An LED luminaire data sheet fits on one page. It lines up some twenty values, five or six of which decide visual comfort, compliance and the bill. You still need to know which ones, and under what conditions they were measured.
Most of the gaps between a data sheet and the reality of a site do not come from bad faith, but from different measurement conventions, laboratory values given without context, and indicators that are not properties of the luminaire. Here are the main ones: what they measure, what to demand, and the question to ask.
Luminous flux: the module’s flux is not the flux that lights
The lumen measures the quantity of light emitted. First ambiguity: emitted by what. Some data sheets give the flux of the LED modules, sometimes the nominal flux of the components measured on a test bench at 25 °C, and not the flux that leaves the luminaire after passing through the diffuser, the optic and the body of the fitting.
The difference is far from anecdotal: an opal diffuser, a low-luminance louvre or a lens absorbs part of the flux. The light output ratio, or LOR, expresses this ratio. The wording to look for is explicit: "luminaire output flux", measured according to the EN 13032 series or the LM-79 procedure. If the scope is not specified, the photometric file decides: it gives the flux actually emitted by the complete fitting.
Lumens per watt: a value that only holds under its conditions
Luminous efficacy relates the flux to the power. Three questions are enough to know what it is worth:
- Which flux? That of the module or that of the luminaire, with the gap described above.
- Which power? That of the LEDs alone or the system power, driver and losses included. Only the second appears on your bill.
- At what temperature? Standardised measurements are made at 25 °C ambient. Under a warehouse roof in August or in a poorly ventilated technical ceiling, you are far from it.
Efficacy also falls as the colour rendering index goes up and as the colour temperature goes down: with identical technology, a 3000 K at Ra 90 cannot show the same value as a 5000 K at Ra 80. A data sheet that gives a single figure for a whole range of colour temperatures is rounding somewhere.
Comparing two luminaires on efficacy alone only makes sense if the light falls in the right place: 130 lm/W badly directed deliver fewer useful lux on the working plane than 110 lm/W whose optic matches the height and spacing. It is the role of the photometric study to check this before purchase.
UGR is not a property of the luminaire
UGR, the unified glare rating, quantifies the discomfort caused by luminaires in the field of view. It depends on the room: dimensions, reflectances of the surfaces, position of the observer, direction of view. In other words, a luminaire on its own has no UGR.
When a data sheet states "UGR < 19", it gives at best the result obtained in a reference room with conventional reflectances. In a longer room, with dark walls or a low ceiling, the real value can fall outside the limit. What you need to ask for is the full UGR table: the value for several room geometries and two directions of view.
NF EN 12464-1 sets the limits by activity: screen workstations and precision tasks go lowest, circulation and storage areas tolerate much more. Second pitfall: a low UGR obtained with a very closed louvre is paid for in lost flux, and therefore in additional luminaires. Finally, UGR says nothing about reflections on screens or the luminance of the walls, which fall under other requirements of the same standard.
Uniformity: it judges the layout, not the fitting
Uniformity Uo is the ratio between the minimum illuminance and the average illuminance over a given area. NF EN 12464-1 requires one per zone, with requirements decreasing from the task area to the back of the room. Like UGR, it does not belong to the luminaire: it results from the mounting height, the spacing, the beam angle and the geometry of the premises. A product data sheet that shows a uniformity is in reality showing the result of a typical layout, rarely yours.
In a study report, two details deserve a glance: the calculation grid, because a grid that is too coarse smooths out the dips, and the extent of the calculated area, because a uniformity measured in the middle of the room, edges excluded, is always flattering. Warehouse aisles are the textbook case, because the racking changes everything: a narrow-beam high bay and a wide-distribution high bay give two opposite results on the same grid.
CRI, and the two indications it hides
The colour rendering index, Ra, is the average of eight pastel samples. Saturated red, referenced R9, is not among them. Two luminaires stated at Ra 80 can render red very differently, which matters as soon as there is visual inspection, painting, food processing or identification of coloured wiring. Ask for the R9 value in addition to the Ra, and aim for a higher Ra at inspection stations, where the standard requires it.
The second useful indication is SDCM, or MacAdam ellipse, which measures colour variation between fittings in the same batch. On a ceiling fitted with LED panels, a high variation shows immediately, especially after a partial replacement a few years later.
