PDA

View Full Version : US Dept. of Energy Report on LED Lamps In Our Industry


JimLewis
01-17-2009, 01:20 PM
I tried doing a search to see if anyone else had posted this already but didn't find anything. So sorry if this has already been posted here.

Here is an interesting report from the US Dept. of Energy concerning claims made by manufacturers of outdoor LED lamps.

---

LED
DEPT OF ENERGY FINDINGS REPORT

• For the nine manufacturers who published overstated performance information, some compared the outputs of their products to incandescent products; some have published explicit, but inaccurate, values for the output or efficacy of their luminaire; and some have published values for output or efficacy, but have omitted to indicate what the values correspond to (i.e., they may be publishing a ‘lamp’ output for the LED devices, without explicitly stating that). A number of manufacturers claim their products provide equivalent output to a 20W halogen MR16 or 15-25W incandescent, yet, based on CALiPER testing, they only produce one-sixth to one-half the output expected from such incandescent or halogen lamps.
As suggested in earlier reports, for the products with overstated published values for performance, the divergence from actual tested values may stem from a number of issues:
• Misinterpretation or lack of experience relative to SSL testing concepts (e.g., LED device performance vs. luminaire performance, lamp efficacy vs. luminaire efficacy, relative photometry vs. absolute photometry)
• Lack of industry standardization in LED device performance testing and reporting and infeasibility of determining luminaire performance based on reported LED device performance
• Confusion or lack of clear distinction in marketing literature between LED device performance and luminaire performance literature which claims an output of 2800 lm for 25 W power draw – which would imply 112 lm/W). Round 5 of CALiPER testing will test SSL T8 drop-in replacements and benchmark fluorescent T8 tubes in troffers to enable direct comparison between the two sources used in fixtures. The Round 4 testing of this drop-in lamp was not performed in a fixture, so a reference ballast was used. Further testing of this and similar products in Round 5 will use typical magnetic and electronic ballasts (in troffers), as well as the reference ballast measurements.

• The SSL MR16 products all claim to be replacements for halogen 20W MR16 lamps, but do not produce output levels or center beam candle power at levels that would typically be seen in 20W MR16 halogen lamps. Considering only 20W, flood-style, 40° halogen lamps, IES files from lamp manufacturers indicate outputs ranging from 200-450 lm, center beam candle power (CBCP) of around 500, and efficacies of 9-19 lm/W. Across the three SSL MR16 products that were tested, the output and CBCP are, in general, less than half of what would be typical of traditional products used in this application: the highest output was 133 lm (averaged over two units of the same product) and the highest CBCP was 283 cd. Two of the SSL MR16 products have efficacies which are indeed higher than halogen MR16s, and the third has an efficacy similar to the high end of halogen efficacies. Surface temperature measurements taken on the heat sinks of these products indicate relatively high operating temperatures ranging from 54-83° C—indicating that in situ and lumen depreciation testing should be conducted to study the reliability of these products.

• While LED devices and hence SSL products are expected to have very long lives, the true rate of lumen depreciation of SSL luminaires and replacement lamps is largely unknown. Lumen depreciation characteristics for LED devices may be available from device manufacturers, but when those devices are integrated in luminaires, the LED device’s long-term performance will be affected by many factors stemming from the luminaire as a whole. These factors include, most obviously, operating point characteristics (such as forward current) and thermal management (such as heat sinking and thermal bonding), but also other factors such as the use of materials during manufacturing which may effect LED device lifetime such as cyanoacrylates (super glue) and O-rings. Lumen depreciation of the LED source is not the only failure mode contributing to the reliability of SSL luminaires, but it is a primary facet of luminaire reliability that is being studied by the CALiPER program.

• Of note in these results:
• For one product, 06-10, the output dropped to less than one-fifth of the initial output in less than 1000 hours. This product used LED chips from an undisclosed source and only a copper core printed circuit board as a heat sink.

• Analysis of the initial color shift results must take into consideration that the uncertainty factors for these spot color shift measurements have not yet been determined and may be relatively large. Nevertheless, for the two luminaires that have shown significant lumen depreciation in the first 2500 hours of operation, the color shift is also significant: the color shift for the undercabinet fixture, 06-10, is .056 (eight times more than the ENERGY STAR lifetime color shift limit), and the shift for the RGB outdoor area light, 06-05, is 0.024 (three times more than the ENERGY STAR lifetime color shift limit).

• Performance Reports in Manufacturer Literature
As evidenced in this round of testing, a preponderance of inaccurate and misleading performance claims on SSL product literature persists in the marketplace. This phenomenon is increasingly worrisome with more LED products becoming available through major retail outlets.
Only two of the 24 SSL products tested in this round provided accurate or somewhat accurate performance data in product literature. For some products, little to no performance information is provided. Most products tested have packaging and/or advertisement material and/or data sheets that make highly overstated claims of wattage equivalencies to traditional sources. Such equivalency statements are published for the majority of products tested, e.g., “Replaces 40W”, “Compare to 60-watt light!”, or “50W LED head is = to 100W Metal Halide.” In every case, these comparisons are inaccurate –often overstated by a factor of 2 or 3.

