New 5W Luxeon rumors...

Wonderful things are in the works. I just
read on sci.electronics.design newsgroup:
----
Re: Blue LEDs
From: Bob Wilson <[email protected]>
Date: Monday 11 February 2002 07:30:23 pm
Groups: sci.electronics.design

In article <[email protected]>,
[email protected] says...

>>
>Care to post a URL for those Lumileds, Bob?

It's www.lumileds.com

More interestingly, I just saw a prototype of their new 5 Watt (yes, that's
right!) blue LED. It simply CANNOT be believed. Even 60 degrees off axis, it
is literally blinding. On axis, no one dared to look at it except through
dark glasses. It was FAR brigher on-axis, than a 20 Watt halogen. Absolutely
incredible! It is not available yet, but is soon supposed to be available in
blue, white and green (their "normal) 1 or 2 Watt red LED that is presently
available is more efficient than the green and blue ones, so it already
outputs the same light that the new 5 Watt ones do.

Lumileds is a combined effort between Agilent and Philips.

Bob.

AWWWWWWWSOME!!! Incandescents beware!

Sheesh...simply unbelievable...

OK 5W luxeons sounds nice -

BUT

I think what would be even greater on a shorter time-frame are the HD lambertian white ones which were supposed to be available "early 2002" - anyone saw one or does now something about ? Wouldn´t that be a nice one to go with the Zetek driver circuit in a MiniMag ?

Klaus

I think it's all hype. Lumileds can't even make a decent true white Luxeon right now, let alone some new improved one. Big deal, a 5W pea green LED that nobody wants to look at, let alone purchase.

Show me; so far I'm seeing green.

I just got off the phone with Lumileds

They said they just changed the web-site and the blue and green lambertians are there now

The white lambertian is probably around April timeframe - hmmh - waiting for that

They confirmed the 5W rumour without timeframe

Klaus

Same here Daniel,

I've got one last Luxeon mod to complete, then that stuff goes into the box too. It's a shame because the Luxeon did have potential, but obviously is no where ready for prime time.

I've gone back to 5mm and 10mm white LED mods also, as people will pay real money for those; unlike the puke green Luxeons that only gets an 'ick' out of customers.

Lambda

Hmm 5W,you`ve also got be wondering how big the heatsink on that`s gonna be......

Remabering that LEDs are not yet the most efficient power to light convertors.

Adam

More on the new Luxeons:

Re: Blue LEDs
From: Bob Wilson <[email protected]>
Date: Wednesday 13 February 2002 08:02:26 pm
Groups: sci.electronics.design

In article <[email protected]>, [email protected] says...
>
>Those LumiLEDs are absolutely stellar!
>Was the demo that you saw using the ones in the TO-66 looking power
>transistor
>package?

No. We have a couple of pre-production devices, and they are in a small (5mm
or so) round SMT package that is surface-mounted on a small square of
aluminum sheet that has a thin layer of FR-4 laminated to it, with circuit
traces on top. They light up a dark room! There are actually 4 dies in the
package, in series/parallel. Maximum current is 1.4 Amps (this is not a
misprint!), so that's 700mA per LED.

What is interesting is how the LEDs are constructed. It appears that they
are microscopically interdigitated. There is NO bond wire at all. All
connections are made directly to the die. Heat sinking is done by seating
the package on a thin layer of Gap-Pad, or similar, on to an area of copper
foil, The heat then can spread out somewhat, so it can pass through the thin
layer of FR-4, then into the 1.6mm thick aluminum backing and from there
into a small heat sink of the user's choice.

Bob.

It sounds to me like Luxeon just put four of their present 350mA dies in parallel, as 350 times 4 is exactly 1.4 amps.

Still, to cram that much luminosity into such a small space is pretty amazing.

I wonder if it will use the same collimator as the current LS/O, and how much it will cost?

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by bikeNomad:
More on the new Luxeons:
....
What is interesting is how the LEDs are constructed. It appears that they
are microscopically interdigitated. There is NO bond wire at all. All
connections are made directly to the die. Heat sinking is done by seating
the package on a thin layer of Gap-Pad, or similar, on to an area of copper
foil, The heat then can spread out somewhat, so it can pass through the thin
layer of FR-4, then into the 1.6mm thick aluminum backing and from there
into a small heat sink of the user's choice.
....
<HR></BLOCKQUOTE>

There's a white paper on this technology on the lumileds web page. They posted a white paper on this that I read and wrote here about mid January. If you're into semiconductor technology, check this out:
Lumiled's white paper particularly pages 8+9 on. There are pictures of the LED die and diagrams of the mounting technique. They are saying these new parts are good for 30-100 lumens and are calling them "luxeon B", compared to 4-25 lumens for the current stuff "luxeon A".

In other papers on this technology they report a 100 lumen device (cyan), and a near UV device with 1 Watt of optical output(!).

As far as the quality of the white light from white units, the new mounting technique will allow them to coat the LED die with an even layer of phosphor before it is mounted (rather than glopping it over the LED die afterward) which they SAY will result in much better color characteristics.

They are also investigating using UV LEDs with tri-color phosphor (like a compact fluorescent lamp), or coating blue LEDs with red and green phosphors. Either way they want to improve the quality of the white light.

