Cree Sets New R&D Performance Record with 254 Lumen-Per-Watt Power LED

Moddoo

Flashlight Enthusiast
Joined
Nov 19, 2008
Messages
1,243
Location
Minnesota
"Cree reports that the LED efficacy was measured at 254 lumens per watt, at a correlated color temperature of 4408 K. Standard room temperature, 350 mA testing, was used to achieve the results."

Interesting, and glad to hear more news from Cree.

I am wondering what the definition of "
Standard room temperature" is?
 

awenta

Enlightened
Joined
Oct 27, 2009
Messages
255
Location
Oxford, CT
Sounds awesome. It would be even better if we could actually get them. They should send us some for testing. Lets forward this thread to Cree. :p

Room temperature is usually 68F. But can be up to 72. So I guess anywhere in between there.
 

AnAppleSnail

Flashlight Enthusiast
Joined
Aug 21, 2009
Messages
4,200
Location
South Hill, VA
Sounds awesome. It would be even better if we could actually get them. They should send us some for testing. Lets forward this thread to Cree. :p

Room temperature is usually 68F. But can be up to 72. So I guess anywhere in between there.

In other words, the old 25C junction test? I hope not :(
 

Optical Inferno

Enlightened
Joined
Sep 28, 2010
Messages
242
Location
Niagara Falls ON. Canada
Yep...Standard Room Temperature is 25 Celsius (I know...metric system). And like AppleSnail says, it probably refers to the junction temperature.

Still, what I have heard from my Cree reps is that the difference between 25C and 80C is about 15%. Although that is probably just a ballpark figure.
 

jtr1962

Flashaholic
Joined
Nov 22, 2003
Messages
7,506
Location
Flushing, NY
At the efficiency this LED is operating at, junction temperature is probably within a few degrees of room temperature. I'm estimating around 90% WPE for the primary blue emitter.
 

slebans

Enlightened
Joined
Mar 1, 2010
Messages
457
Location
Moncton, NB Canada
At the efficiency this LED is operating at, junction temperature is probably within a few degrees of room temperature. I'm estimating around 90% WPE for the primary blue emitter.

Are you basing this on a straight line approximation of the published XT-E Royal Blue specifications?

Cree XT-E Royal Blue
53% WPE
Cree XT-E Cool White
148 lm/w

148 / 254 = .5826
53% / .5826 = 90.97% WPE

Does this not assume that the Royal Blue LED underlying the XT-E White LED has the same WPE as the XT-E Royal Blue LED? I always thought that Cree held back the highest Bins of their Royal Blue LEDs for their own production of White LEDS of the same families. In fact, I'm quite sure I acquired this information from an older post of yours. I found the assumption logical in that Cree would never want a third party to be able to repackage a Royal Blue in a configuration where the efficacy of the resultant White LED would exceed Cree's own product of the same family.

Alternatively, is your estimate simply based on a LER value of 350 lumens per watt?
254 /350 = 72.5% WPE
Add back on original Phosphor/Package losses of 20%(90% x .2 = 18%)
72.5% + 18% = 90% WPE

I hate guessing. I wish these CREE R&D press releases contained more detailed information.
 

BigRiz

Newly Enlightened
Joined
Sep 9, 2010
Messages
50
I did this timeline of how R&D efficacy from CREE has moved up over time.

120413125247671782.png
 

jtr1962

Flashaholic
Joined
Nov 22, 2003
Messages
7,506
Location
Flushing, NY
Are you basing this on a straight line approximation of the published XT-E Royal Blue specifications?

Cree XT-E Royal Blue
53% WPE
Cree XT-E Cool White
148 lm/w

148 / 254 = .5826
53% / .5826 = 90.97% WPE

Does this not assume that the Royal Blue LED underlying the XT-E White LED has the same WPE as the XT-E Royal Blue LED? I always thought that Cree held back the highest Bins of their Royal Blue LEDs for their own production of White LEDS of the same families. In fact, I'm quite sure I acquired this information from an older post of yours. I found the assumption logical in that Cree would never want a third party to be able to repackage a Royal Blue in a configuration where the efficacy of the resultant White LED would exceed Cree's own product of the same family.

Alternatively, is your estimate simply based on a LER value of 350 lumens per watt?
254 /350 = 72.5% WPE
Add back on original Phosphor/Package losses of 20%(90% x .2 = 18%)
72.5% + 18% = 90% WPE

I hate guessing. I wish these CREE R&D press releases contained more detailed information.
I assumed a LER of 340 lm/W. This gives an overall WPE of 254/340=~75%. I further assumed a combined phosphor conversion efficiency/package efficiency of around 83% (this is about the highest number I've heard for phosphor white LEDs). Therefore, WPE of the royal blue emitter ~ 0.75/0.83 = ~90%.

Yes, Cree probably would hold back their best royal blue emitters for their own white LED production but I'l hazard a guess the numbers for those aren't all that much higher than the royal blues they offer for sale (i.e. WPE of the royal blue XT-E used to make the XT-E white might be around 56%). In any case, we're both coming up with pretty similar numbers here. I'd be very surprised if the WPE of the royal blue emitter here was less than high 80s. I'd love more detailed info of WPE versus current also. I'll bet the WPE numbers get into the low or mid 90s at a few tens of mA.

