Carbon dioxide sensors: direct CO2 measurement and device design

How do conventional and photoacoustic NDIR CO2 sensors work?

NDIR sensors evaluate how much selected infrared light is absorbed as it passes through the gas. The source, measurement volume and detection arrangement belong to the same measurement chain. An appropriate reference channel can help reduce the effects of source or optical-path changes. The resulting stability does not make contamination and condensation unimportant.

The photoacoustic method detects pressure changes in the gas caused by absorbed light. Sensirion's SCD4x family uses this principle and measures actual CO2. The method name does not establish equal accuracy across every range. The finished device's response, power consumption and environmental robustness are evaluated alongside the selected model's conditions and integration.

Why is eCO2 not a substitute for carbon dioxide measurement?

eCO2, or CO2 equivalent, is an estimated indicator some air quality sensors generate algorithmically from their response to other gases. For example, Sensirion's SGP30 documentation explains that its equivalent CO2 signal is calculated from ethanol and hydrogen measurements. A ppm display does not prove that carbon dioxide molecules are measured directly.

Changes in cleaning products or other volatile compounds can cause the estimate to diverge from actual CO2. If the requirement is CO2 concentration, select a measurement method defined for that gas. If a device shows both CO2 and VOCs, label them separately; technical documentation should clearly identify the sensor type and how the data is produced.

Read ppm, percentage and accuracy specifications together

A CO2 concentration of 1% by volume equals 10,000 ppm. A module monitoring room-air changes and a probe measuring percentage-level process gas are selected for different expectations. Some sensors provide a wide numerical output range while their stated accuracy applies only over a narrower range.

The fixed ppm term and the percentage-of-reading term in an accuracy specification must be considered together. Repeatability and the final display digit are distinct from accuracy. The acceptance plan should include normal operating points, the highest expected level and recovery behaviour. Returning to the normal range after very high gas exposure may also matter in the application.

Verify the assumptions behind automatic baseline correction

Some sensors' automatic calibration algorithms assume periodic exposure to a known low CO2 level. A continuously occupied space or a process kept rich in CO2 may not meet that condition. An incorrect assumption can cause the displayed value to drift over time even when the actual concentration does not change.

Algorithm settings therefore follow the selected component's current instructions and actual use profile. A forced adjustment using known gas or a reliable reference is a separate operation. Record check conditions rather than assuming that opening a window makes the reading correct. Sensor replacement, installation and calibration coefficient updates can be retained as traceable software events.

How do pressure, heat and airflow affect measurement?

Since gas density changes with pressure and temperature, examine how the sensor compensates for these effects. Not every probe measures its own pressure; some models require external pressure data or an altitude setting. Electronics warming inside the enclosure can also change the sensor's environment.

For room monitoring, direct breath or positions beside an open window or air outlet may produce unrepresentative results. Enclosure openings and internal volume affect how quickly gas reaches the sensor. In line with Sensirion's design guide, heat sources, vibration and unintended airflow effects are examined on the prototype. Condensing processes require a suitable probe and sampling arrangement rather than a room-type module.

What is designed when turning a CO2 module into a product?

An internal connection such as I2C and a long field communication line require different designs. The power supply must handle measurement-related load changes as well as average sensor current. PCB and enclosure layouts can alter results from a previously successful evaluation board, so the final design is tested again.

Embedded software separately handles startup, new-data readiness, communication faults and invalid measurements. A lost connection must not leave the last reading appearing current. Filtering and control outputs define delay, repeated switching around thresholds and restart behaviour. The display, records and server must use the same units and timestamps.

Transferring sensor data from the device to automation systems →

What information can start a CO2 project?

Share the space or process, expected CO2 range, temperature and humidity conditions, mounting position and decision the data will support. Battery life, measurement frequency, display, network connections and maintenance access are part of the design. Specifically identify continuously high-CO2 use so the calibration approach can suit it.

A VOMRO project clarifies sensor selection, electronics, embedded software, enclosure and validation scope together. The comparison location and acceptable deviation from a reference are documented. Source code and design documents are included in the deliverables list. A prototype displaying CO2 does not promise an independently unverified life-safety alarm function.