New Luxeon: Rebel

[mobile1]

I dont know but the 115lu press release and now the rebel dont make much sense to me... than the k2 is still somewhere out there or is the k2 now gone and replaced by the rebel...
To me it seems like they are struggling for news.. the rebel is somewhat on par with the market but not the product that would keep their customers from switching to the crees and ssc of this world...

 

[Effulgence]

Waiting for something solid to arrive... Dare I say... 20+ Rebel emitter light?

 

[IsaacHayes]

28th I think was the official launch date, so that makes sense that they are all out. I remember someone mentioning the 28th.

The max flux green, royal-blue, blue, and all whites show up but call for avail...

The ones that seem to be in stock and ready are the cool-white 40 and 50lm @ 350ma versions. Bummer.

 

[CM]

70 lm/W was groundbreaking just MONTHS ago, not even years! Man, you guys are hard to please. My personal opinion, spoken without much industry inside info, is that the efficacy improvements we see as end-users actually follow innovations in semiconductor manufacturing technology and phosphor technology that are the work of very heavy-duty PhD's. The Rebel isn't doing any better than Cree 7090 XRE or Seoul P4 because they are all up against the same technology ceiling.

Oh, and, FYI, 300 lm/W is never going to happen. The absolute peak efficacy is 683 lm/W but that is specific to a sickly monochromatic shade of yellowish-green at 555 nm wavelength AND it requires absolutely perfect conversion of 100% of the electrical input into photons i.e. absolutely no waste heat whatsoever. For white light, the absolute pinnacle is about 240 lm/W and again that requires 100% perfect conversion of electrical input into photons, which is a long long long way off.

For the record, both lumens and watts are measures of power. Converting between them is kind of like converting from inches to millimeters. Lumens and watts aren't quite as simple as inches and millimeters, but the point here is that the relationship is fixed. One inch will never be more than 25.4 millimeters. Likewise one watt will never be more than about 240 lumens (I am forced to say "about" because that number can vary slightly, depending on exactly which wavelengths are combined to make white... this is because the human eye is not equally sensitive to all wavelengths). The only complicating factor is heat. You don't lose any millimeters to heat when you convert to inches, but you DO lose some watts to heat when you convert to lumens.

I understand what you're saying but this news is like someone saying "we have a new car that gets 12 miles to the gallon" right after someone announces that they have a car that gets 40 miles per gallon (and when people are concerned about efficiency). Yeah I know, not the same exact thing but I think understand my point. A smaller die with less efficacy is not where it's at. The big market is in dies that put out more light per given input power. I used a term lm/w but that's more or less the de facto metric when talking about efficiency. That was not meant to imply that the term watt has any meaning--I agree with you there.

However it does not change the fact that more output and higher efficiency is where things are going. Compact flourescents are soon to be replaced by LED's that are more like Cree's XR-E and less like the Rebel. The flashlight market is small compared to the replacement market for line voltage flourescent and incandescent fixtures. And I can almost assure you we're not going to get there with "advances" like the Lumiled Rebel. To me, it's a ploy by Lumiled to distract people from the fact that they're getting their arses kicked by Cree.

Oh, and saying 300lm/w is "never" going to happen is a strong statement that I think will be proven wrong in due time. It may not happen in the next few months or years but ruling it out absolutely at this point is speculative. BTW, I threw that number out tongue in cheek but since you chose to be argumentative, I thought I should present a counterpoint.

I dont know but the 115lu press release and now the rebel dont make much sense to me... than the k2 is still somewhere out there or is the k2 now gone and replaced by the rebel...
To me it seems like they are struggling for news.. the rebel is somewhat on par with the market but not the product that would keep their customers from switching to the crees and ssc of this world...

Edited to add the above. I think you took the words right out of my mouth. "Struggling for news..." Reminds me of the dotcom days when companies would make worthless press releases to keep the investing public interested. Yep... Sorry lumileds, I need to yawn again.

 

[daveman]

Three pages worth of posts about absolutely nothing.

 

[EngineeringGuy]

Oh, and saying 300lm/w is "never" going to happen is a strong statement that I think will be proven wrong in due time. It may not happen in the next few months or years but ruling it out absolutely at this point is speculative.

I am going to have to side with soffiler on this one. Lumens are a measure of the light output of a device. A lumen is a unit of measure specifically weighted for the response of the human eye, but it is still a measure of output power.

Another way to look at this is that we are converting electrons into photons. Each electron has a specific amount of energy and each photon generated also contains a specific amount of energy. We cannot generate any more photons than we have energy to produce. The lm/w figure is then the efficiency of conversion (ratio of photons to electrons). Basically it is the percentage of input energy that gets converted to the desired output energy, but the units are goofed up. If you were to convert the lumen figures to radiometric output (units in watts) then it immediately becomes obvious that you are expressing the efficiency of conversion. By definition efficiencies can never be more than 100%.

