To read an LED lighting datasheet, start with the test behind it. Find the photometric report the numbers came from, then check each line against that report, and treat any figure without a measurement as a claim. A datasheet is a summary written by the seller. The report is the evidence, and you only know what the summary is worth once you have seen both.
Two “200 W flood light” datasheets can print the same lumens: one measured on the complete fixture, the other a chip rating multiplied by chip count. Below are the 20 specs that matter, a red-flag table, a line-by-line reading of one of our own measured reports, and an RFQ checklist.
Key Takeaways
- A datasheet number only means something if you can trace it to a measurement of the complete luminaire, with the test ambient stated. LM-79 testing is done at 25 °C ± 1 °C ambient.
- Luminaire lumens and chip lumens are different quantities. Summed chip output always runs above measured fixture output, so a datasheet that does not say which one it quotes has told you which one it is.
- Efficacy must be measured flux divided by measured input power on the same sample. You can check it with a calculator in ten seconds.
- Power factor, THD and surge rating are driver specifications. They decide how the fixture behaves on your supply, and they rarely appear on brochures.
- A supplier that publishes a modest measured figure is more useful to you than one quoting a perfect number for every wattage. Our measured 150 W flood returned 142.3 lm/W against 175.4 lm/W for the 200 W build, and both reports are public.
LED Datasheet Explained: Check the Evidence Before the Numbers
Every figure on a datasheet belongs to one of three types:
| Type | Example wording | What it is worth |
|---|---|---|
| Measured | ”35,309 lm, goniophotometer, 25.3 °C ambient” | Evidence, if you can see the report |
| Rated / nominal | ”200 W, 160 lm/W” | The design target. Useful for grouping products, not for calculation |
| Derived / marketing | ”Up to 40,000 lm”, “equivalent to 1000 W” | A claim until someone produces the measurement |
Before reading any line, find out four things.
1. Which sample was measured. A report should name the model code, and that code should match your quotation. A report for a neighbouring wattage or optic describes a different product.
2. How it was measured. LM-79 is the IES method for complete solid-state luminaires, tested as sold in a sphere or on a goniophotometer, with driver, optical and thermal losses included.
3. Who measured it, and when. Instrument, date and laboratory make the result traceable. Ask every supplier, including us.
4. The test ambient. LM-79 specifies 25 °C ± 1 °C. LED output falls as temperature rises, so a flux figure without a temperature cannot be compared.
Light Output: Specs 1–4
1. Input power: rated versus measured
“200 W” is a nominal class. Measured input power, reported with voltage and current, is what the meter bills and the cables carry. Calculate efficacy from the measured figure: a fixture drawing 185 W but sold as 200 W looks unusually efficient if you divide by the wrong number.
2. Luminous flux: luminaire lumens, not chip lumens
This is the line where most confusion arises. Chip lumens are the LED package rating multiplied by the number of packages. That ignores lens and cover losses and the heat inside a sealed housing. Measured whole-fixture output always lands below the summed chip figure, by an amount that depends on the optic and the thermal design, and lumens vs lux vs candela explains the mechanism. Look for the words “luminaire”, “fixture” or “total luminous flux” next to a measurement reference. “Up to” in front of a lumen figure is a claim.
3. Efficacy (lm/W)
Efficacy is flux divided by input power, both from the same complete sample. Divide the stated lumens by the stated watts; if the result differs from the printed lm/W, one of the three numbers did not come from the test. A single efficacy applied to every wattage in a series was calculated, since different builds measure differently.
4. Beam angle, peak intensity and distribution
The beam angle is the cone inside which intensity stays above 50% of peak. On its own it is incomplete. You also need the peak intensity in candela, because that converts directly to delivered light: lux = candela ÷ distance². Beyond both sits the full distribution, supplied as an IES or LDT file. That file is what DIALux or Relux uses to calculate your site. We won’t repeat the file format here; how to read an IES file covers the header, the multiplier trick and the four checks that expose a fabricated file.
For floods specifically, the beam angle drives the whole design. The flood light selection guide shows how throw distance picks the beam before wattage enters the calculation.
