What a Lighting Control System Is Being Asked To Do

Why this matters

Most lighting control callbacks are not device defects. They are assignment problems: one device was handed two duties whose failure states point in opposite directions, and nobody wrote down which duty wins. The relay that saves energy by dropping a zone is the same relay standing between an occupant and a lit egress path. The sensor that holds a corridor on is the same sensor that decides a stairwell landing is empty. You cannot buy your way out of that with a better product. You get out of it by naming the duties, naming what each device does when it fails, and separating the pairs that conflict.

The duties are separate, and a device serves one of them well

A control system is a bundle of duties. Write them out for the space before you pick hardware:

Two duties on one device is normal and fine. Two duties whose failure states conflict on one device is the fault you are looking for, and emergency transfer is almost always one half of the conflicting pair.

The failure state is the specification

For every device in the path, answer three questions in writing: what does the zone do when this device loses power, when its network drops, and when its internal logic faults. A latching relay holds its last state; a mechanically held relay drops out; a driver at a lost 0-10 V signal goes to full output on some products and to minimum on others, and the datasheet is the only place that tells you which. A networked controller that loses its time reference reverts to a default scene, and the default scene is the single most commonly unrecorded value in a commissioned building.

A device sitting in the supply or control path of an emergency-designated luminaire has to fail to full output. That is what a listed emergency lighting control device is for, and the listing (UL 924, "Emergency Lighting and Power Equipment") binds through the National Electrical Code's listing-and-labeling rule as adopted in your jurisdiction, plus the device's own installation instructions. The electrical detail of what that path has to ride through belongs to the sibling card on what an emergency lighting circuit has to survive, and the performance criterion (floor-level illuminance along the egress path at the END of the rated duration) belongs to the sibling on emergency and exit lighting. This card owns the assignment question only.

Topology decides how far a single failure travels

• Line-voltage switching. The control device carries the load. A failure takes the whole zone. Cheap, robust, coarse. The zone is the wiring.
• Low-voltage relay control. The relay carries the load, a separate low-voltage path carries the command. A command failure is now separable from a load failure, which is worth a great deal at troubleshooting time.
• Digital addressable control. Each driver is an address. The zone is software, so it can be changed without rewiring, and it can also be changed by someone who does not know what the zone was for. A driver removed and replaced comes back with no address and joins the default group.

The rule that follows: on the two topologies where the zone is software, the as-built record has to carry the addressing, the group membership and the default behaviour, or the building cannot be maintained by anyone who was not there on commissioning day.

The dimming curve is a convention, and it travels with the number

Two published curves are in common use. Linear means measured light output is proportional to the control signal. Square law means output is proportional to the square of the control signal, which tracks perceived brightness more closely. At a 50 percent control signal the first gives 50 percent measured output and the second gives 25 percent, and on the same slider the two feel nothing alike. Name the curve in the same clause as any percentage you quote, every time, including on a submittal.

"Dims to 1 percent" is likewise three separate claims wearing one number: that the driver reaches 1 percent measured output at some stated control signal, that it starts there rather than snapping on at a higher level, and that it holds there without visible flicker. The marking answers the first. The datasheet may answer the second. Only a measurement in the installed fixture answers the third.

Worked example: the sequence of operations for one floor, filled in

An open-plan second floor, four zones, low-voltage relay control with 0-10 V dimming and a networked time schedule. Six luminaires in Zone 3 are designated emergency. Every illuminance figure below is horizontal at 30 in above finished floor, as-found on the same meter, and is not corrected for depreciation.

Zone 3 fails the check on its own row. The six designated luminaires sit behind a sensor and a dimming relay that both fail to dark, which means the space passes every occupied-hours test and fails the one event the six luminaires exist for. Nothing in the sensor is defective. The assignment is.

