Paul Scheidt, a product marketing manager from Cree Solid State Lighting wrote me a detailed refutation and commentary about my LED lighting post a last week. Firstly, I have to apologize to Paul and all my marketing pals since when I slagged on marketing in that last post I did not make clear I was talking about the snake-oil salespeople that ran the website that compared 30 lumen LEDs to 800 lumen incandescents. Sorry, folks, I never meant to imply my marketing pals at all the analog companies I cover as well as the nice folks at Cree are, shall we say, disingenuous.

I do maintain that as far as general lighting, LEDs are not economical. At least not today. Mr. Scheidt agrees, but promises better LED lamps in the future. I have to cover what is available today, and as you could see from the website I referenced in my other post, the current crop of LED fixtures (LED plus power supply) are pretty crappy. I also say, just as I did in the original post, that there are some really great applications for LEDs— like Christmas lights that don’t burn out, especially if they are at the top of a palm tree like in that Corona commercial. But my discussion centered on using LED lamps that screw into incandescent light sockets. Now, Mr. Scheidt and I may be arguing on orthogonals — he is talking about individual LEDs and I am talking about an LED lamp with a power supply that works off 120 volts wall power. When Cree characterizes their LEDs I have to assume they do it like FET manufactures. They mount them on a cold plate and hold them at 25 degrees to measure life— since they will last longer if they run cool. To characterize luminous efficiency they might let them run them hotter, since LEDs do drop forward voltage with temperature, unlike what a commenter to my original post maintained.

When an LED fixture company comes out with something that costs as much as a CFL and has the same lumens per watt, I will wholeheartedly endorse them. Until then we still live in an analog world and there is a complex decision matrix needed to determine if the longer life and vibration resistance of LEDs justifies the higher price. But not all LEDs have the same lifetimes as the Cree lamps. This website shows a huge drop in luminous efficiency after only 15,000 hours on some LEDs, but I have to believe, as Mr. Scheidt points out, some of these anecdotal results are based on older LED designs.

Paul Scheidt writes:

ADVERTISEMENT

LEDs have come a long way in the past 2 years. However, we still have a lot of potential lighting customers that think the LED performance 2 years ago is still current and indicative of the fact that LEDs will never replace incumbent lighting technology. Articles like yours (with poor research and hear-say) only serve to continue this type of perception in the marketplace. To put it another way, you are helping to stall the adoption of LED lighting.

I just looked and half of your commenters are on to the fact that you are misinformed at best. I am perfectly willing to discuss more about LED lighting with anyone at EDN or Electronic News to help correct this kind of misperception in the market.

Let me go through your email's points in complete detail:

1. "Press release you sent last year ... [with fluorescent efficacy] was a 20 mA LED"

http://www.cree.com/press/press_detail.asp?i=1160427137863

Back in Oct 2006 (which was not last year), we launched the XLamp XR-E cool white LED that got up to 85 LPW @ 350 mA. At that time, XR-E was qualified up to 700 mA. In January 2007:

http://www.cree.com/press/press_detail.asp?i=1169819309344

Jan 2007: We qualified XR-E up to 1000 mA max drive current. This is clearly not a 20 mA LED.

As for "same efficacy as fluorescent", there is wide range of performance in linear fluorescent bulbs. Bad T12 bulbs can be 60 LPW or less. Good T5HO bulbs can be 120 LPW or more. We could split hairs, but 85 LPW is definitely above compact fluorescent (CFL) bulb performance and in the range of linear fluorescent.

2. "The large 90 lumen [per watt] one was still in the lab [last year]"

Actually:

http://www.cree.com/press/press_detail.asp?i=1182948090554

June 29, 2007 we announced commercial availability of XLamp XR-E 100 lumen min order codes in cool white @ 350 mA. This is (100 lm / 1.15 W) = 87 LPW minimum. Not typical, not "up to", but MINIMUM. Since then, we've shipped millions of these parts.

3. Efficacy of halogen

You could pore over the catalogs of the big 3 bulb makers: GE, Philips and OSRAM to get a feel for the market. Wikipedia has a nice table showing Overall luminous efficacy for a variety of light sources with links to their sources:

http://en.wikipedia.org/wiki/Luminous_efficacy#Lighting_efficiency

4. "you will gladly quote efficiency of the LED alone"

Let's be clear, we only make the LED. I don't have control over the rest of the system - that is in the hands of our customers. There will be optical, thermal and power loss associated with a complete fixture that will derate the numbers we publish as LED efficacy down.

Most LED lighting people realize this. Even the US Dept of Energy is aware and has a program called CALiPER to measure the real-world performance of commercially available LED luminaires. The website is here:

http://www.netl.doe.gov/ssl/comm_testing.htm5. "[total LED fixture] efficacy was about equal to halogen"

You will want to read Round 3 & Round 4 of the CALiPER link. In Round 4, a wall light using Halogen was tested using the same methods as used to test the LED fixtures. The halogen light only measured 8 LPW (remember, it has fixture losses, too). Even a relatively poor LED fixture scored 10 LPW in the same study - or 20% better than halogen. All the other LED fixtures in the study tested to at least 16 LPW (or 2x the halogen) or higher. Today's LEDs are apples-to-apples at least twice as efficient in practice as halogen.

6. "how LEDs are more efficient than any other light source when clearly they are not"

As they do [become more efficient], switching out old lighting technology for LEDs will create total-cost-of-ownership savings for lighting customers - through energy savings because LEDs are more efficient and through maintenance avoidance because LEDs last for much longer than any light bulb.

But it's not just the LED makers saying this. The US Dept of Energy is also involved. You can see here their portfolio of projects that have been funded with taxpayer money:

http://www.netl.doe.gov/ssl/publications/publications-reports.htm

DOE and other lighting & LED committees believe that cool white LEDs will reach 150 LPW by around 2013. You can see that on page 55 of this document:

http://www.netl.doe.gov/ssl/PDFs/SSLMYPP2007_web.pdf

The DOE funding serves to ensure that LEDs do hit those efficacy targets and do become the most efficient light source. But, I admit, we aren't there quite yet.

Part 2 of my response is that right now, we are only discussing the total light output of the light source - not the efficacy of the entire fixture. LED light is directional (not omnidirectional like a fluorescent tube or high-pressure-sodium (HPS) bulb), so it is MORE efficient in directional lighting fixtures than the traditional bulbs. (you don't get light loss because of a reflector)

Part 3 is to answer the question you're probably thinking right now: "why isn't that reflected in the CALiPER results"? Lighting companies have fairly long design cycles, about 12 months on average. It takes a while for them to design products around the brightest and most efficient LEDs and get them into the market. CALiPER only tests commercially available fixtures.

To review, there's 3 different development stages to keep in mind with LED lighting:

1. Low performing LED fixtures available today (what CALiPER tests)

2. Higher performing LED fixtures that will be released soon (the LEDs are commercially available)

3. Highest performing LED fixture that will be made in the future (once the LEDs become more efficient)

What you may be hearing is initial results from stages 2 or 3 that will come true soon, even though products from stage 1 are only available today.

Thanks for your comments Paul, and I will be sure to use the latest data when doing my analysis as well as checking in with Cree and other lamp manufactures. I do have to say that I met a chief technologist for a giant lamp manufacturer when I was at the Strategies in Light conference last moth, and he was not worried about LEDs talking over his business, at least not just yet. I do agree with Paul that the future holds promise, I just wish it were here now.