R.I.P. R2H; Long live S4K!!

I just busted my nicest r2h... /ubbthreads/images/graemlins/frown.gif Now it is just barely glowing. BOOOHOOO. Now for the good part; I won an auction at Elektrolumen's forum for a 3watt. It looks just like a 5watt star, except with one die, and the inscription "S4K" on the back, written by hand with a black alcohol marker. Before I broke the r2h (lament) I compared it to the 3w from a BB500. They had about the same output, both a nice purple-bluish white. The s4k had Vf 3.3, and the r2h 3.7V. I probably soldered it to death when I added a BB400 in parallell. I swapped it with another r2h with similar Vf's and brightness as the dead one (at assumed 400 and 500 mA), but which had a greener colour. At what I assume to be 900mA the 3w is brighter, and had Vf 3.59 against 3.85V for the r2h.

I am going to assume that since the new 3W LS's are still the same dice as the 1W, they're probably not really 3W, but the best of the best 1 watters. This assumption is based off of what I've read from Peter G and Elektrolumens evaluations of their samples. So they're not really 3W, just really good 1W being run at 3W like we flashaholics have been doing all along (nearly)...

If the 3 watt bins are on teh same scale as the 1 watters, the extra 10 lumens from the R to S bin doesnt sound like a good increase in brightness for triple the power supplied to the 1 watt. I agree with LedmodMan; it just sounds like lumileds are cherry-picking the best 1 watters off the line and charging more.
Having said that, I'll be first in line for a few when they are made available.

I almost wonder if we're (CPF) the largely responsible party for this? /ubbthreads/images/graemlins/icon3.gif

Well im not too bothered ; we wil in effect, be paying a premium for the best 1 watters...but then again, this will only be true if lumileds improve their QC. Lets hope the 3 watters will be less prone to the "Luxeon Lottery" syndrome.

[ QUOTE ]
If the 3 watt bins are on teh same scale as the 1 watters, the extra 10 lumens from the R to S bin doesnt sound like a good increase in brightness for triple the power supplied to the 1 watt.

[/ QUOTE ]

Though the Vf is correctly binned if it is rated at 900mA, it /is/ brighter than the r2h at that current, while consuming less energy. There is something else, though - I don't think that it's brighter at 500mA.

ok I'm confused. The 3w has the same Vf range as the 1 watts, ~3.x volts. So how can these be the same die if you are to run them at 3.x v @ 900ma? The obviously will have to pull more current right? So you don't need to up the voltage to get them to pull that current right?
Then the die must be configured differently.

Also if a 3w is rated at S flux, it really isn't a premium 3w since 3w=lots more power than 1w. And if there are R 1w's, then there should be a lot higher flux lettered 3W's out there.

Thats like saying a T 5w is premium, well not when your talkign about 5ws...

Am I making any sense?

[ QUOTE ]
ok I'm confused. The 3w has the same Vf range as the 1 watts, ~3.x volts. So how can these be the same die if you are to run them at 3.x v @ 900ma? The obviously will have to pull more current right? So you don't need to up the voltage to get them to pull that current right?
Then the die must be configured differently.

[/ QUOTE ]

I am not sure I understand the question. for the 3watt, there is a 0.3V difference between 500mA and 900mA. Pretty similar to some of the 1watters.

[ QUOTE ]
Also if a 3w is rated at S flux, it really isn't a premium 3w since 3w=lots more power than 1w. And if there are R 1w's, then there should be a lot higher flux lettered 3W's out there.

[/ QUOTE ]

Yes the flux binning is confusing, I do not understand it. Should do more comparisons.

it kind of sounds like Lumileds is overdriving 1W's (because they will take it) and re-binning them as "3W"? /ubbthreads/images/graemlins/confused.gif

that's what i've been thinking all along /ubbthreads/images/graemlins/frown.gif

[ QUOTE ]
Slick said:
it kind of sounds like Lumileds is overdriving 1W's (because they will take it) and re-binning them as "3W"? /ubbthreads/images/graemlins/confused.gif

[/ QUOTE ]
I think that when the first 3Ws showed up (in PG's hands), he looked closely at them and they are 3W dies but with improved packaging in the area of thermal management.

i.e. they improved the heatsinking of the die so that the same die can be rated at twice the current. We already knew it was possible to do so with massive heatsinking, Lumileds just made it somewhat easier by improving the "bottleneck" - The emitter package itself.

