Things that I didn’t know about LED light bulbs

I got hold of LED Lighting: A Primer to Lighting the Future and learned a bunch of stuff that I didn’t know about LED light bulbs.

Many of the bulbs sold are not lighting the room via LED. The LED blasts a phosphor with a harsh narrow spectrum of (fairly blue) radiation and then the phosphor re-radiates a broader spectrum. Partly this is because the phosphor radiates more broadly than an LED and partly this is because it enables all of the electricity-to-light conversion to be done by a blue LED. A blue LED is about 30 percent more efficient than an LED of a redder color. The warmer the bulb desired, the less efficient it will be, due to the fact that more red phosphor must be used and the photons coming out of the red phosphor don’t have as much energy as those from yellow phosphor.

LED bulbs are not especially energy-efficient. They produce approximately 50 lumens/watt, which is the same as compact fluorescent bulbs.

The traditional color rendering index (CRI) to measure the extent to which a bulb puts out a broad spectrum of color is being replaced by Color Quality Scale (CQS). To some extent, this new standard is intended to make LED bulbs measure better.

Haitz’s Law says that every 10 years we are supposed to have a 10x drop in cost-per-lumen and a 20x increase in light from an LED. It is unclear how much more efficiency can be wrung from LED bulbs. This article says that someone the laws of physics have been bent by an LED in a lab at MIT, with more light output than energy in. Other sources say that LEDs are already about 80-percent efficient.

LED bulbs depend critically upon heat sink design. The bulbs don’t radiate waste energy in the form of infrared light so any inefficiency needs to be dealt with by a heat sink especially due to the fact that if the LED gets hot it will produce less light. On the other side of the spectrum, LEDs don’t put out ultraviolet light either and therefore supposedly don’t attract insects.

Household wiring is incompatible with LED bulbs in two ways: the voltage is way too high; the voltage potential rises and falls 60 times per second. LEDs operate at lower voltages and have almost no persistence, i.e., when you take the voltage away the light goes off instantaneously. In order to bring the voltage down and turn the alternating current (AC) into direct current (DC), an LED light bulb is crammed with electronics. Therefore it might not be that much more expensive to add IP-addressability and have all of a house’s light bulbs on the WiFi network.

[I’m not sure if I recommend this O’Reilly book, by the way. The author makes no attempt to explain what an LED is. I’ve found this with most books on technology written for an American audience. The lay person is assumed to be too stupid to understand anything. In this case it might be the author also who has not reached the Britney Spears level of understanding of semiconductors. He says “LEDs and SSL [solid state logic?] can be just as complex as you want to make them, especially if you have a thing for semiconductors [like Britney!] or photmetry. It’s one thing to understand the lighting in terms of heat sinks and LEDs, but it’s entirely another to get deeply involved in the electrical engineering, physics, and material sciences… Many of these elements are outside of our scope and are fully explained in white papers, journals, and textbooks, if you’d like to dig more.”]

My personal experiments with LED bulbs so far have been reasonably encouraging. I spent $38 on a Philips L-Prize Award Winning Bulb in October. It is now down to $30. I installed it in an outdoor fixture that requires a ladder to service. The other bulb is a Mitsubishi Verbatim BR30 at the top of a stairwell. This is a truly challenging destination to reach. The bulb produces a beautiful uniform light. It is rated at 85W equivalent but seems brighter. It impossible to tell whether there are really multiple small LEDs inside and the spectrum seems pretty broad. It was a little more than $30 and I got it from a specialist retailer,, founded by Rick Regan.

The Philips Web site distributes a calculator that shows the payback period for Massachusetts (average electricity cost 14 cents/kwh) is 3.8 years (26.7 percent ROI) when replacing $1 bulbs with $30 LED bulbs that are used 1000 hours per year. As noted above, given that LED bulbs are not substantially more efficient than CFL, it does not seem as though replacing CFL bulbs makes economic sense, though personally my experience with CFL is that the bulbs fail after a year or two, the light is slow to warm up, and disposal of the mercury-laden product is a distressing operation for a parent. A calculation that I have not seen done is of when it makes sense to buy an LED bulb given how quickly the prices are coming down. Even if one is using incandescents it might make the most sense to continue using them for one more year and then switch to a Philips bulb that is selling for $20 rather than $30 today. Philips says that a standard bulb will consume about $7 per year in electricity if operated for 1000 hours. The LED bulb that I bought has already fallen in price by more than $7 in the past month.