I know that there some who are holding out some hope for a cold fusion process which will allow LED lights forever and for ¨free,¨ but the energy still has to come from somewhere (its that whole E=MC² principle). The light conversion process will never be more than 100% efficient... ever. The good thing though is that if we approach 100% efficiency the devices ought to be able to be driven harder because they will produce much less heat than current devices. They can then get brighter for a given size, but not more efficient.

Hope this wasn´t a bore to everyone...

 

[Gryloc]

To all (and not any single person to clarify, just in case),

Its late for me, but maybe area lighting is not the goal of the Rebel LEDs. I keep hearing about these being used for full-color RBG signs, like what you see in the football/basketball/baseball stadiums (even the nice color screen on the rear of the stage on the show American Idol- something I usually prefer not to watch by the way). I suppose you can get a brighter, higher resolution display by using these versus standard LEDs due to their small sizes. I don't know. Don't forget about cell phone lighting (since the Lumiled's "Flash" emitter kinda sucks in terms of brightness. These are tiny, and, don't forget cheap (-er).

Oh, and I can see where these can affect the big-screen LCD television market. More can be crammed into a smaller place for a brighter back-lighting solution, while saving a few bucks. These are much nicer than Luxeon 1 emitters in a way when it comes to RBG backlights and lighting, right? It seems like so.

I bet that Lumileds did throw the Rebel out there for the customers to gnaw on until their better and brighter "Gen X" product is ready for sale. This product, whenever it will come out, will better suit our needs here at the CPF. I think that the Rebel is a nifty little product, and I can think of many nice uses, like for signal lighting for automotive use. Bigger clusters of cheaper and more efficient (because there are more in the cluster to produce the same light) Rebels would make great tail lights, flashers, and even decent reverse lights. Meantime, use the Seoul P4 for headlights.

I agree that it is not the best compared to Cree and Seoul, but Lumileds tried. Imagine creating a new blue LED die from scratch to compete with Cree's. I heard from many on here the Cree always had superb blue (or was it royal-blue) dies, which are great for white LED lighting. I bet Lumileds tried hard but could not get it quite right to meet the specs of Cree's XR-Es exactly. Cree probably had the upper hand a bit for a while, and they could have lost it (maybe all together since the Q2 bin XR-E was just finally released). I bet Lumileds had a vision that the K2 was the future (higher current handling), and while distracted with this, Cree, with a slight and brief burst of speed, leapfrogged Lumileds. No big deal. No company is a loser. Each brand has a product to suit their purposes (and different customer industry types). Companies can screw up once (or twice) here and there. Maybe you will be surprised later... you never know.

Well, this is how I see things... I always try to think optimistically; it does not hurt to do this, by the way. Just sit back and wait patiently for the right LED to be available. Technology takes time. Poor Lumileds is already feeling the heat. Don't forget that they probably still have to compete against their own design manufactured by other foreign companies (clones and look-a-likes). I wonder if Cree or Seoul will face this problem later on. Hmmm... Anyways, Lumileds will work things out. Every company probably slows down in a race at times and have problems, and each handles these problems differently.

Dang, I need to go. It is way too late here and I have class in the morning. Oops...


-Tony

 

[270winchester]

yawn.

back to sleep I go.

 

[chris_m]

Cree probably had the upper hand a bit for a while, and they could have lost it (maybe all together since the Q2 bin XR-E was just finally released).

Not strictly accurate - the Q2 bin has only just been released to consumers. Romour is the Q bins have been out for a while - probably since November when they were originally scheduled - it's just that somebody big has been gobbling them all up.

 

[riffraff]

Its late for me, but maybe area lighting is not the goal of the Rebel LEDs. I keep hearing about these being used for full-color RBG signs, like what you see in the football/basketball/baseball stadiums[...]

Yeah, that's the first thing I thought of...smaller size = higher density = higher resolution screens.

 

[Opto-King]

Yeah, that's the first thing I thought of...smaller size = higher density = higher resolution screens.

You still have the problem with the heat, the heat sink will be much bigger than the LED it self. Seccond, the LED displays that today are using "normal" PLCC2, 4 and 6 LEDs are working well and to a cost of 5% of the Rebel LED. Just take a look at the soccer games in Italy, the comersial LED signs are using "normal" SMD LEDs.

 

[soffiler]

...Oh, and saying 300lm/w is "never" going to happen is a strong statement that I think will be proven wrong in due time. It may not happen in the next few months or years but ruling it out absolutely at this point is speculative. BTW, I threw that number out tongue in cheek but since you chose to be argumentative, I thought I should present a counterpoint.

