Advertisement
Articles
Advertisement

Designing CO2 and alcohol-sensing applications

Tue, 04/17/2012 - 10:22am
Jason Seitz, Staff Applications Engineer, Texas Instruments

Many modern day carbon dioxide (CO2) sensing and blood alcohol detection applications utilize infrared (IR) spectroscopy technology. IR spectroscopy has increased in popularity due to its combination of exceptional accuracy, selectivity, and reliability. IR spectroscopy is based on the principle that most gas molecules absorb IR light (absorption occurs at a specific wavelength), and the amount of light absorbed is proportional to the gas concentration. This relationship is explained by the Beer-Lambert law:

I = Io * e -kcl

Where I is the transmitted IR intensity, Io is the initial intensity, k is the gas specific absorption coefficient, c is the gas concentration, and l is the length of the absorption path. Since Io, k, and l are known, you can determine the gas concentration (c) from the transmitted IR intensity (I).

There are two common types of IR spectroscopy technologies used to measure gas: dispersive and nondispersive. Dispersive IR (DIR) technology employs a prism which disperses the IR light source, and only the selected wavelength light goes through the gas sample. Alternatively, nondispersive IR (NDIR) technology passes all IR light through the gas sample and uses an optical filter to isolate the wavelength of interest (Figure 1).


Typically, a thermopile with a built-in filter is used as the detector to detect the amount of a specific gas. For example, since CO2 has a strong absorbance at a wavelength of 4.26 µm, a filter which removes all light outside of this wavelength is used. With DIR systems, you gain the ability to scan a broad wavelength. However, the tradeoff is a measurement system that tends to be larger, heavier, more complicated and costly than a NDIR system. NDIR systems are better suited for portable applications, which is the focus of this article. Along with CO2 and alcohol detection, NDIR gas sensors also can be used to detect green house gases and refrigerants such as Freon for demand control ventilation, building monitoring, automotive cabin control, and industrial safety and security applications.

Now that we have an understanding of the theory and applications associated with NDIR systems, let’s discuss some of the challenges a designer may face when designing a gas measurement system. When designing NDIR applications, system architects need to consider several design issues such as gas environment, sensor sensitivity, sensor offset voltage, and noise.

How can a gas environment impact a NDIR system design? The measured gas type, its absorption coefficient, and the gas concentration range all impact the amount of incident light on the thermopile detector, as dictated by the previously discussed Beer-Lambert law. Thermopiles used as IR detectors in NDIR systems produce a voltage (V) based on the amount of incident light they receive in Watts (W). A thermopile datasheet will specify this relation as sensitivity (V/W) in Volts per Watt. This specification can vary anywhere from about 10 V/W to over 100 V/W, depending on the type of thermopile selected. The result is in thermopile output voltages, typically in the range of 10 s µV. Therefore, you need to design supporting electronics with the ability to amplify the thermopile output voltage with different gains. This can be handled by an analog front-end (AFE) with a built in programmable gain amplifier (PGA). Gain settings in the range of hundreds to thousands of V/V are required in order to amplify the thermopile signal to full scale of the system’s analog-to-digital converter (ADC) and achieve maximum system accuracy. Having a PGA with a wide range of gain settings provides the designer with the most flexibility and enables you to measure a variety of gases at different concentrations (Figure 2). (Note: For clarity, as the design progresses, the area in red is the focus of discussion for that figure.)


Another factor in NDIR system design is in knowing how to handle the significant offset voltages associated with thermopile sensors. To understand the impact from the thermopile offset voltage, it is helpful to understand the steps of a gas measurement. Note this is one basic example, and measurement methodologies vary in complexity. First, the sensor is measured when no light is detected by the sensor. This measured signal is often called the “dark signal,” and this step or phase is called the “dark phase.” Second, the IR light is turned on and the sensor is measured with no gas in the chamber. In this phase, the transmitted light intensity is at its maximum level, and the designer can capture the full span of the system. Finally, measurements are taken with gas in the chamber; gas absorbs the IR light; transmitted light intensity goes down; and by measuring the shift from full span, the gas concentration can be deduced.

