Purchasing new hardware? Read our latest product comparisons

A tale of two tests: why Energy Star LED light bulbs are a rare breed


April 11, 2013

How do you measure the life of a light source that could last decades? (Photo: Shutterstock)

How do you measure the life of a light source that could last decades? (Photo: Shutterstock)

Image Gallery (4 images)

Just over a week ago we reported that Philips' 22 W LED light bulb, designed as a like-for-like replacement of a 100-W incandescent light bulb, was the first LED bulb of its type to receive the stamp of approval from Energy Star. But looking at the Energy Star requirements reported by Philips in its press release, it seemed a little strange that Philips' product is the only one to have been certified – given that products long on the market appear, at face value, to meet those requirements. Since then, Gizmag has spoken to LED light bulb makers Switch Lighting and other industry players to find out why they're apparently playing catch-up.

"Those who do not remember the past …"

It turns out the reason is simple, but has a little bit of back-story attached. Unlike Energy Star certification for other types of appliances, an Energy Star-certified LED light bulb signifies reliability and performance as well as energy-saving performance. The reason for the difference is down to the problems faced by ordinary people when compact fluorescent bulbs first appeared.

"Originally, the Department of Energy was very concerned about what happened with compact fluorescent lighting when it first entered the market in that there were all kinds of variations of quality, reliability and performance," Gary Rosenfield, EVP of Marketing & National Accounts at Switch Lighting told Gizmag. (Though the lighting aspects of Energy Star now fall under the jurisdiction of the Environmental Protection Agency, when it came to specifying Energy Star criteria for LED products, it was the DoE.)

Compact fluorescent light bulbs are often criticized for being dim and slow to warm up (Photo: Mathew Bajoras)

The result has been something of a minor backlash against compact fluorescent lighting, though in reality this was more down to the shortcomings of individual products and the misleading ways in which they were sometimes sold. The DoE was keen to avoid a repeat performance when LEDs arrived, since LED products also vary enormously in quality. The relevant tests, devised by the Illuminating Engineering Society, are known as IES LM-79 and IES LM-80. A manufacturer that wants to apply for Energy Star certification must first pay an approved independent laboratory to carry out the tests on its behalf. With the results in hand, the manufacturer can then apply to the EPA (provided they're up to scratch, that is).

There is no fee for applying for Energy Star certification, nor for using the label. In fact the tests in themselves do not specify pass or fail results, but instead simply quantify performance and lifespan. It's the EPA that decides the benchmarks. Light output depends on the bulb's spec, but is a minimum of 1,600 lumens for bulbs purporting equivalence to a 100-W incandescent bulb. Among the other criteria are color rendering index of at least 80, and a rated life of at least 25,000 hours – more on that in a moment.

"The LM-79 is basically the test that defines the photometrics, in other words that defines what the light performance is," Rosenfield said. LM-79 quantifies light output and distribution, as well as electrical power, allowing the calculation of the all-important efficacy in lumens/watts. It also identifies the light's color characteristics, such as its appearance, and its ability to render colors accurately.

"That's a fairly short test to run," Rosenfield explained. "It's done with spheres in independent labs. You put the product into these spheres, very expensive test equipment that basically encloses the entire fixture, and it tests all those variables." Easy peasy.

An Integrating or Ulbricht sphere used for photometric testing of lamps and light fittings (Photo: Gamma Scientific)


"The other test […] is purely a life test," he added. "And in our category that test is run for 6,000 hours, so it's effectively nine straight months of testing." Life test is a useful if loose euphemism for what the test actually measures, which is lumen depreciation: the extent to which the lamp's light output falls over time.

"All LED lights at some point are going to start losing the amount of lumens that come out of the light," Rosenfield said. "The industry has set a number, called L70, which basically represents 70 percent of the original light output, as the end of life for an LED." So, the industry defines an LED light bulb as dead when its output falls to 70 percent of what it was to begin with.

