Since most flashlights use LED’s now, I have learned a lot about LED’s over the last couple of years. And in my house I have gotten rid of almost all of my incandescent bulbs and replaced them with compact fluorescent bulbs. I was an early adopter of CFL bulbs, buying them one at a time when they were $15-20 each. Now you can get them way cheaper. Using about 25% of the energy of an incandescent bulb, they pay for themselves in power savings in a year or two. And they rarely burn out, so you don’t have to change them nearly as often.
But there are downsides to CFL bulbs. For one thing the tint of the bulb can be a bluish light. The tint of light is measured in K, from the warm orange light of incans at 2700K to the much harsher “daylight” blue of some CFL’s which is 5500K. Flashlights use LED’s which often are cool white and as high as 6500K. Fortunately, there are now 2700K CFL bulbs available that give off a much softer warmer light than those colder hues.
Another failing of CFL’s is the Color Rendering Index, which is how well the light shows all the colors of the spectrum. White light is really a combination of all colors, and colors are just wavelengths of light. Incandescent light has a CRI of almost 100, meaning all colors are represented accurately. But with a CFL, not all colors are accurate. The value for a CFL light might be 72 and might completely miss some colors. Yellows might appear greenish, which indicates poor color rendering. What is happening is that fluorescent lights actually produce very intense ultraviolet light by using electricity to excite mercury vapor inside the tube. We can’t even see UV light. But they add a fluorescent coating to the inside of the tube and when the UV photons hit that coating, they make it release visible light, though again in a very narrow range. By using different types of coatings, they can add a few different wavelengths and control the overall tint of the resulting light. But the light still doesn’t have the 100% accurate index value of incan lights.
The other thing at play here is efficiency. I have some CFL bulbs that give off the equivalent of a 60 watt bulb but use only 13 watts of electricity. Because watts are just power, the lighting people are trying to get people to think in terms of lumens which is a measure of how much light is actually put out by the bub. So a 60-watt equivalent is 900 lumens. The key to efficiency is how many lumens you can get per watt. Instead of calling this efficiency, though, they call this efficacy. So a 60-watt incan bulb puts out 900 lumens and uses 60 watts. That’s 15 lumens per watt. However, the CFL bulb puts out the same 900 lumens with only 13 watts. That’s 69 lumens per watt. That’s like comparing a car that gets 15 mpg to one that gets 69 mpg.
A couple of other problems with CFL bulbs is they contain mercury and they lose brightness over time (several years). Some bulbs don’t come on at full brightness right away, though all of the ones I have are bright immediately.
So those are the basics of you measure quantity of light (in lumens), the tint (in K), how well it represents colors (in CRI) and how efficient it is (lumens per watt). In order to help spur development of efficient high quality light, in 2008 the federal government offered a $10 million prize to any company that could produce a 900-lumen light bulb (equivalent to 60 watts) that uses only 10 watts of electricity (so efficacy would be 90 lumens per watt!). And the tint has to be a nice warm color of between 2700 and 3000K. It also must have a CRI of 90. It also had to last 25,000 hours (2 years and 10 months) and still produce 70% of the light it originally did (though they only tested for 7,000 hours and extrapolated forward).
This prize is called the “L Prize” and is based on similar reward-based research programs like the “X Prize,” a private fund which awarded $10 million to the first private company that could launch 3 people into space on a vehicle, return them safely, and re-use the vehicle again 2 weeks later. That prize was offered in 1996, dozens of companies started developing rockets, and the prize was finally won in 2004. The idea is by offering a reward, different companies will compete to win and that is more efficient than hiring a company to produce something that may not even be possible.
The L Prize was announced in May 2008 and the first company to enter a product was Philips, in September 2009. Philips already produces a lot of incandescent light bulbs and CFL bulbs, but they also own Lumileds, a company that makes LED’s, originally formed as a joint effort between Philips and HP (actually Agilent, an HP spinoff of all of HP’s non-computer products). In particular, they produce the Rebel LED which is used in some flashlights, including LED versions of Maglites.
LED’s have similar problems to CFL bulbs. They produce light very efficiently, but only in a narrow range of the light spectrum. For flashlights, the LED companies usually apply yellow phosphors to the LED itself. The LED produces intense blue visible light, but the some of the blue light causes the phosphors to produce yellow light. The combination of blue and yellow light appears white, but is usually a cool white. By adding more phosphors and different color phosphors, they lose some more output but are able to stretch the CRI value from 70 up into the 90’s. Since the same watts are producing fewer, higher quality lumens, the LED is less efficient. And because only the most efficient LED’s can even hope to produce 90 lumens per watt as required, they don’t have many lumens to lose.
Another problem with LED’s is they become less efficient the harder you drive them. So while a single Rebel LED can easily produce 300 lumens of light output, it might take 3.6 watts of power (83 lumens per watt). However, that same LED will produce 130 lumens for only 1.3 watts (100 lumens per watt!). LED’s are expensive, but in order to get an efficient bulb, Philips had to underdrive the LED’s and use more of them. So their bulb actually has 18 Rebel LED’s producing 900 lumens. And rather than apply phosphor to the LED’s, they put yellow windows over the LED’s that have phosphors in them. And to balance out all the color, half of the LED’s are actually red instead of blue. However when the bulb is lit, it looks white.
For testing, they had to supply the government with 2,000 bulbs. These bulbs were tested for performance and also installed in real world conditions. After a couple of years of testing, the government announced in August 2011 that Philips, the only company that had ever entered the contest, had won the $10 million prize.
The bulbs aren’t cheap though. They are just now coming on the market with a list price of $60, and available for about $50. One guy bought one and did a lot of testing on it, with all kinds of pictures and graphs here. His write-up shows the LED without the yellow windows and he analyzes the CRI of the light. As a comparison, he tests CFL lights and shows the spiky spectrum they produce and which colors they do not render well (0 for red, so much for having red lips). Then just today I saw an article talking about a rebate that would bring the price down a little further. Part of the intention of the L Prize was that the price would come down pretty quickly, dropping to $10 within a couple of years of production. I don’t know if that is possible, but even since 2009 when Philips entered the contest, new LED’s have come out that are even more efficient and could probably compete. Philips offers a 12.5 watt bulb based on the L Prize bulb that is about half the price.
So the L Prize winner is still expensive, but there are other bulbs available now that aren’t as efficient and don’t perform as well that are much cheaper. In fact I just bought a 30-watt equivalent bulb for $6, which I will write about tomorrow.