Preliminary Luxeon K2 - Concerns that bother you...

I couldn't come up with a good title for this. I'll change it if I can think of something better worded.

From what we've seen so far on the Preliminary Sample K2, which may or may not make it into the final production peices, what bothers you about them, and what do you think we can do to deal with them if it stays that way when released?

This is a discussion of the negative aspects and how we might have to deal with them. I'm not discussing stuff like improved thermal management/etc.

To start off, here is a list of my concerns. :nana:

1), High Voltage Forward. - Hopes for a more efficienct luxeon go out the window. Less efficient than luxIII. So high that for use in small flashlights isn't pratical.

2), Gummy Bear Lens. - Thought the original luxeon lens was fragile and could crush? How about a gummy soft lens that could tear/deform and have dirt stick to it like mad...

3), Photon Management. - Will the die height be different like the recent LuxIII's coming out that change focus in exsisting reflectors/etc. Will the "reflector cup" behind the chip cause artifacts since it's no longer a smooth dish, but a multi faceted hexagon? Will the gold tint/residue it has on it have an effect on beam color?

4), Change in package. - Lights like the Alephs, and exsisting self-centering HS's like the O-sink will not be off the shelf compatible. Revised versions or modification will be nessasary. The 4 leads get confusing, as some are not-connected (true?).

Some of these can be worked around, and some will have a much larger impact and effect on how we power/focus/handle these things.

What else do you see as being a potiential problem if these things make it into the final product? What do you suggest to deal with this? Any additional info/corrections are welcome in this discussion. :wave:

Hi Issac

On any LED, as you increase the current, the Vf increases. On the two I have, the Vf at 700ma was lower than a TWOK, so I am not sure that this is an issue. It certainly was a whole lot brighter than the TWOK at the same driver current, so I interpret that as more efficient, or at least equal. The data I have seen posted, it is still roughly equivalent to the Lux III, and still lower than the Cree part for the same current.

I believe the die height is lower than on the Lux III, but it is a new P/N, not a Lux III, so that is less of an issue for me. I was able to increase the Lux readings of my light using the K2 samples and the exising optics by 25% using the same electronics.

The lens on the ones I have was not very soft at all. I did not push very hard, but at least at room temp, it was a non issue for me.

It is a surface mount device, and that is a challenge in some cases. A particularly useful feature in some ways, is that the alignment is much more specific, and the package is notched to allow it to be held in place, sort of like a star.

If there is a mistake, IMHO, they should not have told customers to back off on the heat sinking. I suspect that the pressure they have from their larger customers to "simplify the heat sink" is the primary issue with the release of the part.

I personally wish the slug length was longer, (like 1-2mm), but that is just me.

Thanks for chiming in Harry. Good news on the VF. From what I've seen people talk about most of the samples where high vF compared to luxIII at the same current level. Hopefully Vf will be comparable to LuxIII typical levels. 1500ma will obviously be higher of course though. Hopefully it won't be too hard to modify reflectors/etc to keep the same tight focus on the die height change.. Good news again on the lens, Don had some that were soft domed... odd...

Could you confirm all 4 leads make contact, or are 2 non-connected? Certianly the SMD tech will alow easier mounting for some, and the heat tolerances are better.. But I do 2nd the longer slug. We'll just need a HS with the holes for aligning it and have a nice flat surface for the slug to contact. Perhaps 2 rods can be inserted through the luxeon/into HS, and then after the epoxy cures, removed. I think that'd be a good solution.

I don't feel so worried now about items 1-2 as those were the most scarry to me.

The high Vf may be due to the apparent change in the die structure compared to the Lux III - instead of current spreading bars there's a grid of dots that appears to perform the same function? Perhaps some of the samples are earlier construction where the technique resulted in a bit more contact resistance resulting in a higher Vf.

andrewwynn in the K2 thread in the mods forum reported that the two leads marked as N/C are indeed N/C - unlike the clipped leads on a Lux I/III/V.

I'm concerned that they took a step backwards with the "warm" luxeons (red/orange/amber) - the K2 versions of these colors only have the smaller, 1mm x 1mm dies, and not the larger dies of the Lux III - I wonder if there's a reason for this? Perhaps the larger dies don't perform any better in the K2 packaging, and thus, should be left as Lux IIIs? It still would make more sense to unify your product line on the one package type - maybe we'll see some -14 parts for the Red/Orange/Amber K2s later that use the larger TIP die - as of now, we only have -12 parts for those colors.

evan, I just started reading the K2 thread again in C&M. That's what I thought too.
Im just now reading what andrew posted. Seems the lens is softer than before, but not too soft.. I wonder how scratch/dirt resistance it is with handling?
Those leads are useless then unless it's for just holding it down with solder?
Perhaps the reason for the warm luxeons using the tradional sized 1W die is because focus is tighter, and these can be driven harder than a 1W. Later they might add the larger 4x as big "lux3" sized dies with even more power handling...

I know a K2 warm will out throw both a Lux1 and a Lux3 in say a mag reflector, as the Lux1 and Lux3 both have the same power/surface area, and the K2 will have more current/surface area. Of course the lux3 will still have a larger hotspot.