The problem, Rosenfield explained, is that LED light bulbs haven't been around long enough for the industry to have a good understanding of how their performance degrades over time. Instead an algorithm is used which, after 6,000 hours of testing, projects the number of hours remaining before the LED will reach L70. The 25,000-hour rated life of Switch's and other LED products has not been tested in full because doing so would take nearly three years. Instead this number is a projection based on the algorithm used. Even so, it's still a test that takes a significant amount of time. Depreciation, by the way, is non-linear, and light output can at times increase as well as fall away, Rosenfield explained.

There is a semi-shortcut, though. LED light bulbs are effectively electronics packages in which the LEDs are but one type of component. Being solid state lighting products themselves, the LEDs are eligible for testing under IES LM-80. If Cree, one of the LED manufacturers Switch has used for its bulbs, has its LEDs tested to 6,000 hours, Switch doesn't have to repeat the entire test. Switch's product can instead undergo a 3,000-hour (4.5-month) test to check that depreciation follows the expected pattern. "It's still an advantage to test for the entire 6,000 hours because we may be able to run longer than what the LEDs actually are claiming," Rosenfield said.

Quality control

Though LEDs may not generate heat to the extent of other light sources, they are very sensitive to heat, so equally, just as a well-designed light bulb might extend an LED's longevity, might a poorly designed one compromise it? "I think that many companies fall out of the Energy Star qualification process because of that, because they don't have sufficient thermal management which is one of our big advantages," Rosenfield said, alluding to the LQD liquid cooling system employed in Switch's bulbs.

"Also companies fall out because they don't have the full light distribution required. For example, with an 'A lamp,' you have to have, to get the full Energy Star standard, 170 degrees of radial flux or light distribution all around the product at generally the same intensity all the way around," he added.

To pass the Energy Star benchmarks, Rosenfield concludes, products must perform robustly as well as demonstrate long life. He also explained that the EPA will pick up Energy Star-certified products in the market place to ensure that their performance matches what was certified – that the products are what they're being sold as, effectively.

The floodgates open?

But why is Philips' product the only "100-W equivalent" to have been certified? "They won't necessarily be the only one, because it's just a matter of when companies are able to start their testing," Rosenfield said. "The longest lead time of this whole process is the four and a half months initially for the LM-80 testing for at least a half the full test […] assuming you're working with an LED manufacturer that has already done their testing."

Are Switch products going to be certified in the near future? "As a matter of fact we'll shortly be announcing the first set of our products that will be Energy Star qualified," said Rosenfield. "We're within days of that."

Though Switch has had products on the market for some time, it's worth bearing in mind that Switch is a start-up company whereas Philips is one of the world's largest lighting manufacturers – not to mention one which manufacturers its own LEDs.

Osram Sylvania is another company with apparently eligible products. We asked it if Energy Star certification was on the cards. "The EnergyStar qualification is a very rigorous test with tight parameters," a spokesperson told Gizmag. "Some of the screw-based LED bulbs need to be tested for at least 6,000 hours (for example) and this can take up to nine months – which is why some of our bulbs have not been qualified yet."

Though we also reached out to GE about its LED products, they declined to respond.

Official line

In an email to Gizmag, an EPA spokesperson confirmed the certification process. "To earn the label, the manufacturer of the product must be an Energy Star partner and obtain third-party certification of the product based on testing in an EPA-recognized lab," the EPA said. "The requirements for LED bulbs address efficiency as well as a number of performance related metrics such as light distribution, durability, lumen maintenance and color quality (both initially and over time). Higher lumen products are eligible for certification, as long as they meet all the requirements and complete all the necessary testing."

It sounds as if a number of Energy Star LED products, 100-W equivalent A-lamps and otherwise, are headed for the market. This can only be a good thing. For LED products, Energy Star has set robust guidelines that should ensure that an Energy Star bulb is of high quality. Provided consumers know what to look for, perhaps LED can avoid the demonization that befell compact fluorescents.

