1W VS Luxeon III - data points

Hi guys,

Well we have been speculating on where the luxeon III fits in relative to the 1W and 5W. We have seen that driven at 700 mA and 1 amp, the III is typically brighter than a 1W driven at similar levels and further, we are led to believe that the III can manage these levels of current better than the 1W.

I have a McL4 1W R2H on a BB400 and I wanted to see how a III on a BB400 compares. I made one (S rank III) and just checked lux of the two; essentially equal. They are so close (same) that variation in LED's will make one the winner over the other. I propose at this point that a III S rank can be considered as good as a R rank when driven at the 1W power levels.

After ten minutes on, both hosted on E1e's, the temperatures of the heads as measured by an IR thermometer show these two lights as having the same thermal output. I had hoped that perhaps the III would run cooler but given the same luminous output on the same level of power, it is no surprise that the wasted thermal energy would be the same; it kind of has to be!!

So, based on two meager data points, I suggest that if one is in hopes of obtaining a R rank 1W and can't, a S rank III will do just fine, thank you very much! If one can get a T rank III (will we see these?) Whoopie!!!!!!! It will be very interesting to see if any R rank 1W's ever show up any more...... If only Lumileds would give us the "real" dirt.

- Don

Hi Don,

Have you done any comparisons at lower current levels, like say, 120ma to 165ma range?

I'd sure be interested how they compare in efficiency at lower current levels for obvious reasons...

wow nice comparison work

Don,
Your observations pretty much support the popular speculation around here right now that these Luxeon III's are just high-binned 1 watters with minor internal tweaks.
Big surprise... (sarcastically)

I bet that we stop seeing R rank 1w LS's, and that we'll have to pay more for the very same ones as 3 watters. /ubbthreads/images/graemlins/icon6.gif

BOHICA! /ubbthreads/images/graemlins/icon15.gif

On the up side, at least they're getting better! /ubbthreads/images/graemlins/icon3.gif

Don, would this basically mean that if I own a certain number of R2H lights at power levels between 400 and 750 mA, those will be just about the same as lights at similar power using the S-bin`ned III? ... and that the III is interesting at 1000mA or at better bins, say "T"?
bernhard

Hi Kevin,

No I haven't but I am suspicious that the III's will behave in the same manner as the 1W's. I think with more data points and testing, we can determine a mapping of Vf from the III's downwards and get a feel for what we need for power starved applications.

OK, quick bench test on a III of rank SV0J:

with current limited at 350 mA, the Vf was 3.2
with current limited at 160 mA, the Vf was 3.0

This III sure behaves like a RxH 1W !

I am willing to go out on a substtantial looking limb and speculate that the premium 1W's we have come to know are now being packaged as Luxeon III's. This not a new idea I know but I am seeing more and more reason to suspect this.

If a new "S" rank is the lower limit on the III's and Vf's like J can be had, I think these will be viable alternatives for the realm of the micro converters and such. Heck I think the III's are just plain premium 1W's, plain and simple! I said it again, didn't I.

- Don

Bernhard,

I am willing to give the benefit of the doubt to the III's at 1 amp as being brighter and perhaps more long lived than a similarly driven R 1W. My preliminary testing seems to hold this as valid. At 700 and 750 mA, I don't know but doubt that going to a III will get you much over a 1W R. IF we se some T rank then all bets are off! So I guess basically I think the answer to your question is yes.

What I take as really good news here is that the III's seem to be being produced in mass quantities so if a Q level 1W is not enough luminous output, step up to a III and still go with the current level you want. You will pay a premium but at least the LED is not unobtanium!

- Don

Don,

Do you have the capability of taking an extreme macro close-up of one of your L-IIIs?

If not, anyone willing to ship one to me so I can? (will ship back right away /ubbthreads/images/graemlins/wink.gif )

evan9162,

first image is a III followed by a 1W. Hope this is close enough.







and then a closer look at the die of the III:




I admit that I have never bothered to look that closely! Good thing it wasn't lit up!!! /ubbthreads/images/graemlins/grin.gif

- Don

Hmm. The III should measure hotter at the heatsink if it is improved in heat management, meaning it is better at pulling the heat away from the die and into the slug/sink.

