Why does sensor-based lighting in smart buildings switch off too soon?

Why does sensor-based lighting in smart buildings switch off too soon?

When lights switch off while a room is still occupied, the problem is rarely solved by simply increasing sensitivity. Premature shutoffs can come from short timeout settings, poor sensor placement, missed movement, or another controller issuing an off command. A sensor may be working correctly while the overall lighting sequence is not.

In smart building lighting, sensor-based controls need to match both the physical space and the way people use it. Walking through a meeting room is not a useful test of whether the lights will stay on during a quiet presentation. The practical approach is to trace what happened between the last detected movement and the lights going dark, then change only the part responsible.

Find out whether the lights were told to switch off

Before replacing anything, record the affected fixtures, the time of failure, the room activity, and how lighting returns. Does a small hand movement restore it? Does someone need to stand up? Does only the wall switch work? These observations help separate an occupancy detection problem from a control override or electrical fault.

Check the pattern across the installation:

• One fixture goes dark while adjacent fixtures remain on: inspect its driver, supply, address, group membership, and local configuration.
• An entire room switches off together: investigate the shared occupancy state, room controller, scene, or schedule.
• Several unrelated rooms go dark at the same clock time: look for a building management system schedule or central command.
• Lights go off after a repeatable interval following movement: compare that interval with occupancy hold times and controller timers.
• Lighting fades down near windows: check daylight control before assuming that occupancy detection has failed.

Where logs are available, compare the sensor event, controller occupancy state, requested light level, and reported output. Confirm that device clocks align before drawing conclusions. If the controller still requests light but the fixture is dark, move toward driver, power, or communication checks rather than extending the occupancy timeout.

Do not disable emergency lighting functions or required escape-route illumination during testing. Electrical inspection and isolation should follow site procedures and applicable local requirements.

Test the occupied position, not just the doorway

A sensor that detects someone entering a room may still miss them once seated. This is especially relevant in offices, toilets, meeting rooms, and inspection stations where occupants can remain relatively still.

Passive infrared, or PIR, sensors detect changes in infrared radiation across their detection zones. They do not continuously confirm that a motionless person is present. Movement across those zones is generally easier to detect than movement directly toward the sensor. A ceiling sensor positioned for doorway traffic may therefore perform poorly at a desk behind a partition.

Inspect the actual sightlines. Tall cabinets, suspended fittings, new partitions, and even a relocated monitor can change coverage. Compare the installed mounting height and orientation with the manufacturer's coverage diagram, paying attention to any distinction between walking-motion and minor-motion coverage. A nominal coverage radius is not a promise of reliable seated detection everywhere inside it.

Run a seated test at the reported problem location. Let the person type, read, or work normally rather than wave at the detector. Watch the sensor's detection indicator, if available, alongside the controller's occupancy status. Check whether the indicator represents raw movement, transmitted events, or a latched occupied state; those meanings differ by product.

Other sensing technologies introduce different trade-offs. Ultrasonic detection can be affected by airflow and room geometry. Microwave and some radar-based sensors may respond to movement beyond the intended boundary, depending on their design and settings. Dual-technology products may require both technologies to trigger initially but use different logic to maintain occupancy. Verify that logic rather than assuming every detection channel behaves identically.

If coverage is wrong, increasing sensitivity is not always the right repair. It can introduce corridor pickup or detection in adjacent areas. Repositioning a sensor, adjusting its orientation, or adding coverage may be more reliable.

Check which timer is actually expiring

The timeout shown in a sensor app may not be the timer controlling the lights. A room can have a sensor hold time, a controller vacancy delay, a standby dimming period, and a separate gateway automation. Their interaction depends on the system's event model.

For example, a controller receiving occupied and vacant state changes may behave differently from one expecting repeated movement events. A reporting interval or retrigger configuration that does not match the controller's expectations can allow its timer to expire even while movement continues.

Check the configured event type, reporting interval, retrigger behavior, and any suppression period. Battery-powered wireless sensors may limit reporting to conserve energy; this is not automatically a fault, but the receiving controller must be configured accordingly.

Temporarily extending the relevant timeout is a useful diagnostic test. If the complaint disappears, that supports investigating missed detection or timer coordination, but it does not prove that the original setting was the only problem. A longer delay can simply hide a coverage gap.

Set the final delay around the room's quietest normal activity and applicable energy-control requirements. A circulation area and a room used for reading should not automatically receive the same settings. Record the original values before making changes, and avoid adjusting sensitivity, timing, and grouping simultaneously.

Look for another system overruling occupancy

When a sensor still reports occupied but the lights switch off, follow command priority. Common competing sources include wall stations, booking systems, access-control integrations, demand-response functions, cleaning schedules, and building management systems.

An access event, for instance, can establish that someone passed through a door; it does not necessarily establish that a room is now empty. Likewise, the end of a calendar booking does not prove that a meeting has finished. Integrated systems need explicit rules for resolving these situations.

Check whether a manual off command creates a temporary lockout, whether an after-hours schedule overrides occupancy, and what clears each override. In manual-on, automatic-off operation, movement may not turn the lights back on after vacancy has been declared. That can be intended behavior, although it may expose an underlying detection problem.

Daylight harvesting deserves its own check. A photosensor exposed to direct sunlight or an unrepresentative bright surface can cause excessive dimming. Observe the measured light value and commanded output during the fault. Any minimum occupied output or daylight-off threshold should follow the commissioned lighting strategy, not an improvised universal setting.

Separate DALI and Zigbee faults from application logic

A communication problem does not inherently mean “lights off.” The outcome depends on device settings and the controller's failure response. Establish what the installation is configured to do when a sensor stops reporting, a gateway disconnects, or a control bus fails.

On a DALI installation

Check bus power, wiring condition, device reachability, addressing, and group assignments using suitable commissioning tools. Review recent fixture or driver replacements: restored hardware can still belong to the wrong group or retain an unintended scene value.

Where supported, compare the application's requested level with control-gear status and fault information. Also review power-on and system-failure behavior. Loss of occupancy information at an application controller is not the same event as loss of the DALI bus, so do not expect one fallback setting to cover both.

On a Zigbee installation

Inspect battery condition, last-seen timestamps, delivery diagnostics where available, and changes to the mesh. Metal partitions, relocated equipment, radio interference, or removal of a mains-powered routing device can affect message delivery.

A battery sensor may sleep normally; it should not be treated like an always-responsive powered router. Check manufacturer guidance before declaring it offline. If its local indicator detects movement but the gateway receives no corresponding event, investigate the radio path and reporting configuration.

Avoid factory-resetting devices as the opening move. A reset can remove bindings and configuration evidence while creating a second commissioning problem. Export settings and capture logs first.

Close the call with a realistic acceptance test

After a change, repeat the activity that caused the complaint. Test quiet occupancy at the weakest covered position, then verify vacancy shutoff. Include normal door positions, furniture arrangements, daylight conditions, and the relevant after-hours sequence. A quick walk-through only proves that the system detects walking.

Document the sensor location, firmware version, control group, timing values, command priorities, and exact change made. If a longer hold time remains necessary, record why. That prevents a later energy-saving adjustment from reintroducing the same fault.

For SHSS, lighting reliability sits alongside physical security and safe building operation: integration is useful only when its behavior can be understood and maintained. The service record should therefore identify the failure mechanism, not merely state “sensor adjusted.” If normal seated activity still produces no detection, correct coverage or sensing suitability before handing the room back.