About the Author
James Holloway James lives in East London where he punctuates endless tea drinking with freelance writing and meteorological angst. Unlocking Every Extend Extra Extreme’s “Master of Extreme” achievement was the fourth proudest moment of his life. All articles by James Holloway

22 Watts isn't much of an efficiency advance over the CFL lamps, but of course they make up for it in longevity.

I'm just not sure it's worth the price difference at the moment.


Well thought out and researched articles like this make Gizmag a real cut above the rest. Good work!!


@Wombat56, if you are only talking about 100w equiv in cost alone CFL is cheaper even over like 5 years of operation. 22 watts vs the last CFL I tested at 25 watts.

A difference of 3 watts if left on for a year would be 26.28 kWh/year. For residential electrical costs of about 11 cents/kWh it would only save about $2.89/year. At current prices (~$55 for LED, $2 for CFL) it would take about 18 years for the LED to end up cheaper.

If you compare the LED to an incandescent 100w bulb however the savings for using the LED is $6/month and even at current prices it would pay for itself in under 9 months.

There are other advantages to LED though like longer lifespan, more durability, no hazardous materials to clean up if you break one etc. I once dropped a LED light from high enough up doing construction to break the lamp but the LED was fine.

100w replacement LED bulbs are still expensive because they are new but you can buy 40w and 60w replacement bulbs for $10 to $14 at home depot. You can buy a 6 pack of 40w replacement bulbs for the cost of the 100w equivalent Philips and most home lighting situations don't need 100w equivalency.

It wasn't that long ago (about 5 years ago I think) that those bulbs were $50 and $60 each so I am sure the 100w equivalent bulbs will drop in price before long as well.


I still don't see a cycle rating. Cfl's were advertised as lasting 5 times as long as incandescent bulbs and if you turned them on and never turned them off they might.In actual use they don't near that long. Cycling is what kills them, it has nothing to do with run time. I would expect led's to do better, but if I was rating light bulbs I would at least acknowledge that fact. If you are going to rate a bulb for 25,000 hrs it should have to go through a 25,000 cycle test. Wouldn't take but 3 days to do it with a 10 second cycle.


I am a fan of LEDs, but I have two small corrections to make to the payback calculation above. 1. You need to consider the interest on the money used to buy the $55 LED. If you had put the money in the bank instead, you would have earned interest. If you had paid the money off your mortgage, you would have saved even more in interest. Suddenly, your saving of $2.89 per year is more like zero. This is a basic piece of economics, and you need to make sure it is taken into account whenever someone tells you about the 'payback' period (for instance on solar panels on your roof). 2. On a more positive note, CFLs only last a couple of years if left on full time, so you would have to buy about 9 of them over 18 years. And that's a lot of mercury to put into the environment.

Nick 1801


That makes sense. The CFL's in out bathroom lighting kept burning out. They are this style: http://i.imgur.com/zGomDJu.jpg

I was thinking they may have been too poorly ventilated or something in that fixture but it didn't really click how they would be cycled more often than other locations in our home.

We replaced 2 of them with LED about a year ago and so far the LED's are still going strong. We have 5 or 6 LED's in total from different manufacturers so far around our home and no failures yet. Not that it means much but we have had CFL's fail in under a year in the same setting.


I'm not sure why but my power bill went down substantially when I switched a handful of the most used lights in our home from CFL to LED. In the range of $20 Canadian a month. Maybe it takes a lot of power to ignite the CFL? Not only that but I have yet to have any failures of the LED bulbs whereas we were replacing CFLs often. I did this approximately 3 years ago and have been happy so far.

Zaron Gibson

Poorly understood simple components in CF and LED's can result in device failure after only a small fraction of their design life. Electrolytic capacitors are often used and these fail early due to temperature (life is reduced by a factor of 2 for each 10degC rise). Hot bulbs kill capacitors and the bulb engineers/manufacturers often pay little attention to these simple passive components resulting in early failures that could have been prevented by different capacitor selection.


