Perimeter first, interior second, and the reason is timing
Alarm sensors do one of two jobs. Perimeter devices — contacts on doors and windows, glass break detectors listening for the sound of a pane failing — tell you that somebody is entering. Interior devices — motion sensors, pressure mats, beams across a hallway — tell you that somebody is already inside and moving around.
Both are useful and they are not interchangeable, because the difference between them is measured in seconds and in feet. A perimeter trip happens at the boundary. An interior trip happens after somebody has crossed a room. Systems sold on price tend to be interior-heavy, because three motion sensors cost less than twenty contacts and cover a floor plan on a diagram. What they cost you is the warning.
If money forces a choice, contacts on everything that opens is the stronger half of the plan.
Counting the openings, including the ones you forget
This is the part that goes wrong, and it goes wrong the same way every time. People count the front and back doors, count the windows they can see from the kitchen, and stop. The list that gets missed is remarkably consistent: the side door into the garage, the door from the garage into the house, basement hopper windows at ground level behind a shrub, a bathroom window that opens wider than it looks, a pet door large enough for a person, and the sliding patio door — which is one opening in the plan and two panels in reality, one of which moves and one of which is a large sheet of glass at ground level.
Walk the outside of the house and write down every hole in it, including the ones that are painted shut, because painted shut is a maintenance state rather than a security feature.
Motion sensor coverage is a wedge, not a circle
A passive infrared sensor watches a fan of beams spreading out from a corner mounting, commonly around ninety degrees wide and reaching perhaps forty feet. The covered area is that wedge — a quarter of a circle at ninety degrees, so about 1,250 square feet at forty feet of range. Most rooms fit inside one comfortably on paper.
In the room they behave less neatly, and the reasons are all geometric. A PIR detects change across its beams rather than motion toward them, so it responds strongly to somebody crossing the pattern and much more weakly to somebody walking directly at it down a hallway. Furniture blocks beams. A sensor aimed across a room hits the far wall well before its rated range, so the specification number is rarely the working number. And the range figure assumes a target with a useful temperature difference from the background, which a warm room in summer erodes.
| Placement | What it does well | Where it fails |
|---|---|---|
| Corner, aimed across the room | Catches anyone crossing the space | Weak on approach straight down the axis |
| Facing a doorway | Anyone entering crosses the beams | Blocked by an open door leaf |
| Facing a window | Catches entry through it | Sunlight and heat sources through the glass cause trips |
| High, aimed down a stair | Everyone uses the stair eventually | Reports the problem after it has moved through the house |
| Above a heat source | Nothing | Rising warm air moves through the pattern all day |
The same geometry governs the outdoor half of the plan, which the motion light and detection layout calculator handles, and if cameras are going in on the same project the pixel density calculator decides whether their footage will be worth reviewing afterwards.
Glass break sensors and what they are listening for
A glass break sensor listens for the acoustic signature of breaking glass — typically a low frequency thump as the pane flexes, followed by the high frequency of the shatter. It has a rated radius, commonly around twenty feet, and it needs a clear acoustic path to the glass. Heavy curtains, a partition, a room turned around a corner all reduce it below the rating.
The useful property is that one sensor covers a whole room of windows rather than one per opening, which makes it economical in a room with a wall of glass. The limitation is that it does nothing at all about a window that is opened without breaking, or a pane cut or levered rather than shattered. It complements contacts; it does not replace them.
Placement is the part people get wrong: the sensor should face the glass, not sit on the same wall as it, and a room whose corner-to-corner distance exceeds twice the rated radius cannot be covered from any single point. The calculator flags those rooms.
Zones, and what the panel can actually tell you
A zone is how the panel reports what happened. One zone per device gives you the exact sensor by name, which is worth a great deal at three in the morning when the question is whether to walk downstairs. Grouping several contacts onto a zone saves panel capacity and costs you that resolution: the panel says the front of the house rather than the dining room window. Wired panels have a fixed zone count with expanders available; wireless panels have a device limit instead. Decide the grouping before pulling wire, because it is much harder to undo afterwards.
The parts that are not engineering
Two things around an alarm system are governed by local rules rather than by design, and both catch people out. Alarm permits and false-alarm penalties are set by city or county in many places, and a system that goes live without a permit can generate a bill rather than a response. What any insurer recognises, and on what terms, is a matter for your specific policy and your specific insurer, not for a general claim. Ask both directly.
The third is the boring one that does most of the work: an alarm is a detection and notification system, not a barrier. A door that resists a shove, decent lighting at the entries, and sightlines that let a neighbour see someone standing at your door all reduce the number of times the alarm ever has to do anything, which the home security guide covers in detail. An alarm that rarely has anything to report is the outcome worth designing for, and no sensor count on its own gets you there.
Questions people ask
How many motion sensors does a house need?
Fewer than most quotes include, if the perimeter is properly covered. A ninety degree sensor with forty feet of range covers roughly 1,250 square feet on paper, so a single unit handles almost any domestic room. The count is driven by which rooms you want interior detection in rather than by area, and the rooms worth covering are the ones somebody must cross to get anywhere: the main hallway, the stair, and any large open space with several ways out of it. Putting one in every bedroom mainly generates false alarms from people getting up at night. If the perimeter contacts are thorough, three or four interior sensors covers most houses.
Do I need contacts on upstairs windows?
It depends entirely on what is under them. An upstairs window above a flat garage roof, a porch canopy, a sturdy fence, a downpipe or a mature tree is a ground-level window for planning purposes. One above a sheer two-storey wall with nothing to climb is a much lower priority. Walk outside and look up at each one asking how you would get there. The other reason to contact upper windows is not entry at all: it is knowing at a glance from the panel whether the house is closed up before you leave or go to bed, which many people find is the feature they actually use daily.
Are glass break sensors better than window contacts?
They answer different questions and a thorough plan uses both. A contact tells you a window has been opened. A glass break sensor tells you a pane has been shattered, which a contact on a closed sash will not report at all. Conversely, a glass break sensor says nothing about a window that is quietly slid open, or a pane cut or levered out rather than broken. The economics favour glass break in a room with a wall of windows, where one sensor covers what would otherwise be eight contacts, and favour contacts on isolated openings. On a slider, the moving leaf wants a contact and the fixed panel is a glass-break problem.
Will my pets set off the alarm?
Standard motion sensors, yes. Pet-tolerant units are designed to ignore movement below a mass and height threshold in the pattern, and within their stated limits they work. The failure is geometric rather than a defect: anything that lifts the animal higher in the beam pattern — a sofa back, a stair, a windowsill, a counter — puts it into the part of the pattern the sensor does watch, and the tolerance is defeated. The practical approaches are to aim sensors away from the furniture an animal uses, to lean on perimeter contacts rather than interior motion, and to leave interior detection out of the rooms the animal has run of.
How many zones do I need on the panel?
It depends on how much detail you want when something trips, and that is a real decision rather than a technical one. One zone per device tells you exactly which sensor fired, which matters when you are deciding at night whether to investigate or call someone. Grouping contacts by area — all the front-of-house windows on one zone — halves the count and gives you a direction rather than a device. Wired panels have a fixed zone count with expander boards available; wireless panels have a device limit instead. Work out the total device count first, decide the grouping second, and buy the panel third, because retrofitting more granularity means pulling wire again.