If you’ve ever looked at a VO2 max report, a resting metabolic rate printout, or a cardiopulmonary exercise test (CPET) summary, you’ve seen the letters “VCO2” sitting right next to “VO2.” Oxygen consumption gets most of the attention in conversations about metabolic testing, but carbon dioxide production is just as critical. Without an accurate carbon dioxide analyzer, you don’t have a metabolic cart you have a very expensive oxygen analyzer with a printer attached.
At AEI Technologies, we’ve spent more than 50 years building and refining the gas analyzers that exercise physiology labs, cardiopulmonary rehab clinics, sports performance centers, and research institutions rely on every day. Our CD-3A Carbon Dioxide analyzer, paired with our S-3A Oxygen analyzer, has been considered the gold standard in respiratory gas measurement for over 45 years. In this post, we’ll walk through how CO2 analyzers actually work, why their accuracy has an outsized effect on your data, and what to look for if you’re evaluating equipment for your lab.
Why CO2 Measurement Matters in Metabolic Testing
Every metabolic test whether it’s a resting metabolic rate (RMR) assessment or a full graded exercise test is built on the same underlying principle: your body consumes oxygen and produces carbon dioxide as it converts fuel into energy. By precisely measuring how much oxygen goes in and how much carbon dioxide comes out, a metabolic cart can calculate:
- VO2 (oxygen consumption) a direct measure of aerobic energy production
- VCO2 (carbon dioxide production) a reflection of substrate metabolism and, at higher intensities, buffering of metabolic acidosis
- RER (Respiratory Exchange Ratio) the ratio of VCO2 to VO2, used to estimate whether the body is primarily burning fat or carbohydrate, and to identify the ventilatory threshold during exercise testing
- Energy expenditure calculated using both VO2 and VCO2 through standard indirect calorimetry equations
Notice that VCO2 isn’t a side calculation it’s baked directly into RER, substrate utilization estimates, and energy expenditure formulas. If your CO2 analyzer drifts, responds slowly, or reads with a systematic bias, every one of those downstream numbers inherits the error. A lab can have a flawless oxygen analyzer and still produce unreliable RER values and threshold determinations if the CO2 side of the system isn’t equally precise.
How a Carbon Dioxide Analyzer Actually Works
Most modern CO2 analyzers, including AEI’s CD-3A, use non-dispersive infrared (NDIR) spectroscopy to measure the fraction of carbon dioxide in a gas sample. Here’s the basic sequence of events inside a metabolic cart:
1. Sample collection. As a subject breathes into a mouthpiece, headgear, or canopy hood, expired air is drawn into the system. In mixing-chamber systems like AEI’s MOXUS and MAX-II carts, the exhaled air passes into an active mixing chamber (a 4.2-liter chamber in AEI’s systems) where it’s thoroughly blended before a small sample is drawn off for analysis. This mixing step matters it evens out breath-to-breath fluctuations so the analyzer is reading a representative, stable sample rather than a noisy, ever-changing stream.
2. Drying the sample. Water vapor in expired air can interfere with gas analysis, so the sample passes through a drying stage AEI systems use a dual-stage Nafion dryer to remove water vapor without altering the CO2 and O2 concentrations themselves.
3. Infrared absorption. The dried sample flows through a sample cell where infrared light is passed through the gas. Carbon dioxide molecules absorb infrared light at a very specific wavelength (around 4.26 micrometers). The analyzer measures how much of that specific wavelength gets absorbed as the light passes through the sample. More CO2 in the sample means more absorption; less CO2 means more of the infrared light reaches the detector on the other side.
4. Signal conversion. The detector converts the amount of absorbed light into an electrical signal, which the system’s software translates into a CO2 concentration typically expressed as a percentage of the sample gas.
5. Calculation. That concentration, combined with expired volume (measured by a separate flow sensor or pneumotach) and ambient temperature, pressure, and humidity readings, allows the system to calculate VCO2 in liters per minute, which then feeds into RER and energy expenditure equations.
It sounds straightforward, and conceptually it is. But the engineering that separates a mediocre CO2 analyzer from an excellent one lives in the details: how fast the sensor responds to a changing concentration, how stable its baseline stays over hours of continuous testing, and how tightly its readings track known reference gases.
