Non-dispersive infrared (NDIR) spectrometers are commonly used to detect gases and measure the concentration of carbon oxides (such as carbon monoxide and carbon dioxide). An infrared beam passes through a sampling chamber, where each gas component in the sample absorbs infrared light at a specific frequency. By measuring the amount of infrared absorption at the corresponding frequency, the concentration of that gas component can be determined. This technique is called non-dispersive because the wavelengths passing through the sampling chamber are not pre-filtered; instead, an optical filter is placed before the detector to filter out all light outside the wavelengths that the selected gas molecules can absorb.
The circuit shown in Figure 1 is a complete circuit of a thermopile gas sensor based on the NDIR principle. This circuit is optimized for carbon dioxide detection, but it can also accurately measure the concentration of various gases by using thermopiles with different filters.
The printed circuit board (PCB) adopts an Arduino expansion board size design and interfaces with the Arduino-compatible platform board EVAL-ADICUP360. Signal modulation is implemented using low-noise amplifiers AD8629 and ADA4528-1 and a precision analog microcontroller ADuCM360, which integrates a programmable gain amplifier, a dual-channel 24-bit Σ-Δ analog-to-digital converter (ADC), and an ARM Cortex-M3 processor.
A thermopile sensor consists of a large number of thermocouples, typically connected in series (or occasionally in parallel). The output voltage of a series thermocouple depends on the temperature difference between the thermocouple junction and a reference junction. This principle is known as the Seebeck effect, named after its discoverer, Thomas Johann Seebeck.
This circuit uses the AD8629 operational amplifier to amplify the thermopile sensor output signal. The thermopile output voltage is relatively small (from hundreds of microvolts to a few millivolts), requiring high gain and extremely low offset and drift to avoid DC errors. The high internal resistance of the thermopile (typically 84 kΩ) necessitates an amplifier with low input bias current to minimize errors, while the AD8629 has a bias current of only 30 pA (typical). This device exhibits extremely low drift over time and temperature, introducing no additional error after calibrating temperature measurements. A pulsed light source synchronized with the ADC sampling rate minimizes errors caused by low-frequency drift and flicker noise.
The AD8629 has a voltage noise spectral density of only 22 nV/√Hz at 1 kHz, which is lower than the 37 nV/√Hz voltage noise density of the thermopile.
The AD8629 also exhibits very low current noise spectral density at 10 Hz, typically only 5 fA/√Hz. This current noise, flowing through an 84 kΩ thermopile, contributes only 420 pV/√Hz at 10 Hz.
Figure 1. NDIR gas detection circuit (Schematic diagram: not all connections and decoupling are shown)
The sensor common-mode voltage, buffered by the low-noise amplifier ADA4528-1, is 200mV. Therefore, the NTC and thermopile signal outputs meet the requirements of the ADuCM360 buffered mode input: the ADuCM360 ADC buffered mode input is AGND + 0.1V to approximately AVDD - 0.1V . The CN-0338 Arduino expansion board is compatible with other types of Arduino-compatible platforms with only single-ended input ADCs.
The chopping frequency range of this circuit is 0.1Hz to 5Hz, selectable via software. A low-dropout regulator, ADP71051, generates a stable 5V output voltage to drive the infrared light source, and its switching is controlled by an ADuCM360. The ADP7105 features soft-start functionality to eliminate inrush current during cold starts of the light source.
The ADuCM360 integrates a dual-channel, 24-bit, Σ-Δ ADC that can simultaneously sample dual thermopile units over a programmable rate range of 3.5 Hz to 3.906 kHz. The NDIR system's data sampling rate is limited to between 3.5 Hz and 483 Hz for optimal noise performance.
Thermopile detector working principle
To understand thermopile, it is necessary to review the basic theory of thermocouples.
If two different metals are connected at any temperature above absolute zero, a potential difference (thermoelectric EMF or contact potential) will be generated between the two metals. This potential difference is a function of the junction temperature (see the thermoelectric EMF circuit in Figure 2).
If two wires are connected at two points, two junctions are formed (see the thermocouple connected to the load in Figure 2). If the temperatures of the two junctions are different, a net thermoelectric flux (EMF) is generated in the circuit, and current flows through it, determined by the EMF and the total resistance of the circuit (see Figure 2). If one of the wires is broken, the voltage at the break point is equal to the net thermoelectric EMF of the circuit; and if this voltage is measurable, it can be used to calculate the temperature difference between the two junctions (see thermocouple voltage measurement in Figure 2). Remember that a thermocouple measures the temperature difference between the two junctions, not the absolute temperature at one junction. The temperature at a junction can only be measured if the other junction (often called the reference junction or cold junction) is known.
