K2 Emitter from Lumileds

I recently acquired some detailed information about Lumiled's upcoming K2 emitter.. so of course i had to make a computer model to see how it will fit in my projects.. it's kinda cute so i wanted to share.. and maybe start a thread to discuss the implications with the new emitter coming.

The specs show that there are two levels coming.. i call them 'high power' versions and 'low power' versions.. the 'high power' models are designed to run at up to 1500mA and the 'low power' models are designed to run at up to 700mA.. actually it's a little bit confusing because it looks from the specs you could run them all at the higher drive levels.. maybe somebody more in-the-know could help explain that..

In any event.. they are going to have 5 brightness bins.. R, S, T, U, and V.. they are similar to the lux 3 in the grouping regarding their output, but the bin rating i think is at the 'base' drive level (which is 350mA for the low-power model and 1000mA for the high-power model.. so this means that the 'S' bin which is rated at 350 actually outputs the light of a Ubin Lux3 when driven at 700mA.

the R and S bin models are in the 'low power' group and rated from 75 to 100 lumens max
the T, U, V models are in the 'high power' group and rated from 120 to 140 lumens max

(at 350mA.. the R and S are 45 and 60.. at 1000.. the TUV are 80, 100, 120).

the estimated power to drive the emitters:

The low-power models range from 1.2W on the low side and 2.5W on the high side..
The high-power models range from 3.7W on the low side and 5.8W on the high side.

Looks like they managed to get the average Vf down, and i'm hopeful that the typical K2 will have a Vf similar to the J-bin Lux3 now.. only time will tell where the bins end up.. and how the color binning situation will work out.

Some very exciting things i think are coming.. with the advent of making them SMD technology and being able to machine-place and solder... i think that the price will be held down from tremendous production quantities possible.

Notice that the emitter leads are much smaller.. and they leave all 4 on.. (the lux 3 is fabricated with 4 leads and two are cut off).. they are going to make K2 emitters with matching leads so they match the footprint of the lux3.. for replacement and possibly easier hand-soldering.

Notice the slug is hexagonal.. i have no idea why, but guessing it makes it easier to machine orient when building, or helps fit into the square body.

Notice the alignment holes kiddy-corner.... for uber-precise placement.. put a hole in your PCB and put a pin through the hole and get within .05mm centering of your lux!

Notice the one corner not chamfered.. that side is the cathode... technically they say that the chamfered corner is the anode.. but there are 3 chamfered ones so it's backwards from how i'd have done it.

Some size restrictions will come into play with some lights or reflectors that are 'used to' the lux 3 being 8mm diameter..

slug size: 5 1/4 mm on the flats.. approx 6 mm to the points.
the chamfered corners fit in a diameter of 8.76 mm
the long corner fits within a 10.35 mm diameter
the emitter leads fit within a 12.15 mm diameter.

The main issue that is affecting my designs is that corner point.. it may not fit in the IMS27 or similar reflector that has a 9mm opening.. it would have to be shaved down to fit, which surely would not be a difficult thing to do.. in a mass-production job like MM.. i would not want to be cutting down each emitter... so i'll be re-designing the reflector to fit outside that 10.35mm diameter... tough call on the IMS 27.. if it's easier to cut the reflector or the emitter, but the emitter would be my bet... fortunately the IMS 17 and 20 only come to the base of the dome, so are unaffected by this issue.

I have no idea when to expect these emitters to hit the market, but the MM is designed to use and take advantage of several advantages of the upcoming K2 emitter, so i hope they are out by the time MM is ready to go to production.. I will make the prototypes with Lux3s so i don't have to wait.

on the low-power side of the equation.. the S bin model is rated at 60 lumens at 350mA.. that is mind-blowing.. means that if somehow i could get that emitter into a nano it could be double the output.. could be a little tricky given that the insides of a peak head are 8mm diameter..

Well that's what i know.. i can see some pretty amazing mods coming from this new emitter.

-awr

Re: K2 Emitter

Who makes them?

-Mike

Re: K2 Emitter

Post updated to include the maker.

Re: K2 Emitter

Interesting stuff Andrew. Is this K2 the same as the LuxIV?
Hum... 120 lm from 3W now were're talking!

Re: K2 Emitter

never heard the phrase Lux IV.. but logically it might be called that..

it takes more like 4-5W to get the full brightness of 120-140 lumen out of the high-power K2, but you can get as much as 100 lumen from 2.5W with the low-power models.. hopefully those will sell for less $.

-awr

Re: K2 Emitter

Andrew,

I noticed in the leaked datasheet that there was no information or warning about not powering the low-power versions at 1500mA.

I have a suspicion that the low power ones merely reflect that they've been tested/binned at that power level.

It's interesting to compare the typical output values of the different bins (all in white here)...

