Luxeon III opinion

I have long suspected that the new Luxeon III LED's are really 1W devices that have been repackaged with some slight process tweaks to handle larger drive currents. In order to support my stance, let's take a look at some numbers, shall we?

Current high flux 1W such as the R bins are rated at 39.8 Lumens minimum and 51.7 Lumens maximum. From the curves that lumileds provides on their 3W parts, you should realize a 60% increase in luminous flux output when going from 350mA to 700mA. Now let us take this 60% number and apply it to the R flux outputs.

39.8*1.6 = 64 lumens at 700mA minimum
51.7*1.6 = 86 lumens at 700mA maximum

Compare these to the 3W specs:

60 lumens at 700mA minimum
65 lumens at 700mA typical

Judging from these numbers, it's hard to discern a 3W part driven at 700mA from an R flux 1W driven to the same level. Remeber that a few lumens difference cannot be detected by the eye, unless of course you're the Bionic Man. Only a calibrated integrating sphere will be able to discern the difference. We're talking a piece of equipment that's out of reach of the typical flashaholic.

My conclusion is that the 3W device is an extremely close twin of the high flux 1W device. Lumiled's has improved their processes sufficiently to get a viable yield on the high flux parts to offer them as "New" parts. I'm not saying it's a marketing gimmick. I am saying that through the evolution in process technology, higher flux (and better thermal management) single die LED's were inevitable. You cannot sell these "new" parts at a premium without differentiating them from the vanilla 1W devices. Better thermal management, I believe, is the primary differentiator and that high flux dies were bound to become more common as the process to make them improved. I also suspect that dies that fail flux output tests will probably find their way into the more mainstream 1W parts, sold as lower flux parts. So those newer P's and Q's coming out of their fabs are most likely a a fallout from the screening process to make the 3W parts. I offer this conclusion as merely my own opinion and not necessarily gospel or truth, as I am not an authority on this subject by any means. Any differing view point is solicited here. No flames, please /ubbthreads/images/graemlins/grin.gif

CM

CM, I am afraid that I don't have a differing viewpoint to offer. I share your suspicions. On a related note, in the Star configuration, the 1W and the Luxeon III have the same specified thermal resistance. The thermal enhancements, whatever they are, would appear to be more directed to heat tolerance than heat removal.

"...would appear to be more directed to heat tolerance than heat removal..."

Absolutely agree, and that's what I meant by process tweaks. I think the Luxeon III's are a positive step towards perfecting the Luxeon and Lumileds is focusing on the right thing. I do wonder about the Luxeon V portable and how it fits into the big picture. I also believe the Luxeon V is a stop gap until they can get the 1-die LED up to speed and the III's are a big step closer.

CM

CM,
I'm not sure where you got the %60 increase. From the graph of Normalized Luminous Flux vs. Forward Current on the 3w star spec sheet I'm looking at, I only see a %40 increase in output from 350mA to 700mA, not %60 (it goes from 0.6 to 1.0).

Also, I think it would be better to compare this to the 1w LS's by derating the 3w part since this spec sheet is for the 3w parts, not the 1w, and as such, the charts should not be applied to the 1w LS.

So, a derated 3w:

60 Lumens (minimun)
65 Lumens (typical)

* 0.6 =

36 Lumens (minimun)
39 Lumens (typical)

In my opinion, this does support your theory that the new 3w parts are high-flux binned 1w parts with some minor tweaks. I have also been thinking this was the case for some time now. The derated 3w part is 39 Lu, which is on the lower end of the R rank flux bin for the 1w parts.

Hmm, money-making scheme for Lumileds??? /ubbthreads/images/graemlins/thinking.gif /ubbthreads/images/graemlins/icon3.gif

***However, the 3w parts will probably take to overdriving past 1A well whereas the 1w part would not.

Hi,
i also think you're right, but i wonder why there are no red etc. colours available as L IIIs.
Any idea?
Joachim

Quoting from the Luxeon III FAQ sheet,
"The technology adopted in Luxeon III does not allow for the creation of LEDs producing these colors" (red, red-orange, and amber)

It then goes on to talk about the tech. difference is in the chip itself allowing it to carry higher current.

