LED Reliability – What you need to know

As everyone knows, LED technology is displacing older lighting technologies across almost all applications. Although LED has a higher upfront cost, energy savings combined with an extremely long product lifetime make LED a compelling choice. But what exactly does it mean when an LED lighting supplier claims a 50,000-hour product lifetime? And what data should the supplier have to support such a claim?

LEDs don’t “burn out” like incandescent bulbs, they slowly degrade over time. This slow degradation in output is called light level maintenance. The industry standard test methodology for quantifying light level maintenance has been defined by the Illuminating Engineering Society of North America (IESNA), and is called LM-80. To be valid, LM-80 must be conducted by a test facility certified by the National Voluntary Laboratory Accreditation Program (NVLAP). Once LM-80 testing is complete, a test report will be issued that quantifies the light level maintenance of the LED.

During the LM-80 test, output intensity from the LED is recorded at various temperatures and input currents. Data must be collected at increments of no less than 1,000 hours, and for a total duration of at least 6,000 hours (although 10,000 hours of test time is recommended). The graph below shows a typical LM-80 light level maintenance result.

But with only 6,000 – 10,000 hours of test time, how can LED lighting suppliers claim product lifetime of 50,000 hours? The answer is another standard from IESNA called TM-21, which extrapolates the data collected from LM-80 into the future. TM-21 applies curve fitting to the last 5,000 hours of collected data, and projects forward the light level maintenance. The maximum extrapolation is 6X the LM-80 test duration – see TM-21 project example below.

But where does a 50,000-hour lifetime claim come from? The LM-80/TM-21 light level maintenance projection is associated with a specific amount of intensity reduction, typically defined as either L₇₀ or L₅₀. L₇₀ is 70% of the original intensity from the LED (30% reduction), and L₅₀ is 50% of the initial intensity. So, if we look at the TM-21 projection at the 70% intensity mark, and identify the hours to get to that point, we can determine the L₇₀ light level maintenance. This is the number that is often used to support product lifetime claims.

However, beware of the fine print. Having LM-80 data with a TM-21 extrapolation that quantifies the L₇₀ point at 50,000 hours does not mean your fixture will last 50,000 hours. You need to know the percentage of LEDs that are expected to fail at 50,000 hours, which is the B-number. The B-number states the percentage of the LED population that will successfully achieve the L₇₀ light level maintenance projection. The full story of light level maintenance is written like this example: L₇₀B₅₀ = 50,000 hours, which says that 50% of the LED population will degrade in intensity by 30% (or less) after 50,000 hours. Another example is L₅₀B₉₀ = 40,000 hours, which says that 90% of the LED population will degrade in intensity by 50% (or less) after 40,000 hours. But remember, the light level maintenance is only projecting the lifetime of the LED – not the fixture. In many cases, the lifetime of an LED fixture will be determined by other components, such as the ballast.

So, if you’re in the market for an LED grow light fixture, make sure you ask for the LM-80 / TM-21 test report. And then ask about the reliability of the other components in the fixture. There’s much more to understand about LED fixture reliability than the simple claim of having a 50,000-hour lifetime.