Nikkit logo NikkitSmart home, explained simply
Buying Guides

Air Quality Monitor Buying Guide for a Smart Home

The first air quality monitor I bought sat on a bookshelf in my office and told me the air was excellent for four months.

A desktop air quality monitor on a bedside table showing a carbon dioxide reading beside a smaller particulate sensor in a kitchen

The first air quality monitor I bought sat on a bookshelf in my office and told me the air was excellent for four months. The room has one window that opens about eight centimetres because of a safety catch, a door I keep shut, and me in it for nine hours a day. Excellent was not plausible. When I finally put a borrowed unit with a proper carbon dioxide sensor on the desk beside my keyboard, it read 1,480 parts per million by mid afternoon, which is the range where people reliably report headaches and dull thinking.

The cheap unit had not malfunctioned. It never measured carbon dioxide at all. It guessed, from a volatile compound sensor, and the guess was wrong in the one direction that matters. That is the whole problem with this category: two devices with identical looking dashboards can differ by a hundred pounds and measure genuinely different things, and the box rarely tells you which you are holding.

The sensor inside decides everything

There are three measurements worth paying for, and each has a real and a fake version.

Particulate matter. The real version is a laser scattering sensor, which pulls air past a laser and counts how light bounces off particles. Almost every monitor above about 35 pounds has one, and even the cheap ones track relative changes honestly. They are less accurate in absolute terms, especially below 5 micrograms per cubic metre, but if the reading jumps from 6 to 90 when you sear a steak, that jump is real. This is the one measurement where budget hardware is adequate.

Carbon dioxide. The real version is NDIR, non dispersive infrared, which shines an infrared beam through a chamber and measures how much the carbon dioxide absorbs. It costs money, it is physically bigger, and it is the single feature that separates a 40 pound monitor from a 170 pound one. The fake version is an eCO2 or equivalent reading, calculated from a VOC sensor on the assumption that human breath and human smells rise together. In a room with a scented candle, eCO2 reads high with the window wide open. In a sealed room with no smells, it reads fine at 1,600 ppm.

Volatile organic compounds. Useful as a relative signal for solvents, new furniture, cleaning products and cooking. Treat the absolute number as decoration. No consumer sensor distinguishes formaldehyde from the limonene in your floor cleaner, whatever the listing claims, and I would not pay extra for a formaldehyde label on anything under 400 pounds.

If you remember one thing from a listing, look for the letters NDIR in the specification table. If the page only says CO2 equivalent, eCO2, or estimated CO2, you are buying the guess.

What each price band actually buys

PriceTypical sensorsWorth it for
Under 40Temperature, humidity, VOC with eCO2 estimateDamp and mould watching, nothing else
40 to 90Laser PM2.5, temperature, humidity, VOCCooking smoke, pollen season, candle habits
100 to 200Laser PM2.5 plus NDIR carbon dioxideBedrooms, home offices, ventilation decisions
Over 250The above plus PM10, PM1 and radon or formaldehyde modulesA specific diagnosed problem, not general curiosity

Most households want one device from the 100 to 200 band and, if the budget stretches, one from the 40 to 90 band in the kitchen. Two monitors in two rooms tell you more than one expensive monitor in the wrong room.

Placement matters more than the extra fifty pounds

I spent six years mounting intrusion sensors for a living, and the habits transfer directly. A sensor reports the conditions where it sits, not the conditions you care about, and people put these monitors where they look nice.

Carbon dioxide needs to sit at roughly the height of your head when you are sitting or sleeping, within a metre of where you breathe. On a bedside table, not on top of a wardrobe. Not above a radiator, which drives a convection current past the sensor and biases both temperature and the gas reading. Not within half a metre of an open window or an air purifier outlet, which gives you a flattering number that describes the air leaving the room rather than the air in it.

Particulate sensors want the opposite of a shelter. They should be in open air, about a metre and a half up, away from curtains and corners where airflow stalls. In a kitchen, three metres from the hob is better than directly beside it, because beside it you learn only that frying produces smoke.

One more thing that caught me out. Several models have an automatic baseline calibration routine that assumes the room reaches outdoor levels, around 420 ppm, at some point every week. In a flat that is never properly aired, that assumption fails and the device slowly drifts low, which is the worst failure mode available because it reassures you. Check whether the model lets you disable that routine or trigger a manual calibration outdoors. The ones that do say so in the manual.

Why the single air quality score is the wrong thing to read

Nearly every app in this category reduces everything to one word and one colour. Good. Fair. Poor. It is the feature people quote in reviews and the one I would remove.