Finally, it seems production LEDs are about 4 years behind what's happening in the lab. The R&D efficacy line looks remarkably straight. If it continues trending as it is, we should approach or exceed 300 lm/W in about 2 years (and also more or less approach 100% WPE for royal blue).
 

yifu

Enlightened
Joined
Oct 15, 2011
Messages
713
Location
Australia
I wished they released the 208lumen/watt revealed 2 years ago instead of wasting R and D money on prototypes. I means prototypes are useless without production. On the plus side, the colour temperature of sub 5000k is very good, a nice departure from cool whites being more efficient.
 

monkeyboy

Flashlight Enthusiast
Joined
Mar 7, 2006
Messages
2,327
Location
UK
I wished they released the 208lumen/watt revealed 2 years ago instead of wasting R and D money on prototypes. I means prototypes are useless without production. On the plus side, the colour temperature of sub 5000k is very good, a nice departure from cool whites being more efficient.

Seriously?

Without prototypes they would have nothing to put into production.
 

AnAppleSnail

Flashlight Enthusiast
Joined
Aug 21, 2009
Messages
4,200
Location
South Hill, VA
I wished they released the 208lumen/watt revealed 2 years ago instead of wasting R and D money on prototypes. I means prototypes are useless without production. On the plus side, the colour temperature of sub 5000k is very good, a nice departure from cool whites being more efficient.

The needs of prototyping and production are very different. You don't trade one for the other except in money, which is to be calculated as an opportunity cost. What does Cree get for being able to pop out reports of higher and higher Lm/W numbers every few months? Routine publicity and advertising. What would they get for abandoning that regular investment in R&D? A few million more for production - which I'm sure is a drop in the ocean of their development teams.
 

Genes

Newly Enlightened
Joined
Aug 24, 2003
Messages
85
Location
Kentucky, USA
Bottom line, the investors aren't impressed. Look what Cree stock has done in the last two years; almost tanked. The investor fervor that drove Cree for a few years appears gone.
 

Freeze_XJ

Newly Enlightened
Joined
Aug 21, 2009
Messages
43
Look what happened to most stocks... Simply said: appreciation for almost everything has dropped quite a bit, with the tech stocks doing a bit worse now due to less demand. Cree still has to care for its customers, and if they cut off R&D, they're going down in 4 years. That's roughly the same time it costs to design and build a new fab to produce more LEDs, so if you talk about shooting your own foot... I'd be quite impressed if they got this tech to the market in 4-5 years in sufficient quantity to make flashlight folks happy. It took roughly a year before XM-Ls got into our lights from the moment they were announced.

ps, don't forget there is no mentioning of die size, so if they got a 10x10 mm die, the luminosity would still suck.
 

blasterman

Flashlight Enthusiast
Joined
Jul 17, 2008
Messages
1,802
If it continues trending as it is, we should approach or exceed 300 lm/W in about 2 years

In two years time we'll be lucky to be at 200 lumens per watt in solid state fixtures, and the department of energy is even more pessimistic. We're just now creeping over 100lumens per watt in the best commercial fixtures at reasonable CRI's, and a lot of that is due to improvements in thermal handling and not so much raw efficacy.

From an industry perspective it's better for Cree's bottom line to drag our efficacy improvements to stay just ahead of their competition anyways. Which is why Genes comment makes the most sense.
 

Optical Inferno

Enlightened
Joined
Sep 28, 2010
Messages
242
Location
Niagara Falls ON. Canada
He may have been referring to just the LED efficacy. But I definitely agree with you in that we won't see 200lm/W in the next two years on the market. And as for the 300lm/W I don't think that can be accomplished due to theoretical limits for efficacy of about 270lm/W. I might be a little off on that but pretty sure it is below 300lm/W.
 

jtr1962

Flashaholic
Joined
Nov 22, 2003
Messages
7,506
Location
Flushing, NY
He may have been referring to just the LED efficacy. But I definitely agree with you in that we won't see 200lm/W in the next two years on the market. And as for the 300lm/W I don't think that can be accomplished due to theoretical limits for efficacy of about 270lm/W. I might be a little off on that but pretty sure it is below 300lm/W.
Yes, I meant 300 lm/W in the lab. Nichia said this was possible last year I think. I know they hit around 265 lm/W at a few mA, and the blue emitter they were using had a WPE in the high 80s.

Seeing the lag of ~4 years between R&D and production LEDs, we probably won't see 200 lm/W in production until at least 2014.
 

MichaelW

Flashlight Enthusiast
Joined
Dec 8, 2007
Messages
1,788
Location
USA
If this marks the 75% efficiency point, I think LEDs can more/less stop improving for flashlights. [waiting is the hardest part]
If we currently use a 50% efficient LED; drive it at 4 watts, rejecting 2 watts, we can now drive it at 8 watts, rejecting the same 2 watts. Assuming no battery limitations, we should have tripled our light?
 

Lynx_Arc

Flashaholic
Joined
Oct 1, 2004
Messages
11,212
Location
Tulsa,OK
Yes, I meant 300 lm/W in the lab. Nichia said this was possible last year I think. I know they hit around 265 lm/W at a few mA, and the blue emitter they were using had a WPE in the high 80s.

Seeing the lag of ~4 years between R&D and production LEDs, we probably won't see 200 lm/W in production until at least 2014.
when the 200 L/W LEDs get out there in mass production perhaps we can start seeing LED bulbs/fixtures sold in stores that sport 120 L/W figures and start comparing themselves to fluorescent lighting instead of incan.
 
Top