However, measuring the voltage generated by a thermocouple is difficult. Assume a voltmeter is connected to the first thermocouple measurement circuit (see Figure 2 showing an actual thermocouple voltage measurement at a cold junction). The wires connecting the voltmeter create additional thermocouples at the connection points. If these additional junctions are at the same temperature (whatever the temperature), the intermediate metal rule states that they make no net contribution to the system's total EMF. If their temperatures differ, errors occur. Since each different pair of contact metals generates thermoelectric EMF—including copper/solder, Kovar/copper (Kovar is an alloy used in IC lead frames), and aluminum/Kovar (soldering within ICs)—the problem is more complex in actual circuits, necessitating extreme care to ensure that all junction pairs in the circuitry surrounding the thermocouple (except for the measuring junction and the reference junction themselves) are at the same temperature.
Figure 2. Thermocouple principle
A thermopile consists of a large number of thermocouples connected in series, as shown in Figure 3. Compared to a single thermocouple, a thermopile generates a much higher thermoelectric voltage.
Figure 3. A thermopile composed of multiple thermocouples
In NDIR applications, filtered pulsed infrared light is applied to a series active junction; thus, the junction heats up, generating a small thermoelectric voltage. The temperature of the reference junction is measured by a thermistor.
In many gases, the centers of positive and negative charges do not coincide transiently or steadily. In the infrared spectrum, gases can absorb specific frequencies, a property that can be used for gas analysis. When infrared radiation enters a gas, and the self-resonant frequency of the molecules matches the infrared wavelength, the gas molecules will resonate with the incident infrared radiation based on the energy level transitions of the atoms.
For most infrared gas detection applications, the composition of the target gas is known, thus gas chromatography analysis is unnecessary. However, if the absorption lines of different gases overlap, the system must handle the mutual interference between these gases.
Carbon dioxide exhibits absorption peaks at 4200 nm and 4320 nm, as shown in Figure 4.
Figure 4. Absorption spectrum of carbon dioxide (CO2)
The output wavelength range of the infrared light source and the absorption spectrum of water both determine the choice of detection wavelength. Water exhibits strong absorption below 3000 nm and between 4500 nm and 8000 nm. If the target gas contains moisture (high humidity), the detection gas will be strongly affected by interference in these ranges. Figure 5 shows the overlap between the absorption spectra of carbon dioxide and water. (All absorption data are from the HITRAN database).
Figure 5. Overlap of absorption spectra of carbon dioxide and water
If infrared light is applied to a dual thermopile sensor and a pair of filters are installed, with one filter having a center wavelength of 4260 nm and the other a center wavelength of 3910 nm, the carbon dioxide concentration can be measured by measuring the ratio of the voltages of the two thermopile sensors. The filter whose center wavelength overlaps with the carbon dioxide absorption wavelength is used as the measurement channel, and the filter whose center wavelength is outside the carbon dioxide absorption wavelength is used as the reference channel. Using the reference channel eliminates measurement errors caused by dust or attenuation of radiation intensity. It is important to note that carbon dioxide and water vapor absorb almost no infrared light at 3910 nm; this makes this region an ideal location for the reference channel.
The thermopile used in NDIR detection has a relatively high internal resistance, and 50Hz/60Hz power line noise can couple into the signal path. The internal resistance of the thermopile can be around 100kΩ, making thermal noise the main noise in the system. For example, the voltage noise density of the thermopile sensor used in the system shown in Figure 1 is 37nV/√Hz. To achieve the best system performance, the sensor output should have the largest possible signal, and a low gain should be used in the circuit.
The best way to maximize the signal from the thermopile sensor is to use a chamber with high reflectivity, which ensures that as much radiation as possible enters the detector without being absorbed by the chamber. Using a reflective chamber to reduce the amount of radiation absorbed by the chamber also reduces system power consumption, as it allows the use of a low-power radiation source.
Beer-Lambert Law for NDIR Gas Absorption
The infrared intensity of the measuring channel sensor decreases exponentially, a relationship known as the Beer-Lambert law:
in:
1. I represents the intensity of the emitted light.
2. I0 represents the incident light intensity.
3. k represents the absorption coefficient of a specific gas and filter combination.
