Lumileds uping the ante on the Rebel

[saabluster]

Of course this is not a revolutionary new offering but more an improvement to the existing line. Still I am impressed to see how well Luxeon is keeping up(passing?) with Cree.

The Luxeon Rebel ES can deliver 100 lm/W efficacy at the maximum 1A drive current, or luminaire makers can lower the drive current and achieve even greater 125 lm/W efficacy.


Datasheet


Of note is the 1500mA max current capability and what seems to be awfully close to no sag with increasing current.

I just wish their binning was tighter so you knew better what you were getting.

A little more info.

[znomit]

Nice update

I was going to guess bigger die... but from the product brief:

Whats new

• - Very small optical source size

But then the pics...
ES:


Old:

[Curt R]

Thanks for the info, will look into them as a back-up to the Cree in our lights.

Still waiting for my XM samples, they may have a little problem getting good production quality samples as the
buildup manufacturing process is different than they have been using on the XP dies.

Curt

[saabluster]

Nice update

I was going to guess bigger die... but from the product brief:

Whats new

• - Very small optical source size

But then the pics...
ES:

I thought that was the case when I saw the pics but didn't have the time when I posted to do an in-depth examination. That has a pretty big ramification on the optics used and results seen from use in certain applications. I would much rather have seen them rename it even though it is the same package more like Cree does. It makes it really confusing to call two products by the same name.

With the larger die size in mind I rescind the remark about them maybe passing up Cree as that was based more on the assumption/hope that the die was still 1mmx1mm. So now Lumileds has an XP-G competitor. Sort of.

[LumenHound]

Lumileds who???

I know, I know....

It's amazing how things have changed in the last few years.
I remember when we were happy to pay $30.00 for a primo UWOK bin star.

[evilc66]

Hopefully it holds true in real life, but the datasheet claims very little difference in lumen output between cool and neutral white (around 5lm at 350mA and 700mA, 10lm at 1000mA). That's pretty interesting.

[blasterman]

From the datasheet:

"Typical CRI for Neutral-White is 65, Cool-White is 70."

Which didn't make sense until I went to Lumileds site and found what I believe is a more accurate chart:

http://www.philipslumileds.com/produ...on-rebel-white

So, the higher flux neutrals take a big hit in terms of CRI - no surprise there. If you want more efficiency at lower CCTs yellow/amber gets sacrificed.

If I'm reading this chart right the best ES takes a serious back seat to the XP-G in true neutral white flavors. This was surprising. However, the higher CCT values keep up pretty good.

[HarryN]

I thought that was the case when I saw the pics but didn't have the time when I posted to do an in-depth examination. That has a pretty big ramification on the optics used and results seen from use in certain applications. I would much rather have seen them rename it even though it is the same package more like Cree does. It makes it really confusing to call two products by the same name.

With the larger die size in mind I rescind the remark about them maybe passing up Cree as that was based more on the assumption/hope that the die was still 1mmx1mm. So now Lumileds has an XP-G competitor. Sort of.

I have been fooled before by assuming die size changes when the actual change can be simply optical properties of the package as a whole. I am not saying the new die is larger or not, but in the end, what matters are the results.

I got a little nervous when you reminded me that this new part could results a different radiation pattern, so I needed to double check it. (on paper for now, later of course with real parts)

I looked at the data sheet spatial radiation pattern (data sheets, typically page 18 or 19) from the older Rebel parts of various CRI types, colors, and the newer upgraded part. I compared 5 points from each of them and they are virtually a match, "indicating" that these will be really close in focus results.

Philips Lumileds has such a large base of customers and related firms banking on being able to use any Rebel in any optic interchangeably, that I can't imagine them not engineering this part to optically match the others. Keep in mind that these are the guys that took a beating in the past for making a very small change to a dimension in the Lux III that was not even an engineered spec dimension.

OK, yes, I am still hoping that it is the same, not proven, although their data sheets are usually pretty darn accurate.

BTW - if you want tighter binning of a CCT or CRI, just look a little closer at the white p/n selection. Some of those are simply tighter binned versions of the more generic white parts. There is a slight price premium for them, but it is relatively small, and much less painful to purchase an "off the shelf" p/n than the old method of being on the hunt for a specific bin.

[znomit]

I looked at the data sheet spatial radiation pattern (data sheets, typically page 18 or 19) from the older Rebel parts of various CRI types, colors, and the newer upgraded part. I compared 5 points from each of them and they are virtually a match, "indicating" that these will be really close in focus results.

Check out XPG/XPE datasheets, different dies, identical spatial radiation patterns but the smaller die XPE throws much better.

[IMSabbel]

Still, its nice to see that the competition is not dead.

While not world-shaking, it still is a huge step forwards in light output per device.

