Cree to release Quad chip LED

[mds82]

I do not know any Specs yet. I heard from my LED supplier that Cree is going to release a 4 die chip q1 or q2 of this year. I hear is supposed to be in the same size package as the XR-E as well.

I will post up more info as soon as i know.

 

[rizky_p]

nice...lets just wait and hope it is true...

 

[Cody Gracie]

I do not know any Specs yet. I heard from my LED supplier that Cree is going to release a 4 die chip q1 or q2 of this year. I hear is supposed to be in the same size package as the XR-E as well.

I will post up more info as soon as i know.

My source is multi-chip product but not sure whether 4/6 chips.
Also they will have a new 0.5W LED that really outperforms Nichia and SSC.

 

[IMSabbel]

My source is multi-chip product but not sure whether 4/6 chips.
Also they will have a new 0.5W LED that really outperforms Nichia and SSC.

That small one might be really interesting for a L0D-style light. If they reduce the emitter size, that is.

 

[TORCH_BOY]

I want one for my M@G

 

[Gryloc]

mds82,

Thank you for the info! That is good news. There are many people that do not like the idea of four die in an emitter like the Luxeon V due to heat. Nowadays, the LED technology is far better and more efficient. Now you could do this and pull it off with less heat. With 4 EZ1000 dies, you could really under power the emitter and get similar brightness of the X-binned Luxeon V! Lets assume that the Q5 binned dies are used (based from jtr1962's data): Power each emitter at around 150mA (~0.5W each, ~2W total) and get up to 200 lumens! Power each one at 400mA (~1.3W each, 5.2W total) and expect up to 460 lumens!!!

I remember after the P3 and P4 bin came out, I dreamed up such an emitter as this (4 x EZ1000 die) and played around in Solid Works for a while (instead of focusing on class work ) and came up with this:

(click for larger versions)

I originally posted these pictures in some of my first threads here at the CPF. This was to be used in a dream automotive headlight cluster (4 of these radially mounted to a special heat sink, then focused in the reflector of a big car headlight). I planned the light if it could use the P4-bin die technology. Now, with the Q5 (widely available) and the R2 (sort of available), this would be a killer emitter! Notice I did not consider using even an external ESD diode, and I did not consider the SiC substrate that usually goes under the XR-E die. I was hoping that the dies could be soldered to the traces like the older Lumiled's Luxeon die+SiC chip was. This first picture shows the green solder mask and a possible lens retaining ring, which would be slightly larger smaller* than the current XR-E's retaining ring (see below), holding a seemingly invisible glass dome. The other two are bare. I have the original files somewhere. I should recapture higher resolution images and re-post them.

Well, I am super happy that the quad die emitter will be around the same size as the existing XR-E. I hope the dies, if they are the more powerful EZ1000 dies, will be tightly grouped together. It seems like Cree is not concerned about space in the XR-E since the bond wires spread out pretty far. Tighter grouping would be great for us Flahsaholics and Modders :candle:. Lumileds could do this easy with their TFFC technology (no bond wires). They could make a Rebel 2-4 times larger in area, group 4 Rebel dies in a tight bunch, and have a great product. They could even update the 4 dies in the existing Luxeon V with the Rebel dies (with a slight emitter redesign). I love dreaming...

-Tony

*EDIT: :sweat:
Oops, I opened my original drawings and found that my drawn emitter was 5mm x 7.95mm. That is much smaller than the XR-E emitter, which is about 7mm x 9mm. The Rebel is 3mm x 4.5mm for comparison. In the above images, just imagine the dies appearing smaller so it can be comparable to what Cree may release...

 

[Sgt. LED]

What a great place, the OP reveals something new and you whip out a pic years old showing virtually the same thing!
Just think what we could do if we had funding for our dreams. I need to win the lottery!

 

[jtr1962]

Great news! It should be possible to get >100 lm/W at the 2 watt power level, and >80 lm/W at the 6 watt level. I do hope they learned from the past mistakes of Lumileds and put the 4 dies in series. Paralleling dies is what probably caused the Luxeon V its low lifetime.

Gryloc,

That picture reminds me of a larger version of the Rebel except with four dies! It looks like you did a 4-series arrangement based on your arrangement of bond wires. Maybe Cree stole your design. :laughing:

 

[2xTrinity]

Great news! It should be possible to get >100 lm/W at the 2 watt power level, and >80 lm/W at the 6 watt level.

A few weeks ago I predicted AAA keychain lights would be able to produce over 200 torch-lumens (on 10440s) within the year. Hopefully this emitter lives up to the specs.

I do hope they learned from the past mistakes of Lumileds and put the 4 dies in series. Paralleling dies is what probably caused the Luxeon V its low lifetime.

I agree that would make things better for longevity of the LED, especailly in fixed applications, but 4s would make the LED essentially impsosibel to use in single-cell applications, as a high-power boost driver with significant gain (almost impossible to do in a small size) would be necessary to make it work. So for flashlights IMO I'd prefer to have parallel, and simple underdrive the emitter.

I'm guessing series is more difficult to do however, as the heatsink/contact pad in many cases acts as a common positive terminal (on the XRE's it's isolated with a layer of ceramic).

