As an Amazon Associate, Akermin earns from qualifying purchases. Product links on this page are affiliate links.
Accuracy on an air quality monitor is a per channel property, not a device rating, because each reading on the screen comes from a different sensor working on a different physical principle. A monitor can carry a defensible carbon dioxide figure and an essentially unspecifiable formaldehyde figure in the same case, and the listing will usually quote one percentage covering both. The channels that support a real accuracy claim are carbon dioxide from an NDIR sensor and temperature and humidity; particulate mass is an estimate derived from scattered light, and total volatile compounds and formaldehyde are indices expressed in concentration units.
Accuracy is a per channel claim, not a device rating
Open any ten channel monitor and you find a cluster of unrelated parts. Carbon dioxide comes from an infrared bench with a lamp, a chamber and a detector. Particles come from a laser diode, a photodiode and a fan. Formaldehyde and volatile compounds come from small cells whose resistance or current changes in the presence of reducing gases. Temperature and humidity come from a single integrated chip. The only thing those four have in common is the enclosure.
| Channel | Sensing method | What a real specification states | What the reading supports |
|---|---|---|---|
| Carbon dioxide | Non dispersive infrared absorption | Range, plus an error as an absolute figure and a percentage of reading | Absolute values, once calibrated against outdoor air |
| PM2.5 and PM10 | Laser light scattering, mass inferred from counts | Size range, counting efficiency, humidity handling | Relative change in one location |
| Particle counts by size | Optical counting into size bins | Bin boundaries and flow rate | Comparison of bins and trends |
| Formaldehyde | Electrochemical or semiconductor cell | Reference gas, cross sensitivity, drift rate | Direction of change only |
| Total volatile compounds | Metal oxide semiconductor | Calibration gas and the equivalence used | Direction of change only |
| Temperature and humidity | Integrated silicon sensor | Error in degrees and percent relative humidity | Absolute values |
Listings do sometimes name the parts, which is a good sign even when no numbers follow. The LifeBasis eleven in one detector states an NDIR infrared carbon dioxide sensor, a laser particle sensor, a semiconductor sensor, a photoelectric sensor and a temperature and humidity sensor, which is an unusually clear account of what is inside. No per channel error figure is stated in the listing bullets, so the parts list is the specification.
What an accuracy figure has to say to be checkable
A figure you can act on names four things: the quantity, the range over which it applies, the error in both absolute and proportional terms, and the conditions. Miss any of them and the number cannot be tested. The form that appears on well specified carbon dioxide instruments reads something like plus or minus 50 parts per million or 5 percent of the reading, whichever is greater, over a stated range and temperature band. That phrasing is doing real work, because it admits that a fixed offset dominates at low concentrations while proportional error dominates at high ones.
Compare that with a single figure spanning everything. The ZEMEOLLO ten in one tester states an error margin of less than plus or minus 5 percent while monitoring carbon dioxide, formaldehyde, total volatile compounds, benzene, PM2.5 and PM10. One percentage covering six chemically unrelated channels, with no reference method, no range and no conditions, cannot be verified in either direction. It is not necessarily wrong about any particular channel; it is simply not a specification.
Partial publication is more common than either extreme. The Temtop M2000C and T1 bundle states a temperature accuracy of plus or minus 0.5 degrees Fahrenheit and a humidity accuracy of plus or minus 3 percent relative humidity for the thermometer, and states data export to CSV over USB for the monitor, which is the feature that makes any accuracy question answerable at all. A carbon dioxide accuracy figure for the monitor itself is not stated in the listing bullets. That combination is typical: the cheapest channel to specify gets a number, the expensive one does not, and the export function is what lets you check the behaviour yourself over time.
The channels where the number is close to meaningless
Two readings on a multi channel display deserve a mental asterisk every time you look at them. The first is total volatile organic compounds. A metal oxide semiconductor sensor heats a film whose electrical resistance falls in the presence of reducing gases, and it cannot distinguish between them. Firmware converts that resistance into a concentration by assuming the gas present is the one the part was calibrated with, often ethanol or isobutylene. A reading in milligrams per cubic metre is therefore an equivalent of a reference gas, not a measurement of the mixture in your room. Spraying cleaner near the unit sends it to the top of its scale, which is a correct response to a real event and a wrong number for it. Our page on what TVOC means on an air quality monitor works through the conversion in more detail.