I have been using cfls for almost 20 yrs. My experience is that they are much more sensitive to heat than incandescent bulbs. If you put them in confined enclosures designed for incandescent bulbs they are prone to fail quickly. CFLs perform very well if you make sure they get adequate ventilation.

No experience with LED lighting yet. Way too expensive compared to well ventilated CFLs.

Charles Hart

@Diachi "...most home lighting situations don't need 100w equivalency." That is a matter of opinion. I would have to differ there. I cannot stand anything less than 75Weq. 100Weq is my target. With 1600+ Lumens I can see details and do not need glasses. With anything less than 75Weq I cannot see small details and my peripheral vision does not work at all glasses or not. Unless you put three or more 40Weq bulbs for every 100Weq, which negates all the savings, it would be analogous to torture for me. It would be Just enough light that my dark vision doesn’t kick in and not enough to keep from bashing my shins into everything in the house. I agree LED replacement bulbs still are not affordable enough to use in every fixture. They do have a use when mixed. I have several dimmable fixtures that could really use them as CFL bulbs will not even light at full power. As others have noted the CFL’s are sensitive to on/off cycles due to the plasma arc eating the internal leads. Mythbusters had a show on “Is it cheaper to leave the light on for a few minutes or turn them off?” One of the tests they ran was on/off cycles as the electric costs didn’t justify it. They set it up to switch on and off every five seconds. The lights failed in this order; CFL, industrial FL Tube, incandescent, and the LED had yet to fail after a month. So if you have a light that does on and off regularly an LED could save considerably over any of the other choices.


These are anecdotal observations only so take them for what they are worth:

I switched to heavy cfl use in 1988 partly to deal with an undersized AC in a new house. I reasoned that in addition to saving on electric with CFLs they would also generate a bit less heat resulting in less AC load.

I have since given up on CFLs ever getting over their limitations and shifted my hopes to LEDs. A few years ago I dabbled in what was available at the time and was disappointed. Cheap no-name LEDs failed way too often, not much light, too big, too ugly for some applications.

We just bought a new house that had no fixtures (common in FL when a foreclosure involved). This gave us the opportunity to buy fixtures that would work with the latest LEDs.

I resolved to only buy Philips in the hope of getting reliability and quality. I didn't worry about price. I bought three of their yellow ones with the three heatsink slots to experiment with and was impressed but dismayed by cosmetics for some applications. OR maybe I should say my halogen and incandescent-loving wife was. Those are now deployed in my office and they do a good job though they do look a little geeky.

Then they came out with these just as we needed to buy a bunch:


I'm pretty impressed with these. They also come in an 8 watt version, two color temps. Smaller, better looking, nice light.

We bought (or I should say "invested") in 25 of these. I can't say what the energy savings might be but they really work well. The trick is picking the right fixtures if you have the chance. Open for cooling of course, large enough shades to cover, and in the case of ceiling fan light kits or fixtures over a dining room table, with enough bulbs to amass sufficient lumens to compete with halogen and incandescent.

That's another issue. We're getting older and that means we need brighter light. over the dining room table we have a fixture with 6 of these things pointing down. Plenty bright.

If these things prove to be reliable and long-lasting I think we will have lucked out and had them come on the market just when we needed to buy a bunch.

Bottom line? Take a look at the new Philips I linked to, and also their new non-dimmable "round" ones that work well in table lamps. And stay away from cheap LEDs.

And if you can, buy fixtures with LEDs in mind.


Not sure why you LED bulbs are $55. I have been buying them for around $20-25 a piece. Best buy carries its own line of bulbs, that are very bright, and work very well. They have many advantages over CFL. I switched when I got sick of having to replace my CFL bulbs every 2-3 years. For some reason no matter which brand of CFL I use, they are always dying.

LEDs come on faster, light up brighter, the light doesn't vibrate like a CFL. Which is better for your vision, and migraines.

Considering I have yet needed to replace a LED bulb yet, and I have had CFLs die with in 6 months, thats my number one reason for switching. Now if a bulb dies, I buy an LED.

