High CRI and its significance

McGizmo

Well-known member
Joined
May 1, 2002
Messages
17,274
Hi guys,

I would like to focus this thread on discussions more in general and of the rambling and brain storming fashion that we sometimes enjoy here. I'll set a little background and then leave the thread open to go where it will.

There have been many threads and discussions already about color rendering, tints and the ability of the various lights to illuminate colors and their relative intensities. Initially when LED's were weak, there was much said on the LED VS Incan comparisons. Now that LED's are putting out significantly more lumens, for many the issues have become less significant if at all.

My house now is essentially all LED in illumination and it has been great for the most part. I did discover that in the bathrooms at least, I needed to add some red LED's to bring the skin tones and colors closer to realistic or what you would see under natural light. Another area I found I needed to improve was in the cooking area where the meat preparation was difficult without some better red rendering. Instead of using red LED's in addition to the typical high power LED's I have been working with, I opted to use a LED I have been silent about*, namely a Nichia 083 High CRI LED. Above my stove, I have a string of these:

LED-String.jpg


complete-installation.jpg


There were two 35 watt incans which were replaced by these Nichias driven by a Xitanium driver.

off-on.jpg


* I feel bad about keeping silent so long on these Nichia High CRI LED's because my motivation has been purely selfish in doing so and not in the spirit of moving the "art" forward as quickly as our community can. I told myself that as soon as I had a complete first article light in hand I would spill the beans on this LED. I completed a first article SunDrop today:

SunDrop-1.jpg


More on the SunDrop in a new thread devoted to it.

Back to High CRI.....

I think for the most part, color rendering is adequate with what we have on hand now and with enough lumens, one can overcome short comings in certain wave bands. However, there are specific applications and times when better or "full color" is important and desired if not required. Further it is my suspicion that typically this need for good color rendition is in flood applications where you don't want radical gradations in intensity across the field as would be the case of a collimated beam like from a reflector.

These Nichia 083 High CRI LED's have a different phosphor in them and unfortunately this phosphor inherently cuts down on the flux similar to how the warmer tint LED's suffer in lumen reduction. However, I have found in talking about CRI and warm tints with the LED people that there are bins for warm and neutral white that are no better in CRI than the higher color temp bins and then in addition, there are bins and phosphors where CRI itself is of primary focus. I have sampled some Seoul LED's which report to approach 90 CRI a. The Nichias are ranked in the mid 90's.

With no idea of true accuracy or level of possible error, I offer a couple spectrums I captured with my spectrometer and integrating sphere. The first is with a sample 083 High CRI and a Cree XR-E that I had handy:

Cree-NichiaHI-CRI-Spectrums.JPG


I was impressed enough with some proto lights I built and walked around with to bite the bullet and order a reel of these Nichias (1400 pieces!) along with modifying a MCPCB that was in the works so that it would be host to both the Dragon and Nichia 083 LED. In the spectrum below, one of the 1400 production pieces in the SunDrop is compared to today's noon sun here on Maui:

SunVS-SunDrop-Spectrum.JPG


You will notice that in the case of this sample, the blue spike is not as high as the max or peak wavelength which is pretty close to the sun sample. Again, I don't consider myself expert or even proficient with the spectrometer and its software so these graphs can't be taken as rock solid. They do support what my take and perception has been in actually using the LED's.

I also have a CreeBar on Makita battery that I mentioned in another thread a while back:

CreeBar-083-LT-2.jpg


I have been using this light for fill in in many of the recent photos and there may be some advantages for high CRI sources in photography as well. :shrug:
 

AW

Active member
Joined
Oct 15, 2004
Messages
33
I have also tested some SSC samples ( S42180 ) with a CRI of 93 and 4000K. Very pleasant tint which compares to a bright incan bulb. Flux is 60 lumens @350mA and about 100 lumens when driven to 800mA ( max. ). It is my EDC keychain light now.


P1000238.jpg




42180warm.jpg
 
Last edited:

souptree

Well-known member
Joined
Dec 24, 2005
Messages
1,174
:twothumbs

This is SO much more interesting than the flux o' the month club. :thumbsup: I am very excited to see color rendition moving to the forefront of the discussion.

It is obvious we are on the brink of many interesting developments.

Personally, I am glad to see you withholding info to protect your own interests and I don't think you owe the hobby any apology for it. Your continued success is pushing this hobby forward in tangible ways and your protecting that which enables you to keep the cycle flowing certainly merits no shame on your part.

