Glass-Break Detectors: Placement, Range and Testing

Why this matters

A window contact catches a window that opens. It does not catch a pane broken out and climbed through, which is the whole reason a glass-break detector is on the job. That makes the glass-break the one device whose failure shows up only during a real break-in, so the only proof it works is a proper test at the farthest pane with the room as it is when the system is armed: drapes drawn, shutters closed. A tech who claps, sees the light flash and moves on has tested nothing a burglar will reproduce.

What an acoustic glass-break listens for

A breaking pane makes a distinctive sound in two parts, and most acoustic detectors want both:

• The flex: a low-frequency thump as the impact bows the pane, before it fails.
• The shatter: the high-frequency burst of the glass breaking apart, arriving within a short window after the flex.

Detectors that require both stages, in order and within the maker's timing window, reject most household noise: a dropped pot is a thump with no shatter after it, keys and dishes are high frequency with no flex before them. That two-stage logic is also why a hand clap or a jingle of keys is not a test. At best it excites one channel and flashes an LED that some detectors light on any sound above a threshold; it tells you nothing about whether the detector would recognize a real break at the far pane. Some single-stage detectors exist; read which kind you are installing.

The detector hears glass; it does not see it. It does not catch a pane cut out quietly or removed from its glazing, and it does not catch a door or window opening. It sits alongside contacts, not instead of them.

Range, and what eats it

The maker rates each detector for a maximum range, measured to the farthest point of the protected glass, and often for a minimum distance too. The rating holds for the glass types and pane sizes the maker lists. Several things reduce effective range or take a type off the list entirely, per the maker's listing:

• Glass type. Plate and tempered glass are commonly covered. Laminated glass (two lites with a plastic interlayer), wired glass and glass with security or tint film change the break sound, and makers either derate them, list them with conditions, or exclude them. Check the instructions for each type before you count a pane as covered.
• Window coverings. Heavy drapes, lined curtains, cellular shades and closed interior shutters absorb the high-frequency shatter before it reaches the detector. The maker may give a reduced range with heavy coverings or require testing through them. What a customer leaves drawn at night is the condition that counts.
• The room. A hard room (tile, glass, bare walls) carries sound well; a soft room (carpet, upholstered furniture, acoustic ceiling) soaks it up. Same detector, same distance, different margin.
• Line of sight. The detector should have a direct acoustic path to the glass. A pane around a corner, in an alcove, or in the next room is not protected because it is "within range" on a tape measure. Mounting rules (opposite wall, adjacent wall, ceiling, and whether the same wall as the glass is permitted) are per the maker.

A living room, planned and tested

A living room 22 ft east to west by 18 ft north to south, with three panes to protect: a plate-glass picture window centered on the north wall, a tempered-glass slider in the west half of the south wall behind heavy lined drapes, and a double-hung window in the west wall. The detector goes on the east wall, centered, at the height the maker specifies.

 LIVING ROOM 22 x 18 ft, top view
 +--------[ PICTURE WINDOW, plate ]--------+
 |        x=8 ft            x=14 ft        |
 |                                         |
 [DH, west wall]                    [GB-A] east wall,
 |   y=7 to 11 ft                     centered (y=9)
 |                                         |
 |  [GB-B on ceiling, added after test]    |
 +-[ SLIDER, tempered, heavy drapes ]------+
    x=2 ft           x=10 ft

Distances to the farthest point of each pane, from GB-A at x = 22 ft, y = 9 ft:

• Picture window, far end at x = 8, y = 18: 14 ft across and 9 ft over, about 16.6 ft.
• Slider, far end at x = 2, y = 0: 20 ft across and 9 ft over, about 21.9 ft.
• Double-hung, far corner at x = 0, y = 11 (or y = 7, the same distance): 22 ft across and 2 ft over, about 22.1 ft.

Say all three sit inside the maker's rated range for plate and tempered glass on paper. Paper is the start, not the answer.

The test. First, the account goes on test with the central station, because a detector in test mode on some panels still reports its trips; see related: How to Put an Alarm Account on Test and Take It Off. Then the detector goes into test mode per the maker, usually with the maker's handheld tester, which sends an activation signal the detector recognizes. The room is set up as it is when armed: the customer said the slider drapes are drawn every night and the picture window has light sheers only.