I did not intend to be argumentative and I'm sorry if you read it that way. I am an engineer and I deal primarily with facts, NOT speculation. I tried to explain the facts to you, and EngineeringGuy did a nice job in post #86 also. Briefly restated, in order to achieve anything above ~240 lm/W of white light you're going to need the equivalent of a perpetual motion machine. Using your tongue-in-cheek number of 300 lm/W, it means an efficiency of 125% - getting OUT fully 25% more power than you put IN. If you understand the laws of physics and thermodynamics, you know that you can never, and I stress never, get more out than you put in.

Personally, I think the confusion arises because people don't grasp that a lumen is simply a measurement of power. It's called "photopic" power because it is weighted for the response of the human eye. If, on the other hand, (as EngineeringGuy also said) we threw away the word lumen and used "radiometric" power instead, the radiometric units of light are watts. Now you can look at electrical power input in watts, and light output in watts, and immediately see that the ratio cannot exceed 100%.

 

[SemiMan]

"Meantime, use the Seoul P4 for headlights."

- Automotive manufacturers are looking more in the 4,500K for headlights. The Seoul P4 would never pass muster on a car due to the color (and lack of color consistency). Car headlights also cram a lot of light into a small area and hence heat would be a concern. I can't see the P4 standing up to this sort of abuse based on their specifications and some of the posts I have seen here.

" Imagine creating a new blue LED die from scratch to compete with Cree's. "

- This is an interesting point. Lumileds did not really need to create a "new" die to compete with Cree, just like Cree did not need to so much create a new die to get the huge leap in performance. Cree was the first to market with a thin-film blue die used in the XR-E. I understand this is what Lumileds is doing on their new high bright products. From what I understand, neither of them pioneered this technology, OSRAM did with their AlInGaP (red, yellow) LEDS several years ago. Since OSRAM and Cree have a close relationship, I expect that Cree was given access to this patent a while ago? Someone posted recently about Lumileds and OSRAM cross licensing patents. I wonder if that influenced this recent product release?

I did just have a great thought! I don't know about anyone else, but the prices for the XR-E's have not been exactly cheap. Perhaps with the new Lumileds parts coming out, we will finally be able to get Cree XR-Es at a better price. Now that would be a change from the past when we were hoping for competition to Lumileds to get their prices down! I can still remember paying $10+ for a 20 lumen part.

Someone also asked what the goniometer/spectrophotometer would do. While this would give spectrum plots like Newbie provides, more importantly, it would give x-y (angular) beam plots. As opposed to beamshots which are impossible to compare and given the way cameras are made, almost impossible to get a linear response, these would show the exact candela measurements at every angle that the beam emits from the flashlight. This is how lighting fixtures are measured.

Semiman


Semiman

 

[CM]

I did not intend to be argumentative and I'm sorry if you read it that way. I am an engineer and I deal primarily with facts, NOT speculation. I tried to explain the facts to you, and EngineeringGuy did a nice job in post #86 also. Briefly restated, in order to achieve anything above ~240 lm/W of white light you're going to need the equivalent of a perpetual motion machine. Using your tongue-in-cheek number of 300 lm/W, it means an efficiency of 125% - getting OUT fully 25% more power than you put IN. If you understand the laws of physics and thermodynamics, you know that you can never, and I stress never, get more out than you put in.

Personally, I think the confusion arises because people don't grasp that a lumen is simply a measurement of power. It's called "photopic" power because it is weighted for the response of the human eye. If, on the other hand, (as EngineeringGuy also said) we threw away the word lumen and used "radiometric" power instead, the radiometric units of light are watts. Now you can look at electrical power input in watts, and light output in watts, and immediately see that the ratio cannot exceed 100%.

Sorry, I need to correct my statement above. I was thinking of the marketing number Seoul uses which is 240lm but achieved at 1A. You and engineeringguy are correct. My typing was much faster in this case than my thinking was :green:

However, I still say the Rebel is an attempt to make news out of nothing.

 

[Kinnza]

I believe the Rebel is intended to be the "cheap" LED in the future LML portfolio. Hopefully, LML will release along this year a "high perfomance" LED to compete with CREEs. In this scenario, the Rebel would compete with the XR-C.

This mean the Rebel isnt the part flasholics was waiting from LML. Not only for perfomance, but for mounting specifications. The Rebel's optical axis isnt refered to the ceramic base, but to two references points in it (cathode contact and a hole in one corner). So centering this LED in a reflector may be a PITA.

I has another concern about if its possible to wire the Rebel from top, using the electrical paths over the ceramic base. Has anybody with samples tried it?

About the theoretical photometric efficacy of white light, 240lm/w (emitted watt) is a bit restrictive. Current photometric efficacy of white LEDs are between 260 to 330lm/w. Reducing color rendering for application were it isnt important, allows to design white LEDs with 400lm/w, by doing it with RGB instead of using phospors (and their implied conversion losses). Although RGB solutions has currently several tech problems, potentially it can reach 400lm/w at 100% eff (wich is really far). Obviously, it will take a long time, but i believe white LEDs with 350-380lm/w (consumed) are possible, and probably ill see them in action.