I had no idea 2008 was going to be such a breakout year. We aren't even halfway through!

:popcorn:
 

Erasmus

Well-known member
Joined
Dec 29, 2004
Messages
1,077
I'm also doing some projects with warm white XR-E's. Very nice warm white tint, and quite bright at [email protected] each. Yesterday I quickly made a 'portable desk lamp' with 4 of these LEDs, total power consumption of 5W, comparable to output of a 40W incandescent. On 9AA NiMH's it gives me plenty of runtime to study outside in the garden during the night :) More info on this later today, when my neighbour returns home and I can borrow his camera :)
 

Chronos

Well-known member
Joined
Jan 27, 2006
Messages
1,757
FINALLY! This is great news, indeed.

Perhaps it is aging... but I've noticed how much I prefer the more natural, warm tints for most EDC-type lighting tasks. Also, when on the trails at night, a more natural tint is wonderful as most other tints tend to "flatten" the landscape.

This probably makes little, if any sense.

I'd love to see the SunDrop in action.
 

chimo

Well-known member
Joined
Sep 16, 2004
Messages
1,905
Great post Don.

Since these are likely going to be used a lot for fixed lighting, (which should see a much longer duty cycle than portable lighting) is there any info on lumen maintenance?

Paul
 

Haz

Well-known member
Joined
Apr 14, 2005
Messages
919
Thanks for sharing this Don. I've been waited for something like this, and its great to see that it is available. I love the colour rendition of Incans, such as the Surefire A2, but love the efficiency of leds. This seems to be the best of both worlds. For outdoor use, it seems significantly better using incans if you want to differentiate the leaves from the branches. The efficiency and also the fragility of the filament bulb, always concerned me abit, especially going on a trip with one light. I sometimes think white as snow led is a bit over-rated, and i use to be sucked into this notion that the whiter the led, the better it is. Gradually over time, i accepted that a bit of yellowish tinge is better for colour rendition, especially for outdoor use.
 

Edwood

Well-known member
Joined
Aug 27, 2006
Messages
994
I want to build a small video light with High CRI LED's. I've tried using Cree LED's, but they really flatten the scene and look less than appealing.

High CRI LED's look very promising.

-Ed
 

McGizmo

Well-known member
Joined
May 1, 2002
Messages
17,274
Hi guys,
I wanted to bring a spectrum over to this thread that I had posted in another:

lightsourcesfigure2.jpg


I borowed (hope it's OK) this image from HERE. As mentioned in the other thread, as I understand it CRI is a means of rating or quantifying a lights color rendering ability as compared to a tungsten lamp. There is a CRI a which I take the "a" to designate average. There are actually a number of CRI measures based on specific bands of light. The "a" I assume is an average of these other bands. By definition, as I understand it,a tungsten lamp is assigned a CRI of 100.

To add some more confusion, I have read or think I read that to compare CRI of one lamp or light source to another, you also need to take the color temp in consideration. This stuff is all fine and dandy when you are dealing with black body radiators. I have also read that the sun is a black body radiator. Well if you compare the spectrums of the sun to that of tungsten in the graph above, you can see that they don't have the same curve. The sun and a tungsten lamp meet with the same relative output at around 525 nm but below that, the sun has more relative output (in the blues) and the tungsten has more output above and by the time you approach the reds at the end of the spectrum, the tungsten is putting out a heck of a lot more relative output than that of the sun!!!

Are our eyes tuned through evolution to sunlight or artificial light? If we could produce any type of light we wanted, what would be our target or what source should we model after?

All of our quantifications and systems of measure seem to be related to the tungsten or incandescent sources and since that is what we have had to work with since Edison's first bulb, this is understandable. However now with SSL sources, fitting non plankian sources to plankian systems of measure or identification is problematic.

I believe we will all agree that color rendering is important in many applications and endeavors but how best to identify what is needed to allow us good color rendering? I dare say that if we had an artificial source that emitted a spectrum following that of the sun, we would be pretty happy with the results. However, our eyes are remarkable in their ability to interpolate and extrapolate and fill in some "blanks" and provide us good information even when it is not readily apparent.

Below is a data comparison I gleaned from my spectrometer software and it relates to the Nichia and Cree LEDs which are compared in a spectrum graph above.

