Another big breakthrough for CREE! 186 lumens per watt!

[Yucca Patrol]

From today's press release:

Cree, Inc. (Nasdaq: CREE), a market leader in LED lighting, announces it has achieved industry-best reported R&D results of 186 lumens per watt for a white power LED.

This result demonstrates Cree’s continued focus on pushing the performance of its LEDs. Cree’s tests confirmed that the LED produced 197 lumens of light output and achieved 186 lumens per watt efficacy at a correlated color temperature of 4577K. The tests were conducted under standard LED test conditions at a drive current of 350mA, at room temperature.


Edit: at the time of posting, their stock gained $1.75 this morning.

 

[clint357]

Any more info on what LED it was, or when it will be available?

 

[dougmwpsu]

XP-I available on sept 30th, 2010 with a top bin of 164 lm/w. Press release 2 months later announcing Amazing New Breakthrough! of 211lm/w. xp-k available sept 30th, 2011 with top bin of 189/w. Warm white light nowhere to be found.

 

[ifor powell]

I don't know what the lag is from R&D to production but it would be a t5 bin which is fairly amazing.....

Press release here

Ifor

 

[HumanLumen]

Looking at the fact that te Vf was aprox 3.03V leads me to conlcude that it was not a 1mm2 die, but somthing bigger like the XP-G

HL

 

[CampingLED]

Was expecting some anouncement, but this came sooner than expected. Why I say so is that the XP-G output per sq mm is much lower than the XR-E R2 EZ900 output per sq mm. So, they had the technology to improve the brightness per sq mm on the XP-G. They may be working on the heat issue in the mean time. Just my $0.02. Other than that, well done Cree, my favorate LED manufacturer.

 

[jtr1962]

I was expecting this announcement fairly soon also given how close the production XP-Gs are to Cree's last lab result ( 162 lm/W IIRC ). 186 lumens per watt is roughly 55% conversion efficiency, meaning only 45% of the power ends up as heat. And 197 lumens at 350 mA would be a T2 bin ( T bins start at 192 lumens if they continue the 30% jump between letters ). If these scale like the XP-Gs, then we're talking ~500 lumens at 1 amp! This all has me thinking U bins ( 250 lumens @ 350 mA ) may be a remote possibility way down the road. :party:

Note also that they achieved this with a fairly warm tint. 4577K is pretty close to ideal for general interior lighting IMO. Looks like we're within shouting distance of that magic 200 lm/W number. A little work on the Vf ( bring it down to about 2.95 volts from 3.03 volts ), plus a 5% increase in the output, and we'll be there.

 

[dougmwpsu]

What makes 200 lm/w so magic?

 

[jtr1962]

What makes 200 lm/w so magic?

It's magic from a psychological point of view only, just as 100 lm/W was. No physical significance to it however. It's not like we would breaking the sound barrier or otherwise doing something formerly thought physically impossible. The only real physically significant milestone would be when/if LEDs are 100% efficient. At that point there is obviously no room for further improvement.

 

[BurmaJones]

Note the "R&D package" . . . interesting.

 

[Al Combs]

186 lumens per watt is roughly 55% conversion efficiency, meaning only 45% of the power ends up as heat.

How is 186 lumens per watt 55% efficient? I thought 683.002 lm/w was the magic number.

 

[okashira]

How is 186 lumens per watt 55% efficient? I thought 683.002 lm/w was the magic number.

Do you want a solid green colored beam for your flashlight?

 

[ampdude]

That is pretty impressive at that color temperature.

 

[jtr1962]

How is 186 lumens per watt 55% efficient? I thought 683.002 lm/w was the magic number.

683 lm/W is the efficacy of monochromatic 555 nm green light. For any other light source other than monochromatic green, dividing by 683 gives totally meaningless figures which have little to do with how much of the input power to the light source is actually emitted as light, which is what the term conversion efficiency refers to. It is how much of the input power is actually emitted as visible light.

 

[blasterman]

A 4500k phosphor mix is very impressive for that efficiency level, and I will assume we have a typical 80'ish CRI as well. Impressive.

Of course, unless the emitter is the size of a pin head we'll get the typical complaints on how it 'breaks' our current flashlight optics.

 

[SFG2Lman]

damn, i meant to start investing in them before i deployed!

 

[Gomer]

A 4500k phosphor mix is very impressive for that efficiency level, and I will assume we have a typical 80'ish CRI as well. Impressive.

Of course, unless the emitter is the size of a pin head we'll get the typical complaints on how it 'breaks' our current flashlight optics.

These people will only be happy when a TiSaph pumped supercontinuum fiber optic based compact flashlight gets made. (aka, a "white" laser)

 

[ltiu]

Noooooo, there goes my credit card debt again when they release the next batch of lights based on this new LED!

 

[Al Combs]

683 lm/W is the efficacy of monochromatic 555 nm green light. For any other light source other than monochromatic green, dividing by 683 gives totally meaningless figures which have little to do with how much of the input power to the light source is actually emitted as light, which is what the term conversion efficiency refers to. It is how much of the input power is actually emitted as visible light.

OK, lets say that 683 lm/w isn't the correct ultimate efficiency ratio. Then what is? You never did say where your 55% efficiency estimate of 186 lm/w came from. The NIST link doesn't explain how they derived the 683 lm/w ratio. But it seems clear it's part of the current day definition for the lumen and the candella. Here is a link with the same 683 lm/w absolute max ratio determined in 1952 with a "standard lamp" and a thermopile to measure the heat loss. Whatever "standard lamp" means, they don't say anything about its being a monochromatic green source. It's more likely an ordinary incandescent light bulb of the day.

If you will excuse the Wiki reference, here's a luminous efficacy table that says LEDs of between 10-150 lumens/watt have between 1.5-22% efficiency. That equals an inferred 100% efficiency ratio of ≈ 681.8 lm/w. The smaller percentage number within limits of resolution is the same thing. There is in the same chart a 500 watt halogen in the chart rated at 19.8 lm/w and 2.9% efficiency. That's a 100% ratio of 682.8 lm/w. I didn't check every number in the chart, but none of the ones I checked were close to your 338.18 lm/w max ratio.

 

[BentHeadTX]

186 lumens at one watt?

Consider a Fenix L2D Q5 puts out 180 lumens at 2.4 watts or 94 lumens at one watt, that means a single AA light can do what the 2AA light does...with the same run time. Throw an optic on that puppy to narrow the beam without making the head larger (assuming it is 2mm sq. die)

Wonder when the XP-? T5 will be available? I hope within 6 months with optics to match. A single tubular AA helmet light? A 2AA 360 lumen light? A single AAA that puts out over 100 lumens out the front with at least 1 hour of run time?

Yes, I can wait awhile for the 180 lumen single AA light...woohoo!