Why Response Time Is a Bigger Deal Than It Sounds
During breath-by-breath analysis, the CO2 concentration in expired air is changing rapidly with every single breath. A slow analyzer effectively “smears” these rapid changes together, blunting peaks and valleys in the data and making it harder to pinpoint exactly when a subject crosses a ventilatory threshold or how quickly their metabolism is responding to a change in exercise intensity.
AEI’s CD-3A analyzer is built for speed: it reaches 90% of its final reading value in less than 25 milliseconds in response to a step change in CO2 concentration. That kind of response time is what makes true breath-by-breath measurement meaningful, rather than just technically possible. If you’re running incremental exercise tests where identifying the ventilatory threshold or RER inflection point matters for athlete performance testing, cardiopulmonary diagnostics, or research analyzer speed directly affects how confidently you can identify that moment in the data.
Why Absolute Accuracy Matters More Than You’d Expect
It’s tempting to think of a 0.1% or 0.2% error in gas concentration as negligible. In practice, small errors compound. RER is a ratio of two measured values (VCO2 and VO2), so a small bias in either measurement shifts the calculated ratio and RER is frequently used to identify metabolic thresholds and substrate utilization patterns that clinicians and researchers are trying to detect down to fine margins. A CO2 analyzer with a wandering baseline or coarse resolution can make a genuine physiological threshold look blurry, or worse, invisible.
This is part of why AEI’s CD-3A analyzer is built to ±0.02% CO2 accuracy across a 0–7% CO2 range (and within 1% of reading in the 7–15% range). That level of precision, combined with excellent long-term stability, is what allows researchers to trust small, real changes in the data rather than second-guessing whether what they’re seeing is signal or sensor noise.
This same standard is why our S-3A Oxygen analyzer, rated to ±0.01% accuracy, is paired with the CD-3A in our metabolic systems the two sensors are designed to work together so that neither one becomes the weak link in your data.
Mixing Chamber vs. Breath-by-Breath: Does the Sampling Method Change the Accuracy Requirement?
There are two common approaches to sampling expired gas for analysis, and it’s worth understanding how each one interacts with your CO2 analyzer’s performance.
- Mixing chamber systems collect several breaths of expired air in a chamber before drawing a sample for analysis. This averages out the natural breath-to-breath variability that occurs even when a subject is working at a steady state, producing smoother, more stable data. AEI’s MOXUS and MAX-II systems use a 4.2-liter active mixing chamber for exactly this reason it’s a proven approach for producing dependable VO2, VCO2, and RER values, particularly during steady-state and incremental protocols.
- Breath-by-breath systems analyze each individual breath in real time, without the averaging effect of a mixing chamber. This can reveal more detail about rapid physiological transitions, but it places a much heavier burden on the analyzer’s response time and noise characteristics, since there’s no averaging to smooth out sensor imperfections. This is precisely why response time not just steady-state accuracy is such an important spec. AEI’s analyzers are fast enough to support true breath-by-breath measurement, while our mixing chamber architecture also gives labs the option of averaged, highly stable data when that’s what the protocol calls for.
Neither approach is universally “better” the right choice depends on your research question, your testing protocol, and how your field typically reports data. What matters is that whichever sampling method you use, your CO2 analyzer’s accuracy and response characteristics need to match the demands of that method. A slow, drifting analyzer will underperform in a breath-by-breath system in ways that might not be obvious in a mixing chamber system, and vice versa a mixing chamber can mask (but not eliminate) accuracy problems that would otherwise be more visible breath by breath.
Common Sources of CO2 Measurement Error
Even with excellent hardware, a handful of common issues can quietly degrade the accuracy of CO2 data in the field:
- Skipped or infrequent calibration. Sensors drift. A system that was accurate last month isn’t guaranteed to be accurate today without a fresh calibration against known reference gases.
- Leaks in sample lines or fittings. A small leak can dilute the sample with room air, artificially lowering the measured CO2 concentration and skewing RER and energy expenditure calculations.
- Incomplete water vapor removal. Residual moisture in the sample line can interfere with infrared absorption readings if the drying stage isn’t functioning properly, which is why a reliable dryer (like AEI’s dual-stage Nafion system) is a core, not optional, part of the gas path.
- Contaminated or expired reference gases. Calibration is only as good as the reference gas used. Cylinders that are past their certified shelf life, or that have been improperly stored, can introduce calibration error rather than removing it.
- Ambient CO2 interference. In poorly ventilated testing rooms, elevated background CO2 can affect certain testing configurations, particularly canopy or hood-based resting metabolic rate testing, if the system and protocol aren’t accounting for it.
