Based on work conducted at the Pacific Northwest National Laboratory (PNNL), the US Department of Energy (DOE) is readying a report on another round of LED retrofit lamp tests including A19, G25, MR16, PAR20, and PAR30 products. Over the 14 months since the prior round of solid-state lighting (SSL) retrofit testing, PNNL reported that the mean cost in terms of dollar per kilo-lumen dropped from $130/klm to $63/klm.

PNNL lighting engineer Michael Royer presented the results as part of the LEDs in Lighting track at the Strategies in Light (SIL) conference in Santa Clara, CA. The DOE will post details of the testing in a report due in a few weeks. The testing included 38 lamps from 18 manufacturers that were bought anonymously by PNNL personnel from 9 retailers (summarized in nearby table).

The PNNL bought and tested three samples of each lamp. The lab performed LM-79 photometric testing using an integrating sphere. The lab did not test the samples with a goniophotometer. The lab bought the lamps last summer and performed the testing during the fall of 2011.

Efficacy rising but range is wide

Royer begin the presentation of results reporting that both the minimum and maximum efficacy in tested lamps continues to rise, but that the range in efficacy is still broad. In this round of testing lamps ranged from 36 lm/W to 72 lm/W with a mean of 53 lm/W. The DOE has consistently found such wide ranges in efficacy in its varied SSL testing programs and Royer said the range is getting wider. He attributed the range to the different product types and application requirements that lamp manufacturers are pursuing.

Royer discussed cost versus efficacy across the products tested noting that the mean was $0.69/lm/W in the prior round of testing dropping to $0.49/lm/W in the new round. Generally, he said, consumers are getting more "bang for the buck."

A19 lamps deliver consistent efficacy

Moving to specific lamp types, the latest tests include evaluation of 11 A19 lamps. Royer said those lamps were grouped rather closely in terms of efficacy with the mean at 60 lm/W. Lamps that dissipate less than 10W were in the 50-55 lm/W range with those dissipating more than 10W in the 55-70 lm/W range. The prior round of testing had a much broader spread with some lamps below 20 lm/W.

The latest round of testing marked the first time the DOE had evaluated G25 lamps. Royer said the LED G25s performed "better than halogen but not quite as well as CFL based on what was on store shelves." That comment referred to efficacy that ranged from 35-55 lm/W.

The MR16 and PAR20 LED lamps provided a significant efficacy advantage over incumbent lamps. In the case of PAR30 lamps, the 8 LED lamps tested were evenly spread above and below the benchmark CFL efficacy in the 50-55 lm/W area.

Royer said that the PNNL tests did not cover all of the criteria in Energy Star testing. But he noted than in the areas that PNNL covered, "A vast majority of the products are meeting the Energy Star criteria." He said that many meet requirements such as efficacy even though they haven't been qualified to Energy Star. In total 23 of the 38 met the portions of Energy Star requirements that were evaluated by PNNL.

CCT and CRI

The lamps tested ranged broadly in CCT from 2500K more than 6000K. But Royer noted that most of the products were in the 2700-3000K range similar to incandescent and other incumbent sources.

Color rendering varied significantly across the entire test range. Royer said, "The lower cost products are the ones that aren't performing quite as well in color rendition." Some lamps fell below 70. But he noted a general improvement over prior tests with many lamps in the 80 CRI range.

The PNNL color rendering tests also included the R9 red sample that isn't used in the composite CRI score. Royer said that LED lamps are "a little better than CFL on this front."

Royer also discussed the trends in terms of lamp manufacturers participating in programs such as Energy Star and the DOE's Lighting Facts Label. In the prior testing 6% of the lamps were qualified to Energy Star and in the latest round 15% were qualified. The percentage of lamps carrying the Lighting Facts Label went from 29% to 61%.

Size, shape, and conclusions

There are still some issues with size and shape according to Royer. In the case of the A19 lamps, 6 of 11 were longer than the ANSI specification. In the PAR20 set, 2 of 5 were too long, and in the PAR30 sample 4 of 9 were too long. Obviously the length matters in some applications and not in others.

Summarizing the testing, Royer said, "We are definitely getting more for our money these days in terms of lumen output." That may seem like good news, but it isn't always so according to Royer. He said gains in LED chip efficacy are in many cases resulting in too much light output. That can lead to glare or simply too much light for an application. He said that the trend is lamps that hit the low wattage goals but are brighter than the equivalent incumbent source targeted by the manufacturer.

Royer also noted that lower lumen retrofit lamps are "disappearing for some unknown reason." He said that the majority of manufacturers seem to be focused on high-lumen lamps. Royer noted that apartments and small rooms need less lighting output, but that buyers can have a hard time finding lamps in levels such as 400 lm.