The second is any channel claiming a specific named compound on a device that also claims total volatile compounds. Formaldehyde cells have genuine selectivity, but low cost versions are strongly cross sensitive to alcohols and share the same sensing element family as the general channel. Benzene is harder still: separating benzene from toluene and xylene requires chromatography, so a benzene figure on a handheld monitor is derived from the same broad response, not measured. Treat it as a relabelled index.
None of this means the channels are worthless. A volatile compound trace that climbs every time the printer runs, or stays elevated for weeks after new flooring, is telling you something true and actionable. The error is reading the digits as a dose.
Calibration is the half of accuracy you control
Factory calibration sets the starting point; everything after that is drift, and drift is the part you can manage. Four habits do most of the work.
- Reset the carbon dioxide baseline against outdoor air using the unit’s documented routine, rather than relying on an automatic baseline that assumes your room empties regularly.
- Let the device stabilise. Infrared benches and heated semiconductor films need warm up time, and a reading taken in the first minutes after power on belongs to the sensor, not the room.
- Note the humidity alongside any particle reading. Hygroscopic particles swell above roughly 70 percent relative humidity and scatter as though they were larger, inflating the mass figure.
- Keep the inlet clear and clean. A fan drawing through a dusty grille reads low because the flow has fallen, and a unit sitting in a cabinet reads its own microclimate.
Comparison is the cheapest sanity check available, provided you are honest about what it shows. Two units of the same model agreeing closely proves repeatability within one design. Two different technologies agreeing in the same place is better evidence. Our reviews of the Vetouch monitor, the Sculpriv monitor and the HSCLPM monitor record what each listing does and does not publish, which is the information you need before deciding what a side by side result proves.
Reading a specification sheet before you trust a reading
Work down the channels you actually care about and ignore the channel count on the box. For each one, ask whether the listing names the sensing method, gives a range, gives an error in a form that includes both an absolute and a proportional term, and says how calibration is performed. A device that answers those four questions for carbon dioxide and temperature, and stays silent on formaldehyde, is more trustworthy than one that quotes a tidy percentage for everything.
Then decide what accuracy you need. For ventilation decisions, a carbon dioxide channel within a hundred parts per million is plenty, because the thresholds are hundreds of parts per million apart. For deciding whether a filter is working, relative change on a particle channel is enough and absolute accuracy is irrelevant. For anything where a number will be quoted to somebody else, an export function and a documented calibration routine matter more than the headline tolerance, because they are what let you show your working.
Frequently Asked Questions
Is a single accuracy figure for a whole monitor meaningful?
No. A multi channel monitor holds several unrelated sensors, and an error figure that applies to all of them at once has no physical meaning. A carbon dioxide channel can be specified against a reference gas while the volatile compound channel on the same board cannot be specified at all, so one combined percentage is a marketing simplification.
What does plus or minus 50 ppm or 5 percent of reading mean?
It means the tolerance is whichever of the two is larger at the reading you are taking. At 600 ppm, 5 percent is 30 ppm, so the 50 ppm floor applies. At 2,000 ppm, 5 percent is 100 ppm, so the percentage applies. Specifications written this way are honest about the fact that absolute error dominates at low readings and proportional error dominates at high ones.
Why is percent of full scale worse than percent of reading?
Because the error stays the same size as the reading shrinks. A channel spanning 0 to 5000 ppm with an error of 3 percent of full scale carries plus or minus 150 ppm everywhere, which is a reasonable tolerance at 4,000 ppm and useless at 450 ppm. Always check which of the two a figure refers to.
Can a consumer monitor measure formaldehyde properly?
Not to a standard you could defend. Low cost formaldehyde channels use electrochemical or semiconductor cells that respond to a family of reducing gases, including alcohols from cleaning products and perfume. The number moves for the right general reason and cannot be read as a specific formaldehyde concentration. Laboratory measurement uses sampling onto a cartridge followed by chromatography.
Does agreement between two monitors prove they are accurate?
It proves they are consistent, which is not the same thing. Two units of the same model share the same calibration assumptions and the same systematic bias, so they can agree closely while both reading high. Agreement between different technologies in the same place is more informative, and only a reference instrument settles the matter.
Which channels are worth paying attention to?
Carbon dioxide from a true NDIR sensor, and temperature and humidity, because those are specified against established references and drift slowly. Particulate mass is useful for relative changes in one place. Volatile compound and formaldehyde channels are best read as trend indicators that something in the room has changed.