However, the thermopile is expected to have an offset component larger (up to 1 mV) than the actual signal. Therefore, during the dark phase, a significant offset component can exist, which limits the system’s dynamic range. For example, given the thermopile offset and the system gain, the output of your AFE during the dark phase sits at midscale. The thermopile is a unipolar device, so the designer can only use the range between midscale and full scale. As a result, when connecting this range to an ADC, half of your codes will be lost; in effect losing 1 bit of resolution on your ADC.

A way to minimize this problem is to integrate offset compensation into the system’s electronics. One solution is to use a digital-to-analog converter (DAC) to compensate for the measured offset. During the dark phase, the system microcontroller (µC) can capture the level of offset and remove the offset by programming the DAC to shift the output towards the negative rail, zero scale. This solution allows you to utilize the complete dynamic range of the ADC, minimizing ADC resolution requirements (Figure 3).


Also due to the thermopile’s offset voltage, you need to bias the thermopile above ground. You can do this with a common-mode generator which applies a common-mode voltage to the sensor. This level shifts the thermopile sensor signal away from the negative rail allowing for accurate sensing in the presence of sensor offset voltages (Figure 4).


In nearly every electronics application, noise can be a factor and needs to be accounted for. NDIR systems are no different and their performance can be impacted by common noise sources such as 50/60 Hz power (mains) line noise, high-frequency electromagnetic interference (EMI), and noise from the thermopile itself. Depending on the noise you suspect in your system, you may want to include a filtering scheme to help reduce that noise. Filtering could include low-pass, high-pass, or band-pass options based on the noise profile of the measurement environment. Avoid placing the filter at the PGA input because you may load the resistive thermopile, which can have a source impedance up to 100 kOhm, inducing an error voltage. Do not place the filter on the PGA output either as the output typically interfaces with a switch capacitor ADC, which should be driven by a low impedance source. This is done to avoid impacting settling performance. Ideally, place this filtering at an intermediate stage between the PGA input and output. This can be accomplished by breaking up the PGA into two stages and placing the filter in between (Figure 5).


Typically, a designer spends months searching and analyzing components, laying out and assembling a board, then validating their design to create a discrete NDIR solution. However, today Texas Instruments has developed configurable AFEs to tackle some of the more common sensor applications. For example, the LMP91050 is a configurable Sensor AFE for NDIR sensing applications that incorporates all electronics mentioned thus far onto one chip [1]. It has a two-stage PGA, 8-bit DAC for dark signal offset cancelation, common-mode generator, and supports external filtering (Figure 6). This device provides a complete analog signal path solution between the sensor and the µC.


Online design tools are also available to help system designers develop, estimate performance, and optimize their solution. These tools can help familiarize you with the chip and help reduce the pain of having to scour through a massive datasheet. An example of such a design tool is TI’s WEBENCH Sensor AFE Designer [2], It supports CO2 and alcohol detection applications associated with the LMP91050 and many other types of sensors such as temperature, pressure, gas, and pH. Pick a sensor or create your own. The tool configures the AFE accordingly and displays the estimated device performance based on the actual tester and bench data. Finally, when coupled with an evaluation system, a designer can go online, easily configure the virtual AFE to meet their needs, then load that configuration directly into the actual AFE chip via the evaluation hardware.

Summary
Gas detection systems that employ NDIR technology come with benefits as well as challenges. Benefits include a combination of excellent accuracy, selectivity, and reliability. However, obtaining these benefits is not straight forward. The system designer needs to account for the measured gas environment, sensor sensitivity, sensor offset voltage, and noise. With a proper signal conditioning circuit, you can overcome these challenges. However, it may be time consuming to research, design, and validate a discrete solution. Fortunately for today’s system designers, configurable AFE solutions such as the LMP91050 and online design tools such as WEBENCH Sensor AFE Designer [2] are available to greatly simplify the design process and help get products to market faster.

References

1. Download the LMP91050 data sheet: www.ti.com/lmp91050-ca.

2. Free access to WEBENCH Sensor AFE Designer tools: www.ti.com/webench-ca.

About the author
Jason Seitz is a staff applications engineer for Texas Instruments Precision Systems group where he works on precision, low-power and low-voltage analog systems. Seitz received his BSEE from the University of California at Davis, and his MSEE from Santa Clara University, California. Jason can be reached at ti_jasonseitz@list.ti.com.

Advertisement

Share this Story

X
You may login with either your assigned username or your e-mail address.
The password field is case sensitive.
Loading