4. l represents the equivalent optical path length between the light source and the detector.
5. x represents the gas concentration.
For the sensor output of the measurement channel, there is a corresponding output voltage change V0–V:
in:
1. FA represents relative absorption rate.
2. V0 represents the sensor output corresponding to the incident light intensity.
3. V represents the sensor output corresponding to the emitted light intensity.
Simplifying the formula and combining it with the previous two formulas, we get:
If k and l remain constant, FA can be plotted relative to x, as shown in Figure 6 (where kl = 115, 50, 25, 10, and 4.5 ). The FA value increases with c, but eventually saturates at high gas concentrations.
Figure 6. Typical relative absorption rates (kl = 4.5 , 10, 25, 50, 115)
This relationship indicates that, for any fixed setting, the gas concentration has a greater impact on the relative absorbance at low concentrations than at high concentrations; however, k and l can be adjusted to provide optimal absorption for the desired gas concentration range. This means that longer optical paths are better suited for low gas concentrations, while shorter optical paths are better suited for high gas concentrations.
The following describes a two-point calibration procedure, which is necessary when determining the kl constant using the ideal Beer-Lambert formula. If b = kl, then
The first step in calibration requires applying a low concentration of carbon dioxide gas (or pure nitrogen gas, i.e., 0% carbon dioxide gas) to the sensor assembly.
1. ACTLOW represents the peak-to-peak output of the sensor in the measurement channel in a low-concentration gas environment.
2. REFLOW represents the peak-to-peak output of the reference channel sensor in a low-concentration gas environment.
3. TLOW indicates the temperature of a low-concentration gas.
The second step of calibration requires applying carbon dioxide gas at a known concentration (xCAL) to the component. Typically, the xCAL concentration level is chosen as the maximum value within a concentration range (e.g., 0.5 % volume concentration for industrial air quality ranges).
1. ACTCAL indicates the peak-to-peak output of the sensor in the measurement channel when the calibration gas concentration is xCAL.
2. REFCAL represents the peak-to-peak output of the reference channel sensor when the calibration gas concentration is xCAL.
This allows us to write the following simultaneous equations containing two unknowns (I0 and b):
Solve for I0 and b in the two equations .
Then, for a gas of unknown concentration (x), where:
ACT represents the peak-to-peak output of the sensor in the measurement channel in an unknown gas environment.
REF represents the peak-to-peak output of the reference channel sensor in an unknown gas environment.
T represents the temperature of the unknown gas, measured in K.
The coefficient T/TLOW compensates for the effect of temperature changes on gas concentration (the ideal gas law is used here).
Modified Beer-Lambert Law
For practical reasons, when using NDIR, it is necessary to modify the Beer-Lambert law to obtain accurate readings, as shown below:
Because not all infrared radiation reaching the thermopile undergoes ideal gas absorption (even at high gas concentrations), a SPAN coefficient is introduced. Due to the fine structure of the filter bandwidth and absorption spectrum, SPAN is less than 1.
The changes in optical path length and light scattering require the addition of an exponential term c to ensure that the equations accurately match actual absorption data.
The values for b and SPAN constants also depend on the concentration range being measured. Typical concentration ranges are shown below:
1. Industrial Gas Quality (IAQ): 0 to 0.5 % vol . (5000 ppm). Note that the concentration of carbon dioxide in ambient air is approximately 0.04 % vol . , or 400 ppm.
2. Safety protection: 0 to 5% vol .
3. Combustion: 0 to 20% vol .
4. Process control: 0 to 100% vol .
The actual values of b and c for a specific system are usually obtained from a data point on the curve of FA versus concentration x using a curve fitting program.
For a given system where the constants b and c are determined, the values of ZERO and SPAN can be calculated using the two-point calibration method.
The first step in this process is to inject a low concentration of xLOW gas and record the following:
1. ACTLOW : Peak-to-peak output of the sensor in the measurement channel in low-concentration gas environments.
2. REFLOW : Peak-to-peak output of the reference channel sensor in a low-concentration gas environment.
3. TLOW : Temperature of a low-concentration gas, measured in Kelvin (K).
The second step of calibration requires applying carbon dioxide gas at a known concentration (xCAL) to the component. Typically, the xCAL concentration level is chosen as the maximum value within a concentration range (e.g., 0.5 % volume concentration for industrial air quality). Record the following:
1. ACTCAL : Peak-to-peak output of the sensor in the measurement channel when the calibration gas concentration is xCAL.
2. REFCAL : Peak-to-peak output of the reference channel sensor when the calibration gas concentration is xCAL.
This allows us to write the following simultaneous equations containing two unknowns (I0 and SPAN):
Solve for ZERO and SPAN in the two equations:
Then, for a gas of unknown concentration (x), where:
ACT represents the peak-to-peak output of the sensor in the measurement channel in an unknown gas environment.
