In the Long Run: Testing SSL Reliability

June 25, 2018

You are here

Three photos showing accelerated stress testing of LED products and devices.

Solid-state lighting (SSL) is bringing new capabilities to lighting, in part because light-emitting diodes (LEDs) can be easily packaged together on a printed circuit board to form sources that are multicomponent and multifunctional. But the long-term performance of such systems hasn’t yet been fully investigated, despite the fact that next to energy efficiency, long life is the most publicized advantage of SSL. That’s why the U.S. Department of Energy is funding an ongoing series of studies on the reliability of LED lighting products conducted by RTI International, an independent research firm.

SSL reliability is a complex matter. Since LEDs have expected lifetimes in the 10s of thousands of hours, it would take many years to test the performance of a given LED. That’s why we use stress testing (with higher temperatures, humidity, frequent cycling, etc.) to accelerate the degradation and get reasonable indications of “real life” life expectancy and causes of failures. One has to consider lumen depreciation of the LEDs, color shift, catastrophic failure, voltage rise, flicker changes, and efficacy drop in the luminaire. But the recent advent of tunable luminaires has complicated things even further because instead of static depreciation of one type of LED, it introduces the possibility of dynamic depreciation of many different types in the same luminaire – with multiple failure mechanisms (e.g., lumen depreciation and color shift) coming into play with each source type.

The latest RTI study is the second in a series about accelerated stress testing of drivers used for SSL luminaires. The new findings continue to support the observation that many of the drivers used in SSL devices are highly robust. The extreme conditions (75° C and 75% relative humidity) produced minimal changes until failure was imminent. The findings also suggest that degradation of SSL drivers may be monitored through changes in parameters such as power factor, inrush current, and flicker of attached LED loads at low dimming levels.

Another recent RTI study looked at the long-term behavior of commercial products utilizing different types of mid-power LEDs integrated into the same LED module. Among other things, it found that the chromaticity of phosphor-converted LED devices utilizing blue LEDs will ultimately shift in the blue direction, and that this is likely to occur sooner for higher correlated color temperature (CCT) devices than for those with lower CCTs.

As lighting becomes more complicated, new failure mechanisms are becoming increasingly apparent. With today’s color-mixed luminaires, not only can you have different thresholds of L70 (the point at which lumen output has decreased to 70% of initial output) for the different types of LEDs (white, red, green, and blue), but the differential aging of the LEDs can also cause color shift in the luminaire. On top of that, if the product is dim-to-warm, for example, the warm-white LEDs will be on more than the blue and green LEDs and thus degrade sooner – and so on and so on. So you can see why the issue of reliability has become even knottier than it was when SSL first came on the scene.

The good news, though, is that with dogged, in-depth research such as the work RTI is doing, we’re learning more and more about SSL failure mechanisms, what causes them, how to predict them – and, most importantly, how to prevent them.