Light Quality: Specs 5–8
5. CCT and its tolerance
“4000K” names a target. The specification is the target plus a tolerance, stated in SDCM steps (MacAdam ellipses). Batches at 5 steps show visible differences between adjacent fixtures. For a continuous row or a facade that matters, and our LED colour temperature guide covers it in detail. If a datasheet gives CCT with no tolerance, ask for one in writing.
6. CRI: Ra and R9
Ra, the general colour rendering index, is the average of eight pastel test colours (R1–R8) defined in CIE 13.3. Saturated red is R9, and it is not part of that average. A fixture can therefore show a respectable Ra while rendering reds poorly. That matters for skin tones, food, safety signage and colour-coded cabling. Ask for R9 whenever reds carry information. Also check whether the CRI figure was measured on the luminaire or copied from the LED package datasheet.
7. Glare: UGR and its conditions
The CIE Unified Glare Rating depends on room, viewing position and layout, so a datasheet UGR comes from a table for a standard reference room. Use it to compare fixtures and confirm the real value in your calculation. Outdoor area and sports lighting use a different metric (GR, CIE 112), so a UGR on a flood datasheet describes conditions the fixture will not work in.
8. Flicker
Flicker comes from the driver and is invisible in photos. It is expressed as percent flicker or flicker index, with recommended limits by frequency in IEEE 1789; EU Regulation 2019/2020 uses Pst LM (visible flicker) and SVM (stroboscopic effect). Near rotating machinery or slow-motion cameras, ask for data at full output and at your dimmed levels, because ripple often worsens as a driver dims.
Electrical: Specs 9–13
9. Input voltage range
AC100–277 V covers 120 V, 230 V and 277 V supplies with one driver. Check for headroom on weak grids, and that flux and power were stated at your voltage. Our SJTG floods are specified at AC100–277 V, with 347 V and 480 V driver options.
10. Power factor
Power factor is real power divided by apparent power. Low-PF equipment draws extra current that heats wiring and loads transformers. Cable and breaker sizing follow apparent power, not watts. Quality industrial drivers state PF 0.9–0.95 or better at rated load. Look for the load condition, because PF often falls when a driver is dimmed.
You can also check a reported PF: PF = watts ÷ (volts × amps). If a report gives all three electrical values and the PF does not close within rounding, ask why.
11. Total harmonic distortion
Total harmonic distortion measures how far the input current departs from a sine wave. Good power-factor correction keeps industrial drivers below roughly 15–20%. On three-phase systems with hundreds of fixtures, harmonic currents add in the neutral conductor. IEC 61000-3-2 sets harmonic current limits for lighting equipment. As with PF, read the load condition.
12. Surge protection
Surge immunity is specified in kilovolts at IEC 61000-4-5 test levels, and a datasheet line of “surge protected” with no number tells you nothing. Look for three details: the kV figure, whether it applies line-to-line (differential mode) or line-to-earth (common mode), and whether protection is built into the driver or supplied as a separate replaceable module. Pole-mounted and exposed positions generally need a higher rating than indoor fixtures on protected circuits.
13. Dimming interface
“Dimmable” is not a specification. The interface is: 0–10 V, DALI, a timed step profile, or a photocell or sensor input. Also check the dimming range and the minimum level. Specify the interface when you order. Retrofitting a dimmable driver into a sealed IP65 body later means opening the housing and disturbing the gasket.
Durability: Specs 14–17
14. IP rating
The IP code (IEC 60529) has two digits: solids, then water. IP65 means dust-tight and protected against water jets. IP66 adds powerful jets, and IP67 covers temporary immersion. The number is only as good as the weakest seal, which is usually the cable gland or the entry point rather than the glass. Ask how the rating was tested and where the cable enters.
15. IK rating
The IK rating (IEC 62262) grades resistance to impact energy. IK08 corresponds to 5 joules and IK10 to 20 joules. It is independent of IP: a fixture can be waterproof and fragile. Low mounting positions, sports halls, car parks and vandal-exposed sites justify an IK line. If a datasheet omits it, the fixture has not been rated, not rated low.