REF represents the peak-to-peak output of the reference channel sensor in an unknown gas environment.
T represents the temperature of the unknown gas, measured in K.
This equation assumes that TLOW = TCAL .
The influence of ambient temperature
Thermopile sensors detect temperature by absorbing radiation, but they also respond to changes in ambient temperature, leading to increased stray and interference signals. For this reason, many thermopile sensors integrate a thermistor within the package.
Radiation absorption is related to the number of target molecules in the chamber, rather than the absolute percentage of the target gas. Therefore, absorption is described using the ideal gas law at standard atmospheric pressure.
It is necessary to record temperature data in both calibration and measurement states simultaneously.
in:
x represents the gas concentration without temperature compensation.
TLOW indicates the gas temperature at the time of calibration, in Kelvin (K).
T represents the temperature at the time of sampling, in K.
xT represents the gas concentration at temperature T.
Under the ideal gas law, in addition to the concentration changing with temperature, SPAN and FA also change slightly with temperature, which may require calibration when performing highly accurate concentration measurements.
This article does not cover SPAN and FA temperature correction, but details can be found in SGXSensortech's Application Notes 1, 2, 3, 4, and 5, and Alphasense Limited's Application Notes AAN-201, AAN-202, AAN-203, AAN-204, and AAN-205.
Thermopile Driver
Each channel of the HTS-E21- F3.91 / F4.26 thermopile (Heimann Sensor , GmbH) has an internal resistance of 84kΩ. The equivalent drive circuit for a single channel is shown in Figure 7. The internal 84kΩ thermopile resistance and the external 8.2nF capacitor form an RC low-pass noise filter with a -3dB cutoff frequency of:
Changing the C11 and C15 values of different thermopile components alters the noise performance and response time.
Figure 7. Equivalent circuit of the thermopile driver, G= 214.6
The 22-bit settling time for the step function of an 84kΩ/ 8.2nF filter is approximately:
The AD8629 non-inverting amplifier gain is set to 214.6 , and the -3dB cutoff frequency is:
The setup time for a 22-bit system is approximately:
The maximum chopping frequency of NDIR is 5Hz, therefore the minimum half-cycle pulse width is 100ms. The 22-bit setup time is approximately 0.1 times the minimum chopping pulse width.
The AD8629 has an input voltage noise of 0.5 μVp-p from 0.1 Hz to 10 Hz . Ignoring sensor voltage noise and AD8629 current noise, the thermopile's 1 mVp-p signal output has the following signal-to-noise ratio (SNR):
One thermopile is connected to the ADuCM360 ADC1/ADC3 input pins in pseudo-differential mode, and the other is connected to the ADC2/ADC3 input pins. The ADC3 input pin is connected to a 200mV common-mode voltage, driven by the low-noise amplifier ADA4528-1. The ADA4528-1 has an input voltage noise of 99nVp-p from 0.1Hz to 10Hz. To keep the ADC input pin above 0.1V , a 200mV common-mode voltage is required.
The AD8629 stage has a gain of 214.6 , while the ADuCM360's internal PGA gain is automatically set via software, ranging from 1 to 128, ensuring the input signal matches the full-scale range of the ADC input (i.e., ± 1.2V ). The peak-to-peak signal from the thermopile ranges from several hundred microvolts to several millivolts. For example, assuming a full-scale thermopile signal of 1mVp-p, a PGA gain of 4 can generate an ADC input signal of 860mVp-p.
Different thermopile sensitivities may require different gains in the AD8629 stage. Higher gain may be required if the CN-0338 Arduino expansion board is to be connected to other Arduino platforms that do not have an integrated PGA in their ADC.
The simplest way to change the gain of the AD8629 is to change R6 and R10; this will not affect the dominant pole frequency determined by R5/R8 and C9/C10.
The software allows users to select a data processing algorithm for the thermopile output. Users can choose between a peak-to-peak algorithm and a mean algorithm.
For more detailed information on signal acquisition, light source pulse timing, and temperature compensation processing algorithms, please refer to the CN-0338 source code in the CN-0338 Design Support Package and the CN-0338 User Guide .