The correction is a re-basing of the Zone 3 control boundary, not an addition to it: the six designated luminaires come out of the sensor-controlled group and go behind a listed emergency lighting control device that senses the normal supply ahead of the relay and forces those six to full output on loss of it. They keep their dimming and their sensor behaviour during normal operation, because the listed device passes normal control through and only overrides on the sensed event.

Dimming check on the same floor, printed rather than asserted. Zone 1 drivers are marked "dims to 1 percent" on a linear curve. Measured at the 5.0 V control point: 49 percent of the full-output reading at the same grid point, 24 fc against 49 fc. That is 1 percentage point under the 50 percent a linear curve predicts, inside a meter stated at plus or minus 5 percent of reading. At the 0.5 V control point the zone visibly stepped rather than faded, and the measured output was 4 fc, roughly 8 percent of full, not 1 percent. The low end on the submittal is not the low end in this ceiling, and the number that goes in the record is 8 percent measured at 0.5 V, not the marking.

How to verify you got this right

Run these as re-derivations with the figures printed, not as ticks.

• No device serves two duties with conflicting failure states. Zone 3 as found: 1 device (the ceiling sensor) held both auto-off and the emergency-designated group, and its power-loss state is dark against a required full. Fails. After correction: 0 such devices on any of the 4 zone rows.
• Every default behaviour is recorded. 4 zones, 4 recorded loses-power states, 4 recorded loses-network states, 4 recorded normal-supply-lost states. Zone 2's default scene is written as 100 percent, which is a value someone chose, not a value the controller documents about itself.
• Every dimming percentage carries its curve. Two percentages appear above: 49 percent measured at 5.0 V on a linear curve, and 8 percent measured at 0.5 V on the same linear curve. Neither is quoted without the curve or the control voltage.
• No illuminance figure appears without its plane, its height and its basis. Four appear: 24 fc, 49 fc, 4 fc and the implied full-output reading, all horizontal at 30 in AFF, as-found, uncorrected. None of them is a maintained figure, so no light loss factor is applied to any of them.
• The emergency claim is cited, not restated. This card states the assignment rule and names the two siblings that own the circuit and the performance criterion; it states no duration and no floor-level number of its own.

Hazards this work creates

Commissioning means forcing outputs, and a forced output is the same hazard as moving the equipment by hand, reached one step upstream.

• Never force, override or place in test any device in an emergency or egress lighting path while the space is occupied. Darkening an egress route is the hazard your own command creates. Do that testing out of hours, or with the area cleared and a person with a light stationed at each end of the route.
• Do not drive a zone to minimum or off in an occupied work area without telling the people in it first and keeping the adjacent zones up; a shop floor going to 8 percent under a stated 1 percent marking is a trip and machine-guarding problem, not a lighting curiosity.
• Verifying a relay or a control transformer inside an open panel is energized electrical work. Establish an electrically safe work condition first: de-energize, lock and tag under 29 CFR 1910.333(b)(2) for electrical work on utilization equipment, and prove dead with a meter checked on a known source before and after the test (NFPA 70E-2021, 120.5, which binds through your employer's electrical safety program or your contract, not on its own). If the sequence genuinely requires a live reading, it is energized work with the boundaries and PPE your program sets, done by a qualified person.
• Operating a breaker to prove a zone's supply is done at the handle with the deadfront on, standing to the hinge side.

References

• 29 CFR 1910.333(b)(2), electrical work on utilization equipment conductors, which 29 CFR 1910.147 expressly excludes from its scope
• NFPA 70E-2021, 120.5, as adopted through an employer electrical safety program or by contract
• NFPA 70 (National Electrical Code), Articles 110 and 700, in the edition your authority having jurisdiction has adopted
• UL 924, Emergency Lighting and Power Equipment, which binds through the listing-and-labeling rule as adopted and through the device's own instructions
• See related: What an Emergency Lighting Circuit Has To Survive; Why Occupancy Sensing Fails in Predictable Places; What Daylight Harvesting Needs To Work