Guest post by Jim Brodrick, U.S. Department of Energy
Reprinted with permission from “Postings: From the Desk of Jim Brodrick.”
For more information about the DOE’s solid-state lighting program, click here.
As you probably know, long lifetime is considered to be one of the major advantages of solid-state lighting over other lighting technologies. It’s something that gets talked up almost as much as SSL’s energy efficiency and is also a key parameter in cost assessments. Most other lighting lamps have filaments or electrodes that cause them to “burn out” between 1,000 and 30,000 hours of use. However, LED “lamps” — typically referred to in the industry as modules, packages, or arrays — do not have filaments to burn out and therefore have the potential to last far longer, with some manufacturers claiming lifetimes of 60,000 hours or more.
But the question of LED product lifetime and overall reliability is not a simple one. For one thing, the lifetime of any lighting product is a matter of overall system reliability that includes the life of the “lamp” itself plus the reliability of all other components, including ballast or driver, electrical connections, and the fixture housing itself (including lenses). Secondly, LED products haven’t been on the market long enough for those impressive lifetime claims to be put to the test. Compounding this is the fact that, as an emerging technology, SSL is advancing so rapidly that by the time any lifetime test can be completed, it’s a sure bet that the product will have been superseded by a newer and more advanced version. Add to all of this the proliferation of inaccurate or misleading information for LED products, and the question of reliability becomes even more muddled.
But the question of LED product lifetime and overall reliability is not a simple one. For one thing, the lifetime of any lighting product is a matter of overall system reliability that includes the life of the “lamp” itself plus the reliability of all other components, including ballast or driver, electrical connections, and the fixture housing itself (including lenses). Secondly, LED products haven’t been on the market long enough for those impressive lifetime claims to be put to the test. Compounding this is the fact that, as an emerging technology, SSL is advancing so rapidly that by the time any lifetime test can be completed, it’s a sure bet that the product will have been superseded by a newer and more advanced version. Add to all of this the proliferation of inaccurate or misleading information for LED products, and the question of reliability becomes even more muddled.
A recent U.S. Department of Energy (DOE) webcast that examined the issue of LED lifetime and reliability confirmed that determining how long LED modules themselves last isn’t a simple matter, because they don’t have a clear “end of life” the way other light sources do. LEDs will simply continue to operate while reducing their light output over time. For advanced LEDs integrated into well-designed luminaires, this can occur slowly over a very long period of time. Some not-so-well-designed LEDs and systems can, however, degrade relatively quickly.
How, then, do we determine what an appropriate “end of life” is for an LED module? Because LEDs can continue operating until its light output is very low, their life is usually defined as lumen depreciation to a particular point. The webcast explained that the point most frequently used is where lumen output has declined by 30 percent from its initial output. This point is chosen primarily because the human eye doesn’t detect decreases in light levels less than this percentage difference. This point is often expressed as “L70”, meaning when lumen output has fallen to 70 percent of initial output.
But the industry doesn’t yet have a standard method for fully determining this “L70” point. There is a newly published test method, IES-LM-80-08, that prescribes how to measure the lumen depreciation of an LED module through at least 6,000 hours of testing at various temperatures, which can be related to complete luminaire operating conditions. However, the missing element is a way to confidently estimate lifetime (L70) well beyond the 6,000 hours of collected data. A method to do this is being actively worked by standards-writing groups, but it is still under development.
The LED module life is, of course, just one element of overall LED system reliability, which goes beyond simple LED module life and encompasses the expected performance of all other components. For example, fluorescent ballasts have been produced to last well beyond the life of typical fluorescent lamps. However, there is limited data available to verify that LED drivers will last as long as or beyond the expected long life of the LED module. Similarly, while there is historical precedence for many fixtures and their components lasting for relatively long periods of time, there is limited direct data for the reliability of these products as part of a heat-producing LED luminaire.
LED lifetime and reliability are two of the issues getting close attention in DOE’s Commercially Available LED Product Evaluation and Reporting (CALiPER) testing program. Since it began in late 2006, CALiPER testing has shown a steady improvement for commercial LED products on many parameters. But even after seven CALiPER testing rounds (Round 7 was completed this January), there’s still a wide range in performance, and the potentially long life of these products hasn’t been completely explored.
Of the 13 products tested for reliability and lifetime to date, some have maintained output levels over the first 6,000 hours of operation (seven are producing over 96 percent of their initial output), while others have shown rapid lumen depreciation within the first 1,000 to 2,000 hours. In addition, some products have had a significant color shift over the first 6,000 hours of operation. In summary, no generalizable patterns have emerged.
The bottom line is that we don’t yet have a consistent method to estimate lifetime or reliability and won’t have definitive confirmation of any estimates until we’ve gotten more field-testing under our belt with a wider range of products.
In the meantime, how can LED luminaire reliability be assessed? For starters, it pays to be a skeptic. If it sounds too good to be true, it probably is. About half of the LED products tested in CALiPER Round 7 were found to have inaccurate or misleading literature. So you should be proactive, digging to find out as much information as you can about the products you’re interested in. Here’s a handy checklist to help you do that:
• Ask who made the LEDs in the product you’re assessing. Be sure it uses high-quality LEDs from manufacturers that publish reliability data.
• Ask for a luminaire warranty, which should be at least comparable to the warranties for traditional luminaires used for the application under consideration.
• Ask to see a luminaire photometric report of LM-79 testing conducted by a National Voluntary Laboratory Accreditation Program (NVLAP) accredited or CALiPER-recognized testing laboratory.
• Ask to see temperature measurement point data for the luminaire, LM-80 depreciation data for the LEDs, and information about how these data are used to estimate a lifetime.
• Ask to see any test data available about longer-term performance of the complete LED luminaire, such as CALiPER testing, manufacturer in-house testing, or field tests conducted by DOE, utilities, or other parties.
The wisest course for now is to take a conservative approach when it comes to the reliability and lifetime of LED products on the market. And keep in mind that the SSL learning curve is steep, and that it’s a challenge for manufacturers just as much as for consumers. While some manufacturers have ascended to the top of its slope, others are just beginning their climb — and some may even slip and fall. So stay alert and keep informed.
In the meantime, I’ll keep you posted on the progress of a new Quality Advocates task group that will focus specifically on this issue of reliability. We’ve also included a panel session on this topic during the DOE SSL Market Introduction Workshop, scheduled for July 13-15 in Chicago. More details on the workshop will be available soon.