16. Operating temperature range
The rated ambient range (Ta) is the air temperature the fixture is designed to run in at full output for its rated life. It is a different number from the photometric test ambient. Our SJTG floods are specified at -20 °C to 45 °C. The report measures output at 25 °C. For a site whose summer ambient under a canopy or a roof exceeds the top of the stated range, ask the question before ordering rather than after the first driver failure.
17. Lumen maintenance: L70, LM-80 and TM-21
“50,000 hours” means nothing without its definition. L70 is the point at which output falls to 70% of initial. The chain behind a credible claim runs like this. LM-80 and TM-21 data come from the LED package, tested at stated case temperatures. The TM-21 projection is capped at six times the test duration. That projection is then tied to the package temperature inside the real fixture. A life figure with no LM-80 reference and no temperature is a round number.
What to read in the warranty
A warranty is terms, not a spec. Read what counts as failure (dead fixture, LEDs out, or depreciation past a threshold), whether the driver is covered for the full term, who pays freight and labour, which hours or ambients are assumed, and whether surge is excluded. Then ask whether replacement drivers are stocked.
Physical: Specs 18–20
18. Weight
Weight is a structural input for brackets, cross-arms and mast heads, and a handling input for installation at height. It should be the fixture as installed, with the bracket, not the bare body.
19. Wind area (EPA)
The effective projected area is what the pole or mast designer needs to check wind load. US pole makers typically work in EPA. European column design under the EN 40 series uses the fixture’s windage area. A high-mast ring of flood heads adds up quickly, and the figure belongs in the structural check, not in the assumptions.
20. Mounting
Look for the bracket type (yoke, slip-fitter, cross-arm, wall), the spigot or bolt pattern, the tilt scale and locking method, and any secondary retention for work above people. A flood without a readable aiming scale will be aimed by eye and will drift from the photometric design.
Datasheet Red Flags
None of these proves a problem on its own. Each is a reason to ask for the measurement before comparing prices.
| What you see | What it usually means | What to ask for |
|---|---|---|
| Lumens with no “luminaire” or “measured” qualifier | Chip lumens multiplied by chip count | Whole-fixture photometric report |
| ”Equivalent to 1000 W” with no lumen figure | A comparison with no measurable basis | Measured luminaire lumens and the IES file |
| ”Up to” or “max” in front of flux | Best-case figure, sample unknown | The report for the quoted model |
| Identical lm/W for every wattage in a series | Calculated, not measured | Reports for at least two wattages |
| Efficacy above any report the supplier can send | The number came before the test | The report that produced the figure |
| No test ambient on photometric data | Output cannot be compared or reproduced | Test conditions page of the report |
| Beam angle with no peak candela | Delivered lux cannot be checked | Peak intensity and polar curve |
| ”50,000 h” with no L-value or LM-80 reference | A round number | LM-80 source and TM-21 projection |
| ”Surge protected” with no kV figure | Unknown rating, possibly minimal | kV rating, mode, and driver datasheet |
| PF or THD with no load condition | Possibly measured at the best point | Values at full load and at minimum dim |
Worked Example: Reading Our Measured 200 W Flood Report
Here is the same reading process applied to one of our own reports: a 200 W-class flood (model code CDNFC9185), measured on an HPG1800 distributed goniophotometer. The full 16-page PDF is downloadable from our LED flood light range.
Test conditions first. The report header records a C-Gamma test at 5°/s, a test distance of 11.8 m, an ambient of 25.3 °C at 65% humidity, and a date of 30 April 2026. The ambient sits inside the 25 ± 1 °C window LM-79 specifies. That alone does not make a report LM-79 compliant, but it means the flux was measured at the reference temperature.
| Report line | Value | The check a buyer can run |
|---|---|---|
| Nominal power | 200 W | Class only; do not calculate with it |
| Measured input | 201.3 W at 221.1 V, 0.923 A | Within 1% of nominal |
| Power factor | 0.986 | 221.1 V × 0.923 A = 204.1 VA; 201.3 ÷ 204.1 = 0.986. It closes |
| Total luminous flux | 35,309 lm | Whole luminaire, not chip sum |
| Efficacy | 175.4 lm/W | 35,309 ÷ 201.3 = 175.4. It closes |
| Beam angle (50%) | 60.1° H × 59.3° V | Symmetric medium beam |
| Peak intensity | 34,444 cd | 34,444 ÷ 10² = 344 lx at 10 m, matching the report’s 344 lx centre value |
The last row is the check most buyers skip: the inverse-square result agrees with the report’s own illuminance table (344 lx centre at 10 m, 11.38 m spot). Fabricated summaries rarely survive that cross-check.