[Illum]

um...Luxeon is a product series, not the name of a company

Luxeon line died long ago. "Luxeon uping the ante on the Rebel" would be analogous to "XPE uping the ante on the XRE"

Unless you meant it as above I think it would be more correct to consider "Lumileds uping the ante on the Rebel"

But then the pics...
ES:


Old:

Considering the small footprint of the rebel, I wonder if the ESD diode would now be more susceptible to solder damage now that it is within almost direct contact to the soldering points

[fyrstormer]

I have been fooled before by assuming die size changes when the actual change can be simply optical properties of the package as a whole. I am not saying the new die is larger or not, but in the end, what matters are the results.

I got a little nervous when you reminded me that this new part could results a different radiation pattern, so I needed to double check it. (on paper for now, later of course with real parts)

Focusing optics only work perfectly on light emitted from the exact focal point. Larger dies are by definition less point-like, in that more of the light comes from off-center, and from further off-center. Nothing anyone does can change this. There's no law of physics that allows Philips to skirt around the problem that the minimum divergence of a focused beam from a larger die is going to be wider than from a smaller die.

That being said, I don't care about having a perfectly-collimated beam or really anything close to it, as I rarely run into situations where I really need to be able to pinpoint individual rocks on the Moon from my backyard, so it doesn't affect me any. Most likely what Philips is banking on is that most of their customers won't care either, at least not regarding whatever products earn them most of their money, and for the products that do actually need smaller dies, they can just buy the top-of-the-line emitters to get the flux they need. As long as the radiation pattern from the center of the die is the same from one model to the next, the die size shouldn't make any noticeable difference in most applications.

[bshanahan14rulz]

ESD is on the top side, it is that little block outside the dome.

Solder pads are on the bottom of the device.

From the 3D view:
Also, possible barcode on the bottom will make it easier to determine bins, with a little bit of work. Unlike the guessing game of Cree LEDs...
Die looks bigger, has more internal connect points, 36.

Ok, you got me interested.

Oh, and pushing that Vf lower? highest voltage bin is 3.25-3.5V, lowest is 2.5-2.75, at 700mA

[HarryN]

I wonder if what we are thinking is a "larger die" is actually a different phos approach. Lumileds has been working on what they called "Lumeric" or something like this for a while, which is actually molding the phos into a ceramic plate. The idea was to test each die for output, and match it to a "Lumeric" ceramic cover to obtain the same final color bin result.

This is just a guess, but it would make more sense than Lumileds increasing their die size.

In order for this to work, I would think you also need to cover the edges, which would make the die "appear" to be bigger in the package, but may or may not affect what the emission looks like.

In their early days of getting into the power led business, Cree didn't cover the sides of their die with phos. The very narrow blue ring of light emitted from the sides was nearly un noticed - until you put it in a reflector. The resulting beam color was an ugly cloudy mess. Obviously, they have come a long way.

[blasterman]

As long as the radiation pattern from the center of the die is the same from one model to the next, the die size shouldn't make any noticeable difference in most applications.

Most likely scenario. Changes in the LED package or dimensions are typically done to decrease fab costs and be transparent to customers.

The Rebel line seems more focused on OEMs that use large volumes of emitters and need something with a bit more performance than SMD but not state of the art 3watt performance.

Given the choice of unified lumens over a CCT range it's pretty obvious who the customer base is.

[HarryN]

Most likely scenario. Changes in the LED package or dimensions are typically done to decrease fab costs and be transparent to customers.

The Rebel line seems more focused on OEMs that use large volumes of emitters and need something with a bit more performance than SMD but not state of the art 3watt performance.

Given the choice of unified lumens over a CCT range it's pretty obvious who the customer base is.

Really? 19 p/ns for "white" (not including binning) plus 7 p/ns for "colors", not including binning versions. That covers a pretty large range of customers.

I agree that they most likely are focused on large OEM type customers, just like everyone else is. That is why LEDs are so cheap now.

[zzonbi]

different (larger) die

although it trails the cool white xpg, Vf and thermal drift seem lower, also the beam a tad tighter

300lm+-6.5% 1A 25C 3.1V 4100K CRI>60

I like the lack of wires, hope the tales about K2 ruggedness are true here (>70% lm <50kh, 1A, 135C)

[MichaelW]

How else exactly would they increase the maximum drive current to 1.5 A without an increase in die size.
The thermal pad looks unchanged. The thermal resistance is less, 6 vs. 10.
and the 'dot grid' is 5x5 instead of 4x4

I was looking at the Rebel & Rebel ES datasheets a month or so ago, and they said 700mA maximum drive current for regular Rebel, and 1,000mA for Rebel ES. I went and looked at older datasheets, and the Rebel WAS rated to 1,000mA. So, a retrograde derating of drive current?


