There is a small white wedge in the upper corner of almost every commercial room in this province. It is the cheapest device in the alarm system and the one the system leans on: if it does not report, nothing else in the chain gets a chance to.
Nearly everyone who owns one believes it watches the room. It does not. It watches a specific pattern of wedges of space, and reports one thing only — that the infrared arriving from one of those wedges changed relative to its neighbours.
That distinction is why a detector can be working perfectly, pass every self-test, show healthy on the panel, and still cover rather less of your building than the drawing says. This article covers how the device works, the four ordinary conditions that shrink its real coverage, and the one check that finds out — which almost nobody performs at the moment it matters most.
What "passive infrared" means
The P is the important letter. A PIR detector is passive: it emits nothing. It does not sweep the room with a beam and there is nothing to bounce back. It sits there and absorbs the infrared radiation that every object in its field of view emits continuously as a consequence of having a temperature.
Behind the plastic window sits a sensing element that responds to changes in the infrared falling on it. In front of that element sits a lens — the faceted, segmented plastic cover you can see if you look closely — which does the real design work. Rather than focusing the whole room onto the sensor as one smooth image, the lens carves the space in front of the detector into a fan of separate zones, alternating between regions the sensor can see and gaps it cannot.
The result is not a picture. It is closer to a set of tripwires made of light you cannot see, splayed across the room like the ribs of a fan.
Now imagine a warm object crossing that fan. It leaves one zone, passes through a blind gap, enters the next. The sensor sees the infrared arriving rise, fall, then rise again. That alternation is the signal, and the processor is looking for exactly it: a change of thermal energy, of a particular size and speed, moving from one zone into another.
Which gives the sentence this article rests on: a PIR detector does not detect heat, and it does not detect motion. It detects a moving difference in heat crossing between the zones of its pattern. Each of the four conditions below is a case where that specific event fails to happen cleanly, even though a person is unmistakably in the room.
None of this is a defect, and none of it is fixed by buying a more expensive detector. It is what the physics of the device is — manageable, but only by someone who knows the four conditions exist.
Condition one: the geometry of the approach
Because the signal is a crossing between zones, the direction of travel through the pattern matters.
Movement that cuts across the fan sweeps through zone after zone, generating the strong alternating signal the detector is built around. Movement that runs along the axis of the pattern — directly towards the detector or away from it — stays inside one zone far longer and produces a weaker, slower change. The detector is not blind to it. It is working with the least favourable version of its input, and may take more travel before it decides something has happened.
For an owner, this is a siting consideration, not a mystery. It is the reason a detector should look across the paths people take through a room rather than straight down them, and the reason a device aimed squarely along a long corridor is doing its hardest possible job.
Walk your building with one question in mind: for each detector, what is the natural line of travel through that space — a doorway to a desk, an aisle, a stairwell landing — and does that line cross the detector's fan or run down its length? Where a corridor must be covered, mounting the device so its pattern lies across the corridor converts the least favourable geometry into the most favourable, and it is a bracket-and-a-screwdriver change, not a redesign.
Condition two: not enough thermal contrast
The device measures a difference. A person registers because they are warmer than the wall, floor and air behind them.
Narrow that gap and you narrow the signal. As the ambient temperature of a space climbs towards the temperature of a human body, the same person crossing the same pattern produces a smaller change at the sensor, and the device's response becomes correspondingly less certain.
This has a practical BC shape, and not the one people expect. Our climate rarely produces a building at body temperature — but individual rooms reach it routinely. A west-facing office under afternoon glazing in July. A communications room with the cooling fighting the load. A kitchen or production space with process heat. An uninsulated mezzanine above a warehouse floor in a heat wave. A metal roll-up door still holding the afternoon.
A demonstration in February tells you nothing about any of these. If a detector protects a space that gets genuinely hot, test it in the hot part of the year.
The same effect has a second-order version: an air handler, a unit heater, or direct sun tracking across the field of view all push moving thermal energy through the pattern themselves. That shows up as nuisance activations rather than misses — but nuisance activations are how a zone ends up permanently bypassed by somebody tired of it, at which point coverage is not reduced, it is gone. Treat a regularly bypassed zone as an open fault, not a habit.
Condition three: something is in the way
This is the most common cause of lost coverage and the cheapest to fix. Nobody looks for it, because it does not arrive as a fault. It arrives as ordinary business.