Isaac,
Soft is a relative term. Compared to the acrylic dome of the LuxIII, these domes are soft. I would not push donw on them to press out the bonding epoxy, for instance. Although these are in fact SMD parts, this does not preclude mounting them with thermal epoxy instead of soldering them. The Vf issue is an issue with samples I have seen but then there are LuxIII LED's that are binned L as well so only in production parts will we really know. :shrug:

I think the K2 is a good step forward for manufacturers who can take advantage of what the K2 has to offer in terms of assembly alternatives and mounting references. For the little guy, I agree with Harry that if the slug were to extend down further, it could be used as a centering reference. As is, it is of little aid. I wish the package had three full corners and one chopped as opposed to the three chopped and one full. I understand that the package is not to be used for reference but just how far off dimension is it going to be?!?!?! Certainly no more than some of the dies have been on the pedestals of earlier Luxeons?!?! :shrug:

I think the biggest concern I have in regards to the K2 is simply the uncertainty of what it will be and when it will be. This is not a big concern as the LuxIII is viable now as are other contenders becoming. I see the K2 as being in a long distance race with some other contenders and the concern is which to put one's "money" on this early in the race. I don't want to place a bet until I have to. Lumileds unwillingness to provide concise information to the Future reps who are in turn in the trenches with the customers does not bode well or at least endear me to their yet to be released parts. I want to replace the above mentioned biggest concern with the statement that my biggest concern with the K2 is that it is a yet to be production item from Lumileds!

Don, well said. Knowing that you shouldn't press down on the dome to squish out epoxy is something good to know about the delicacy of the lens.

I guess I should just wait for them to come into exsistance, but what else can you do in the meantime. Don't answer that.

My only concern about the K2 is when are we going so see more than a few samples?
We can engineer lights and drivers around them but need more of them in hand so that these new designs can be made and refined. I am quite sure that there are several people right here on CPF with K2's in hand right now . Now for production of the K2 and new related products.
I really just want to know when so that I can get started on the CU version of the O-Sink designed for the K2.
Yaesumofo

If I were going to do a K2 version of the O sink, I would approach it with the idea of using the intended corners for alignment and holding the LEDs down (such as a small screw). There is no appreciable slug length below the package to allow useful centering, at least with my level of dexterity.

As soon as I receive samples from luxeon I will go into the redesign phase
of production. Beleive me we have been looking at the drawings. I Don't beleive these are final yet. I lan to make the best O-sink possible for the K2.
I am sure that People will be interested in making mag mods with the K2. I certainly hope so anyway.
Yaesumofo

HarryN said:

If I were going to do a K2 version of the O sink, I would approach it with the idea of using the intended corners for alignment and holding the LEDs down (such as a small screw). There is no appreciable slug length below the package to allow useful centering, at least with my level of dexterity.

Click to expand...

I will. Hopefully I can get a K2 with a Vf to pull 1.5amp at 3.6v for my NiMH modular mag. And hopefully it'll be brighter than a U luxIII. A copper O-sink would be icing on the cake.

I think you could have the HS have holes, and put removeable pins to hold the lux in place while the epoxy cure. Then remove them when done. The pins could be re-useable, or just the right height to be flush with the luxeon case. The HS below the lux slug could be perfectly flat (no centering raised areas) to insure even contact.

Dear Friend,
I have been in HP/Agilent/Lumileds for a while in Stuttgart and in Italy and let me tell you few things.
1. Normally a big improvement like the K2 from the previous and existing one, gives always big headhache. Lumileds had no other ways than to be at the same point of manufacturing methods of the Golden Dragon.
For this reason the package had to be completely re-designed.
2. Lumileds is now part of Philips Lighting.
Philips priorities may be different from the previous ones in Lumileds.

Saying that, I do not approve -as I did before was I was part of their company- the way of showing new items and let the people/markets waiting for so long time, before to see them released.

Hi Sololuce - Welcome to the CPF and thank you for your insights. There are a lot of different business models in the world, and I am not going to criticize one over the other. The LED business in many ways resembles the DRAM marketplace, and a few players, notably Lumileds, Cree, Osram, and Nichia are attempting to work in a space more like the Intel / AMD microprocessor market. - at least that is my humble perspective. I am sure there is more to it than that.

At least for me, I am glad Lumileds started early to release preliminary data sheets and small numbers of engineering samples. I would gladly pay for preliminary engineering samples if that helped me obtain more, but I am guessing that this opens up a liability door. It is not uncommon in the "bleeding edge" of technology for companies to ship and book revenue from engineering samles.

The reason I am happy to get this information, is that regardless of whether Lumileds makes some changes to the package, output, efficiency, colors, or release date, etc, - I can deal with that. The way I dealt with it, is that I designed my PCB to accept the Lux III, Lux V, Cree Xlamp, and K2. It was not entirely straightforward, but is possible. Now I am ready whenever the K2 is actually available, without impacting my product plans.

As a person who builds products that use LEDs, I need time to experiment, plan, think through potential integration challenges, optics testing, and circuit board design. It is a substanital investment for a small firm to have a quality PCB engineered and built in sufficient quantity to be cost effective, and this stock of boards may last quite some time. If it were not for this preliminary information, the conversion cost to incorporate the K2 would have been painful.

If I have a personal wish about the K2, it would be that they released some preliminary package side / solder pad info earlier, as I had to kind of "guess" at its design, and luckily, was close enough.

I am sure there is pressure internally on Lumileds / Phillips regarding the K2. There is nothing that could have stopped this, as the competitive pressures were going to be there either way.

Thank your for joining us, and we welcome your contributions.

.

I had a meeting with a Lumileds engineer yesterday. These will be 350ma. Lumens output will be very close to the Luxeon III. pricing and availability should be out early January, and they claim there will be plenty of production.