THE LIGHTING DESIGNER’S PERSPECTIVE Michael Hennes, Senior Associate
Cline Bettridge Bernstein Lighting Design, New York
Most lighting designers can probably point to a number of installations they have been involved with where LEDs have failed much sooner than their rated life, leaving the designer to question the reliability of LEDs while dealing with unhappy clients. This reality underscores the need for accurate, consistent, and reliable testing of LED technology.
LM-80-08 was issued with these concerns in mind, and it takes a number of good first steps, starting by defining how to report LED lumen maintenance. LEDs do not die like other lamp sources, but instead become dimmer and dimmer until barely a glow remains, making it difficult to define when LED end-of-life actually occurs.
LM-80 provides two measurement standards (conceived by the Lighting Research Center) for defining LED life: L70 and L50 define the time to 70 percent and 50 percent lumen maintenance left, respectively. Next, LM-80 addresses the many variables that affect LED testing results (and eventual real-life performance), such as ambient temperature and humidity, airflow, case temperature, voltage and current regulation, and drivers. In turn, it establishes a clear methodology of how to standardize these factors in the testing process.
Another major issue is an LED’s sensitivity to high temperatures. LM-80 calls for testing at three case temperatures: 50 C (122 F), 85 C (185 F), and a third temperature chosen by the manufacturer. While 85 C is a good high-temperature target, the third temperature, which is selected by a manufacturer based on its expectations of customer applications, provides an opportunity for testing at an even higher temperature.
Yet LM-80 leaves many questions unanswered. The standard is based on testing LED packages and modules, so it is not taking into account the impact of the luminaire. This would seem to be a big gap, since luminaire design plays such a large role in LED performance. Testing also must be carried out for 6,000 hours, but it is not clear whether this is sufficient for accurate prediction of full life. The problem is further complicated by LEDs’ atypical curve of light degradation, which can have an early bump upward in output, making extrapolation of a full life estimate curve from the test results difficult. Also, LM-80 strongly recommends chromaticity testing to measure color temperature shift over life, because excessive color shift is another measure of end-of-life. Yet it is not clear how this information can be integrated into the lumen maintenance reporting data or what amount of color shift over life is acceptable. Lastly, the testing only addresses lumen maintenance, or light loss over time. While LM-80 requires that catastrophic failures be reported in the test results, it is not clear how that information would be factored into the final reported life data, and it does not take into account the potential for system component failure at the luminaire level.
Clearly, additional research and standards are needed to reach the goal—data that accurately reflects the LED consistency and reliability over time—that lighting designers and their clients expect to receive. Without this type of information, lighting designers will continue to question the reliability of LED luminaires and their ability to assuredly specify these products.
THE ELECTRIC UTILITY’S PERSPECTIVE
Philip F. Keebler, Senior Project Manager
Electric Power Research Institute, Knoxville, Tenn.
LM-80-08 defines a most important test procedure for determining the lumen maintenance for LED packages, arrays, and modules. The standard establishes clear definitions needed to guide users of the test procedures. Regarding test procedures, LM-80 nicely addresses elapsed operating time and uncertainty, and it requires use of video and current monitoring of LED samples during the test period. The standard also properly specifies photo-metric operating conditions and measurements.
The use of LM-80 will identify the drive current to the LED samples and the lumen output they produce within the test duration of 10,000 hours. However, as utilities become more involved in advanced light sources and begin to include more of them in rebate and incentive programs, further documentation of energy and power performance—especially as LEDs reach their expected lifetimes—will help ensure that LED systems save energy as they age.
Adding a requirement to measure the input power to the LED samples in the next revision of LM-80 will enhance data collection and provide a means for further documenting power performance of sample operation during the test duration. The next revision also should consider addressing photometric performance beyond the 10,000-hour duration, provide additional guidance on specifying operating cycle times, and guide users on documenting driver characteristics and operating methods. These considerations for the next revision will help manufacturers and users better understand LED performance beyond typical operating periods to determine any drifts in photometric and energy performance as LED systems age beyond the 10,000-hour point.
