Even though a decade has already passed since they have been first applied for general illumination, the topic of assessing the reliability of LED lighting products continues to be an on-going topic of debate, and in some cases, confusion. Below you’ll learn about some metrics and processes for reliability designing LED lighting systems.

The lifetime of incumbent lighting technologies like incandescent or fluorescent lamps is usually characterised by average rated life, which is the time in hours at which half a test population of continuously running lamps fails (i.e., light is no longer emitted). For incandescent lamps, average rated life is about 1,000 hours; for compact fluorescent lamps, about 12,000 hours; and for linear fluorescents, about 25,000 hours. In contrast, LEDs, operating under appropriate environmental conditions can emit light for 100,000 hours or longer, making an average rated life measurement impossible.

Instead, LED lighting lifetime is characterised using the L70 metric, which is the time in hours at which the light output measured in lumens (L) drops to 70% of initial. The shortcoming of L70 from a reliability engineering perspective is that the test methods, published by the recognised authoritative body for lighting, the Illuminating Engineering Society, are based on lumen output measurements for the LED packages only and don’t take other system components such as the driver, thermal management design and even optics into account. Incumbent lighting manufacturers have not focused heavily on overall system reliability because the lamp is nearly always the first component to fail. With the advent of LED lighting, manufacturers and end-users have come to understand that other parts of the lighting system can be the limiting factor in overall life as the table below illustrates.

20170807_EDNA_LED-lighting-test_01 (cr) Table 1: Typical 2015 performance of SSL compared to other lighting technologies (Source: DOE Solid-State Lighting R&D Plan, June 2016)

Reliability, in its formal sense, is the probability that a part or system will perform as intended for a specified time period under specified operating conditions. Product reliability assessment is nearly always based on accelerated life test statistical modelling and projections. For LED lighting systems, both lumen depreciation and failures of other components contribute to overall reliability, as the figure below shows. But a statistical model for combining these two different aspects of lighting system performance has proved elusive, which is the main reason the focus has been on robustness testing instead.

20170807_EDNA_LED-lighting-test_01_01 (cr) Figure 1: Example of LED lighting system overall reliability (Source: U.S. Dept. of Energy, Building Technologies Programme Solid-State Lighting Technology Fact Sheet, August 2013)

In the context of field operation, robustness is an unquantified indication of the ability of a product or system to operate as intended in the specified environment by confirming that design and production processes yield products that can withstand higher levels of stress than would be expected in normal operation. The electronics industry has for decades employed robustness testing protocols as part of their verification programs. The concept has only recently been embraced by the greater LED lighting community, as the realisation has taken hold that LED-based lighting products are not much different from other types of electronics, except, of course, that they produce light.

Robustness testing typically involves subjecting production samples to a variety of overstress conditions, making it more part of an overall QA programme as opposed to reliability testing that uses design prototypes. And unlike reliability testing where the goal is to test to failure to learn about failure modes, the expectation for robustness tests is that all products will operate correctly at the conclusion of the test.

Robustness tests for LED lighting are similar to tests for other types of electronic equipment published by JEDEC, IEC and other organisations, but tailored to the special characteristics of LED products. The specific test protocol for a given product would depend on the intended application, for example, the test panel for indoor office products would probably not need to include high temperature operation or vibration.

The tests listed in the table below are representative of some of the more common robustness tests recommended for LED lighting products.

20170807_EDNA_LED-lighting-test_02 (cr) Table 2: Typical tests for LED lighting system robustness.

A number of standards bodies have recently addressed the subject of LED lighting system robustness, with further efforts underway. Next time, we’ll review existing and planned standards and how each addresses LED lighting system reliability.

Yoelit Hiebert has worked in the field of LED lighting for the past 10 years and has experience in both the manufacturing and end-user sides of the industry.

First published by EDN.