Preston Garrison

Charles, I think time will bear the same characteristic of LED bulbs - confined, poorly ventilated fixtures (cans specifically) will cause LEDs to fail more quickly, though at a much reduced rate as the electronics from reputable manufacturers seem to be more robust. The only two LED bulbs that have experienced failure in our house over the past four years have been two in our kitchen cans. Both were super cheap (around $12 iirc) from Costco around three years ago. CF floods is the same cans also experienced fairly rapid failure (compared to incandescent). Philips LED floods have proven to last this duration (plus) with no failures in the same fixtures. I'm confident that GE, et al would perform equally well.

We've also been much happier with LED performance in outdoor floods since in cooler temps they switch on to (seemingly) full power instantly where CF floods routinely took 30 seconds to a full minute to reach useful output depending on the air temp.

It also seems as though LED bulbs from different (reputable) manufacturers are far more consistent in color temp side by side than CFs. It was often a pain to swap CF bulbs in multi bulb fixtures because, not only were different manufacturers' bulbs different colors, but the same bulb when newer also created a different color temp.

Especially now that LEDs are becoming more affordable, we're slowly getting the entire house converted - but more rapidly. We're about 2/3 there...

Vince Pack

LEDs are not affected by rapid cycling, CFLs like to stay on for 30 min per cycle minimum, excessive re-strikes will kill CFLs quick. More important is the light color, LEDs are great at reproducing warm white produced by incandescent or the cool white of halogen, CFL color, not so much! With all the positive advantages, LED is the only choice that makes sense

Rob Led
I tend to make my own light fixtures so using CFLs is super easy. If I need more light I install more sockets or buy a CFL that draws more watts. I do suspect that CFLs that are inverted such that the heat rises up over the electronics will tend to fail more often than CFLs installed upright or sideways. The CFLs that I leave on 24/7/365 tend to get three years or better before failure. When I went to all CFLs in my home from old fashioned bulbs my power bill dropped at least $20 per month and being that I install far less CFLs than one a month I know it is a real savings. Jim Sadler

CFLs contain mercury, so I'm skipping them entirely and purchasing LEDs as my incandescents die. BTW, if you leave at least one incandescent in a circuit (of pot lights) the LEDs will dim without any special new switch.

Fritz Menzel

The health benefits of LED lights are greater. The light spectral power distribution (SPD) curve is closer to natural light than fluorescent light. (SPD shows the radiant power emitted by the source at each wavelength or band of wavelengths over the visible region (380 to 760 nm))

Fluorescent light has narrow spikes at certain wavelengths and hardly has any red light. LED is smoother and more consistent across the whole range. Recent research has shown that red light is necessary for sleep and for regulation of the Circadian Rhythm which in turn regulates our hormone levels. Not all light is equal.


@ Diachi "For residential electrical costs of about 11 cents/kWh it would only save about $2.89/year" Are you in the USA? If so I didn't realise electricity is so cheap for you. That would explain why Americans appear to be so wasteful with leaving lights and airconditioning on all the time ( no offense intended, just a relative statement). In Australia we are paying over double the cost per KW/H (25.378 cents including tax), despite having a slightly stronger dollar than the $US and average incomes being comparable. Thus the eneregy savings gained by the use of more efficient technology such as these LEDs will be far more pronounced. Given electricity is so cheap in the USA, I find it very surprising that there hasn't been a large uptake of electric vehicles. You even get government concessions to buy them. I assume it is because your gas (petrol) prices are still relatively cheap. It would seem that an answer to reducing energy consumption in the USA would be to make it more expensive. That would in turn increase the efficiency of vehicle and appliance manufacturers. It would also be far better for the environment. Of course, the same principle would be beneficial for Australia and the rest of the world too. Energy is still too cheap and we don't respect it enough. A small car can easily produce more than 100KW (134 HP) these days. An average adult cycling on an exercise bike can produce a constant 50W (0.067 HP). Yet we think nothing of burrying our foot into the accelerator as the lights go green, nor the fuel we are burning, despite it being 2000 times greater than what a person produces (for this example). One thing is for sure. We cannot continue indefinately pumping oil out of the earth to burn so casually. It will start drying up and costing a lot more. Not "if", just when. I sincerely hope people start to think about reducing their energy consumption and move more towards electric vehicles. Otherwise, as world oil supplies dwindle, there will be more wars and more bloodshed. The decisions we make now will affect tomorrow. Any energy efficiency focus is welcome and that's why I am delighted that the energy star testing is becoming prominent.