CRI-Nichia-Cree-Data.JPG


I have no idea how accurate this comparison is nor do I have a clue as to just what the data is supposed to tell me! :eek:

Note that there is a CRIa shown and incidentally the measured CRIa is in accord with what I would expect given the info on these LED's. Further the CCT measured or assigned is also seemingly in keeping with what one would expect. You can also see that the various CRI measures (R1-R14) are identified with the CCT and seems to show the dependence of CRI on CCT.

I can't seem to find it now but somewhere I saw a mapping of the R1-R14 in terms of what specific band of color they are based on. For some disciplines or areas of interest, the color rendering across the full spectrum may not be as important as the rendering in specific bands. I.E. a high index in say R10, R13 and R4 may be more important to the viewer than a high CRIa which is weak in these bands.

I have been exposed now to enough glimpses into the science of light and color to the extent that I realize it is quite complicated and there are so many dimensions and considerations regarding the light itself let alone how we humans process it!

In my ignorance, I think a high CRIa has meaning to be sure but the number doesn't tell me much more than a Flux, Lux, CCT or other single number in regards to how a light will actually perform. We seek simple answers and simple means of qualifying something as good, better best.

I have been approached by a small company in the medical field who is interested in a 6000K LED light source. I am confident that they have expectations based on this 6000k source which may or may not be met by a 6000k source I might provide. They attach more to this number than it can insure or guarantee.

From what I can gather, the differences for all of the LED manufactures when it comes to high CRI LED's is the phosphor. I have no idea if there is a difference in lumen maintenance based on the phosphor itself or not. These Nichias I am using have an obvious short term glow after the power is shut off.

For you guys who hope to have collimated beams high in CRI, you may need to wait for a while when you consider that these high CRI sources are typically 25-40% lower in flux than there higher CCT and lower CRI brothers. Greater intensity of a lower CRI source, on target may tell you more information or adequate information. :shrug:

Oh yeah, out of curiosity, I checked the CRI of the sunlight sample I took yesterday and again without knowing how off my calibration might be or how it effects the measurements, I was shocked but not surprised to see that the sunlight measured CRIa in the high 80's. :D

It seems that so much of the equipment available to measure our lights does give you a handle with which to get a grab on them but I don't take it as a given that this handle can support the weight if you start swinging the light around by it! :nana:
 

BigHonu

Well-known member
Joined
Apr 1, 2002
Messages
1,242
Don,

Did you hint at the notion that an LED may have a highish color temp AND a good CRI rating?

Initially, I presumed that a lower temperature (between 4500K-5000K) would inherently yield better color (assuming overall saturation was sufficient) than a higher one (6500K and above).

That Nichia looks impressive. What is it rated on the flux side of things?
 
Last edited:

Icebreak

Well-known member
Joined
Aug 14, 2002
Messages
5,001
25 - 40% less flux tells me I can still get good image return from a target 30 feet away in a dark environment even in competition from more traditional low ambient light.

Looks like some really good stuff in that oh-so-important 620nm to 660nm range.

I'm going to guess that a pediatrician is going to like this for throat exams.

I'm also going to guess that this has potential as a blood tracker.

Really exciting progress shown here for solid state illumination.

I'm full goose bozo interested in where you take this, collimated or not. I'll be following along. Thanks for bringing this to us.
 

starfiretoo

Well-known member
Joined
Oct 13, 2006
Messages
1,265
Interesting conversation. I've been a fan of multi-emitters of different white tints to get a "full" spectrum of white. After the brightness factor wears off color rendering becomes increasingly important. Thanks for leading us flock in that direction.
 

McGizmo

Well-known member
Joined
May 1, 2002
Messages
17,274
Big Honu,
As I understand it, the sun can produce light at 8500K and still be considered good at color rendition. It seems with the LED at least so far, it is difficult for them to be considered good on CRI and get above 5000K. And what defines good CRI? Some minimum of the CRIa as I have seen mentioned before? :shrug:

There is so much I find confusing and the more I try to delve into this, the more confused I become. I have seen with my own eyes that there is need for better spectral output than some of these LED's give us if we want to do well in distinguishing certain colors but I don't think there are any easy numbers to find that tell us what we want.

It would be great if some experts chimed in but I can imagine it could be daunting to try to play to our ignorance.

Here's one I have been pondering for example. Lets say you have an object that absorbs most photons except for a narrow band of blue and a narrow band of red. (I don't know if this is even a reasonable assumption). Such an object I believe would appear to be purple to us. However, if we were to match the color of purple with a narrow band light source and shine it on this object, the light would be absorbed and the object would appear black or dark and yet if we were to shine a pair of lights on it that matched its reflective bands, it would appear to be purple. If we shined only the specific blue, it would appear to be blue and if we shined only the specific red, it would appear to be red as in both cases, light would be reflected back. If this example is possible, what color would you say the object is? Purple? Red/Blue?

