The cheapest air monitors on Amazon are technically functional. They have a real laser-scatter PM module, a real NDIR-claimed CO₂ chip, a real screen. They sit on the kitchen counter showing green LEDs.
The trap is in three places.
Zero is a measurement. A £40 PM2.5 monitor that reads 0 in a quiet room at 6 AM is rarely reading zero — it is below its noise floor. Real air is never PM-zero; the global background sits at around 5 µg/m³. If your monitor reads 0, it is not telling you the air is clean. It is telling you that values below ~10–15 µg/m³ are indistinguishable from nothing. That is also where most of the interesting variation lives.
NDIR-equivalent is not NDIR. Many sub-£100 "CO₂" monitors are not running non-dispersive infrared at all. They are running a metal-oxide VOC sensor, applying a heuristic, and printing a number with "ppm" on it. The number tracks human breath roughly — but it also tracks frying bacon, spray deodorant, and a damp dog. Look for the sensor part number on the spec sheet, not the marketing copy.
Calibration drifts and nobody tells you. All gas sensors drift. Electrochemical CO drifts about 5% a year. Laser PM optics foul. NDIR CO₂ baseline-drifts in a closed room over months. The cheap monitor calibrates itself once at the factory and never again. Six months in, the readings are wrong, and there is no signal to tell you they are wrong.
What good measurement looks like
- A noise floor stated honestly. Envora One quotes ±10 ppm CO₂ and ±5 µg/m³ PM2.5 — because it cannot reliably distinguish smaller differences. Other manufacturers print three-decimal numbers and hope you don't ask.
- Auto-calibration when the conditions allow it. Background CO₂ baseline-correction during long quiet periods. PM optics that flag fouling rather than masking it.
- A signed certificate per serial number, not a marketing claim.
The number on the screen is not the measurement. The trust in the number is.