The problem is that these composite scores are weighted by the manufacturer, undocumented, and frequently dominated by whichever sensor is cheapest to include. I have watched a monitor report Good in a bedroom at 1,250 ppm carbon dioxide because particulates happened to be very low that night. The score was arithmetically defensible and practically useless, since the only actionable problem in that room was ventilation.

Read the raw numbers instead, and learn three thresholds rather than a colour. Carbon dioxide under about 800 ppm means the ventilation is keeping up. Between 800 and 1,200 is tolerable but worth a cracked window. Above 1,400 in a bedroom is the one I would act on this week. For PM2.5, a 24 hour average under 15 micrograms per cubic metre is the current World Health Organization guideline, and short cooking spikes to 150 matter far less than a persistent 40.

This is also the honest limit of what these devices do. They tell you when to open a window or run an extractor. They do not diagnose mould, and a humidity reading over 65 percent for days is a prompt to look behind furniture, not a substitute for looking. I have a humidity sensor running the bathroom fan for exactly this reason, and it does more for that room than any monitor would.

Read the data policy before the spec sheet

An air quality monitor knows when your house is occupied, when you cook, when you sleep, and how often you leave. That is an unusually complete occupancy record, and I have yet to see one of these companies describe it that way in their own marketing.

Three questions decide this for me. Does the device keep working and keep logging if the cloud account is deleted or the service shuts down, or does it become a thermometer? Is there a local API or a Matter integration, so the data can land in your own system instead of only theirs? And does the policy name a retention period, or does it say data is kept as long as necessary, which means forever.

Two brands in this category publish genuine local endpoints and one of them will export a CSV without an account. Those are the ones I recommend to family, even when the sensor list is slightly worse, for the same reason I am careful about what a smart plug knows about you. A device that only reports upstream is a subscription you have not been billed for yet.

If you want one action this week, borrow or buy an NDIR monitor, put it on your bedside table for three nights, and look at the reading at 6am. That number decides whether you need anything else at all.

FAQ

Do I need one in every room?

No. Two is the sensible ceiling for most homes: the bedroom you sleep in and the kitchen. Air quality problems are room specific and fixed room by room, so a third unit in a hallway mostly generates charts nobody reads.

Is a monitor with a display better than an app only model?

A display is worth the small premium, because a number you walk past is a number you act on. App only models work if you already have a dashboard you check, but in practice I stopped opening the app after the second week.

Will an air purifier fix a high carbon dioxide reading?

Not at all. Purifiers filter particles and some gases but do not remove carbon dioxide. Only fresh air does. A high reading is a ventilation problem, and the fix is a window, a trickle vent or an extractor.

How long do these sensors last?

The laser particulate sensors are usually rated for about eight years of continuous use and get noticeably noisier before that if the room is dusty. NDIR carbon dioxide modules last longer and drift slowly. Budget on replacing the whole unit at the five to seven year mark, which is one reason I would not spend 300 pounds. There are more honest comparisons in the rest of the buying guides.

AK
Ada Kessler

Ada installed physical security systems professionally for six years before switching to writing about the smart version of the same job. She reads a device's privacy policy before she reads its feature list.

More posts by Ada

More in Buying Guides

01doorganico 084goad 3hot8 4465456 after5internprogram ahandfulofblessings ahmadpc ahomcybersecurity allodialresources annodex antiquesandperiodcollectibles aotc asianworldkashmir beedlectrics bloggertemplatefree bluefeatheragency bodyweightsolution campaign08blog carfixadvisor catstreetcollective cbstvsports cnfanspurchasing committeeof500yearsofdignityandresistance covertopsdjs cs-kaiyuangame cuestacomunicaciontotal cxoutlet dancebetgame deskgearhub destinyvillastay detectmagazine di33 diamondsteelbuildings donmo dyanadropship emarsysfinance emimattl escapechan eskitarih evautoplay everythingweddings-orlando farmzone feedviewerapp foxstudiosscratch gamevui123 gleasongrade greetingsday hacklabasu haskell iw-kaiyun jagoanproperti jilidskripsi joltbusinessfinance juegodepokergratis juegospucca kaeru-festival latestmehndidesigns lestanzediottaviano litemagazines localcommunityhero lustrsazan mariashebets menswheels newwebstechs1 ootheday oxigenvpn plurfitness prefetial probudgetcarpetcleaning psychicmrsmartin pyaten qvantor rejoignezlogis rexzbot riselayouts ruchin rwxqsgd5 sakaltime shibiaowang shop-bot skrtips spotlesshaven strcoding tetapatrium topselller uneplanteartificielle vipertechcr webtech1223411 wo-kygaming yamaha-mekar yeknava yoybuysheets zenunwa yamaha-mekar jilidskripsi ankertoto ankertoto ankertoto