Based on my experience as a shopper it is a good thing that both Mitt Romney and Barack Obama constantly reassured me that America is the world’s most innovative economy. Otherwise it would be disturbing to realize that all of the innovative products that I have encountered were designed by people outside the U.S. The leaders in this market seem to be Dutch, Japanese, South Korean, Taiwanese. Does anyone know of a U.S.-based company that will be competitive in this market? General Electric has a 60-watt replacement LED bulb that consumes 30 percent more power than the Philips bulb and costs nearly twice as much.

Finally the advent of these bulbs makes me realize how poorly adapted to the modern world is the wiring in the orgy of new residential and commercial construction that the U.S. indulged in during the last two decades. We spent trillions of dollars building new structures that are wired with the wrong voltage and that have the dumbest possible controls. One would think that every new outlet would have USB power (example) and that every lighting fixture would have a 3.3V or 5V output as well (see this IEEE Spectrum article that explains more of the engineering behind a bulb).

[Separately, I have been shopping for retrofit power strips that contain USB power outlets. Despite the supercharged price and size of some of these devices (up to $50), they often put out a feeble 0.5 amps of power rather than the 2.1 amps required by popular tablets such as the iPad. has some that are spec’d at 2.1 amps but the typical name brand surge protector is down at 0.5 to 1 amp shared between two USB outlets. Monoprice retails a four-port charger with a total drive capability of 2.1 amps for $7.80. Why can’t a $50 surge protector include the same capability?]


  1. Bonnie Albright

    November 23, 2012 @ 6:17 pm


    You may want to look into getting bulbs through your business, as that $30 bulb will cost you about $5 if you buy it in the next year or so. The seller manages the rebate so you only pay the $5.
    I’ve been trying to get LEDs to work for the gallery but am having some difficulty getting the right spread and color temperature.
    Since I use so many bulbs through the gallery and store I’ve been looking since I opened 3.5 years ago for ways to lower my electric bill (and not overheat in the summer). I have a horrible time with the durability of compact fluorescents, so LEDs are better in that regard.

  2. Jay C

    November 23, 2012 @ 6:31 pm


    Check out Cree’s LR6 can lights. Cree is a us based company They have a CRI of 90. I installed them in my house As for electricity not being wired for LEDs its not wired for compact fluorescents either, as they require a ballast to operate.

  3. Dan Wallach

    November 23, 2012 @ 6:57 pm


    One American player is Cree, based in North Carolina. They mostly sell LED components that others build into finished products. They do sell some nice LED downlights. Several years ago, I bought a bunch of LR6’s, which are “60W equivalent” and a very good match for a halogen bulb. Mine consume 10.5W and put out 600 lumens. The latest consume 11W and put out 1000 lumens. For places where high CRI is a must, they have a more expensive “LRP” line which have 94 CRI.

  4. Tiago

    November 23, 2012 @ 7:01 pm


    I second the comments on the durability of CFLs. I’ve replaced all the lights with CFLs in a couple of houses and several died after a year or so.

  5. dg

    November 23, 2012 @ 9:23 pm


    If you are already ordering stuff from Monoprice, get a pair of 8323’s for an extra $23. Googe the reviews, everybody loves them and nothing can touch them for the price.

  6. Dave

    November 24, 2012 @ 12:07 am


    I’ve not read the book you cite, however SSL is a lighting industry term for ‘Solid State Lighting’

  7. Joe R

    November 24, 2012 @ 12:19 am


    There are no LEDs that are 80% efficient.

    1 watt equals approximately 683 lumens (for 555nm light).

    So a 100% efficient bulb would produce 683 lumens per watt. The best bulbs I’ve read about seem to be at about 160 lumens per watt, so about 23% efficient. 77% of the energy is still converted to heat.

    What people often don’t realize though is that incandescent bulbs are between 1% and 3%.

  8. Josh Volchko

    November 24, 2012 @ 12:19 am


    To those posting about Cree, they seem to be the bulb of choice for every flash-light nerd’s torch. In this case, they care more about lumen output, but everyone seems to swear by them as a pretty solid light that I imagine could transfer over to normal lighting situations.

    Candle Power Forums may be your best bet if you’re interested in delving deeper into this bizarrely intriguing light culture.