LEDmodMan,

It's actually 66.7% (I was trying to do this in my head without a calculator). The formula is (To_Value-From_Value)/From_Value. Comparing the two by overdriving the 1W or underdriving the 3W yield the same results. It's works bi-directionally.

Yes, money making scheme for Lumiled. However, they need to be paid for their development efforts and I see nothing wrong with that. Their progress in this area means we as consumers win. Elektrolumens is selling these parts for about $12 which is very reasonable and I certainly hope he makes a decent margin by catering to our addiction. I don't think anyone can accuse them (Lumiled) of price gouging in light of the fact Wayne is a lower tier level purchaser of these parts and he also needs to make some level of profit for selling them. If these turn out as advertised flux-wise and we get decent color temperatures, $12 is a bargain.

CM

[ QUOTE ]
On a related note, in the Star configuration, the 1W and the Luxeon III have the same specified thermal resistance. The thermal enhancements, whatever they are, would appear to be more directed to heat tolerance than heat removal.

[/ QUOTE ]

Could you elaborate on that? There is at least a two celsius/watt improvement in the LIII emitter if I remember correctly.

Have you reasons to believe that the 3w is more tolerant of heat?

Have you reasons to believe that the 3w is more tolerant/efficient at higher currents?

I agree with your guess that the lousiest 3w's will be tossed (to the 1wpile?), based on the small difference between the "typical" and "minimum" flux. That also leads me to the conclusion that the max flux bins might not be a whole lot better than the "typicals" for a while..

[ QUOTE ]
AilSnail said:
[ QUOTE ]
On a related note, in the Star configuration, the 1W and the Luxeon III have the same specified thermal resistance. The thermal enhancements, whatever they are, would appear to be more directed to heat tolerance than heat removal.

[/ QUOTE ]

Could you elaborate on that? There is at least a two celsius/watt improvement in the LIII emitter if I remember correctly.



[/ QUOTE ]

You are correct that the datasheets show a 2 C/W difference at the emitter level but this difference does not carry over to the Star level which implies that the Luxeon III slug to board resistance is 2 C/W *higher* than that of the 1W. It is a mystery why this would be the case.

[ QUOTE ]
AilSnail said:
Have you reasons to believe that the 3w is more tolerant of heat?




[/ QUOTE ]
Conjecture based only on Lumileds statements about "thermal enhancement" of the Luxeon III product.

[ QUOTE ]
AilSnail said:
Have you reasons to believe that the 3w is more tolerant/efficient at higher currents?



[/ QUOTE ]
I have seen nothing to suggest that the efficiency vs current characteristics of the Luxeon III are better than that of the 1W.

[ QUOTE ]
AilSnail said:
That also leads me to the conclusion that the max flux bins might not be a whole lot better than the "typicals" for a while..

[/ QUOTE ]

I suspect this is true.

I will roll my R2H for now,
The thing I want to see in the LS III is a high ranking, say a T3H would sway me. Take a high R rank and blast it at 700 to 1000mA and I bet it would hit the T ranking. Just pointless ponderings from my weird little mind at this point.
Give me those bin codes and low voltage drives and life will improve. I have some BB400 Q3L low domes that could use a bit of a boost in lumens (spoiled by my BB500 R2H's) My hope is that a 3W driven at 400mA would be brighter and run cooler than a Q3L.
Time, all we need is more time to see how this game plays out. /ubbthreads/images/graemlins/popcorn.gif /ubbthreads/images/graemlins/popcorn.gif

I believe what you guys believe is true. However the release of Luxeon III is the time when Lumileds could promise to manufacture consistant quality luxeons. Maybe 3W is what 1W meant to be. Dispite the shorter lifespan.

Vince.

Well I hear the 3W parts are going to be in Elektrolumens hot little hands next week. Waiting shouldn't be that long now.

BentHeadTX,

I've been overdriving R2's in KL1's at 700mA for a while now, so those who have my overdriven R2's actually have "T" fluxes /ubbthreads/images/graemlins/grin.gif . From what I've seen, it is noticeably brighter but my guess is that most people would be disappointed. I have not gotten any bad feedback from anyone about the lights not being bright. I do suspect a good number of owners of those overdriven LED's were not "blown away". I could be wrong though. I really like the one I've got and it's a mediocre Q3 since all the other lights I've put together were at least as bright as mine and most have been brighter in my comparison beamshots. I'm awaiting the III's with cautios optimism.

