98 CRI from BLUE emitter

Anders Hoveland

Enlightened
Joined
Sep 1, 2012
Messages
858
Apparently it is possible to get very high CRI rating using a blue emitter. claims 97-98 CRI

"Gallery White" from LED Engin
http://www.mouser.com/ds/2/228/LZ9-00GW00-255475.pdf

It has fairly good blue-cyan coverage, looking at a spectral graph the dip in cyan wavelengths are only 66% of the blue maximum 465nm peak.

This is good news. I know they already make (>96) high CRI white LEDs using violet emitters, but I have a rare [eyes/skin] sensitivity to violet wavelengths, so these are not really compatible with me.

gallery_white_spd.png



Typically they have to use a violet emitter to excite a blue LED phosphor to get good coverage in the blue region of the spectrum. I am not sure what LED Engin is doing here. The 465nm emitter they are using is bit longer wavelength than others that have been used for white LEDs, might help provide a little more coverage directly (the blue emitter has some bandwidth), but there are not going to be any shorter wavelengths that could stimulate a cyan phosphor.
 

Anders Hoveland

Enlightened
Joined
Sep 1, 2012
Messages
858
There are nine dice in that emitter. Maybe they are not all the same, so they fill in each other's holes?
That would be a multi-emitter array (and LED Engin does make some other chips which use multi-emitter arrays), but no, that is not what they are using here.
They are just using a different type phosphor composition. It is easy to see from the spectral graph. The only reason there are nine emitters on the board is because it is high power. It would be hard to get 20 watts out of a single emitter!

I am not sure exactly what is different about the phosphor composition they are using, how they get more coverage in the blue-cyan.
 

Xe54

Enlightened
Joined
Sep 12, 2005
Messages
201
It's remotely possible that they have snuck in a cyan chip, covered with phosphor so you don't notice. Some other tricks could yield the same result too. Unless you image each individual chip onto a spectrometer and analyze them all it would be nearly impossible to tell by eye.

Does anyone know for certain that the chips are all the same in LEDengin's Gallery White emitters?
 

Anders Hoveland

Enlightened
Joined
Sep 1, 2012
Messages
858
Perhaps part of it has to do with the blue emitter itself ?
Some blue emitters have wider bandwidths than others. It might be possible they chose to source wider bandwidth emitters for this purpose, and this might have an effect on CRI.
Although the peak is at 465, some of these LEDs can trail off much farther towards increasing wavelengths than decreasing, so the average wavelength is often a little greater than the peak wavelength. It has to do with how the semiconductor layer is formulated. If the emitter they are using does trail off farther towards the right, we would not be able to see it because it is obscured by the phosphor coverage in the graph.
 

WeLight

Enlightened
Joined
Sep 28, 2005
Messages
221
I dont know how you would get a Hi CRI Led without blue(or Violet). Fundamentally HI CRI is an RGB outcome, whether you tweak the led colour mix or phosphor. We have built 98CRI 4400k modules and the blend of these colors is the easiest way to get that result. A histogram of R1-R15 is a useful starting point
 

SemiMan

Banned
Joined
Jan 13, 2005
Messages
3,899
Xicato has been doing this for ages. Bridgelux has very high CRI as well and recently very large gamut lamps that eschew trying to mimic halogen and their hobbling and not necessarily relevant cri to provide a light that more effectively saturates a wider color range.
 

Anders Hoveland

Enlightened
Joined
Sep 1, 2012
Messages
858
So on the topic of some blue emitters having a wider bandwidth of spectral emission, there does not seem to be any blue-green dip in the new Electrospell RGBW white LEDs:
http://www.compoundsemiconductor.net/article/90300-broadband-leds-enhance-colour-fidelity.html

The secret is they are using a blue emitter with broader band coverage.
http://www.led-professional.com/pro...all-kinds-of-high-definition-colour-rendering

Spectrafill blue LED is a broadband (wide emission spectrum) solidstate
light emitter. Its spectral width is roughly three times wider
than that of usual narrow band LEDs. The polychromatic emission
from this device contains correspondingly more shades of blue
colour, making it suitable for high fidelity illumination applications.

Just in case anyone is curious, the wavelength emission peaks at 430nm, trailing further out towards longer wavelengths, and is down to half peak amplitude at about 480nm.