Have you tried running the 1W R at 1 amp and the III S at 1amp and comparing the output? As I've heard people say much over 700ma is moot point with 1W'ers. I wonder if the luxIII does better over 700 than a 1W does? being the thermal upgrades might help it out....

[ QUOTE ]
IsaacHayes said:
Hmm. The III should measure hotter at the heatsink if it is improved in heat management, meaning it is better at pulling the heat away from the die and into the slug/sink.



[/ QUOTE ]

NO, NO, NO !!! Same amount of heat input to the heatsink = same temperature of heatsink.

hmm. I was thinking along the lines of better thermal "connectivity" gets the heat away quicker/better. I was thinking along the lines if say your star doesn't have the emiter glued down well the heatsink doesn't get that hot and the emiter might burn up since it's not being drawn away...

Interesting,
So, my four R2H Luxeons are actually 3 watters in disguise? /ubbthreads/images/graemlins/faint.gif /ubbthreads/images/graemlins/thumbsup.gif So I should not feel bad hammering them with 800 to 900mA with my Madmax feed off of 2 D cells. Sounds good to me so I will just keep beating the little guy up and not worry about it.
Thanks for the info, now I can buy S rank LS III's really knowing what they are.

[ QUOTE ]
IsaacHayes said:
hmm. I was thinking along the lines of better thermal "connectivity" gets the heat away quicker/better. I was thinking along the lines if say your star doesn't have the emiter glued down well the heatsink doesn't get that hot and the emiter might burn up since it's not being drawn away...

[/ QUOTE ]

Isaac, apologies if I sound rabid on this point but I am rabid. This misconception that having a lower thermal resistance from a heat source to a heat sink somehow results in a higher temperature of the heatsink sink is *very widely* held and this for some reason drives me nuts. All that the lower thermal resistance accomplishes is to cause the temperatures on the *upstream* side of the thermal junction to be lower.

BentheadTX,

I would not assume that the R bin 1W will perform as well as the III at elevated current levels. I actually suspect this to be false. I ramped a 1W up some time ago on a bench supply and somewhere just shy of 700 mA, the yield in lux as current increased started diminishing. OTOH, the specs on the III show a "reasonhable" increase in flux from 700 to 1000 mA. I consider the III's as 1W's that can take the high current but don't consider the 1W's as the same animal. In other words, the III can do what the 1W can and more.

- Don

This would come to follow since Lumileds claims one of the improvements was in current handling, but I'm curious as to how they did this. Maybe improvements in the dice themselves?

Don,

Yes, that's perfect. I see no difference. I was expecting, at least, a larger bond wire and/or bond wire connection point. Seeing that they are almost completely identical, the only difference may be a slightly improved thermal bond between the die and the heat slug.


BTW - Here is my contribution - 1W Q2J


In your and my images, you can see "tooling" marks on the die where the bond wire is attached from whatever plugs the bond wire onto the die.

Yep, I posted about my opinions on the III's about a week ago and I think we have a consensus forming. On paper and with Don's quantitive data points, the III's appear to be upper end 1W's. Interesting point about R's disappearing to go into the III's and sold at a higher premium. We'll have to see if that does transpire. The 3W's may be Lumiled's way around the price erosion commonly seen in semiconductor parts.

[ QUOTE ]
McGizmo said:
BentheadTX,

I would not assume that the R bin 1W will perform as well as the III at elevated current levels. I actually suspect this to be false. I ramped a 1W up some time ago on a bench supply and somewhere just shy of 700 mA, the yield in lux as current increased started diminishing. OTOH, the specs on the III show a "reasonhable" increase in flux from 700 to 1000 mA. I consider the III's as 1W's that can take the high current but don't consider the 1W's as the same animal. In other words, the III can do what the 1W can and more.

- Don

[/ QUOTE ]

Ah but Don, remember that the increased flux is based on the junction temperature remaining at 25C. In some real world uses, the effect of the increased junction temperature due to going from 700 to 1000mA may mostly offset the gains that might have been realized. I *do* conceed that some of your creations have the thermal balls to benefit from the increase. It is also true that if the emitter has a good thermal connection to at least some mass, for short bursts benefit should be seen at the higher current even if it is not true at the ultimate thermal steady state condition.

Powered up Don?




Only a fraction of a mA.