 

[Gryloc]

A few weeks ago I predicted AAA keychain lights would be able to produce over 200 torch-lumens (on 10440s) within the year. Hopefully this emitter lives up to the specs.

2xTrinity,

Do you mean in production lights, or do you mean experimentally? For a long running production light to produce 200 bulb lumens that uses a 10440 sized cell is within reach. The reason I say this is because I modified a standard Solitaire (actually "overhauled" is a better term) to run a Rebel 0100 from a Li-ion 10440 cell at over 1000mA (it has a super low Vf of 3.32V at 1A) to produce an excess of 200 lumens. The Rebel 0100 is mated with McGizmo's wonderful McR-10R reflector, and it is stunning! This was built in the late November of 2007 and I have been too busy (and lazy) to create a thread about it on the CPF. I will do that soon. I am tired of having non-flashaholics say, "Ooh, thats a bright.... and so cute", or, "My big Maglite is still brighter".

So, experimentally, it is easily possible. The poor 10440 li-ion is over taxed, so the burn time is less than 20 minutes to dim (it is direct driven). So, your prediction is true, so let us hope that this Cree comes out and makes 200 lumens for a Fenix LOD (or Solitaire) much more realistically than my little freak. I can see this become a hit for the small flashlight market, as well as the monster flashlight market (where multiple emitters are used). So, will McGizmo's McR-19XR work well with this new emitter? If so, I will reserve a Mag for a quad emitter mod. :devil:

Again, bring on the Quad die Cree soon!

-Tony

200lm Solitaire apart:

Assembled and on:

More later...

 

[PhantomPhoton]

Mmmmm the plot thickens.
So thats SSC and Cree with Multi-Die emitters. LumiLEDs must be coming in with one soon as well. The things is this tells me that the manufacturers are starting to run out of headroom for higher binned LEDs. They'll probably be able to squeeze out another flux bin or three on this current technology but that is probably it imho. (hopefully I'm wrong) Things might be slowing down for the next year while they ready new white LED technologies.

 

[Calina]

Mmmmm the plot thickens.
So thats SSC and Cree with Multi-Die emitters. LumiLEDs must be coming in with one soon as well. The things is this tells me that the manufacturers are starting to run out of headroom for higher binned LEDs. They'll probably be able to squeeze out another flux bin or three on this current technology but that is probably it imho. (hopefully I'm wrong) Things might be slowing down for the next year while they ready new white LED technologies.

Maybe you're right, maybe not. What that 4 dies LED tells me is that Cree is going after the car headlights market. They're going where the money his (for now).

 

[IMSabbel]

A few weeks ago I predicted AAA keychain lights would be able to produce over 200 torch-lumens (on 10440s) within the year. Hopefully this emitter lives up to the specs.

Sorry, for AAA-lights, this emitter will be no use at all.
The size will just be too big to get any kind of reflector in a non-self-shadowing bezel.

 

[rizky_p]

Mmmmm the plot thickens.
So thats SSC and Cree with Multi-Die emitters. LumiLEDs must be coming in with one soon as well. The things is this tells me that the manufacturers are starting to run out of headroom for higher binned LEDs. They'll probably be able to squeeze out another flux bin or three on this current technology but that is probably it imho. (hopefully I'm wrong) Things might be slowing down for the next year while they ready new white LED technologies.

didnt someone just announce a new method using nano-blah-blah technique to increase die efficiency up to 300 lumens? i am sure with this technology manufacturers has some ways to increase lumens even further.

 

[CM]

If it's 2s2P like the Lux V, I can see myself working OT with KL6 upgrades

 

[daveman]

Mmmmm the plot thickens.
So thats SSC and Cree with Multi-Die emitters. LumiLEDs must be coming in with one soon as well. The things is this tells me that the manufacturers are starting to run out of headroom for higher binned LEDs. They'll probably be able to squeeze out another flux bin or three on this current technology but that is probably it imho. (hopefully I'm wrong) Things might be slowing down for the next year while they ready new white LED technologies.

I agree with your take on these quad die emitters, but I believe Cree will reach 140 lumens at 350 ma before the ez1000 has been tapped out. Beyond the 140 lumens figure, they'll definitely need the next big breakthrough to reach that.

 

[easilyled]

Sorry, for AAA-lights, this emitter will be no use at all.
The size will just be too big to get any kind of reflector in a non-self-shadowing bezel.

For a pure flood light, no reflector is necessary.

The McGizmo mules are a prime example.

Since Crees project most of the light forward anyway, this could be
an ultimate flood light.

 

[Bullzeyebill]

If it's 2s2P like the Lux V, I can see myself working OT with KL6 upgrades

You got that right.

Bill

 

[lyyyghtmaster]

I believe what's happening here is that, by packaging 4 (or more) reduced-current dies within one housing which shouldn't get that much more expensive as a result, the companies are trying to be more economical at efficiently creating lots of lumens that don't need to be focused very tightly. Where would this be most useful? In general lighting, of course! Though mostly cost-prohibitive now, in the ever-less-distant future we're going to see LEDs take major strides toward replacing incandescent bulbs and fluorescents! This sounds like fun!