Nobody wants to waste energy or money to light their home or workspace. It seems logical to simply repopulate every fixture with the most efficient lamps available. But that may be a poor solution because it ignores a fundamental principle of lighting design: Make only as much of the right kind of light as you need, and shine it only where it is needed.
New energy-efficiency standards for electric lights, based on “lumens per watt,” have been imposed in the United States and elsewhere. Lamps that don’t meet these standards are being phased out and replaced by more efficient (and more expensive) products. Unfortunately, standards based solely on lumens per watt don’t tell the whole story—they ignore that first fundamental rule of lighting.
I’ve been shopping for lighting based on lumens per watt since I first moved off the grid in 1991. With my first tiny solar electricity and battery system, it was critical to use as little energy as possible. That need remains, even now that I’m lighting a bigger space with a bigger PV system. I’ve had many lighting failures. Besides a shorter-than-expected lamp life and poor light quality, many of the lamps used more energy than claimed for the amount of light produced. Running two “efficient” lights to replace a single inefficient one doesn’t make sense. So I asked acclaimed lighting engineer Howard Brandston, an outspoken critic of government lighting bans based on lumens per watt, what I had been doing wrong all these years.
“All lumens are not created equal,” Brandston says. “Those metrics are simply agreements amongst scientists, and have very little to do with the human perception of light. Eyes are extremely sensitive—your visual system, all the way from eye to brain, works at 85% effectiveness in only 7.5 foot-candles of illumination. That is not much light. Mozart wrote his music under the light of only two candles, and the average person can read The New York Times...in only two foot-candles.” To put this into perspective, full sunlight has an illuminance of about 10,000 foot-candles; an overcast day, about 1,000 foot-candles.
Both biology and psychology are at least as important as physics. The lumens per watt standard was developed in 1924 to compensate for the difference in sensitivity of the human eye to various frequencies of light. But it doesn’t compensate for how the eye shifts its color sensitivity in different brightness conditions, or for the fact that most energy-efficient lamps emit light in a very narrow spectrum compared to the broad spectrum of sunlight. These effects can make a room seem brighter or dimmer, and also result in lighting that seems inadequate or is unpleasant in quality.
The graph (See ITA) shows the spectral distribution of light from different sources. Both sunlight and incandescent light have an even distribution of frequencies. Fluorescent (FL) and compact fluorescent (CFL) lighting have large spikes at certain frequencies, with light-emitting diodes (LEDs) somewhere in between. This is why the light from FL and CFL lamps can seem harsh and fatiguing, and also why Brandston says that incandescents are “the gold standard of lighting.”
Both the lighting industry and the U.S. Department of Energy (DOE) are aware of the limitations of lumens per watt and are developing new metrics. I asked Mark Rea, director of the Lighting Research Center at Rensselaer Polytechnic Institute, why the situation is so confusing for consumers, and what is being done to improve it.
“We need to do a better job at recasting the problem as a benefits-versus-cost decision,” Rea says. “There is a lot of effort in the industry right now focused on ‘bundling’ the numerous benefits of different lighting choices into an easy-to-understand format. For example, if a lightbulb has good color quality, no detectable flicker, and is dimmable, it could be labeled as a Class A light source.”
Rea and other industry experts are working with the DOE and its Energy Star program to make things easier, but it will take time. “It’s still the wild, wild West out there with residential lighting,” says Rea, “and consumers still have to choose by trial and error.”
Advertisement
The spectral distribution of common light sources graph has two errors that should be corrected. First, the red peak beyond 600 nanometers is missing for FL and CFL lights. FL's are heavy in the green area, but do have significant red output. Second, the color wash behind the graph is backwards. The graph reads blue to red from left to right. Longer wavelengths are redder, until you get into infrared that we can't see. That's where the maximum output of the incandescent bulb is headed on the graph.
Hi David,
You are correct about the rainbow color was being reversed in the graph's background. As the article's designer, that was my mistake...I was thinking "Frequency" when the x-axis is clearly marked "Wavelength". As for the spectral distribution of the CFL, I looked at a lot of curves for "typical" CFLs, and while they all had similar shapes (spikes), they also varied a lot in their distribution. Perhaps I picked a poor one to be "indicative". See http://ledmuseum.candlepower.us/led... for the varieties and similarities...a neat site.
I would like to see additional coverage of "quality of light." The article does a good job of identifying the limitations of the 1-dimensional Lumens/watt rating, but could go farther in explaining color temperature and color rendering index. Halogen incandescent lamps can do an excellent job of replicating the color temperature and CRI of sunlight - but they are not very efficient.
So I am continually trying to find more efficient lamps that have a 3500ºK color temperature, and a CRI of 90 or better. These are rare and very expensive lamps typically. One of the biggest disappointments is finding a lamp with good quality of light, paying a fortune for it, and then having it fail after 6 months of use. Both LED and fluorescent-ballasted lamps do not survive in closed fixtures where heat can not escape. I would love to see a lifetime study of quality lamps in typical application fixtures - I doubt anyone is getting more than a few thousand hours (no better than the incandescent lamps). Moreover, I find that the fluorescent lamps (CFL) have a significant shift in output spectral content as they age - maybe you could quantify that in a future article?
Really enjoy the magazine - keep up the good work!
Thanks for your comments, EP.
When I interviewed Dr. Rea, we in fact did discuss CRI. But it, too is very subjective. Rea's research found that 65 to 70 percent of people in their studies preferred light more towards the 'daylight' side of the spectrum than the 'warmer' side . Very-high-CRI lamps are indeed very expensive, and are mostly used in photography, cinematography, and microscopy applications. For more info about CRI (and CCT) , the Wikipedia has a good summary with lots of math:
http://en.wikipedia.org/wiki/Color_...
Personally...I like my lighting towards the warm side of the spectrum, which does not give a high CRI. But it does make my light-brown wood interior walls and ceilings look cozy, and makes my two white and brindle-colored cats' coats look more lustrous.
I also hope that some research institute will undertake a long-term study of lamp life, but with new LED lamps and fixtures entering the market at a furious rate, compared with the estimated lamp life, I think such metrics would quickly become obsolete; by the time the lamp dies (even if it dies young) multiple generations of new and improved-lifespan models will have already replaced that discontinued lamp being tested for lifespan.
I have also noticed the change in CFL spectral distribution with age that you mention. I hope that PEARL will be able to continue their research as lighting and luminaire technology progresses. At the end of the day, though, lighting is still at least 50 percent psychology and phisiology.
Best regards, and keep your lamps trimmed and burning;
DAN FINK