You just painted a wall, assembled a new bookcase or gave the kitchen a thorough clean, and the air quality display on the shelf jumps from “good” to “poor.” So what is it actually reacting to? A VOC sensor does not measure one specific chemical. It responds to a whole group of volatile organic compounds at once and tells you, reliably, that the air has changed and whether there is more or less of that mix than before. It cannot tell you which substance is in the air or whether that substance is a health concern.
For everyday use, that is exactly the information you need. A reading that climbs sharply after you varnish a table and drops again once you open the windows tells you when the air is back to normal. Reading such numbers well means knowing how these sensors work, what units they report and where their limits are.
How does a VOC sensor work?
Almost every affordable indoor VOC sensor is built around a metal oxide semiconductor, usually called a MOX gas sensor. A tiny chip carries a thin metal oxide layer that a micro hotplate heats to several hundred degrees Celsius. At that temperature, gas molecules react at the surface: organic compounds are oxidized there, and the electrical resistance of the layer changes. The electronics measure that resistance and turn it into a value that rises with the amount of reactive gas in the air.
This explains the most important property of these devices: they do not tell substances apart. Ethanol from a disinfectant spray, solvents from paint and fragrance from a candle all change the resistance. Gases that are not strictly VOCs, such as hydrogen or carbon monoxide, can trigger a signal too. What you get is a sum, weighted by how strongly the chip responds to each compound, and that sensitivity is not the same for every compound.
Modern MOX chips are very small and frugal. Sensirion specifies a response time of under ten seconds for its SGP40, built-in humidity compensation and a device-to-device variation of about 15 percent. That makes a volatile organic compounds sensor fast and suitable for continuous use, but not substance-specific.
What is a good TVOC reading?
TVOC stands for total volatile organic compounds, the sum of all VOCs in the indoor air. One of the most widely cited reference schemes comes from Germany’s Federal Environment Agency (Umweltbundesamt, UBA) and its indoor air committee. It sorts TVOC concentrations into five levels. The table below lists them in mg/m³ along with a rough figure in ppb, which you need for sensors that report in parts per billion. For the conversion, Milesight’s datasheets give a factor of about 0.5 (1 mg/m³ is roughly 0.5 ppm).
| UBA level | TVOC in mg/m³ | Approx. TVOC in ppb | Assessment |
|---|---|---|---|
| 1 | up to 0.3 | up to about 150 | hygienically safe (target value) |
| 2 | over 0.3 to 1 | about 150 to 500 | still acceptable if no single-substance guide values are exceeded |
| 3 | over 1 to 3 | about 500 to 1,500 | hygienically conspicuous |
| 4 | over 3 to 10 | about 1,500 to 5,000 | hygienically concerning |
| 5 | over 10 | over about 5,000 | unacceptable |
A good TVOC reading, in other words, is 0.3 mg/m³ or below. Values up to 1 mg/m³ are still considered acceptable as long as no individual compound exceeds its own guide value. From level 3 upward, the UBA recommends looking for sources and bringing the concentration down clearly within a few months. Such readings appear mostly in new buildings and after renovations.
One caveat applies to the ppb column. Converting between mg/m³ and ppb depends on the molar mass of whatever is in the air at that moment. A sensor does not know that mix, so it calculates against a fixed reference gas. A ppb value from your sensor is therefore an estimate in “equivalents,” not the same thing as a laboratory measurement graded against the UBA scheme. The levels are still useful for orientation, as long as you read them that way.
TVOC vs VOC index: what is the difference?
Datasheets show two fundamentally different kinds of output. A TVOC sensor with a concentration output reports a value in ppb or µg/m³. That value comes from a factory calibration against a reference gas and is an estimate. A VOC index, by contrast, is a relative number with no physical unit.
Sensirion describes its index like this: the algorithm builds a moving average over the past 24 hours, called the learning time, and always maps that average to a value of 100. An index above 100 means more VOCs than the recent average, for example while cooking or cleaning. A value below 100 means fewer, for example after airing out the room. The scale runs from 1 to 500. Sensirion compares it to the human nose, which also gets used to its surroundings and mainly notices change.
That leads to a limit worth knowing: an index cannot tell you whether a room is persistently polluted. If new flooring off-gasses evenly for weeks, the algorithm learns that state as normal and the index settles back around 100. To answer “Is my air fundamentally okay?” a concentration output is the better tool. For “When should I open a window?” both work.
The Milesight AM307 uses a third approach that sits between the two: it reports TVOC as an IAQ level from 1.00 to 5.00, with a resolution of 0.01 and an accuracy of ±1 level, and Milesight refers to an IAQ rating guideline for it. When you compare two devices, check which quantity each one really reports. Otherwise you are not comparing two accuracies, you are comparing two measurement concepts. The same goes for combined monitors: TVOC and CO2 describe different things and cannot stand in for each other.
What causes high VOC levels indoors?
Most spikes trace back to a handful of typical sources. The most common include:
- cleaning products and disinfectants, especially alcohol-based sprays
- paints, varnishes, adhesives and new flooring after renovation work
- new furniture, particularly pieces made from particleboard, engineered wood and plastics
- cooking and baking, especially frying in hot oil
- scented candles, air fresheners, incense, perfume and hairspray
- solvents from hobbies and the workshop, such as thinners or spray adhesive
A cleaning spray produces a steep, short spike that is usually gone after half an hour with the window open. New furniture and flooring, on the other hand, off-gas for weeks or months and raise the baseline. That second case is hard to catch with an index alone, because the algorithm gradually learns the higher level as normal.