The maintenance factor: discreet and decisive
The standards do not impose an illuminance at commissioning, but an illuminance to be maintained: the level still available at the end of the maintenance period. Between the two lies the maintenance factor, the product of four terms: lumen depreciation of the sources, survival rate, dirt on the luminaire, dirt on the surfaces.
The maintenance factor is the only parameter of a study that can be changed without changing anything in the project, and that makes the number of luminaires vary. This is exactly why it must be written, justified and dated in the report.
A study calculated with an optimistic factor in a dusty workshop cleaned every three years is compliant on paper and below the values two years later. Conversely, an overly cautious factor leads to oversizing. The orders of magnitude have nothing in common between a clean office floor and a foundry or a cutting workshop. The only defensible answer is a factor justified by the environment, the protection rating of the luminaire and the cleaning interval chosen. It is one of the points settled during the lighting audit, on site.
L70, L80, B10: a lifetime without a temperature means nothing
The notation is read in two parts. The L indicates the remaining flux: L80 means 80 % of the initial flux. The B indicates the proportion of fittings that do not reach this threshold: B10 means that 10 % of the batch falls below it. A complete value therefore looks like "L80B10 at 50,000 h".
The essential is still missing: the temperature. The depreciation of an LED depends on the junction temperature, and therefore on the ambient conditions and the quality of heat dissipation. A lifetime stated at 25 °C has no relevance under a warehouse roof or near a furnace. The usable indication always includes an ambient temperature, and the LM-80 and TM-21 extrapolations must relate to the component actually used.
One last blind spot, the driver: its lifetime is most often shorter than that of the LEDs, and it is the first to fail. A data sheet that announces 100,000 hours without a word on the power supply or the availability of spare parts leaves the question open. Ranges that include lighting management and control add electronics whose warranty conditions must also be known.
No photometric curve, no comparison
The photometric file, in IES or LDT format, is the only document that makes it possible to simulate a room and check everything above: it contains the intensity distribution, the flux, the power and the test conditions. Without it, no promise can be checked.
Three quick checks: does the file correspond exactly to the reference ordered, optic and colour temperature included; is its flux consistent with that of the data sheet; are the laboratory and test date mentioned. A manufacturer who makes its files available for download in its professional LED luminaire catalogue accepts in advance to be checked.
Summary table
Indicator | What it measures | What to demand | Classic pitfall |
|---|---|---|---|
Luminous flux (lm) | Quantity of light emitted | Luminaire output flux | Flux of bare LEDs, diffuser not deducted |
Efficacy (lm/W) | Flux per watt consumed | System power, at real temperature | Bench value at 25 °C, on the module alone |
UGR | Discomfort glare | Full UGR table, then calculation in the room | A single figure without geometry or reflectances |
Uniformity Uo | Evenness of illuminance | Value per zone, grid and extent stated | Calculation limited to the middle of the room |
CRI (Ra) and R9 | Colour rendering | Ra suited to the activity, R9 and SDCM stated | Ra 80 with a very low R9 |
Maintenance factor | Loss of performance over time | Value justified by the environment and upkeep | Optimistic factor in a dusty room |
L70 / L80 / Bxx | Useful lifetime | Threshold, proportion, hours and ambient temperature | Lifetime stated without Ta and without the driver’s life |
IES / LDT file | Real intensity distribution | Dated file matching the exact reference | File missing or from another version |
The secondary indications that come back to haunt you
- IP and IK: sealing and impact resistance, to be compared with real use, high-pressure washing included.
- Flicker: since September 2024, the European ecodesign regulation has governed flicker and the stroboscopic effect. A data sheet silent on these points needs completing, especially near rotating machinery.
- Warranty: its duration matters less than its conditions, annual hours of use, maximum permitted temperature and parts actually covered.
None of this requires rare expertise: you just need to ask two or three questions before comparing prices, rather than deciding on a table that lines up lumens from different conventions. In an airport arrivals hall, this reading allowed us to go from 398 to 315 luminaires while holding the levels: consumption went from 12,197 to 3,340 euros a year, for a budget of 27,405 euros and a payback in 37 months.
A dubious data sheet on your desk, a quote to compare, a room whose levels no longer hold? Write to us: the audit and the photometric study are free, site visit and 3D modelling included. We are paid on the solution we supply and have installed, not on the engineering, and we give you a calculation you can check, zone by zone.