What about EMI? Most LED lights are terrible RFI emitters and more than anything else they have made life miserable for radio amateurs. Either start designing quiet ones or incur the wrath of hams worldwide.


Fantastic article, with the exception you are missing some manufacturers with stringent QC practices and EnergyStar rated products. We have come across a few great LED products over the years through detailer research, trial and error. One of my preferred brands is Zenaro, an LED available in the US and Canada. They have a line up of Energy star stamped products that I use frequently and the prices are comparable to that of Philips and GE.


Australian: I and (unfortunately) too few other Americans greatly appreciate your frank observations & comments - all of which hit the nails right on the head. The only reply I can provide is that we have an endless supply of brainwashed, uneducated, what's-in-it-for-me and/or unimaginative citizens in our country. That is why, despite all our potential and many enlightened citizens, we are headed in the direction we are. We are powerless in the face of monsters like FoxNews, and there's absolutely nothing we can do but to intercede to the best of our limited ability and hope.

Fritz Menzel

Cheap LED lamps die early because of heat that kills other components such as the capacitors used in their current control circuits. Testing only light output is inadequate. NASA does accelerated life tests on electronics by increasing temperatures and input voltages to their expected maximums and have done so for 50 years. A LED bulb may be relatively simple but poor component choices will make them an expensive mistake.


Nice to know that there actually are reliable tests being done rather than just rely on consumer feedback and company propaganda. LEDs have indeed come a long way since it was first used for conventional lighting. But even then it was already clear that the LED would soon outpace CFL for the simple reason that CFL has very little more room to grow plus its health and environmental issues.

Cassandra Allen

I will never buy another CFL, or allow in my house. I had one in a reading lamp, and when I would sit by the lamp the side of my face facing lamp would start to feel like it was being sunburned. I admit I am chemical sensitive, but clearly there is a problem with the material used in these bulbs that has yet to be addressed. I feel we can do better. I would like to add we had one that absolutely melted down, and we are not sure why.

I have replaced all the CFL bulbs with LEDs and so far no problems, and my energy bill has come down. I am happy.

Andrea Nash

" Hot bulbs kill capacitors and the bulb engineers/manufacturers often pay little attention to these simple passive components resulting in early failures that could have been prevented by different capacitor selection. "

grtbluyonder, you are overlooking the obvious. Manufacturers aren't failing to pay attention to the simple passive components. In fact, they are paying very close attention to those components - right down to cost per piece. The key to reducing consumer price tag is to minimize manufacturing costs. When production runs in the millions, the tens of millions, the tiniest fraction of a cent adds up quickly.

Noel K Frothingham

I've heard the concerns on how consumers failed to adopt CFL's after experiencing failed products and light quality, but I am afraid the same will happen to the LED bulb as manufactures race to the bottom with $1.00 LED bulbs. I have seen LED warranties 10 years out, which is complete BS as the life cycle of LED product is very short.

I would like to see all the relevant certification printed on each box listing chips used.

Some manufactures cannot be botherd with testing, as testing takes time and each third party test cost money.

I have seen people in Costco purchase three bulbs with equilivant to a 60 watt light bulb on the package over a single more expensive bulb because they don't understand things like lumens per watt.

The success or failure of the A19 LED bulb will come down to marketing and price, not on quality or value.

Rick Hyne
Post a Comment

Login with your Gizmag account:

Related Articles
Looking for something? Search our articles