Home / Blog / Homeowners
Homeowners

Why Your Motion Sensor Keeps Triggering: Fans, Curtains, Plants and Mirrors

An older Malaysian terrace living room with a small white sensor on the ceiling corner, a compact hub on a sideboard and a door sensor on the frame, with no cables run.

Answer up front — A sensor that keeps firing is usually not broken. Four things cause most of it: a ceiling fan or an air conditioner blowing on the sensor, a curtain or a plant moving in the breeze, a reflective surface such as a mirror or a glass wardrobe, and pets. Passive infrared sensors react to moving heat, so they are fooled by fans and by animals; millimetre-wave radar reacts to movement of any kind, so it reacts to curtains and plants. The fix is in this order: move or shield the sensor, shorten the detection range, reduce the sensitivity, lengthen the absence delay, and only then move the automation. Changing the automation first hides the problem and leaves the sensor reporting things that are not there.

This page is for anyone who has watched a light switch on by itself, or a fan that runs for an empty room. Almost every case is one of the four causes above, and almost every case can be fixed without changing the device you bought.

1. The four causes, in the order they are worth checking

Moving air. A ceiling fan pointed at the sensor, or an air conditioner outlet in the same wall cavity, moves the air the sensor is watching. Passive infrared sensors read that as a body crossing the field.

Moving objects. A curtain next to an open window, a plant with large leaves, a washing line, a drying rack, a ceiling fan blade itself. Millimetre-wave radar detects the movement of these objects as readily as it detects a person.

Reflective and metal surfaces. A mirrored wardrobe, a glass partition, a metal wardrobe rail, a stainless steel kitchen surface, a window frame. Radar waves bounce off these and return in a pattern that can look like movement.

Animals. A cat on the sofa at 2 m, a dog that walks past the sensor on the way out. Passive infrared sensors detect animals because they read heat, not because they identify species.

Everything below is about separating these from a genuine detection.

2. Passive infrared and radar fail in different ways

This distinction decides which of the four causes you are dealing with, so it is worth being precise.

A passive infrared motion sensor does not look at a person. It looks at changes in infrared radiation across a field of view, which in practice means it detects a moving heat source. Aqara's own motion sensor documentation describes it as detecting human motion through a moving heat source, using a Fresnel lens to divide the field into zones.

That is why passive infrared sensors trigger on a cat and on a dog, and why the recommended mounting height is 1.2 m to 2.1 m — high enough that a pet on the floor is below the field, and low enough that a person walking past fills enough of it. Aqara's manual is explicit that the detection area reduces below 1.2 m and that blind spots appear above 2.1 m.

Millimetre-wave radar does not care about heat at all. It detects movement of reflective objects in its field, whatever they are. This is why it can tell that a person is sitting still — and also why a curtain moving in the afternoon breeze reads as a person in the room.

So: if your triggers correlate with heat sources, animals or fans, you are looking at a passive infrared sensor. If they correlate with anything that moves, including curtains and plants, you are looking at a radar sensor.

3. The four controls, and which one you should reach for

Every Aqara motion and presence sensor family gives you a set of adjustment controls. The order in which you use them matters, because each one narrows the problem rather than hiding it.

Detection range. The first and most effective control. Reducing the range of a radar sensor, or reducing its detection angle, removes the part of the room where the interference is. A sensor watching a corridor does not need to see the far end of the hall.

Sensitivity. Aqara's newer radar sensors publish adjustable sensitivity levels. On the FP400, Aqara describes three levels, with high sensitivity for quiet rooms such as a study where micro-movements matter, and lower sensitivity for high-traffic areas where environmental noise would otherwise cause false triggers. The FP310 uses the same three-tier idea, with its low level aimed at filtering plant movement and curtain drift.

Absence delay. How long the sensor waits before it declares the room empty. Lengthening it stops the light flickering off while somebody is still and reading. Shortening it stops the light staying on long after the last person left. This is the control people reach for first, and it is the wrong one to reach for first — it changes the symptom rather than the cause.

Interference and exclusion zones. The FP2 allows you to define interference areas in the app and exclude objects such as plants, curtains and pets, with a filter height for ignoring fan movement or crawling pets. This is the most surgical fix available, because it tells the sensor exactly which part of the room to distrust.

Detection timeout, on the passive infrared models with a configurable timeout such as the Motion and Light Sensor P2, sets how long the sensor reports movement for after it is triggered. It is useful for a short trigger that a light needs, and useless for a trigger caused by a fan.

4. Position fixes that cost nothing

Before touching any setting, check where the sensor physically is. Most false triggers are a placement problem that a settings change then papers over.

  • Angle it away from the source. A passive infrared sensor on a wall corner should not face the ceiling fan. Radar sensors have a defined field of view — the FP2's wall-mounted coverage is published at 120 degrees horizontally, and the P2 at up to 170 degrees horizontally and 45 degrees vertically — and you can rotate or aim the device away from the interference.
  • Check the height. 1.2 m to 2.1 m is the published band for passive infrared motion sensors. A sensor at skirting height will either miss people or trigger constantly.
  • Move it off the reflective surface. A radar sensor mounted on a mirrored wardrobe is reading its own reflection.
  • Watch the doors. A sensor that covers a door picks up every time the door opens, because air and curtains move with it.

5. The learning step, and how long you should give it

Aqara's newer radar sensors describe an AI spatial learning step that scans the space and learns its background interference, identifying reflective materials such as metal and glass as sources of false reports.

The learning is a first-run behaviour, and the honest expectation is that a newly installed radar sensor will report things for a few days before it settles. The mistake people make is deciding on day one that the sensor is faulty and replacing it. Give it time in the room, and spend a commissioning session watching the app while you walk through the space and deliberately provoke the failures you have seen.

Aqara publishes no learning period and no false-positive rate for any of these models. We will not invent one.

6. When it is not a false trigger

Some triggers are real, and they are worth identifying before you spend an evening changing settings.

  • A neighbour's cat on your roof, or a dog that crosses the garden.
  • A passing car, if the sensor covers a window or an open gate.
  • A curtain or a floor mat moved by the air conditioning.
  • A second person in the house who you did not account for — a family member getting up at night is a detection, not a fault.

The diagnostic that separates these is simple and it is worth doing before anything else: check the app's activity log and see whether the trigger correlates with something you can see in the room. If you can reproduce it on purpose, you can fix it deliberately. If you cannot reproduce it at all, the sensor is probably seeing something outside its stated field, which is a placement problem.

7. A short order of operations

  1. Open the activity log and find the time of day and the direction of the trigger.
  2. Walk the room deliberately at that time and see what you can reproduce.
  3. Remove the source: aim the sensor away, move the curtain, close the door, cover a reflective surface.
  4. If the source cannot be removed, reduce the detection range so the sensor stops watching it.
  5. Then, and only then, adjust sensitivity, absence delay and timeout.
  6. Re-test over several days, including during the hottest and wettest parts of the afternoon when the air conditioning is running.

Most cases end at step three or four. Steps five and six are the ones that make a sensor stable.

What this page does not claim

  • Aqara publishes no false-positive rate, no learning period and no test result for any of these sensors, so budget a commissioning session into the job.
  • We do not claim that a presence sensor can be reconfigured into a motion sensor, or the other way round.
  • We do not claim that pet filtering eliminates every animal trigger.
  • We do not claim that changing an automation is an acceptable substitute for fixing a genuine false trigger.

Planning a project?

Tell us about your space. Our B2B team will reply within one business day with a recommended setup and quotation.

WhatsApp us →
[email protected]
+603-5880 5486