CO2 sensors that incorporate an auto-calibration system that periodically resets the sensor’s calibration based on the lowest measured CO2 concentration during a prior time period, should be deployed in the occupied space for approximately two weeks before an initial calibration check. During this initial period of deployment, the sensor’s self-calibration system may be making rapid adjustments in the sensor’s output.

Some of the CO2 sensors sold for DCV applications have no provisions for convenient calibration in field settings. Other sensors can be provided with a port or apparatus for introducing a calibration gas and the manufacturer provides a calibration procedure and may market a calibration kit. Some calibration systems can be located by searching the web for “CO2 sensor calibration kits”. If the user plans on performing calibrations, which is recommended, they should select the second of these two types of CO2 sensors and follow the manufacturers’ calibration procedures. Because the output signal of CO2 sensors is affected by air pressure, it is important to introduce a calibration gas in a manner that does not significantly change the air pressure in the sensor.

Even when a direct calibration is not possible, the accuracy of installed CO2 sensors can be checked using a calibrated portable reference CO2 instrument. The reference instrument should be calibrated using calibration gas mixtures and then placed near installed CO2 sensor being checked for a period sufficient to allow the instrument outpour to stabilize, often less than one minute. The concentration of CO2 indicated by the sensor can then be compared to the “true” concentration indicated by the reference instrument. In some cases, the manufacturer will provide a procedure for adjusting the calibration of the CO2 sensor, if such a procedure is not available, a correction can be made using software. During this calibration check process, it is critical to avoid exhaling high CO2 breath toward the CO2 sensor and reference instrument. Quality portable CO2 calibration instruments usually cost at least $1000. Less expensive portable instruments might be usable for this application if their calibration is checked before and after the period of use and found to be stable.

Few data are available for development of recommended calibration intervals for CO2 sensors. Calibration intervals recommended in manufactures literature range from 3 years to never needed calibration, with five years common. Sensor tests over a one year period by Shrestha [8] found that many sensors had a change in output less than 75 ppm at a CO2 concentration of 1100 ppm. Other sensors had a less stable calibration, with a change in output of in excess of 100 ppm over one year [10]. 

If neither a direct calibration nor a check with a reference instrument is possible, sometimes faulty sensors can be detected via a simple examination of sensor output data. In an intermittently occupied space, a properly operating CO2 sensor will report a concentration that decreases, usually to about 400 ppm, when the building has been unoccupied for many hours and will report a concentration that increases during occupancy. Also, sensors that indicate CO2 concentrations very different from surrounding sensors should be checked for accuracy. Using sensors with a visual display of CO2 concentration can facilitate identification of faulty sensors and faults can sometimes be noticed by walking by sensors and noting the displayed CO2 concentrations.