A PIR detector needs a clear line from its lens to the space it is meant to cover. Anything solid in that line does not soften the coverage behind it — it removes it. And the pattern spreads with distance, so an obstruction close to the detector removes a far larger share of the room than its own size suggests. A pallet stacked shoulder-high near a corner-mounted device can take out a substantial wedge of the floor beyond.
The list of ordinary things that do this is short and predictable:
- Stock and inventory. Seasonal peaks are the classic case — the racking that was chest-high at commissioning is ceiling-high in November.
- Shelving, racking and cabinets added after handover. The system was designed against a floor plan. The plan changed and nobody re-read it against the detectors.
- A tenant fit-out or a partition. New walls, new millwork, new display units.
- Plants. Large ones block. Ones that move in an air current cause activations. Ones in front of a window do both.
- Signage, banners and seasonal decoration, hung in good faith from the ceiling and wall positions detectors occupy.
- Stacked boxes against a wall — the temporary arrangement that has been there for two years.
None of these generates an alert, a trouble condition or a line on a report. The panel is satisfied. The detector is working exactly as designed, on the reduced volume of space it can now see.
Condition four: glass and solid objects stop infrared
This is the one that surprises people most, and the most absolute of the four.
Infrared in the band these devices work in does not pass through ordinary glass the way visible light does, and does not pass through walls, doors, partitions or millwork at all. So a PIR detector cannot see through a window, a glazed partition, a display case, a glass office wall or a glass door.
The assumptions this rules out are common ones:
- A detector inside a room, aimed at a window, is not covering the parking area outside. It is covering the window.
- A detector whose pattern crosses an interior glazed partition — increasingly normal in modern offices — covers the side it is on and nothing beyond.
- A device behind a protective glass cover, inside a display case, or looking through a glazed vestibule is doing markedly less than it appears to.
- The reverse is useful: a PIR detector is not triggered by a passer-by on the other side of your storefront glass. The glazing that limits it also insulates it from the street.
The same property is what makes interior doors matter. A door left closed turns a covered area into an uncovered one; a door propped open can extend the pattern somewhere it was never meant to reach — occasionally onto a warm exterior surface or a heater, which is how a stable installation starts producing false activations after a change nobody connects to the alarm.
This same physics appears elsewhere in security, and the two ideas are worth linking rather than confusing. A thermal camera also senses emitted infrared and is also stopped by glass — but it builds a continuously updated image of a scene, while a PIR detector produces a single yes-or-no output from a change signal. Different instruments, one shared piece of physics. If that distinction bears on a decision you are making, thermal is not night vision sets out what a thermal camera can and cannot deliver. Axis puts the underlying physics plainly: "Conventional cameras work in the range of visible light, which is radiation with short wavelengths. Thermal cameras, on the other hand, detect radiation with mid-length or long wavelengths, which is called infrared radiation."
The check that finds all four: the walk test
Every condition above is invisible from the panel and obvious on the floor. There is one procedure that catches all of them, it takes minutes, and your system almost certainly already supports it.
A walk test puts detectors into a mode where each activation is indicated locally — commonly by the detector's own LED — so one person can walk the space and watch the device respond. Your installer can enable it, and on many systems a manager can. Then walk the room the way people actually move through it: in through the door people use, down the aisle, behind the racking, into the corner by the stairwell, along the glazed side. Watch where the device answers and, more importantly, where it does not.
The part that matters is a scheduling point rather than a technical one.
The walk test that gets done is the one at commissioning. The one that gets skipped is the one after the change. The fit-out that moved the shelving, the tenant improvement that added the glass partition, the stock build that raised the racking, the reorganisation that turned an open floor into cubicles — those are the moments coverage actually changed, and the moments nobody thinks about the alarm, because none of them are alarm work. They appear on a facilities schedule, not a security one.
So make the trigger a rule about your building rather than a date on a calendar: if the layout of a protected room changed, walk-test that room. Ten minutes, and it either confirms the coverage or shows you a hole that has been open since the trades left.
Write down two things while you are in there: which zones are bypassed regularly and why, and whether anything has been mounted, stacked or planted in front of a detector since the last check. Both are answerable by looking, and both beat anything the panel tells you.
The short version
A PIR detector is not a camera and does not watch a room. It reports a moving difference in heat crossing between the zones of a fixed pattern, and four ordinary conditions weaken that event: an approach along the pattern rather than across it, a space warmed close to body temperature, anything solid in the line of sight, and glass or partitions, which stop it completely.
None of the four is a fault, none is fixed by spending more, and none will appear on your panel. All four are found by one person walking the room and watching the light on the device — and the walk test worth having is not the one at handover, but the one after the day the shelving moved.