• Picture window, 16.6 ft: tester held at the far end of the glass, aimed at GB-A, set to the glass type per the tester's instructions. Detector indicates a detection. Pass.
• Double-hung, 22.1 ft: same procedure at its far corner. Pass.
• Slider, 21.9 ft, drapes drawn: tester held between the drapes and the glass, at the far end of the slider. No detection on three tries. Drapes opened: detection on the first try.

Reading the result. Distance did not fail the slider; the double-hung is farther and passed. The drapes did. With them open, the slider is covered; with them drawn, which is how the house sits armed at night, it is not.

The fix. Three options, in order of reliability:

• A second detector with a short path to the slider. GB-B went on the ceiling about 6 ft in front of the drapes, per the maker's ceiling-mount rules, mounted from a stepladder rated for the tech plus tools and set on the floor, not on furniture, with the cable fished from the attic rather than pulled down a finished wall. Retested with the drapes drawn: detection on every try. Pass.
• Ask the customer to leave the drapes open when armed. Unreliable. People close drapes at night for a reason.
• Protect the slider's glass a different way, with a shock or vibration sensor on the frame or glass where the maker allows it. Reasonable on a single large pane; another device to maintain.

GB-B went in. Both detectors were taken out of test mode per the maker, the trips were confirmed at the central station under the right zones, and the account came off test.

Where to mount, and where not to

Placement steps belong to the job's survey, but four positions cause most problems:

• In a kitchen, or facing into one. Pans, dishes and blenders produce the frequencies the detector listens to. The two-stage logic rejects most of it; nearness and repetition defeat the margin. Mount where the kitchen noise is behind a wall or at a distance.
• Too close to the glass. Some detectors have a minimum distance; a detector inches from a pane can misjudge a break. Per the maker.
• Behind the coverings. A detector mounted inside a window bay with the drapes hanging in front of it is listening to fabric.
• Near a loud fixed source. A doorbell chime, an intercom, a speaker for a whole-house audio system, a barking dog's favorite spot. Each is a candidate for a false trip, and the event history will show it.

Testing, and why the clap persists

The habit of clapping persists because it gives an answer fast and the answer is usually "a light flashed". Name the problem plainly for newer techs:

• A clap is broadband noise with no flex stage. It does not test the two-stage logic the detector uses to alarm.
• A clap from the middle of the room does not test range to the far pane.
• A clap does not test the room as it sits armed, with the drapes drawn.

The maker's tester is designed to stand in for a real break at a known level. Use it at the farthest point of every pane, with the coverings in their armed position, and record each result. If you do not have the right tester for the detector on the wall, you cannot certify that detector; say so on the work order rather than writing "tested OK". Many detectors also have to be taken back out of test mode, by the tester or by timing out; leaving one in test mode may leave it indicating, not reporting.

When the glass changes after the install

Coverage is a property of the glass on the day of the test. Replacement windows often arrive laminated or coated, and a customer may add security or solar film to a pane years later, either of which can take that pane off the maker's list or shrink the range. A new set of blackout drapes does the same to the slider above. Put a line in the customer's handover asking them to tell you about new windows, film or heavy coverings, and recheck those panes with the tester on the next inspection.

How to verify you got this right

The survey sheet lists every protected pane with its glass type, its covering and the measured distance to its farthest point. Each pane carries a tester result taken with the coverings in their armed position. Any pane of a type the maker does not list, or that failed through its coverings, shows the alternative protection that replaced it. Every detector is out of test mode, and the central station has received a trip from each under the right zone before the account comes off test.

References

• Detector manufacturer's installation instructions (rated range, minimum distance, listed glass types and sizes, mounting positions, covering derating, test mode)
• Manufacturer's tester instructions (glass-type settings, activation, test-mode exit)
• See related: Motion Detector Coverage and the Walk-Across Rule; Door and Window Contact Types and Gap Ratings; How to Walk Test Every Zone with the Central Station