• These erroneous and misleading comparisons, like other inaccurate performance claims, may stem from a number of factors:
• Lack of understanding of SSL testing concepts (e.g., that LED device performance does not translate to product performance once the LED device is integrated in a replacement lamp or luminaire);
• Manufacturers’ product literature that does not clearly indicate what specific product configuration was tested to produce the published performance values (e.g., rapidly evolving LED devices and a tendency to publish data for the highest performing version of a product—often that with the highest color temperature);
• Differing interpretations of benchmark values for products using traditional light sources (note that the CALiPER program will soon be issuing benchmark reports for a number of applications to provide a clearer basis for comparison); and
• Inflation of performance claims (e.g., from selection of test conditions not representative of actual use, such as chilled or pulsed device testing or testing without optics, or from testing of prototype units that are not representative of production units).

• In addition to false performance claims, SSL replacement lamps purchased from major retail stores may have questionable reliability. For one product, 2 out of 4 units failed before testing could be completed. Another product from the same manufacturer and purchased from the same retail chain had 1 out of 2 samples fail, and yet another type of lamp from the same line had 22 out of 80 LEDs in the lamp fail during testing. For another replacement lamp, purchased from a different retail chain, 1 out of 3 units failed. On visual inspection, the design and construction (as well as quality control) of these products do not appear to be robust—thermal management is not apparent and structural defects are visible. While the CALiPER program recognizes that these are not statistically large samples, these failures can be seen as reasons for caution.
Manufacturers and retailers need to be wary.

• The replacement lamps tested in Round 4 provide examples of products which do not necessarily have poor performance results (their efficacy is better than similar halogen and incandescent products), but which have been misrepresented in product literature. The MR16 and candelabra lamps that were tested may fill needs in market niches (for example, for low wattage devices in decorative applications), but cannot serve as direct replacements for 20W halogen or incandescent products as claimed in marketing publications because their output levels are too low. The T8 drop-in replacement product that was tested produces far less light output than claimed in its product literature, but with its fairly good efficacy, it may be a suitable replacement for fluorescent tubes for specific applications.

• The intermediate results from in situ and lumen depreciation testing serve as a reminder that the reliability of SSL products is still largely theoretical. Claims of consistent color maintenance in luminaires and 50,000-hour lifetimes (or more) are at this point only speculative. Decisions which take into account the expected life of an SSL product should consider product guarantees, knowledge about which LED devices are used in a product, the reputation of the manufacturer of the LED devices used in the luminaire, and the reputation of a luminaire’s manufacturer.

• Because of the wide variation in performance and the immaturity of this industry, it is essential for buyers to request explicit indications of luminaire output and luminaire efficacy and to be informed enough to question the information provided by manufacturers carefully.

irrig8r
01-17-2009, 02:13 PM
Very interesting report.

INTEGRA Bespoke Lighting
01-17-2009, 05:12 PM
Jim, please indicate the date that report was first published. I recall reading this about a year or so ago and noted then that it was already out of date and not representative of 'current technology'. Also, there should be an appendix that indicates exactly what products were tested.

The efficiency and efficacy of HB LEDs and their associated applications, be it in a lamp format or a fixture format are constantly improving. There is a lot of 'junk' out there in the market, but there are also a few gems. Don't throw the baby out with the bathwater.

JimLewis
01-17-2009, 06:28 PM
You probably did read it about a year ago, because that's when the article was written; January, 2008. But it looks like it was new then, not outdated.

Here is the link to the article from the actual DOE website;

http://www.netl.doe.gov/ssl/PDFs/CALIPERROUND4summaryFINAL.pdf


And it looks like the article I first posted is some sort of summary of the larger article found at the link above. Because the article at the link above has the same text as the article I posted originally, just interspersed inside of a much larger report, from what I can tell. I didn't read much of the larger report.

That's all I can find on this report. I wish I had more to give you. It was passed on to me by my lighting rep. with no other info. attached other than what I pasted above....

JimLewis
01-17-2009, 06:39 PM
Here is another article that has a lot of the text above in it;


http://ecmweb.com/design_engineering/electric_standards_deviant/

Mike M
01-17-2009, 06:57 PM
All I know is the part about their samples coming from "retail stores" really stood out.

Alan B
01-17-2009, 07:16 PM
Exc post Jim. I'll trust the US Dept of Energy report over manufacturers (or people who resell LED's on the internet).

I came to a very similar real life conclusions. (wonder why real intergrated LED fixtures are not out there? There's a reason).

Not to put LED's down, as Interga says some are better than others.

I personally think retro LED bulbs are more of a statement (of being green/progressive)--than a solution. Sometimes when I read the claims (especially as applied to retro bulbs), I think it borders on deceit.

David Gretzmier
01-18-2009, 03:57 PM
Wow, to think the government testing agrees with my tests about all the LED claims out there- amazing.

I do however, again, thank all those folks out there who are testing Leds on thier cusomers to determine which ones work and which ones don't. In 5 years, we'll all know what to use and what not. I'm all for testing, as I am updating my ryco trans test today, but I think this report supports my belief these Led's are not ready for prime time.

try to keep in mind I tested my first LED mr16 5-6 YEARS ago. I remember it had multiple ( 21? 36? ) 5mm leds and was claimed a warm white. I also tested several 3 watt luxeon single led's and a 5 watt luxeon. all burned out in less than a year.

They've come a long way, as the 3- cree 1 watts warm whites are pretty dang close to a 20 watt and have lasted over a year in my tests. OOH, at 30 plus bucks a bulb, they SHOULD. At 5 years I'll be impressed.

JoeyD
01-21-2009, 10:44 AM
Those are bullet points pulled from these links.....

http://www.netl.doe.gov/ssl/comm_testing.htm

http://www.lrc.rpi.edu/programs/solidstate/assist/pdf/AR-LEDLightEngine-May2008.pdf