- Sensor aging. Like any precision instrument, infrared CO2 sensors have a service life. A system engineered for long-term stability, and serviced on a reasonable maintenance schedule, will hold its accuracy far longer than one that wasn’t designed with longevity in mind.
Most of these issues are preventable with routine maintenance and a disciplined calibration schedule but they’re also exactly the kind of issues that separate labs that trust their data from labs that are constantly double-checking it.
Where Accurate CO2 Measurement Makes the Biggest Difference
Not every application demands the same level of precision, but several common use cases are particularly sensitive to CO2 measurement accuracy:
- Ventilatory threshold and RER inflection detection. Coaches, sports scientists, and exercise physiologists frequently use the point where RER crosses 1.0, or where ventilatory equivalents for CO2 shift, to define training zones or assess aerobic fitness. A sluggish or noisy CO2 analyzer makes that inflection point harder to pinpoint with confidence.
- Substrate utilization studies. Because RER is used to estimate the relative contribution of fat versus carbohydrate oxidation, small systematic biases in CO2 measurement can shift substrate utilization estimates in ways that matter for nutrition and metabolic research.
- Resting metabolic rate testing. RMR testing typically involves smaller absolute gas volumes and subtler changes in concentration than exercise testing, which means measurement noise or drift has a proportionally larger effect on the final energy expenditure calculation.
- Longitudinal and multi-site research. Studies that track the same subjects over time, or pool data across multiple testing sites, depend heavily on consistent, well-calibrated equipment. Analyzer drift that goes unnoticed at a single site can quietly undermine the comparability of an entire dataset.
- Clinical cardiopulmonary exercise testing. CPET is used to help differentiate cardiac from pulmonary limitations in patients, and to assess the severity of established conditions. Because clinical decisions can be informed by these results, the underlying gas analysis needs to be dependable across every test, on every patient, without exception.
Calibration: The Other Half of the Accuracy Story
Even the best-engineered CO2 analyzer will drift over time sensors age, ambient conditions change, and gas cylinders themselves have a shelf life. Regular calibration against known reference gases is what keeps an analyzer’s real-world readings matched to its rated accuracy specifications.
A well-designed metabolic cart makes this process fast and repeatable rather than a burden. AEI’s systems are calibrated using primary standard reference gases (for example, a certified cylinder with known CO2 and O2 concentrations balanced in nitrogen), and the calibration procedure is built into the system software so operators aren’t left guessing whether a calibration “took.” Some AEI systems also include an independent calibration verification feature, giving labs an extra layer of confidence that their equipment is reading accurately between full calibrations something worth asking about when evaluating any metabolic cart, not just ours.
A good habit for any lab: calibrate before each testing session (or at minimum, each testing day), track calibration values over time, and flag any analyzer that shows a pattern of increasing drift between calibrations. That trend is often the first sign a sensor needs service catching it early protects your data quality and your equipment.
What This Means for Your Lab
Whether you’re running VO2 max tests on athletes, measuring resting energy expenditure for a clinical nutrition program, or collecting CPET data for cardiopulmonary research, your CO2 analyzer isn’t a background component it’s central to the accuracy of nearly everything the report generates. When you’re comparing metabolic carts or gas analyzers, it’s worth asking vendors directly:
- What technology does the CO2 sensor use, and what’s its documented accuracy?
- What’s the response time to a step change in concentration?
- How is the system calibrated, and how often should that happen?
- What’s the sensor’s expected service life, and what does long-term drift look like?
- Is there independent published validation of the system’s accuracy against a reference method?
On that last point, AEI’s MOXUS Modular Metabolic System has been independently evaluated against the Douglas-bag method long considered the reference standard for gas exchange measurement and published research has demonstrated its accuracy and reliability across a range of exercise intensities.
The AEI Standard
For more than 45 years, our CD-3A and S-3A analyzers have been the sensors that NASA, Olympic training centers, top research universities, and hospitals around the world have trusted to produce data they can stand behind. That reputation wasn’t built on marketing it was built on decades of consistent, repeatable accuracy in labs that can’t afford to guess.
If you’re evaluating a new metabolic cart, gas analyzer, or building out a lab from scratch, our team of engineers is available to walk you through exactly how our CD-3A and S-3A analyzers stack up, and to help you find the system that fits your application and budget. Contact AEI Technologies for a personalized quote or to talk through your lab’s specific testing needs.