  9. J. Peterson

    November 24, 2012 @ 2:54 am


    One other note about incandescent-replacement lights like the Philips: thanks to the necessary heat sinks, they’re heavy. This can be problematic for things like articulated (Luxo) lamps or flimsy fixtures designed for lightweight incandescents.

  10. Paul Houle

    November 24, 2012 @ 10:21 am


    Surge protectors are a scam. They trip GFCIs, catch on fire, fail competely or partially and often ruin the equipment you plug them into in the latter case.

  11. Larry Gebhardt

    November 24, 2012 @ 11:14 am


    I bought two of the L Prize bulbs for $42 on October 25th (shipped Prime, so free for me). I now see they are down to $32, but with $5 in shipping from Amazon. So they are now $5 cheaper than a month ago. Looking at that I certainly see it making sense to wait to buy more.

    I have also been using the Philips AmbiLED bulbs in 8 and 12.5 watt varieties for the last year. I’ve had two of the 8watt (40 watt equivalent) bulbs in my bath room fixtures and they were brighter than the compact fluorescents rated at 60watt equivalents. More importantly these bulbs haven’t burned out yet. In this room (probably because of high humidity and frequent on/off cycling) the CFL bulbs were lasting about 6 months. The instant on and improved color make these well worth the cost all on their own.

    The 12.5watt AmbiLED is just a hair dimmer than the 10watt L Prize bulb. Given the price difference it seems like the bulb to buy for the time being (if you absolutely need one today). At $.35 more per year in electricity costs it will take a while for the L-Prize bulb to pay back the extra cost. Longer than the lifespan of the bulbs at the current prices.

    I have also tried some other LED bulbs besides the Philips, but they are noticeably inferior in color spectrum and brightness to the Philips.

  12. goods

    November 24, 2012 @ 3:28 pm

  13. philg

    November 24, 2012 @ 3:38 pm


    Paul: Thanks. I am not looking for surge protectors because I want surge protection (I’m now 49 years old and have never been in a house or office building that experienced a voltage surge (despite having spent my childhood in lightning-plagued Maryland), nor have I ever had any electronics damaged from the mains voltage despite not generally using any surge protectors). It seems that the only devices that have USB power also claim to offer surge protection. I would rather have a straight power strip with USB.

    If you think about it, pretty soon the U.S. will be spending 100 percent of GDP protecting against potential bad events. The military protects us against the possibility of invasion. The courts and prisons protect us against loss from crime. Insurance protects us against various kinds of losses. Health care, Medicare, Medicaid, etc. protect us against loss of health. Anti-virus software and firewalls protect us against losing our computers and data. Considering how much we spend to protect what we already have, it is kind of surprising that we have any money or energy left over to try to build anything new.

  14. philg

    November 24, 2012 @ 5:40 pm


    Joe: Thanks for pointing out the efficiency stats. I’m embarrassed for having believed the articles that I’ve seen on this subject. Plainly if a device that consumes 10 watts overall needs a huge heat sink it is not very energy efficient!

  15. Larry Gebhardt

    November 24, 2012 @ 5:43 pm


    I have been in an office that suffered a lightning strike about 20 years ago. The equipment with surge protectors came out fine. The couple of computers and printers without surge protectors all needed repair. The few computers that needed repair that were protected all suffered dead network cards, most likely because the hub wasn’t protected. The insurance covered all the fried equipment. So in a sense the cost of the surge protection was wasted.

    Also, that long time ago and I wonder how much money I’ve spent on surge protection that I didn’t need since that time. It’s really hard to tell if they are a good investment.

  16. From new scientist

    November 26, 2012 @ 1:10 am


    Natural Light
    – Morning 500-590 nm (Red-Orange) light
    – Midday – 450-475 nm (Blue) light
    – Afternoon – 590-620 nm (yellow, orange, red)

    Incandescent Light 620-750 nm (Red)

    Compact Fluorescent Light 500-570 nm (Green)

    LED Light 450-475 nm (Blue light)

    New tunable LED Light bulbs may be capable of improving on nature.

  17. drew

    November 26, 2012 @ 2:53 am


    LED’s are not ‘bulbs’.. they are diodes that emit light.

  18. Dan Weber

    November 26, 2012 @ 10:46 am


    How important is broad-spectrum if used in a mixed lighting environment?

    Right now I am lit by natural light, incandescent, and florescent light sources. As long as *some* of my light is broad-spectrum, is it really noticeable if some of the rest if only coming out in a limited spectrum?

