Anything brewing at Lumileds?

First we had the 1W, then the 3W, plus minor improvements in output and color. (I'll ignore the 5W.)

Shouldn't there soon be a high output/efficiency 1w or 3W with say, double the lumens/watts we're seeing now? I know they've got to be working on it for the automotive industry.

Personally, im hoping they apply the same process they did with the 1-3W to the Luxeon V.

I know Nichia and Lumileds signed an agreement to share technology, and Nichia plans to start mass producing the 60 lm/W white LEDs in 2005 which they lab tested last year. Therefore, while I don't think we'll be seeing any huge efficiency improvements this year, next year or the following year look very bright (no pun intended). Additionally, we have no idea what Nichia or Lumileds may have tested and not publicly announced (so as to keep the competition guessing). I for one am far more interested in efficiency improvements than in power increases. The higher you get the efficiency, the less percentage of your power goes into waste heat, and this means that you can increase the power level in an existing package. Lumiled's 3W Luxeon package can deal with ~3W of heat with no problems. At present LED efficiency is maybe 10%. Were this to increase to 50% then you could drive an LED in the L3 package to over 6W, and since the efficiency per watt is five times higher, the light output per package increases by a factor of about 10. The same reasoning applies to the more common T-1 3/4 LEDs. Current white LEDs produce 1 to 2 lumens but if efficiency increased to 50% they could be driven at 40 mA instead of 20 mA, and produce 10 to 20 lumens-more than enough for many small lighting applications. While this year doesn't look too exciting, the future looks bright. /ubbthreads/images/graemlins/grin.gif

I agree with the common content of your post, but I do not understand a few technical details:

[ QUOTE ]
jtr1962 said:At present LED efficiency is maybe 10%.

[/ QUOTE ]
What do you mean with an efficiency of 10%? Is that 1W of electrical input power gives 100mW of radiation?
Or do you think that they are now at 10% of the theoretical limit (White or monochrome?)?

[ QUOTE ]
and since the efficiency per watt is five times higher,

[/ QUOTE ]
What is the efficiency per Watt?

[ QUOTE ]
jtr1962 said:At present LED efficiency is maybe 10%.

[/ QUOTE ]
What do you mean with an efficiency of 10%? Is that 1W of electrical input power gives 100mW of radiation?
Or do you think that they are now at 10% of the theoretical limit (White or monochrome?)?

[ QUOTE ]
and since the efficiency per watt is five times higher,

[/ QUOTE ]
What is the efficiency per Watt?

[/ QUOTE ]

10% efficiency means that 90% of the energy you feed it is lost as heat and 10% is photons. 10 percent is a fancy way of saying 10 per hundred. If it's 10% efficient with a million watts input, you get 100,000 watts worth of photons out.

This is similar to why Americans refuse to accept a flat tax. If the tax rate was 10% for everybody, most people think the rich pay the same as the poor. It may be the same rate but 10% of $10,000 is $1000 and 10% of a million is $100,000.

[ QUOTE ]
PeLu said:
What do you mean with an efficiency of 10%? Is that 1W of electrical input power gives 100mW of radiation?
Or do you think that they are now at 10% of the theoretical limit (White or monochrome?)?


[/ QUOTE ]
I meant that 1W of input power gives 100mW of optical output and 900mW of waste heat. Translating into lumens per watt, 10% efficiency means roughly 25 lm/W for white light, which is about where Luxeons are now, although the efficiency for some of their colored LEDs, especially red and red-orange, is much higher (maybe 20 to 25%). Actually, both definitions you gave mean exactly the same thing since 100% efficiency is the theoretical limit, which incidentally will never be reached in practice (although that would be nice since it would mean a light source producing no heat whatsoever). However, I feel we will reach 50% efficiency within a decade at most, and perhaps eventually 80% to 90% efficiency.

[ QUOTE ]

What is the efficiency per Watt?

[/ QUOTE ]
Actually, that's a mistake on my part. There is no such thing as efficiency per watt. I meant lumens per watt.

I think Cree is putting a big push on to get more products into the market soon, so perhaps Lumileds will announce something this year. Most of the bright lights using Luxeons require either CR123 cells or a lot of AA cells to really light things up. High efficiency LEDs would allow really impressive output using just a couple of AA cells.
/ubbthreads/images/graemlins/popcorn.gif /ubbthreads/images/graemlins/sleepy.gif

[ QUOTE ]
jtr1962 said:
I meant that 1W of input power gives 100mW of optical output

[/ QUOTE ]Thanks for the clarification. [ QUOTE ]
Translating into lumens per watt, 10% efficiency means roughly 25 lm/W for white light

[/ QUOTE ] Depending on which 'quality' of white light you want to have. I remember 199lm/W as the theoretical limit for white light with a CRI of 100.