Now the point of the example. On the same basis we published a 150 W unit that measured 20,422 lm at 143.4 W, or 142.3 lm/W (59.4° × 57.8° beam, PF 0.960), and a compact economy 200 W build at 127.9 lm/W. Different constructions measure differently, which is why our range-level efficacy is stated as 120–160 lm/W and the 175.4 lm/W flagship is one build’s result, not a figure for every model. When comparing quotations, ask which build the price refers to, then ask for that build’s report.
The Datasheet Request Checklist
Paste this into your RFQ. It asks for the evidence behind all 20 specs:
For the exact model, wattage, optic and CCT on this quotation, please send:
1. Photometric test report of the complete luminaire (model code, instrument,
test date, test ambient temperature)
2. Measured input power, voltage, current and power factor at rated load
3. Total luminaire luminous flux and efficacy (not LED package figures)
4. Beam angle, peak intensity (cd) and the IES or LDT file
5. CCT with SDCM tolerance; CRI Ra and R9
6. UGR table (indoor) or glare data (outdoor), and flicker data at full
and minimum dim
7. Driver datasheet: input voltage range, THD at rated load, surge rating
in kV (differential and common mode), dimming interface and range
8. IP and IK ratings with test basis and cable entry detail
9. Rated ambient temperature range (Ta)
10. LM-80 source and TM-21 projection behind the stated life (L70 hours)
11. Weight with bracket, effective projected area, mounting and aiming detail
12. Warranty terms: failure definition, driver coverage, claim freight/labourFor how these documents fit into the wider design process, from lux targets to maintenance factors, see industrial lighting design fundamentals.
The Bottom Line
A datasheet shows what the seller emphasises; the report shows what the fixture did when measured. Trace every line to a measurement, run the two arithmetic checks (lumens ÷ watts, candela ÷ distance²), and ask for the driver data brochures omit. Definitions live with the Illuminating Engineering Society (LM-79, TM-21), the IEC (IP, IK, surge levels) and the CIE (colour rendering, glare).
Next action: send the checklist to every shortlisted supplier and compare what comes back. To run it against us, send the model, quantity and site details and our engineers return the reports, IES files and a DIALux layout for your geometry.
FAQ
What is the most important spec on an LED datasheet? Measured luminaire lumens with the input power they were measured at, because every design calculation starts from those two numbers. Neither means much without the test report, the test ambient and the model code. After that, the beam angle and peak candela decide where the light goes, and the IES file lets a designer check the layout.
What is the difference between chip lumens and luminaire lumens? Chip lumens are the LED package rating multiplied by the number of packages. Luminaire lumens are measured from the complete fixture, after lens, cover and thermal losses. Measured fixture output always runs below the summed chip figure, so a datasheet that does not say which one it quotes should be treated as quoting chip lumens.
How do I check whether a datasheet efficacy figure is real? Divide the stated luminous flux by the stated input power. If the result does not match the printed lm/W, the numbers did not come from one measurement. Then ask for the report and check that the flux is luminaire output measured at a stated ambient, close to 25 °C for LM-79 testing, rather than a package rating.
What power factor and THD should an LED fixture have? Quality industrial drivers typically state power factor of 0.9–0.95 or better and THD below roughly 15–20%, both at rated load. The load condition matters, because both figures usually get worse when a driver is dimmed. For large installations, ask for values at full output and at the minimum dimming level you plan to use.
Does a 50,000-hour rating mean the fixture lasts 50,000 hours? Not necessarily. A credible life claim names the L-value, usually L70 (70% of initial output), and rests on LM-80 package data projected with TM-21, capped at six times the test duration. It should also relate to the package temperature inside the real fixture at your ambient. A bare hours figure with none of that is a design target, not a measured result.