(I really wish they would make a spectrafill blue-cyan 488nm peak LED, but that's just wishful thinking :sigh:)



On the topic of super high CRI LEDs, if anyone is wondering why 95 CRI LEDs are not good enough...
In terms of just color rendering, 95 CRI is excellent, really there is no color rendering problem unless it is for art use and there is some turquoise light blue color that needs to be rendered perfectly. But another reason higher than 95 CRI is important does not actually have to do with color rendering. Blue-emitter LEDs, which are 95CRI, emit light that still feels a little like "LED light". A better spectrum will make them seem more "natural". Difficult to describe, but it does not actually have to do with color rendering. 95CRI LED might be closer in feeling to halogen/sunlight than 80CRI, but it still has a slight yellow-blue-purple tint, and possibly just a little difficulty focusing the eyes under this type of illumination.

To me, a 95CRI LED seems like 60% halogen mixed with about 40% LED light.

An important reason for super high CRI LEDs is that once the CRI starts getting to above 95 or 96, the light starts to feel much more natural, almost like it is not "LED light" anymore. This does not really have to do with actual color rendering. This is based on my observations with 3 different types of super high CRI emitters.

My hypothesis is that more cyan-azure phosphor coverage in the spectrum means the blue spike does not need to be quite as high. The shorter blue wavelengths tend to be harsher on the eyes.

The LED does not necessarily have to have an exceptional CRI value for this effect to be observed. Some time ago, GE developed the VIO LED, which used a 405nm violet emitter in conjunction with a blue phosphor. This design was primarily intended to allow dimming in LED lamps without any color shift, but it had the unintended benefit that the quality of the light seemed surprisingly natural, despite the CRI being only 85.
Although the VIO never caught on and was discontinued, one of the members in this forum tested a 4W GE Vio in 2009 and commented at the quality of light (you have to remember, there were not a lot of super high CRI LED options available at this time).
I just received a set of GE 4W LEDs to test. I've got to say their product literature does them a great dis-service- they look nothing like what's on the market right now. In fact they look ALOT like real light bulbs- at least the light emitting part.

Jason
Sorry, I know we are discussing blue emitters, and violet-emitter white LEDs are a whole separate category unto themselves, but just wanted to quickly bring this up.
 
Last edited:

SemiMan

Banned
Joined
Jan 13, 2005
Messages
3,899
There does not seem to be any blue-green dip in the new Electrospell RGBW white LEDs:
http://www.compoundsemiconductor.net/article/90300-broadband-leds-enhance-colour-fidelity.html

The secret is they are using a blue emitter with broader band coverage.
http://www.led-professional.com/pro...all-kinds-of-high-definition-colour-rendering



Just in case anyone is curious, the wavelength emission peaks at 430nm, trailing further out towards longer wavelengths, and is down to half peak amplitude at about 480nm.


(I really wish they would make a spectrafill blue-cyan 488nm peak LED, but that's just wishful thinking :sigh:)

__________________________________________________________________________________________

But to change the topic, if anyone is wondering why 95 CRI LEDs are not good enough...
In terms of just color rendering, 95 CRI is excellent, really there is no color rendering problem unless it is for art use and there is some turquoise light blue color that needs to be rendered perfectly. But another reason higher than 95 CRI is important does not actually have to do with color rendering. Blue-emitter LEDs, which are 95CRI, emit light that still feels a little like "LED light". A better spectrum will make them seem more "natural". Difficult to describe, but it does not actually have to do with color rendering. 95CRI LED might be closer in feeling to halogen/sunlight than 80CRI, but it still has a slight yellow-blue-purple tint, and possibly just a little difficulty focusing the eyes under this type of illumination.

To me, a 95CRI LED seems like 60% halogen mixed with about 40% LED light.



Old and very wrong news.

I am not sure this company even really exists or at least their products.

Their so called "flat" 95+ CRI emitter is listed to have color coordinates of 0.38,0.32. That is so far off the black body it's not funny. It's CRI would by definition be terrible.
As well, they call the light "white", neither cool nor warm bias by their wording, BUT, a flat spectrum would be cool biased, at least 5000K if not 5500K.

They have a semiconductor RED led that magically has a peak in the blue ... just like in the human eye. Just how do they accomplish this? .... oh ya, that would be a blue peak from a phosphor converted LED at 455nm ... the ones you do not like.

Oh, not new either .... they announced that all in early 2012 .... and I don't think they have released a new piece of data yet.