[ QUOTE ]
efficiency for some of their colored LEDs, especially red and red-orange, is much higher (maybe 20 to 25%).

[/ QUOTE ]A green LED might be 'only' 30% efficient by your definition and still have more than 200lm/W.
For me this kind of efficiency is a little bit misleading (should not be any critics, just my feeling, for you it may be the other way round as you can easier tell the wasted energy).

[ QUOTE ]
However, I feel we will reach 50% efficiency within a decade at most, and perhaps eventually 80% to 90% efficiency.

[/ QUOTE ]

As I wrote somewhere else before, I remember similar claims from 7 years ago. And the climb was much slower than predicted. Of course I wish I would be wrong here.

Does anybody remember what's the phosphor's efficiency?

And BTW, is the Lumiled development center still in the Netherlands?

[ QUOTE ]
PhotonBoy said:
I think Cree is putting a big push on to get more products into the market soon, so perhaps Lumileds will announce something this year. Most of the bright lights using Luxeons require either CR123 cells or a lot of AA cells to really light things up. High efficiency LEDs would allow really impressive output using just a couple of AA cells.
/ubbthreads/images/graemlins/popcorn.gif /ubbthreads/images/graemlins/sleepy.gif

[/ QUOTE ]
You can drive a high-power LED with any power source that can produce sufficient current.

Boost converters aren't terribly complicated, and permit driving a 3.5V LED from voltages down to, say, 800 mV.

Unless white LED Vf's miraculously drop, you'll always need a boost converter for most single-cell applications.

[ QUOTE ]
PeLu said:
Depending on which 'quality' of white light you want to have. I remember 199lm/W as the theoretical limit for white light with a CRI of 100.


[/ QUOTE ]
You're probably right here if you mean a spectrum comprised of equal energies across the entire visible range (380 to 700 nm). However, if you drop out the deep red and deep violet ends of the spectrum and perhaps have only frequencies from 420 nm to 730 nm then your maximum theoretical efficiency jumps to about 250 lm/W while the CRI only drops to 98 or so. I would imagine for non-CRI critical applications like streetlighting you could drop out even more frequencies and obtain white light in excess of 300 lm/W, but maybe with a CRI of only 70 or so (incidentally way better than the sodium vapor streetlights currently in use which have a CRI of around 22).

[ QUOTE ]
A green LED might be 'only' 30% efficient by your definition and still have more than 200lm/W.
For me this kind of efficiency is a little bit misleading (should not be any critics, just my feeling, for you it may be the other way round as you can easier tell the wasted energy).


[/ QUOTE ]
Lumens per watt is easier to use in that you know you'll get a certain amount of light out per watt in. As an engineer I also like to know the efficiency so as to calculate waste heat. Right now I assume that most of an LED's power ends up as heat but that will not be so in the future.

[ QUOTE ]

As I wrote somewhere else before, I remember similar claims from 7 years ago. And the climb was much slower than predicted. Of course I wish I would be wrong here.


[/ QUOTE ]
Efficiencies of 45% have already been demonstrated in the lab for red-orange LEDs, and lesser efficiencies for other colors (Cree's new production XThin blue chip has an efficiency of 28%). While it will take time to develop production LEDs based on these lab samples, the demonstrations at least show that efficiencies around 50% can eventually be attained in the relatively near future (< 10 years). I'll admit my eventual prediction of 80% to 90% may be a stretch, but I think gradually as processes are refined over the decade or two after 50% efficiency is attained we will get there.

[ QUOTE ]
PeLu said:
Does anybody remember what's the phosphor's efficiency?



[/ QUOTE ]
Peter, as I am sure you know, phosphor efficiency differs depending on whether you are speaking of quantum efficiency [1 photon to 1 photon is 100%] or energy efficiency [1 watt to 1 watt is 100%]. I don't remember the exact values but I believe that the quantum efficiency is >60%.
FWIW, I remember reading in a Lumileds paper that the typical spectrum of a white LED has a luminous *effacacy* of around 300 lm/W.