100,000 hours —
what it actually means.

LM-80: measuring actual LED lumen maintenance

IES LM-80 (Approved Method: Measuring Lumen Maintenance of LED Light Sources) is the industry standard for measuring how much of its initial light output an LED source retains over time at a defined temperature. The test runs LEDs at three temperatures (typically 55°C, 85°C, and a third temperature chosen by the manufacturer), measures lumen output at regular intervals, and reports the results as a lumen maintenance curve.

The minimum duration for a valid LM-80 test is 6,000 hours. Most tests in practice run to around 10,000 hours — approximately 14 months of continuous operation. A handful of high-investment programmes reach 15,000 hours.

At 10,000 test hours, the data represents roughly 14 months of measured performance. A 100,000-hour service life claim requires projecting that data to a point ten times beyond what was actually observed.

TM-21: projecting beyond the test data

IES TM-21 (Projecting Long-Term Lumen Maintenance of LED Light Sources) describes how to project lumen maintenance beyond the duration of the LM-80 test. It fits an exponential decay curve to the measured data and extrapolates forward — mathematically.

The standard contains an important constraint: the projection is only stated to be valid to a maximum of six times the duration of the LM-80 test. If the LM-80 test ran for 10,000 hours, TM-21 allows a projection to 60,000 hours. No further.

A 100,000-hour claim based on 10,000 hours of LM-80 data is therefore projecting to ten times the test duration — well outside what TM-21 itself supports. The TM-21 standard requires a footnote or a capped statement when projection exceeds the six-times limit. In practice those caveats are rarely prominent in commercial data sheets.

Reading the notation: L70, L80, L90, B10, B50

Lumen maintenance claims combine an L-value (the percentage of initial flux that remains) with a B-value (the percentage of a luminaire population that has fallen below that L-threshold) and an hour figure. A complete notation looks like:

The B-value matters significantly. An L70B50 claim means that half of the population will have fallen below 70% by the stated hour figure — which may not be acceptable for road lighting applications under EN 13201. L70B10 is a far more conservative and useful benchmark for infrastructure procurement.

Why the headline figures inflate

Several factors combine to produce headline service life claims that overstate likely real-world performance:

1. Test temperature vs operating temperature: LM-80 tests are run at fixed temperatures. A luminaire with inadequate thermal management will operate its LED module at a higher temperature than the test condition — and lumen depreciation accelerates with temperature. The test result does not represent that luminaire's actual performance unless the operating temperature has been measured and matched to the appropriate test temperature.
2. Component ageing beyond the LED: the LM-80 test measures only LED lumen output. Driver degradation (especially electrolytic capacitor dry-out) and optical degradation (yellowing diffusers, dust contamination) also reduce effective light output but are not captured in the LM-80 curve.
3. Projection beyond TM-21 limits: as noted above, many commercial claims extend projection well beyond the six-times cap. The projection model remains mathematically continuous — but its physical validity has not been verified.
4. B-value not stated: a claim of "L70 > 100,000 h" without a B-value is, technically, an L70B50 claim — meaning 50% of units may be below threshold by that point. Requiring an explicit B-value in procurement specifications closes this ambiguity.

What to ask instead

Rather than accepting a headline hour figure, a more productive procurement approach is to ask for the underlying data:

• LM-80 test report: how many hours did the test run, at what temperatures, and how many samples were tested?
• TM-21 projected values at L70B10 and L80B10 — stated separately — with the projection capped at six times the test duration.
• Operating temperature of the LED module in the actual luminaire at rated drive current — measured, not calculated.
• System-level lumen maintenance, accounting for driver efficiency over time and optical component ageing.

These questions are not unreasonable for infrastructure procurement with a 20-year planning horizon. A manufacturer who cannot provide the underlying test report is asking you to accept a projected figure with no independent basis for verification.

Summary

A 100,000-hour claim in a street luminaire data sheet is, in nearly all cases, a mathematical projection built on test data that covers a fraction of that duration. IES LM-80 defines how LED lumen maintenance is measured — with a minimum test duration of 6,000 hours. IES TM-21 defines how that data may be projected — capped at six times the test duration.

A projection to 100,000 hours on the basis of 10,000 hours of test data significantly exceeds what the standard endorses. The figure is not falsified — but its precision is far lower than the headline number implies. Requesting the underlying LM-80 test report, a B-value alongside the L-value, and evidence of the operating temperature in the actual luminaire gives a considerably more reliable basis for comparing competing products.