  19. Perpetual Motion Boi

    November 26, 2012 @ 3:05 pm


    “Scientists Create 230-Percent Efficient LED Bulbs”

    Now if they could only create a 50% efficient photocell we would be set. It would be better than Mr. Fusion.

  20. GK

    November 27, 2012 @ 9:42 am


    ‘Now if they could only create a 50% efficient photocell we would be set. It would be better than Mr. Fusion.’

    I think the original article at implies that local environmental energy is utilized in the process. Good way to cool/stabilize a machine, does not look like that 18th century thermodynamics yet refuted. Probably larger setup will need energy to stay warm enough. This can be used for Mr. Fusion’s cooling.

  21. Mark L

    November 28, 2012 @ 12:08 pm


    Being a professional physicist / electrical engineer, I read the original article on this “230% efficient” LED. The absolutely crucial point, stated clearly in the scientific paper but left out of every general audience article I have seen, is that they didn’t observe the claimed effect until they HEATED THE LED TO 135 deg C.

    What they are doing is running their LED not in the normal manner (where the input electrical energy is directly converted to light) but in the manner of a tiny thermoelectric engine. The input electrical power is not directly converted to light, but is used to help extract some heat from a hot (135 deg C) reservoir, convert a fraction of that extracted heat to coherent work (the light emitted by the LED), and dump the unused (wasted) heat into a colder reservoir (the ambient environment, quoted in the paper as approximately 25 deg C). All this is perfectly fine with the first and second laws of thermodynamics.

    The “230% efficiency” comes when they divide the LED light emission power (69 picowatts) by the electrical power input directly to the LED (30 picowatts). That’s not good bookkeeping. If they account for the power input needed to heat the LED as well, then the true efficiency drops precipitously. With the LED at 135 deg C, they measure that they are generating 40 nanowatts of wasted heat power. A better estimate of true efficiency is then the ratio of LED light power (69 pW) to an total input power (electrical 30 pW + LED light of 69 pW + wasted heat of 40 nW), which is 0.17%. Unfortunately, that is not an impressive number, so the authors chose to trumpet the 230% figure to the press.

  22. Charles E Flynn

    November 29, 2012 @ 11:41 pm


    Here is a fun use of LEDs and some programming:

  23. presidentpicker

    November 30, 2012 @ 10:15 am


    are you suggesting that houses should be wired for 3.3 and 5v DC? Not really feasible unless you install a transformer every few outlets and light fixtures and/or use heavy gauge car battery style wiring to carry huge current and prevent significant voltage drop. That was covered in a Britney Spears electronics course 🙂

  24. philg

    November 30, 2012 @ 3:09 pm


    presidentpicker: Since LEDs are so efficient, I don’t think the wiring would need to carry heavy amperage. An iPad draws 2.1 amps at 5V from a USB outlet. That’s 10.5 watts or about the same as a Philips L-Prize bulb. The charging cable supplied with an iPad is not like a car battery jumper. Either one would standardize on something like 12V DC with the expectation that bulb designs would use multiple LEDs in series or a lower voltage designed for LEDs to be used in parallel. Perhaps there would be one DC power supply per room of a typical house? That is a lot more streamlined than what we’ve got right now (wall wart plugged into every socket; power supply built into every bulb).

  25. Alan C

    December 6, 2012 @ 5:57 pm


    I have to say I’m not convinced of the case for LED lights and not just because they don’t produce a natural feeling light.

    Europe introduced new technical standards for lighting on environmental grounds. The main effect though has been to protect Philips (of the Netherlands) and Siemens (of Germany) from competition from China which was eating up the market for incandescent lamps. This is a form of trade protection of course paid for by consumers of much more expensive alternative lighting equipment.

    The European industry promotes the much greater efficiency of LEDs compared to incandescent lamps of course. You mention some of the figures yourself. What they are largely talking about is electricity consumption to light output ratio. They blithely assume all the rest is wasted. In fact it is largely emitted as heat. In cool temperate England where I live the background heat from incandescents contributes to home heating. More efficient lamps means I turn up the gas boiler (furnace) to replace it. Of course, if you live in a climate where you need air conditioning, then its all goodness.

    Anyways, in England at least we no longer have a choice, the sale of incandescent lamps of 100W has been banned and lower wattages are to follow. I hope its worth it.

Log in