One Datasheet, Two Lumen Figures: What Australian LED Specs Actually Tell You

Eagle Lighting publishes two lumen figures for the same 28W batten. 4,450 and 3,915. Both are correct. The gap between them is the part most datasheets never show you.

August 2026 · 10 min read · All states · By Mark Riley, Cutter Electronics

Open the RapidLED datasheet and the Optical Performance table has two columns sitting side by side. Systems Lumens: 4,450. Luminaire Lumens: 3,915. Same product, same 28W, same 1200mm fitting.

Neither number is wrong. The first is what the LED module produces. The second is what actually leaves the fitting after the diffuser has taken its cut. The difference is 12%, and it is the light you paid for that never reaches the floor.

What makes that datasheet unusual is not the 12%. Every diffused fitting loses something. What makes it unusual is that Eagle prints both numbers and lets you see the loss. We went looking at 47 LED luminaires sold in Australia, across high bays, panels, battens and street lighting, and it is the only one that does.

Three ways to quote a lumen, and they are not the same number

A lumen figure on a datasheet can mean three different things, and they are not interchangeable.

Chip lumens are what the bare LED emitters produce before anything gets in the way. No optics, no diffuser, no driver losses. It is the most flattering number available and the least like anything that ends up installed.

Module lumens include the optics but exclude the driver. Closer, still not the whole picture.

System lumens are what leaves the luminaire, divided by the watts drawn from the mains. This is the number that describes what you actually get, and it is the only one you can safely put into a lighting calculation.

Between chip and system you can lose 20% or more. Driver efficiency is typically 85% to 92%. Optical efficiency through a diffuser runs 85% to 95%. Stack them and a fitting quoting 150 lm/W at chip level is delivering something closer to 115.

The finding: over half of them do not say

Across the 47 products, 21 datasheets state clearly that their figure is system level. Not one states that it is chip or module level. The other 26 publish a bare Power and Lumens pair and never say which of the three they mean.

Panels are the worst category in the set. Ten out of ten publish a bare Power and Lumens pair with no statement of basis.

That matters more for panels than anywhere else, because panels are overwhelmingly remote-driver products. Haneco's Matrix G4 datasheet even tells you the driver output is 30 to 42V DC. So the question of whether the quoted watts are mains input or DC module power is live, material, and unanswered on every Australian panel datasheet we could find.

What the ambiguity is worth in wattage

Take a panel quoting 3,600 lumens at 30W. That is 120 lm/W on the sheet.

If that figure is already system level, you get 120 lm/W and your design works. If it is chip level, apply a typical 0.90 optical and 0.88 driver derate and you are delivering closer to 95 lm/W. Same product, same datasheet, and the honest answer is somewhere in a 25 lm/W band that the manufacturer could close with one sentence.

Scale that to an office. You are replacing 100 twin T8 battens and your lighting design calls for 3,500 delivered lumens per fitting to hit 320 lux at the working plane. Spec on the sheet figure and you order 30W panels. If the figure was chip level, each fitting delivers around 2,850 lumens and you are 19% short across the whole floor. The room is under-lit, it will not meet AS/NZS 1680 for office work, and the fix is either more fittings or higher wattage. Both cost more than asking one question before you ordered.

Six datasheets that do not agree with themselves

Separate from the basis question, some datasheets publish an efficacy figure that does not match their own published lumens divided by their own published watts. Six of the 47 differ by more than 1%.

None of this is evidence of anything sinister. The likely explanations are dull: a figure rounded to a convenient number, a value carried over from a different colour temperature or wattage variant in the same family, or one number revised while another was left alone. The largest discrepancy in the set runs against the manufacturer's own interest. Pierlite quotes 160 lm/W for a product whose own published figures compute to 170.5, understating itself by 6.6%.

The point is not that anyone is cheating. It is that these are the documents your entire lighting design rests on, and the arithmetic on six of them does not close.

Who states it plainly

Some manufacturers make this easy, and they deserve the credit for it.

Eagle Lighting publishes both figures side by side, which is the gold standard and currently unique in this dataset.

Ektor defines the term on the sheet: "luminaire lumens refers to the exit lumens or delivered lumens from the luminaire", and publishes a Total System Efficiency that reconciles exactly.

Pierlite heads its ordering table columns SYSTEM POWER and EXIT LUMEN. No ambiguity at all.

Thorn uses "Luminaire luminous flux", "Luminaire input power" and "Luminaire efficacy" as its literal field labels. So do Trilux, Signify, Dialight, Ledvance and Sylvania.

Nobody has to invent anything here. The vocabulary exists and roughly half the market already uses it.

What to ask before you specify

One question closes most of this gap:

Three more worth asking on any commercial job:

• What ambient is the lumen figure rated at? Almost every datasheet quotes at 25°C. Not one product in this set stated the ambient its lumen figure was measured at. A high bay near roof level in an uninsulated shed is not operating at 25°C in February.
• Is there a photometric file? An IES file from an independent test lab settles the question in a way a marketing figure cannot.
• Does the stated efficacy match the lumens and watts on the same page? Divide one by the other. It takes five seconds and it caught six products here.

The numbers are published, so we published them

Every figure in this article comes from a manufacturer's own datasheet, retrieved and read, with the URL recorded against each row. We put the whole thing online as the Australian LED Efficacy Index: 47 products, four categories, claimed against realistically delivered efficacy, with a stated method you can reproduce and a free CSV download.

Where a datasheet does not state its basis, the index does not guess. It publishes a range, from best case if the figure is already system level to worst case if it is chip level. The width of that range is simply how much the published data leaves open. Haneco's SKYLUX X3 sits somewhere between 132 and 167 lm/W depending on what its lumen figure means, and the datasheet does not say.

Summary

One Australian datasheet publishes both the module and the luminaire lumen figure for the same fitting, and the gap is 12%. Across 47 products, 26 never say which figure they are quoting, and all ten panels fall into that group. Six state an efficacy that does not match their own arithmetic. None of this requires bad faith to explain, and all of it lands on whoever specifies the job. Ask what the number means before you divide by it.