CM

CM,

I must wholeheartedly agree with you.

Here's the thing -- someone very deeply into the major flashlight manufacturing business had already told me a while back that at least one of the precursors to the uprating (and that's what it is, pure and simple) of the 1W to 3W was due to the maniacal bleeding edge the LGI was run at.

"Maniacal" is a relative term, as today it is considered the norm for the 3W.

I guess Lumileds saw that 1Ws indeed weren't blowing up when run at 4½V. Heavy overdriving, as long as it is accompanied by heavy heatsinking, does in fact allow the 1W to be run at a full amp. Over four hundred LGIs still running fine and dandy in daily use and even as EDCs (for those of us with, er, longer pockets) out there attest to this. Maybe not for 100,000 hours, but any guy (or gal) with a 1W Luxeon light that is perceptibly too dimmed-out to the human eye will probably only be noticing this while on their deathbed. Because in real-world flashlight usage, that's likely how long several thousand hours would translate out to be.

Of course other tweaks include augmented thermal pathways (Don told me this; I have yet to see it with my own eyes till the "few" 3Ws I ordered get here hopefully this week) such as the connection leads themselves.

But at the rate LED technology is growing, even the 3Ws may be in museums sooner than we expect...

So why is this a bad thing? It's marketing and incremental improvement based on real world experiences.

Whatever allows us more light output sounds good to me. All semi manufacturers tweak recipes, use different packaging for different applications and the rate or derate accordingly.

I guess I'm missing something here, can I buy a clue?

GregR

I would say that some of us are looking for more revolution from Luxeon III rather than small evolution.

The 5W was definitely revolution, with many teething issues. It may turn out that the current 5W design is destined to be an intermediary step.

One may think that Lumileds was "holding out" on us, if there aren't any big changes to the 1W design that makes it a 3W. Many of us achieve luxeon III power levels in our own lights, without having the luxury of modifying any parts of the manufacturing, and with good levels of reliability/dependability.

Overclockers in the computer enthusiast world aren't overly impressed with "offical" CPU speed grades from CPU makers that match what they've achieved with their extreme overclocking measures. I would say that some of that may be present with the luxeon III release.

Just my rambling

I am excited about the III's because of the promise of luminous output from a single, photon manageable, die. What I am really excited about though is the continuing increase in flux of all of these, including the 1W and even the 5W! At some point, we will hopefully have enough "horsepower" so that we won't be having to necessarily use it all the time. The varible output is just around the corner. I have a small one cell R2H light that is pocketable and only running at 400 mA. This thing is amazing compared to the NX05 lights we were excited about last year! When this point in time has become history, we or others, will marvel at the fact that we had to use Al and other thermally conductive materials in sizeable mass and surface area to bleed off wasted energy in the form of heat. In the mean time, I am looking forward to using these newer LED's driven at or even below spec for great light as well as pursuit of max lux. /ubbthreads/images/graemlins/grin.gif

Point of comment to Charlie: The III star uses the wider portion of the lead frames but the sample III emitter I saw uses the skinny lead frame??!?!? I wish Lumileds would supply the emitters the way they come of the line with both leads still attached!

- Don

mine has the thinner connectors cut off.

It looks as the gold wire is thicker, but I can't say for sure.

evan,

The Luxeon III *IS* a revolution in many ways. Look at the original spec for the 1W. Remember the minimum flux rating of 13.9 lumens, and typical of 25. Now with a similar die, the numbers are 60 and 65 respectively. That to me is pretty remarkable. How long did that take? What's in store for us in 18 months? I can't wait to see.

CM

I can't help but feel guilty about overdriving an "old tech" /ubbthreads/images/graemlins/smile.gif 1w R2H at 900ma+ (MM+ sandwich). I won't feel as guilty driving a III at that level.

The III's Lumen Maintenance at 1a isn't stellar but I'd wager it's much better than the 1w's at 1a...

Jon

all lumileds did was see our success in overdriving the 1w parts and seeing them take it they uprated them.we did the r&d work for them.
btw i am still running a celeron 300a@504
thats a 66 mhz fsb running 112!