The data on the web ... years old, looks like it was done by a 8 yr. old who just learned how to use MS Paint. That picture of the LED with the reflector is embarrassing ..... not remotely straight. What a joke.

I hard a hard time believing that 430nm peak as well unless they are pumping with a 405nm LED which seems doubtful.

All in all I am calling BS .... likely why no new data.
 

Anders Hoveland

Enlightened
Joined
Sep 1, 2012
Messages
858
I hard a hard time believing that 430nm peak as well
What is so strange about that? In fact, the emitter used in the Philips Luxeon Lime is 430nm (and a tiny bit of it gets through the phosphor).
It is very common for the wavelength distribution of an LED spectral emission to be lopsided. The peak wavelength is not the "dominant"/average wavelength. The wavelength distribution often trails further out towards the longer wavelengths, and this type of distribution tends to be more pronounced in LEDs with a broader bandwidth. For this reason, the actual color of an LED can be different from the color of its peak wavelength value. Despite the 430nm indigo-violet peak, probably that spectrafill blue LED is more like a royal blue, just with more coverage across the blue wavelengths.

In the multi-emitter combination Electrospell is using here, no single LED is producing most of the cyan. The cyan/azure blue wavelengths are produced by a combination of the spectrafill blue (which trails off at a lower intensity far into the cyan), the small 465nm blue spike from the red phosphor emitter (which trails off, half peak intensity at 480nm), and a small contribution from the green phosphor (this contributes some 490nm but rapidly falls off so barely contributes any 480nm). It is interesting how the combination of 2 different emitters and a phosphor combines to fill in this range of the spectrum, despite none of the primary emissions being centered very near here, everything fits together.
 

SemiMan

Banned
Joined
Jan 13, 2005
Messages
3,899
I dont know how you would get a Hi CRI Led without blue(or Violet). Fundamentally HI CRI is an RGB outcome, whether you tweak the led colour mix or phosphor. We have built 98CRI 4400k modules and the blend of these colors is the easiest way to get that result. A histogram of R1-R15 is a useful starting point

Getting 98 cri with RGB is well ... Impossible unless what you are really using is broadband phosphor converted emitters ... Which are not really red green or blue at that point. Then you add in the issues of mixing and lifetime control.
 

sinner-cpf

Newly Enlightened
Joined
Apr 20, 2014
Messages
183
Location
Dubai, UAE.
Blue has the shortest of wavelengths hence it produces higher resolution in microscopy applications, in the early scientific research.
 

Anders Hoveland

Enlightened
Joined
Sep 1, 2012
Messages
858
Blue has the shortest of wavelengths hence it produces higher resolution in microscopy applications, in the early scientific research.
Yes, for imaging microscopic cells. But when it comes to visualizing ordinary things it is more the reverse, since the lens in the eye has difficulty focusing the shorter wavelengths. Ever looked at a fluorescent illuminated blue sign at night? It appears noticeably blurry, difficult to properly focus on. In fact, it is easier to focus on longer blue wavelengths than shorter ones, even noticeably so.
 

SemiMan

Banned
Joined
Jan 13, 2005
Messages
3,899
Try again .... Low CRI approach? .... Its 92 ... Nothing low CRI about, just that they know how to adjust spectr for CRI and you do not.

People who don't know what they are talking about talk about CYAN valleys in LEDs without knowing the typical 80 cri LED actually had an excess of CYAN to balance out the rest of the spectrum, hence why True white has less Cyan.
 
Last edited:

Anders Hoveland

Enlightened
Joined
Sep 1, 2012
Messages
858
I have been looking at the spectral graphs (from the data sheets) of different 95-97CRI LEDs (all blue-emitter) and they do NOT all have exactly the same type of spectrum. Some of them have more coverage into the deep red long wavelengths, some of them have more greenish-cyan coverage, and some of them have better bluish-cyan coverage than the others. This has to do with both the exact characteristics of the phosphor being used and, to a lesser extent, the spectral characteristics of the blue emitter itself. There might be two LEDs rated 95 CRI and one has slightly better red saturation, whereas another has slightly better cyan saturation, so different engineering designs seem to be pursuing slightly different strategies.

Some of these other blue emitter super high CRI LEDs are the Bridgelux Décor Series Ultra, the Oslon SSL150 family, the Xicato Artist Series module, Hexolux CineWhite, and others, such as the Ledzworld MR16 Chameleon (though I cannot find the spectral graph for that one).
 
Last edited:
Top