Sensor Technology


Fig 3: Gas Sensing’s high-range explorIR, ultra-fast sprintIR, and ultra-low-power COzIR families of NDIR sensors implement the above techniques for use in a range of IoT and other CO2


sensing applications


sensor and system considerations that can be implemented for a low-power system.


monitoring in IoT systems Moving beyond the tuning-fork mechanism, CO2


CO2 concentrations can be


measured by shining NDIR (non-dispersive infrared) light through a sample and measuring the frequencies absorbed, with CO2


absorbing the mid-IR 4.25 µm wavelength.


For IoT, arguably the key requirement of any sensor is for it to be low power. IoT systems often need to operate away from a main’s power source and therefore need to run for many months or even years from an AA battery / coin cell power source. If it can be run from an energy harvester, even better.


Before we get to power, it’s also important to note that this isn’t the only consideration. Accuracy is obviously vital and depending on the application an accuracy of 70 ppm is more than enough for most applications. Similarly, range should be a consideration – does the system need to have a full (0-100 per cent) range, or does it need increased accuracy within a limited range - for example, 1 per cent (10,000 ppm).


The sampling method – diffusion versus flow through – are also key considerations, with flow through delivering faster response times, but requiring a pump and therefore having larger power requirements. And then there’s operational temperature and the choice of digital interface such as I2


C vs UART. Moving back to power. Traditionally, www.cieonline.co.uk


CO2


IR sensors worked by illuminating gas via filtered incandescent sources of IR light and measuring absorption via a thermopile, often requiring temperature stabilization. The power required for these sensors was significant and prevented their use in anything other than mains- powered IoT equipment.


More recently, LED-based sensors have been developed. But even using LEDs, the average power consumption is in the order of 50-150 mW, with peak of 300-400 mW. This is still too high for use in battery powered IoT systems running for several months.


Sensor design for power savings With the light source being the biggest power drain, reading frequency matters. A key consideration is therefore duty cycle. How often does the sensor need to activate? As above, CO2


isn’t toxic at even


10-times atmospheric levels. And even at 100-times atmospheric levels it still isn’t immediately dangerous to health. As such, a high-frequency sample rate is (usually) less of a priority than for – say – radon or carbon monoxide detectors.


For an application that isn’t in rapid flux, such as in-building / greenhouse air quality monitoring, every 10-15 minutes (or even less frequent) should be fine. This


isn’t true for every application and a CO2 incubator, or a wearable CO2


alarm worn


by those working in brewing and industrial settings, for example, would need to have a much higher sample frequency. By considering the application, the power source, as well as the ease of replacing/


recharging batteries the duty cycle can be tailored to improve efficiency to the point where it’s possible to run a CO2


coverage makes it a more natural choice for many applications.


sensor for


over several years on AA batteries. Of course, this isn’t the only way and the sensor’s phantom load needs to be minimized too. To achieve this, the effi ciency of a sensor’s sleep mode and wake-up cycle is critical. Also important is the optimization of the sensor’s signal processing algorithms and its additional hardware, which will reduce the computational load and further save power. By applying these techniques, it’s


possible for an LED-based NDIR CO2 sensor to run at sub 1.5 mW levels and with a peak of 0.1W. This is considerably more than an order of magnitude lower than existing LED-based sensors.


System level gains


Moving beyond the sensor, we need to consider the system as a whole in order to maximize power savings.


If we look at a typical IoT device, we will (typically) see four major building blocks: the sensor itself, a microcontroller, radio module and a power source – yes, I’m ignoring the handful of peripheral items such as a timing crystal, PMIC etc. for simplicity. And yes, there will be some systems – such as wearable sensors – where the radio is replaced with an alarm or display module.


Based on this design, however, the next biggest drain on the power budget is the wireless protocol. The choice will be dictated by the application, but LoRaWAN’s efficiency and wide-area


The MCU can also significantly affect power budgets and, with NDIR sensors being so power intensive, it’s even more vital in applications with tight constraints on power to select an MCU which has been developed for exceptionally low power functionality. This includes its sleep mode, wake up mode and general low- power design. A good example of such an MCU is ST Micro’s STM32L0.


Conclusion – what is possible CO2


sensors have existed for many years; however, their high-power consumption levels make them badly suited to IoT applications which rely on batteries or energy harvesting and are intended to run for several years.


However, our work has shown that by adapting solid-state LED sensors – via changes to the duty cycle, the implementation of more efficient algorithms, and with a power-focused system level view that assesses every on- board component – it is possible to reach the power levels required. Indeed, it is possible to deliver a sensor that draws just 60 µJ per reading while still delivering an accuracy of 30 ppm. And this sensor could therefore be powered (depending on the duty cycle) for many years via either AA batteries or the budgets available to a piezo / or solar energy harvester with a relatively small supercapacitor.


www.gassensing.co.uk Components in Electronics October 2025 45


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44  |  Page 45  |  Page 46  |  Page 47  |  Page 48  |  Page 49  |  Page 50  |  Page 51  |  Page 52  |  Page 53  |  Page 54  |  Page 55  |  Page 56  |  Page 57  |  Page 58  |  Page 59  |  Page 60