NTC thermistor driver
The characteristics of the integrated NTC temperature sensor in the thermopile are as follows:
RTH=100kΩ
β=3940
The Thevenin equivalent circuit for the thermistor driver is shown in Figure 8. The voltage divider resistors R3 and R4 provide a 670.3mV voltage source, which is connected in series with a 103.6kΩ resistor. The drive voltage is 670.3mV - 200mV = 470.3mV .
Figure 8. Equivalent circuit of NTC thermistor driver
When RTH=100kΩ (25°C), the voltage across the thermistor is 231mV, so the PGA gain is set to 4 during measurement.
The flexible input multiplexer and dual-channel ADC in the ADuCM360 support simultaneous sampling of thermopile signals and temperature sensor signals to compensate for drift.
Infrared light source driver
The International Light Technologies MR3-1089 was selected as the infrared light source. It features a polished aluminum reflector and requires a drive voltage of 5.0V at 150mA to maximize infrared radiation and achieve optimal system performance. Heat from the lamp keeps the reflector temperature above ambient temperature, helping to prevent condensation in humid environments.
When the temperature is low (lights off), the filament has low resistance, which causes a current surge when the light is turned on. A voltage regulator with soft-start capability is useful for solving this problem.
The low-dropout regulator ADP7105 has a programmable enable pin; connecting it to the DuCM360's general-purpose input/output pins allows for switching control of the light source. The 10nF soft-start capacitor C6 provides a soft-start time of 12.2ms , which is approximately 0.125 times the minimum chopper step time of 100ms .
The lamp has a relatively large on-state current (~150mA), so careful circuit design and layout are required to prevent the lamp's switching pulses from coupling to the tiny thermopile output signal.
Carefully ensure that the lamp's return path does not flow through the sensitive thermopile sensor's ground return path. The lamp's current loop must not overlap with the processor's current loop, otherwise voltage offset errors may occur. It is strongly recommended to use separate voltage regulators for the lamp driver and the system's signal conditioning section.
The ADP7105 light source driver is powered directly by an external power supply connected to the EVAL-ADICUP360 board.
Software considerations
Synchronous chopping and sampling
To measure gas concentration, the peak-to-peak signal values in both the reference and measurement channels must be sampled. The ADuCM360 integrates two 24-bit, Σ-Δ ADCs that operate in continuous sampling mode. The ADCs are driven by programmable gain amplifiers with gain options of 1, 2, 4, 8, 16, 32, 64, and 128.
The default chopper frequency is set to 0.25Hz , and the default sampling rate is set to 10Hz. However, the chopper frequency can be set in the software, ranging from 0.1Hz to 5Hz; the ADC sampling rate can also be set, ranging from 3.5Hz to 483Hz. The software guarantees that the sampling rate is at least 30 times the chopper frequency.
For the default chopping frequency of 0.25Hz , thermopile data is acquired at a sampling rate of 10Hz within the last 1.5 seconds of the 2.5 -second cycle to ensure complete signal establishment. The first 500ms of data (blank time) is ignored. The blanking time can also be set in software, with separate settings for the rising and falling edges. Note that NTC thermistor data is acquired during the blanking period.
Calibration procedure: Ideal Beer-Lambert equation
Because lamps and thermopiles have different characteristics, the circuit must be calibrated when using them for the first time or when changing a thermopile or lamp.
It is recommended to place the entire assembly in a sealed chamber and inject carbon dioxide gas of a known concentration into it until all the original gas in the chamber is expelled. After stabilizing for several minutes, measurements can begin.
The calibration method and algorithm for the ideal Beer-Lambert equation are shown in the following steps:
1. Enter the following command: sbllcalibrate (standard Beer-Lambert calibration).
2. Inject low-concentration (xLOW) or zero-concentration gas (nitrogen) and stabilize the gas in the chamber.
3. Enter the carbon dioxide concentration at the terminal.
4. System measurement ACTLOW, which represents the peak-to-peak output of the sensor in the measurement channel in low-concentration gas.
5. The system measures REFLOW, which represents the peak-to-peak output of the reference channel sensor in low-concentration gas.