When they are able to pull some of this off - I'll be ready with my hard earned $$

I read the white paper and it simply confirms my original suspicions. They are simply cramming four dies into the same emitter package that currently holds just one die.

From a pure size perspective, this is very exciting, especially for flashlight use. But there's no efficiency improvement, and the device will generate four times the heat, and substantial heat dissipation will be required.

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by bikeNomad:

Lumileds is a combined effort between Agilent and Philips.
<HR></BLOCKQUOTE>

And guess where Agilent comes from?

hint: starts with an 'h' and ends with
a 'p'

--Al

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by php_44:

As far as the quality of the white light from white units, the new mounting technique will allow them to coat the LED die with an even layer of phosphor before it is mounted (rather than glopping it over the LED die afterward) which they SAY will result in much better color characteristics.

They are also investigating using UV LEDs with tri-color phosphor (like a compact fluorescent lamp), or coating blue LEDs with red and green phosphors. Either way they want to improve the quality of the white light.

When they are able to pull some of this off - I'll be ready with my hard earned $$

<HR></BLOCKQUOTE>

To create true white light with good color rendition, wouldn't it make more sense to use 3 chips in the basic colors (RGB) and add up their light? I think this way a higher total lumen output per watt should be possible, because the 3 basic outputs are just added up. Phosphors have efficiencies of less than 1, so there will always be some loss compared to the original (blue or UV) chip's output.
Also, we know from the 3-, 5- or wideband phosphors used in fluorescents that

a) the color rendition even of the best phosphors is limited (CRI of the best fluorescents is around 95, but 75 to 80 is more typical),
b) phospors with good color rendition have lower output (about 20 to 30% less than high output, worse color rendition types),
c) color rendition changes with temperature (and the temperatures in most fluorescents doesn't even go as high as on LED chips!), and
d) phosphors don't have absolute long term stability, resulting in small but visible changes in color rendition over their life time.

For these reasons, I don't think we will soon see a white LED with good and stable color rendition, and the life time you can expect from non-white LEDs.

In contrast, additive white light generation could have several advantages:
a) higher total output per watt, as there is no loss in any phosphor
b) stability typical for LEDs over a longer lifetime than usual for phosphors
c) if the output generated by the 3 colors is made separately adjustable, color, and, if white light is to be generated, color temperature could be adjustable.

What do you think?

Woo Hoo!!! Almost time for a direct drive LS "B" mod!!! Can't wait, I'm about to pee my pants!

Eric

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by wus:
To create true white light with good color rendition, wouldn't it make more sense to use 3 chips in the basic colors (RGB) and add up their light? <HR></BLOCKQUOTE>

Actually, the Lumileds white papers mention that going with red, green, and blue LEDs to make white light would be significantly more efficient. But, they want a white lamp to sell (ie- apply power and it makes white light), rather than an RGB LED, so the disadvantages they site for this approach are:
<UL TYPE=SQUARE><LI>Mixing the light within the package to produce a uniform white output.
<LI>Practically being able produce lamps with consistent color temp given the variability of LED die light lumen output in a production line. (ie- what if you are consistently getting really bright reds, but have a low percentage of blues that can keep up, and the greens are all over the map).
<LI>As the lamp ages - being able to somehow keep the color temperature constant given the different age based lumen fall for each color LED. (ie- perhaps as the lamp ages, green LEDs loose lumens faster than reds - resulting in a white lamp that ages to a magenta color - YUK!)
<LI>The RGB solution has a deficiency or hole between the red and green part of the spectrum.
[/list]

Apparently they think a phosphor based solution, though less efficient, will yield a more consistent product from their manufacturing process - and result in a lamp that is more stable with age than an RGB solution; for the cost target they want to meet.

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR> Hey, I can hear your "WOO HOO!!" all the way out hear in Sunny Southern California! <HR></BLOCKQUOTE>

LOL!!

I can confirm the existance of the 5W Luxeon but my rep asked me not to publish anything.

It's a good thing you guys posted this because otherwise I wouldn't have known to ask him about it.

It seems they had asked some of their big customers if they should make such a device and their response they got was lukewarm. I told him that he should have asked one of the flashlight manufacturers. I told him about 6 times during the call that the LED would be a definate winner and he would sell lots of them. He was suprised to hear that. I guess the feeling was that we thought that the 15 lumen LS was plenty and our customers preferrred more battery life. I said, "oh no. Definately nix the battery life in favor of more brightness everytime!".

So who knows? He will try to send me a sample and maybe they will decide to manufacture the part sooner rather than later? This will probally appear in lights early next year?

Peter Gransee

Gee, I hate to sound negative, but will this super white luxeon be pink, violet, blue, or pea green?

It's nice they're working on new stuff, but sure would be nice if they could produce a decent white Luxeon in the current form.......

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Gransee:
I guess the feeling was that we thought that the 15 lumen LS was plenty and our customers preferrred more battery life. I said, "oh no. Definately nix the battery life in favor of more brightness everytime!".
<HR></BLOCKQUOTE>

How could they be so out of touch with what their customers want? It's a good thing you told them that.