Article from the U.S. Department of Energy
http://www1.eere.energy.gov/buildings/ssl/luminaire_efficacy.html

The use of light-emitting diodes (LEDs) as a general light source has forced changes in test procedures used to measure lighting performance. Click on the topics below to learn about the concept of luminaire efficacy and the technical reasons for its applicability to LED-based lighting fixtures.

a) Overview of Luminaire Efficacy

Lighting energy efficiency is a function of both the light source (the light "bulb" or lamp) and the fixture, including necessary controls, power supplies and other electronics, and optical elements. The complete unit is known as a luminaire.

Traditionally, lighting energy efficiency is characterized in terms of lamp ratings and fixture efficiency. The lamp rating indicates how much light (in lumens) the lamp will produce when operated at standard room/ambient temperature (25 degrees C). The luminous efficacy of a light source is typically given as the rated lamp lumens divided by the nominal wattage of the lamp, abbreviated lm/W. The fixture efficiency indicates the proportion of rated lamp lumens actually emitted by the fixture; it is given as a percentage. Fixture efficiency is an appropriate measure for fixtures that have interchangeable lamps for which reliable lamp lumen ratings are available.

Currently, lamp rating and fixture efficiency measures have limited usefulness for LED lighting, for two important reasons:

1.         There is no industry standard test procedure for rating the performance of LED devices or packages.

2.         The luminaire design and the manner in which the LEDs are integrated into the luminaire have a material impact on the performance of the LEDs. This makes it difficult to compare LEDs to other types of lighting.

Photometry

The measurement of quantities associated with light, including luminance, luminous intensity, luminous flux, and illuminance.

Integrating sphere

A device that enables geometrically total luminous flux to be determined by a single measurement. The usual type is the Ulbricht sphere with associated photometric equipment for measuring the indirect illuminance of the inner surface of the sphere.

Goniophotometer

An apparatus for measuring the directional light distribution characteristics of light sources, luminaires, media, and surfaces. Goniophotometry can be used to obtain total luminaire flux (lumens) and efficacy (lumens/watt), but not the color metrics (chromaticity, CCT, and CRI).

Spectroradiometer

An instrument for measuring radiant flux (visible and non-visible) as a function of wavelength. Visible radiation measurements can be converted into luminous intensity (candela) and flux (lumens).

Lamp or light source

A generic term for a device created to produce optical radiation.

Luminaire

A complete lighting unit consisting of a lamp or lamps and ballast(s) (when applicable) together with the parts designed to distribute the light, to position and protect the lamps, and to connect the lamps to the power supply.

b) Comparison of LEDs to Traditional Light Sources

Standard methods for rating lamps and fixture efficiency are inappropriate when comparing LEDs for two important reasons:

1.         There is no industry standard test procedure for rating the performance of LED devices or packages.

2.         The luminaire design and the manner in which the LEDs are integrated into the luminaire have a material impact on the performance of the LEDs.

Given these limitations, how can LED luminaires be compared to traditional lighting technologies? As an example, the table below compares two recessed downlight fixtures, one using a 13-watt CFL and the other using an array of LEDs. The table differentiates data related to the light source and data resulting from actual luminaire measurements. Luminaire photometry shows that in this case the LED fixture has input wattage and light output similar to the CFL fixture, and matches the CFL product's luminaire efficacy. This example is based on a currently available, residential-grade, six-inch diameter downlight. LED downlight performance continues to improve rapidly, with some LED retrofit products surpassing CFL downlights in luminaire efficacy.

Example: Comparison of CFL and LED Downlight Luminaires

Items in italics are not based on industry standard test procedures.

*Depends on specific LED used. Estimate is based on "typical luminous flux" declared by LED manufacturer on the product datasheet, which assumes 25°C LED junction temperature.

c) The Impact of LED Luminaire Design

For all light sources, there is a difference between rated luminous flux of the lamp and actual performance in a luminaire. However, traditional light sources installed in luminaires operate relatively predictably because the performance of traditional light sources in a wide range of luminaire types, applications, and use conditions is well documented and understood. LED technology is at a far earlier stage of development, so experience and documentation of performance within luminaires is lacking. The efficiency of LEDs is very sensitive to heat and optical design, which increases the relative importance of luminaire design.

Ensuring necessary light output and life of LEDs requires careful thermal management, typically requiring the use of the fixture housing as a heat sink or at least as an element in the heat removal design. Luminaires therefore have a fundamental and typically large effect on the luminous flux produced by the LEDs, and on the rate of lumen depreciation over time. LED "drop-in" replacement lamps, such as Edison-based reflector lamps or MR-16 replacements, are in theory designed to provide the necessary heat sinking for the LEDs, but given their installation in fixtures not specifically designed for LEDs, good heat management will be a challenge.

d) Why a New Testing Standard Is Needed for LEDs

Traditional light sources (incandescent, fluorescent, and high-intensity discharge) are rated for luminous flux according to established test procedures. In contrast, there is no standard procedure for rating the luminous flux of LED devices. LED light output estimates (as reported on manufacturer datasheets) are typically based on a short (<1 second) pulse of power applied to the LED chip, with junction temperature held at 25 degrees C. This is because LED chips must be binned for luminious flux and color during the manufacturing process. To run them longer without a heat sink would damage them. LED manufacturers usually list "minimum" and "typical" luminous flux on their product datasheets. There is no standardization of the test conditions, or the meaning of "typical." Further, there is no standard test procedure for measuring the luminous flux of LED arrays, such as multiple LEDs mounted on a circuit board.

In summary, luminous flux—and by extension, luminous efficacy—must be measured at the luminaire level for two primary reasons: 1) no standard procedures are available for rating LED devices on their own, and 2) the amount of light emitted by a fixture cannot be predicted reliably based on available information about LED devices and fixtures. The lighting industry has adopted luminaire efficacy as the preferred measure of LED performance, as evident in the development of a new test procedure based on this approach..

e) A New Test Procedure: LM-79-08

The lighting industry looks to the Illuminating Engineering Society of North America (IES) for lighting measurement test procedures. These test procedures are designated "LM" for lighting measurement, followed by an ordinal number, and the year of adoption or revision. They are developed by the IES Testing Procedures Committee, whose members include representatives of industry, research institutions, and testing laboratories. The document entitled "IES Approved Method for the Electrical and Photometric Measurements of Solid-State Lighting Products," designated LM-79-08, was developed by a joint IES-ANSI committee on SSL and published in 2008. Key elements of the document include:

1.         Covers fixtures incorporating light sources as well as light sources used for fixtures (e.g., LED retrofit products)

2.         Provides test procedures for photometric measurements using an integrating sphere, goniophotometer, and spectroradiometer

3.         Photometric information measured may include: total luminous flux (lumens), luminous intensity (candelas) in one or more directions, chromaticity coordinates, correlated color temperature (CCT), and color rendering index (CRI)

4.         Electrical information measured includes: current, voltage, and power

5.         Products must be stabilized until they reach thermal equilibrium before testing