There is currently no way to measure or estimate the lifetime of SSL luminaires. LEDs do not typically “burn out” or abruptly fail at end-of-life, like an incandescent lamp. Instead, they get dimmer over time, the speed of which depends on exposure to heat and other variables. Standards have made progress on measuring and predicting lumen maintenance (the relationship between temperature, operating time, and light output) for individual LED packages, but the life of an integral lamp or luminaire is determined by more than the LEDs. For example, the failure of an electrical component or darkening of an optical component may limit the lifetime of an SSL luminaire. Predicting and measuring the lifetime of the integration of varying subcomponents makes this topic very technically complicated.
FINDING: There is no standardized method for measuring the lifetime of SSL products.
The CRI is the internationally accepted metric for the evaluation of a light source’s color rendering abilities (CIE, 1995) and was developed in response to the advent of fluorescent lamps. Fluorescent lamps had spectral power distributions (SPDs) unlike anything the lighting industry had used before, and the quality of color rendering from these light sources was highly variable. The calculation of the CRI requires only the spectral power distribution of the light source of interest and is basically a series of colorimetric simulations. In these simulations, the appearance of a predefined set of reflective samples (object colors) is compared when illuminated by the test source and when illuminated by a reference illuminant (blackbody radiator or daylight simulator). If the samples appear identical in both cases, the test lamp would receive a general CRI (Ra) of 100. Deviations in the appearance of the test sample colors lower the score. A number of flaws of the CRI have been recognized for years, but the problems have not been considered important enough to warrant change (CIE, 1999). The problems of the CRI include the use of outdated and obsolete colorimetry (the math used to calculate the appearance of the reflective samples), a set of reflective samples that do not detect certain color rendering problems, and, according to some, an underlying definition of color rendering that does not correspond to actual users judgments of color rendering quality (Davis and Ohno, 2009).
For much of the history of electric lighting, a small number of very large companies produced nearly all of the light sources sold in the world. Each of these companies had the resources and expertise to understand the limitations and flaws in their measurements and ensure that metrics did not lead them to inadvertently create poor products. However, the lighting industry has changed considerably, largely because of SSL. Many smaller companies are now developing lighting products, some with very little experience in lighting, and the metrics used to evaluate light sources need to give users accurate predictions of performance. Furthermore, the problems of the CRI are particularly pronounced for some SSL sources (CIE, 2007). For example, certain LED spectra can render the reflective samples of the CRI very well, but render other object colors poorly.
There is now widespread agreement that a new method is needed to evaluate the color rendering quality of light sources (CIE, 2007), and many different approaches and methods have been proposed (reviewed in Davis and Ohno, 2009; Guo and Houser, 2004). DOE has publicly supported one proposed metric (DOE, 2010), the Color Quality Scale (CQS) (Davis and Ohno, 2010). Although it still uses the CRI, it is the current standard. A technical committee in the CIE was formed in 2006 to recommend a new procedure for evaluating the color rendering of light sources. Because the CIE committee consists of a diverse international group of stakeholders, consensus has not yet been achieved. This is problematic for SSL manufacturers, particularly small companies incapable of performing detailed colorimetric simulations, as they optimize their products to the metric.
FINDING: The CRI does not always yield results that predict or evaluate performance well, so manufacturers cannot rely on it to guide product development.
DOE has actively supported the consensus process for the development of testing and measurement standards. DOE has provided experts and supported their work time for numerous SSL standards committees. The agency also organized and sponsored an SSL standards workshop, as well as several related round-table meetings. Furthermore, DOE uses its demonstration projects to provide input on additional needed standards and provides financial support to the U.S. national committee of the CIE. DOE also funds measurement and standards research at the National Institute of Standards and Technology.
The Commercially Available Light-Emitting Diode Product Evaluation Reporting (CALiPER) component of the DOE program includes verification testing that produces extensive data on individual products. DOE funds independent laboratories to conduct the testing, each focused on one product type (e.g., high-bay luminaires; small replacement lamps (MR16, PAR lamps, and so forth)). The testing follows the IESNA LM-79-08 method of electrical and photometric measurements to verify that lamps are performing according to specifications.
ALA (American Lighting Association), IES (Illuminating Engineering Society), and IALD (International Association of Lighting Designers). 2010. What’s Your Quality of Light? The Lighting Industry’s Call for a Practical Lighting Energy Policy. Dallas, Chicago, and New York: ALA, IES, and IALD.