Regarding the CRI of the new Rebel ES, is it possible that someone made an error?
Lumileds used to always order things in their PDFs cool-white, neutral-white, warm-white. In this PDF, it goes cool-neutral for minimum, then switches to neutral-cool for typical. That does not seem right.
Also the SPD graph for neutral-white doesn't look like it was compromised to get enhanced output.

[Black Rose]

I was looking at the Rebel & Rebel ES datasheets a month or so ago, and they said 700mA maximum drive current for regular Rebel, and 1,000mA for Rebel ES. I went and looked at older datasheets, and the Rebel WAS rated to 1,000mA. So, a retrograde derating of drive current?

The absolute maximum DC forward current is still listed as 1000 mA.

[MichaelW]

I have a datasheet from last year, and it says 1 amp.
I looked at the datasheets from 1-2 months ago, and it say 700mA for Rebel, 1,000mA for Rebel es

Now the Rebel ES [new] is 1,500mA, and Rebel is back to 1,000mA.


I just want to know if I can get this for my aging Fenix RB100 L2D & P3D.

[saabluster]

um...Luxeon is a product series, not the name of a company

Luxeon line died long ago. "Luxeon uping the ante on the Rebel" would be analogous to "XPE uping the ante on the XRE"

Unless you meant it as above I think it would be more correct to consider "Lumileds uping the ante on the Rebel"

Indeed I should have written Lumileds. Change made. Thanks.

I don't know why you say the Luxeon line is dead though as it is clearly plastered all over the new datasheet.

[saabluster]

How else exactly would they increase the maximum drive current to 1.5 A without an increase in die size.

By improving the performance of the die or improvements in the thermal characteristics of the package. Enlarging the die is not the only route to increasing current capability but it is a bit easier at this point since the other two avenues are at a relatively advanced stage.

[jeffosborne]

Each day since this thread appeared, I've been checking the Rebel ES listing at Future Electronics, eager to purchase a few. And just today the listing show both the cool and neutral devices in stock! 2,276 of the neutral whites in stock, it says. I bought 12 of them at $4.85 each.

I have several 80-lumen per watt Rebel emitters in lights around the house. I'll upgrade some of them to this 125-lumen per watt newbie.

I will be sure to report the BIN received, as it is a roll of the dice on small quantities from Future.

Hooray!
Jeff

[jeffosborne]

I received 12 Rebel ES emitters today from Future Electronics! I am eager to get 300 lumens from one emitter on a 1-amp budget, especially in neutral white. But the packing list shows a BIN of 'JUL', which does not tell me anything about the parts. The Phillips doc calls out a BIN scheme that does not follow this BIN at all. I wrote to Future and asked them to give me a correct BIN.
Here's a comparison photo:

The original neutral white Reble to the left, the more recent high-CRI part in the middle, and the new Rebel ES to the right. The ES die looks to be much larger than the other two. I've covered the bottom terminals on one of the ES LEDs with a thin coat of enamel, and will mount it into the Aurora light on the left side, below. I'll be booting an XPE to put it into place. What fun!

Jeff

[Illum]

I don't know why you say the Luxeon line is dead though as it is clearly plastered all over the new datasheet.

huh...I recieved word on here that they're discontinuing luxeon 1s...my guess is that information was faulty

[jeffosborne]

The BIN code for the 12 Rebel-ES emitters I received yesterday is:

UTOS

U = 220-240 lumens at 700ma

TO = 4100-4500K at the black body locus :-)

S = 3.0 - 3.25 Vf

I am happy with this, and hope to fire one up this evening.

Jeff

[Black Rose]

huh...I recieved word on here that they're discontinuing luxeon 1s...my guess is that information was faulty

Nope, not faulty info. The Luxeon I, III, V, K2, etc have been killed off.

PhilipsLumileds is still using the Luxeon brand name for their active LEDs (Luxeon Rebel, Luxeon C, Luxeon Flash, Luxeon Altilon)

[croled]

New 120lm "old" rebels?
http://www.futureelectronics.com/en/...UMILEDS1000732

[saabluster]

New 120lm "old" rebels?
http://www.futureelectronics.com/en/...UMILEDS1000732

That's much more interesting than the new ES. Thanks.

[jeffosborne]

But, the newly listed 'old' Rebel is a dollar more than the ES, and a few less lumens per watt! Is it your view, Dr. saabluster, that the smaller die of the 'old' Rebel is an advantage for throw, perhaps? I did notice that the ES I put into the Aurora flashlight has a brighter and slightly larger center spot than the XP-E I removed. Also I love the neutral-white tint of the ES part. Cheers, Jeff