 

[PhotonFanatic]

The Rebel, especially in the higher flux bins, which aren't even listed on Future's website, will be great for what it is--a reflow solderable, high junction temperature, small form factor LED, but it isn't what will make life sweet and easy for CPFers.

What we need is the Gen-X and 200 lumens at 700 mA.

 

[soffiler]

...About the theoretical photometric efficacy of white light, 240lm/w (emitted watt) is a bit restrictive. Current photometric efficacy of white LEDs are between 260 to 330lm/w.Reducing color rendering for application were it isnt important, allows to design white LEDs with 400lm/w, by doing it with RGB instead of using phospors...

Hi Kinnza:

This is the first I've heard of a conversion factor for WHITE LED's (blue with yellow phosphor) that exceeds the ballpark of 240 lm/W(emitted). Likewise, the numbers I've heard for RGB tend to be around 270 lm/W(emitted). I can see where the RGB number could be tweaked if you stretch your definition of white and let CRI drop, but 400 lm/W seems like a mighty stretch. I'd like to learn more... got any references?

 

[Mike Painter]

I did not intend to be argumentative and I'm sorry if you read it that way. I am an engineer and I deal primarily with facts, NOT speculation. I tried to explain the facts to you, and EngineeringGuy did a nice job in post #86 also. Briefly restated, in order to achieve anything above ~240 lm/W of white light you're going to need the equivalent of a perpetual motion machine. Using your tongue-in-cheek number of 300 lm/W, it means an efficiency of 125% - getting OUT fully 25% more power than you put IN. If you understand the laws of physics and thermodynamics, you know that you can never, and I stress never, get more out than you put in.

Personally, I think the confusion arises because people don't grasp that a lumen is simply a measurement of power. It's called "photopic" power because it is weighted for the response of the human eye. If, on the other hand, (as EngineeringGuy also said) we threw away the word lumen and used "radiometric" power instead, the radiometric units of light are watts. Now you can look at electrical power input in watts, and light output in watts, and immediately see that the ratio cannot exceed 100%.

You seem to be saying that 240 lumen is about equal to 1 watt.
It has been demonstrated that modifying the *shape* of the surface of a LED with concentric rings can imorove light output 7X This is because less than 2% of the available light in the LED gets out. "Light" is emitted through out the LED and little of it reaches teh surface.
If your definition is correct this means that a LED can be about 50x smaller and emit the same light if properly designed.

I've posted references a couple times and should have kept them. I don't have time today to find them again but, and sadly, because of the volume "nano-holes" will lead you to it.

 

[2xTrinity]

Hi Kinnza:

This is the first I've heard of a conversion factor for WHITE LED's (blue with yellow phosphor) that exceeds the ballpark of 240 lm/W(emitted). Likewise, the numbers I've heard for RGB tend to be around 270 lm/W(emitted). I can see where the RGB number could be tweaked if you stretch your definition of white and let CRI drop, but 400 lm/W seems like a mighty stretch. I'd like to learn more... got any references?

Well, I think the 330lm refers to the current Phosphor LEDs that that don't have a lot of red, and tend to have slightly "yellow-green" tint. Not using any scientific reference, but just looking at the luminosity function graph on wikipedia and some LEDs I have around, it seems like 400 lm/W(radiated) would be possible using a slightly greenish blue LED (about 500nm), a yellow-green LED (555nm -peak on the graph), and a red-orange LED (620nm) would make something that looks reasonably white, with a bit of a tint. Holding up LEDs at approximately those wavelengths looks pretty reasonble to me to use for something like a flashlight, or street lighting where CRI isn't crucial.

 

[soffiler]

You seem to be saying that 240 lumen is about equal to 1 watt...

Let's be clear that is emitted watts, as Kinnza pointed out. This side of the equation has nothing to do with how much electrical power you supply on the input side. This is a conversion on the output side that compares photopic power (tied to the sensitivity of the human eye, which varies substantially with wavelength) versus radiometric power (which is independent of wavelength).

The conversion factor reaches a peak at 555 nm wavelength, and at this point it is an impressive-sounding 683 lm/W(emitted). However, that's a yellowish-green color, nothing like white, not very useful for illumination purposes, with a color rendering index of zero.

Kinnza also said you could go a lot higher than my stated 240 lm/W(emitted) depending on how you mix your "white". The typical wavelengths of Red, Blue, and Green (RGB) which are mixed to get a reasonable approximation of white will get you more like 270 lm/W(emitted). Kinnza has suggested that if you monkey around with your choices of wavelength for R, and G (most likely using something close to 555 nm for Green), and B, you can still get something more or less recognizable as white (at reduced CRI) and greatly exceed my stated 270 lm/W.