6. The system measures the temperature TLOW of low-concentration gases .
7. Inject high-concentration carbon dioxide (xCAL) into the chamber.
8. Enter the carbon dioxide concentration at the terminal.
9. The system measures ACTCAL, REFCAL, and calibrated temperature TCAL.
10. The system calculates the ZERO and b values:
To measure an unknown concentration of carbon dioxide gas using the ideal Beer-Lambert equation, please follow these steps:
1. Inject an unknown concentration of gas into the chamber and stabilize it.
2. Measure ACT, which represents the peak-to-peak output of the measurement channel sensor.
3. Measure REF, which represents the peak-to-peak output of the reference channel sensor.
4. Measure temperature T, unit K.
5. Use the calibrated ZERO value.
6. Use the calibrated b value.
7. Calculate the relative absorption rate:
Calculate the concentration and apply temperature compensation under the ideal gas law:
This step assumes TLOW = TCAL .
Note that the CN-0338 software will automatically execute steps 2 through 7.
Calibration Procedure: Correcting the Beer-Lambert Equation
If the values of constants b and c are obtained through measurement, then use the following steps.
1. Enter the following command: mbllcalibrate (corrected Beer-Lambert calibration).
2. Input constants b and c.
3. Inject low-concentration (xLOW) carbon dioxide gas (nitrogen gas) and stabilize the gas in the chamber.
4. Enter the carbon dioxide concentration at the terminal.
5. System measurement ACTLOW, which represents the peak-to-peak output of the sensor in the measurement channel in low-concentration gas.
6. The system measures REFLOW, which represents the peak-to-peak output of the reference channel sensor in low-concentration gas.
7. System temperature measurement TLOW .
8. Inject high-concentration carbon dioxide (xCAL) into the chamber.
9. Enter the carbon dioxide concentration at the terminal.
10. The system measures ACTCAL, REFCAL, and calibrated temperature TCAL.
11. System calculation of ZERO and SPAN:
To measure an unknown concentration of carbon dioxide gas using the modified Beer-Lambert equation, please follow these steps:
1. Inject an unknown concentration of gas into the chamber and stabilize it.
2. Measure ACT, which represents the peak-to-peak output of the measurement channel sensor.
3. Measure REF, which represents the peak-to-peak output of the reference channel sensor.
4. Measure temperature T, unit K.
5. Use the calibrated ZERO and SPAN values.
6. Use the previously determined values for b and c.
7. Calculate the relative absorption rate:
Calculate the concentration and apply temperature compensation under the ideal gas law:
This step assumes TLOW = TCAL .
NTC Thermistor Algorithm and Calculation
The equivalent circuit of the NTC thermistor is shown in Figure 9.
Figure 9. NTC thermistor circuit
The voltage across the thermistor is:
in:
1. VCC is 3.3V .
2. RNTC is the thermistor value.
The value of an NTC thermistor can be expressed as:
in:
1. RTH represents the thermistor value at temperature T0.
2. β is a parameter in the NTC thermistor datasheet.
3. RNTC represents the thermistor value at temperature T.
Combining the two equations above, we get:
During the chopping interval of each lamp, the ADC switches to NTC sampling, as shown in Figure 10.
Figure 10. NTC and thermopile sampling timing and lamp chopping
User interface
The EVAL-ADICUP360 platform board connects to a PC via a USB port. The board appears as a virtual COM device. Any type of serial terminal can interact with the EVAL-ADICUP360 board for development and debugging. For detailed information on software operation, please refer to circuit note CN-0338.
Figure 11 shows the relative absorbance (FA) of a typical EVAL-CN0338-ARDZ plate as a function of carbon dioxide concentration.
Figure 11. Relationship between relative absorbance and carbon dioxide concentration of a typical EVAL-CN0338-ARDZ plate .
The complete design support package for the EVAL-CN0338-ARDZ board includes layout files, a bill of materials , schematics, and source code. Please refer to www.analog.com/CN0338-DesignSupport .
The functional block diagram of the test setup is shown in Figure 12, and the photos of the EVAL-CN0338-ARDZArduino expansion board and the EVAL-ADICUP360Arduino compatible platform board are shown in Figure 13.
Figure 12. Test Setup Function Block Diagram
Figure 13. Photos of the EVAL-CN0338-ARDZ board and the EVAL-ADICUP360 board.
Summarize
The analog electronics required for NDIR measurements include precision low-noise amplifiers and high-resolution analog-to-digital converters. The circuit described in this article is a highly integrated solution that utilizes the precision analog microcontroller ADuCM360 to perform precision PGA functions, precision Σ-Δ ADC conversion, and digital control and processing.
Arduino's extended compatibility supports rapid development of NDIR design prototypes and custom software tailored to specific application requirements.