To narrow down a source, keep a simple log of time, activity and ventilation. If the reading rises every time you clean, the cause is fairly clear. If it rises with no obvious trigger, look at new items in the room, at the heating, since warmth makes materials off-gas faster, and at adjacent spaces like a kitchen or workshop. The sensor shows you the correlation. It does not prove which substance is responsible.
Where should you place a VOC sensor?
Placement decides whether you measure the room’s air or just the air right next to one source. For a representative reading, put the sensor where people actually spend time, at roughly three to five feet (one to one and a half meters) above the floor, on an interior wall or a shelf. Keep it away from windows, radiators, direct sunlight and ventilation outlets. In those spots it measures drafts and temperature swings rather than mixed room air.
Typical mistakes include placing it right next to a fragrance diffuser, on the counter beside the stove or above a printer. A sensor in those positions shows constant spikes that say little about the rest of the room. That is fine if you want to watch exactly that source, as long as the reading is labeled that way.
Two things are especially harmful to MOX sensors. The first is spraying cleaning products directly onto the device, which hits the chip with an extreme concentration. The second is silicone vapor, for example from freshly applied sealant or silicone sprays, which can permanently reduce the sensitivity of the sensing layer. If you are redoing caulk, take the sensor out of the room for a few days.
Finally, keep the position fixed if you want to compare trends; a moved sensor produces a new curve even when the air has not changed.
Do VOC sensors need calibration?
In everyday use, a MOX sensor is not calibrated with a test gas. Concentration outputs are set at the factory against a reference gas, while index outputs continuously adjust themselves. The algorithm gradually shifts its baseline to compensate for slow drift in the chip and for the room’s background level. With Sensirion’s index, that is the 24-hour learning time mentioned above. The Milesight AM307 has a dedicated calibration section in its configuration; which measurements it covers is described in the manual for the firmware you are running.
More important than calibration is the warm-up phase. Sensirion lists a switch-on time of one hour for the SGP40 before readings stabilize, and an index adds its learning time on top. So give a newly started or reset sensor at least a full day before judging its readings. Keep a note of restarts, firmware updates and changes to the output mode, because after any of these the curve is no longer directly comparable to what came before.
Are low-cost VOC sensors accurate?
They are accurate at what they are designed for: tracking change. They are not accurate in the sense of a laboratory analysis. The reference method works completely differently. An indoor air consultant draws room air through a tube packed with an adsorbent, usually Tenax TA, over a defined period. In the lab, the collected compounds are separated and identified one by one using thermal desorption and gas chromatography with mass spectrometry. The ISO 16000-6 standard defines TVOC as the sum of compounds eluting between n-hexane and n-hexadecane, calculated as toluene equivalent. The result is a snapshot, but it names and quantifies every compound.
A continuous VOC meter does the opposite. It measures around the clock, shows trends and costs a fraction of a lab survey, but it only delivers the sum, weighted by the chip’s sensitivity. The datasheet numbers make this clear: the AM307 specifies ±1 level on a five-level scale, and Sensirion states about 15 percent variation between individual SGP40 chips. The pattern over time, how fast a value rises and how quickly it falls again, is far more reliable than any single reading.
For offices, apartments or rental units you want to keep an eye on remotely, a VOC sensor with LoRaWAN is a practical option. The Milesight AM307 measures temperature, humidity and barometric pressure alongside TVOC, shows the values on a 4.2-inch e-ink display and transmits them over LoRaWAN to a gateway. The reporting interval defaults to ten minutes and can be set anywhere from 1 to 1,080 minutes. Milesight rates its four lithium batteries for up to three years. One common mistake when reviewing that data: between two transmissions you only see the latest value, so a short spike can vanish from the graph if the interval is long. If short events matter to you, shorten the interval for a while and see how the curve changes.
What TVOC level is considered high?
The simplest rule of thumb follows the UBA levels. If your sensor stays above 0.3 mg/m³ (about 150 ppb in equivalents) most of the time, air the room out more often. If it climbs above 1 mg/m³ (about 500 ppb) and stays there even after ventilating, it is worth searching for a source. Short peaks after cleaning or cooking are normal; what matters is how quickly the value comes back down.
With an index sensor this works differently, because there are no fixed concentrations. Here you react to clear rises above the average. Sensirion gives the example of starting an air purifier at an index of 150. On the AM307, you can work with the IAQ levels and configure the traffic light LED or the buzzer to go off when readings cross the “Bad” threshold. Milesight notes that you need to adjust the thresholds if you switch the output mode.
The most effective way to ventilate is to open windows wide for a few minutes rather than leaving them slightly ajar all day. The graph shows how long your room needs to recover. When you set up alerts, give them a minimum duration so that not every short spike triggers a notification, and treat missing data as its own state. A sensor that has stopped reporting is no proof of clean air.
What the sensor cannot do is make a health assessment for you. A high total can consist of harmless ethanol, and a moderate one can come from a problematic single compound. The curve is a good reason to open a window or take a closer look, not a diagnosis.
Where can you buy a VOC sensor?
Before you pick a VOC sensor, decide which question it needs to answer. If you only want to know when to ventilate, an index or an IAQ level is enough. If you want to follow the baseline load of a room over months, for example after a renovation, a concentration output in ppb or µg/m³ is the better choice, because an index filters out long-term pollution. Check the datasheet for the output the device really provides, and keep that label in your dashboard.
Then come the practical points: how long the sensor needs to warm up, how its baseline is tracked and whether you can set thresholds for alerts. For several rooms or buildings without an outlet where you want the sensor, a battery-powered device with LoRaWAN that sends its readings to a gateway for years is convenient. A display on the device lets everyone in the room see the values without an app.
If you are looking for a suitable VOC sensor, you can find our models in the VOC sensors category.