some part of the day (e.g., at night), the CO2 levels in the vacant space will drop to outdoor levels, which are generally around 400-450 parts per million (ppm). The sensor ABC algorithm then sets the lowest level observed by the sensor over the ABC period to 400 ppm. This is similar in concept to re-zeroing a scale when no weight is present. In order for ABC to work correctly, the space with the CO2 sensor must experience regular periods with no occupants. ABC is not recommended in continuously occupied spaces such as hospitals.
In the dual beam approach, the CO2 sensor contains two detectors. The first detector measures the intensity of light transmitted through the sample at the wavelength that correlates to CO2 gas concentration. At the same time, the second detector measures the intensity of an alternate wavelength that is not impacted by gas absorption. The measurement of the second detector is used to correct for aging of the light source and particle buildup on the sensor.
An accurate measurement from an NDIR sensors relies on calculation of the CO2 concentration, which requires knowing the density of the air inside the measurement chamber. Density of the air is affected by temperature and pressure, which varies with altitude. While most NDIR sensors measure and automatically correct for temperature, only some include an altitude correction, which is either measured automatically by a barometric pressure sensor or is programmed by the installer. Without altitude correction, roughly each 1000 ft increase in altitude will result in an additional-30 ppm error in measurement at a CO2 concentration of 1,000 ppm . Sensors with ABC partially compensate for altitude; however, they will still be impacted by a lesser degree . For locations at high altitude, purchase sensor models that include an altitude correction method and ensure they are configured correctly.
CO2 sensors should be installed on the wall of the room being monitored at the height of occupant’s heads (3-6 feet above the floor) and ideally away from doors and openable windows. If possible, install the sensor near the HVAC return grill. Since specific products may have additional installation considerations, review and follow all manufacturer specific installation instructions.
1. Schell, Mike and Dan Inthout. ASHRAE Journal. Demand Control Ventilation Using CO2. February 2001. https://www.krueger-hvac.com/files/white%20papers/article_demand_control_ventilation.pdf
2. BAPI. The effects of temperature and altitude on CO2 measurement. https://www.bapihvac.com/wp-content/uploads/2011/04/Altitude_Temperature_CO2_ALC.pdf
3. Sensair. Pressure Pressure Dependence of SenseAir´s NDIR sensors. http://www.co2meters.com/Documentation/AppNotes/AN149-Senseair-Pressure-Dependence.pdf