R: 45 @ 350mA, 75@700mA
T: 80 @ 1000mA

So an R and T bins are nearly the same, with R being a tad more efficient and brighter at a given current

S: 60 @ 350mA, 100@700mA
U: 100 @ 1000mA
V: 120 @ 1000mA


An S bin is actually more efficient than a U bin, and therefore, brighter at a given current...and it looks like the S and V bins will produce the same output at a given current.


So, quite honestly, I'd rather have an S bin -12 part over anything else, especially if there's a premium charged for the "high power" -14 parts...

Re: K2 Emitter

Okay, when are these hitting the market? All else is moot until I get my digits on a few. /ubbthreads/images/graemlins/grin.gif

oO

Re: K2 Emitter

Only info I can find online: K2 preview

Pablo

Re: K2 Emitter

I've asked for the URL of the source of the detailed PDF, hopefully it'll be posted soon..

Evan makes an astute observation.. the S bin -12 is 40 lumen/W at rated output, while the V bin -14 at 140 lumen max output, though brighter costs almost 6W to run so it's 25 lumen/W.

Think of the implications in multiple-led lights like the tesla-6 etc..

oh.. the -12 and -14 are lumiled's 'tag' to describe the 'low power' and 'high power' models.. '12' for 350-700mA and '14' for 1000-1500mA.. that makes perfect sense to me (not)... that's why i call them 'low power' and 'high power'.

-awr

Re: K2 Emitter

The details of the K2 have been discussed here before.

Check out lamperich's post for a link to a "leaked" preliminary datasheet.

Nice model btw Andrew /ubbthreads/images/graemlins/happy14.gif

Re: K2 Emitter

evan: I'm confused. An S bin should be 100lumens @ 700ma, but a T bin would only be 80lumens @ 1000ma!? I thought T was greater than S. Do you have the letters switched? But I get what you're saying.

Andrew: does this mean I'll have to make another emitter in photoshop? (see signature) /ubbthreads/images/graemlins/smile.gif

Re: K2 Emitter

Issac,

Go find the leaked datasheet that lamperich posted (in the thread Rossitron mentioned).

It looks like the -12 parts will be binned/rated at 350mA, whereas the -14 parts will be binned/rated at 1000mA. So, yes, a S-bin -12 part will actually be brighter than a T-bin -14 part, when both are driven at the same current.

gg lumileds - way to confuse the binning game even more.

Re: K2 Emitter

Hmmm.... Does yaesumofo know about the new slug shape??????

Re: K2 Emitter

Evan: ah, I forgot where the cut off was for the low/high power bin letters. Makes sense now. I guess they are doing this binning and part numbers for these reasons:

1) The K2 is more like an improved luxeon replacement, not really a new product line. Therefore you have the same part that can tolerate 350-1000ma. You can use this part in a 1W application or a ~5W application. But since this is such a wide area of power, there is bound to be some color shifting.... So: They test some at 350, and bin them for color/etc there and name them 12. Then they do some at 1000ma, or whatever, and name them 14. So when a developer needs a luxeon, and wants to drive it at say 1000ma, it won't shift blue b/c it was binned at 350ma, or the other way around and turn yellow...
But technicnally you can take whatever part and run it at whatever you want, but the color may shift from what it was binned at.

2) Seems the new PART number, has spots for the binning structure, and you order with that part number! Like it's built into the partnumber. I wonder if that means anyone could call and say I want this P/N.. ok not avalible, what about this one?

Thats kinda cool, as that's how other companies do it. I know I want a color bin 6 blue, or color bin 1 cyan for sure!!! /ubbthreads/images/graemlins/drool.gif !! That would make it great to nab one!!

UncleFester: not sure, but for certain you'll have to use a hotlips or wait for a new o-sink with the K2's. I wonder if yaesumofo will make a hex slug depression, or have 2 holes and push 2 pegs into them to slide into the emitter's notches???

Re: K2 Emitter

I've created a monster :-D

rossitron: thanks re: model.. I can make very precise models.. it really helps when trying to shoehorn things into small places.. i wouldn't do it any other way, they end up looking so cute i have to share.

Isaac.. well the 5W is still a beast.. but i'm really likeing the new K2.. at max power it takes almost 6W to run, so i don't think they will be called 3w or 5w anymore..

evan.. like i said.. and you corroborated.. i found no mention you can't run the -12 models at 1500mA.. i think they might just not output the lumens required to binpart them into the TUV group... they do have a much better lumen/watt which helps a lot in small lights with multiple emitters.. or big lights with LOTS of emitters.. i have revisited the 'gigasonic' design and i'm working on a 1200 lumen 12 emitter design.. will have 2hr runtime and i'm going to have a mix of 20 and 27IMS reflectors.. i'm aiming for 40,000 to 50,000 lux.. it will definitely be a force to be reckoned with.