And as far as both having Vfs of 3.x volts - Remember, LEDs have VERY nonlinear I/V curves. A few tenths of a volt can mean double the current. This is why current regulation is needed.

Ok, I understand the voltage deal now. I was thinking more along the lines of like 5w's being run at 9v is around double the current.. So 1w's can pull a lot with very little voltage changes. got it.

Re: the packaging. It seems the 3w stars are now directly on the metal star with the pcb cut around it, rather sitting on top of the copper pcb like current 1ws. So they are like a 5w in that regards. Is that the only thermal improvement? Or did they do something to the actual DIE it's self? In other words :
1w emitter same as 3w emitter?
1w star diff than 3w star in re: to mounting to the aluminum star.

In regards to the flux bins, I guess we will just have to wait until documentation is released so we can see where the range of bins starts and stops with 3Ws. Who knows how far it will overlap into 1w/5w teritory... /ubbthreads/images/graemlins/icon3.gif
One thing though, since it's getting 3x more power, shouldn't it be 3x as bright as a 1w? But if it's just an overdriven 1w then wouldn't the effeciency be less than when using 3 seperate 1w stars?

[ QUOTE ]
IsaacHayes said:
Ok, I understand the voltage deal now. I was thinking more along the lines of like 5w's being run at 9v is around double the current.. So 1w's can pull a lot with very little voltage changes. got it.

Re: the packaging. It seems the 3w stars are now directly on the metal star with the pcb cut around it, rather sitting on top of the copper pcb like current 1ws. So they are like a 5w in that regards. Is that the only thermal improvement? Or did they do something to the actual DIE it's self? In other words :
1w emitter same as 3w emitter?
1w star diff than 3w star in re: to mounting to the aluminum star.

In regards to the flux bins, I guess we will just have to wait until documentation is released so we can see where the range of bins starts and stops with 3Ws. Who knows how far it will overlap into 1w/5w teritory... /ubbthreads/images/graemlins/icon3.gif
One thing though, since it's getting 3x more power, shouldn't it be 3x as bright as a 1w? But if it's just an overdriven 1w then wouldn't the effeciency be less than when using 3 seperate 1w stars?

[/ QUOTE ]
The 5Ws run at double the voltage because of the fact that they run a 2series/2parallel configuration.

A lot of people have speculated that the die in the 1W units and 3W units is the same, and the packaging is merely different. In the case of whites, they might tweak the phosphor ratio a little bit to prevent the unit from being overly bluish.

Chances are the 3Ws are indeed less efficient. (Not sure exactly about this.)

I believe the 1W blue/green/cyan Luxeons actually use 1.25W or so. The actual input power of the "1W" units varies depending on the Vf. They're binned by Vf, color temperature, and brightness. All are rated at 350 mA. The average power is probably 1W, but it varies with Vf. (Which is why R2Hs are so desirable - Because of the low Vf, they are consuming the least power, but they're very efficient so have a high luminosity bin.)

Hmm. Another thing that has come to mind. Wayne @ EL's put 3 1w's in a tristar flashlight, and also 3 of his sample 3w's in another. The 3w was definatly brighter. He was not using any resistance in either one. So that goes back to the flux bining of 3w's. If they are the same thing as a 1w, then will all 3w's out do 1w's when both are ran at the same power? Hmmm. So maybe 3w's are only high-binned 1w's.
In that case does the binning structure make sense? As you could have a 1w R2H that is rated R in flux @ 350ma, and a S4K 3w that is rated at S @ 900ma.. So this gets confusing.

So if 1w=3w, physically, then a R2H 1w would be brighter than a 3W S4K, since if you ran the S4K 3w at 350ma it wouldn't be an S rank...

I'm probably confusing myself more than anyone else.... /ubbthreads/images/graemlins/smile.gif

1w emitter same as 3w emitter?
1w star diff than 3w star in re: to mounting to the aluminum star.

Not same emitter, at least cosmetically. It has the same leads as a 5w.
Same mounting and pcb as a 5w, too.

I compared the remaining r2h to the 3w with BB400. The 3w is MUCH brighter, just like it is with BB900.

I also tested a w4w with Vf7.0@BB400 against the S4K with Vf3.5@BB750. The 5watter thus got 2.8watt, the 3watter 2.4watt. The 5watter was ofcourse brighter. Wish I could think of a way to equalize their power more.