Issac.. re 100 lumen @ 700 for S and 80 lumen for T bin at 1000.. that is correct.. the difference is that the T when cranked up to 1500 will output 120 lumens.. and i don't believe that will happen with the S.. else the part would be binned higher.

uncle fester: re: slug shape.. i'm guessing that has to do with the likes of o-sink (not sure if he's responsible for that).. the good luck and possibly 'thinking ahead' goes to the fact that a 5 1/4mm hex shape works out to the points fitting in a 5.8 or 6mm circle.. i don't have the exact measurement because they didn't specify the radius.. but looks like the circle that would enclose the hex slug matches the circle of the lux3.. if not somebody should be punished for sure.

Issac.. re binning and color.. that's pretty likely part if the equation.. i have a feeling the lumen output is also part of the equation.. i.e. when they bring a part up to 1500mA and it doesn't output 120 lumen it's not bright enough to be called a "T", so they crank the mA down to 700 and see if they get 100.. if they do they call it an 'S'.. if it's less than 100 at 700 they call it an 'R'.. something like that.

if CNC milled, the posts could be part of the solid, no need for pins.. pins obviously are easier. For my designs i'm using temporary pins.. the PCB is fabricated with holes.. and pins in most likely a wood board (but i do have a CNC milled aluminum design also).. hold the pcb and emitters in perfect alignment for soldering.

Taking a big of a cue from Mr. B.. and going one step further.. the K2 can handle a LOT higher temps.. which means that the connection to the heat sink is not nearly as critical.. i.e. no need for adhesive contact .. normal arctic alumina will suffice as long as the emitter is held in place.. and you can also get away with sinking the slug to the copper on your PCB as long as the PCB is cooled.. i.e. similar to the 'star'.. you can heat-sink the star.. the emitter is on the star.. but unlike the star.. the PCB does not have to be aluminum.

I have a very preliminary order in the works to get in line for K2s in quantity for the MM project.. they are claiming 'early summer'.. but I won't be surprised if that means august /ubbthreads/images/graemlins/wink.gif.. I would like to get 900 to 1000 of them.. for the 300ish MMs i'm aiming to make.

-awr

Re: K2 Emitter

Issac,

Likely true on both counts. Me, I'd be happier with blue shift than yellow/green shift.

As for the part numbers, I could see the distributors only using the first 8 digits (minus the binning info) - otherwise, they could be stuck with less-than-ideal parts in their inventory. Take for example, the Sandwich Shoppe. Wayne has had TV1K Luxeon III stars there forever - it seems like he can't sell them very easily - but with TWOK or TWOJ stars come in, they sell out rather quickly.

So distributors could get stuck being out of inventory of select binned parts if that's all they're able to sell...or they'd have to leave the binning info off their part listing, and bring the lottery back.

If Lumileds insists on integrating binning info into their part numbers, they should do an appropriate price laddering scheme to give incentive to purchase the lower performing parts - much like processor manufacturers do with different speed grade CPUs.

Re: K2 Emitter

Andrew: I said to drill holes and push in pegs because that would cost less and take less material & time to machine on a CNC. Oh and what is the MM?

Evan: yup, good points. I can't recall the datasheet right now but didn't the 12/14 designation in the P/N come after the binning info in it?

BTW, why does everyone spell my name wrong? I've gotten kind of used to it. Andrew you got it right the first time though, lol.

Re: K2 Emitter

I think Isaiah and Isaac are about the two most difficult to spell names.. that's funny (and sad) that i managed to typo the second two times. mea cupla.. oops.

holes and pegs is definitely cleaner i think.. very similar to the concept of my PCB solution..

the MM is my minimighty project: http://mm.rouse.com about 2000 hrs in the design phase.. just about all that is left is laying out the mainboard circuitboard.. it's quite a complex project.. I'm on design #20.. and many of the numbered projects have had an a, b, c..

Check out that thread, there will be some very exciting news there soon.

thirteen CNC parts.. 4 or 5 pcbs.. it uses fatman circuitry with 7 1/2 W capability, but is 1/2 inch thick.

I have some new pics and hope to post them soon.

-awr

Re: K2 Emitter

12 = 1 to 2 (useful range of) Watts Application
14 = 1 to 4 (useful range of) Watts Application

i.e. "Relatively" "No" Color Shift during the said Watts range. This seems more logical to me... However, I do not have any clue to what they really mean. Its just some what Logical from my POV.

Re: K2 Emitter

That leaked datasheet (if I'm looking at the same one) also lists the red, red-orange, and amber as having a max drive current of 700mA. I wonder if that's going to be updated in light of the recently released Lux IIIs.

Also, pin (lead) 1 and 3 of the 4 lead package are not connected.

The new binning structure is going to be complicated, but I guess we just pick based on what our intended drive current will be, and work from there.

On a related note, do the coloured LEDs also experience colour shifts with drive current? Thus far, my understanding is that this was an issue with whites because of the phosphor.