I noticed in the software for my spectrometer, there are options in some setups for choosing reflective or emissive. With a flashlight, within reason, we can control what we send out. We can't control what gets bounced off the real world out there and comes back for us to see.

With a RGB source, we can make the source look "white" but this is not full spectrum and I would guess that color rendition may suffer as a result. My assumption is that ideally, we would have a light source with similar relative spectral output to that of the sun. When our source falls short of this mark, I am not sure what the difference would be or how significant it would be. We clearly not deficiencies in some of our sources of light but what would it actually take to make them acceptable? I have simply added a red in a cluster of whites and found the results acceptable with no obvious short coming. Maybe I haven't looked hard enough or maybe it's right before my eyes and yet I can't see. I am convinced that we all don't see the same way with the same levels of acuity and discrimination.

In regards to collimation of these high CRI LED's, I put a couple of the Seouls behind some reflectors and I'll tell you one thing, they won't win any white wall contests!! :green: Think early Luxeons with warm and bluish tints both in the beam and obvious in contrast to each other. I found with one Seoul LED behind a McR-20S that by bead blasting the inside of the sapphire window to blur the image, the results were certainly acceptable in real world. Again, on the white wall not very pretty.

My thinking, after playing with the Nichias and actually seeing a difference, was that I wanted to go for max CRI if I was going to do a light based on High CRI. Although some of the other LED manufacturers do bin by high CRI, I got the impression and this even from some of the competition, that Nichia has the edge at present when it comes to high CRI. Everyone takes a hit on flux when they go for CRI and that helps level the playing field as well or so it seems. I personally feel that the flood beam is where color rendition is likely most important and the 6 dice of the Nichia are a non issue in a flood type beam.

The subject of this thread mentions the consideration of significance and it is not because I have any real feel for the significance! It is my hope that as a group, we might be able to put some dimension to any significance?!?
 

BigHonu

Well-known member
Joined
Apr 1, 2002
Messages
1,242
As far as the significance of a high CRI handheld lighting device, I'd say that from a practical standpoint...not particularly significant, especially when output (lumens) is relatively low. I'd imagine that tasks requiring good color rendition at close ranges already have lighting options more suited to the task, even for 'in the field' work.

High output needs may benefit more as that that blue and red toy gun in someones hand at 100' may look dark and 'real' with a light with a bad CRI, but easily recognized as a blue and red toy gun with a light with a high CRI.


From a hobbyist standpoint, I'd say it is VERY significant. We have gone beyond just wanting to see. We want to hold the sun in our hands and be able to apply as little or as much as needed. Any sort of development or application that gets us closer to that, is a step in the right direction in my book. Damn the practical thinkers!
 

starfiretoo

Well-known member
Joined
Oct 13, 2006
Messages
1,265
I've always thought about replacing the incandescent bulb in my stereo microscope with an LED. So far none of my LED flashlights tints seemed right. Sounds like this unobtanium Nichia with an unobtainium SF Titan converter sounds like the trick:thumbsup:

Was wondering how the color rendering of the Nichia changed for the better or worse by over or under driving it?
 

Icebreak

Well-known member
Joined
Aug 14, 2002
Messages
5,001
I had these thoughts before I saw your actual Sundrop thread. I knew they would marginal in their contribution but now it appears they may be even less so.

For collimating maybe a consideration of duel strategy used in tandem. IOW's a reflector and a clear optic.

This reflector uses concentric waves to broadcast many instances of many rays from many points.

IMG_0919.jpg


This optic further collimates the rays.

IMG_0917.jpg


This Lamina Ceramics is a 7 package array with each package containing 6 die.

IMG_0920.jpg


This is a white wall of the array with no collimation.

IMG_0916.jpg



Then collimated with the wavey reflector.

IMG_0915.jpg



With the optic uninstalled but at the focal points.

IMG_0914.jpg


Then finally the optic defocused but installed.

IMG_0926.jpg


This is from CentraLED's worklight. Bob Pond and Steve Offiler are CPF members, Soffiller and bpond. Bob is the engineer and with great effort on his part and with the help of optical engineers came up with this system.

The concept seems sound. Of course, I'm not suggesting this exact optical engineering set-up but maybe some elements would be viable. I think the waves in the reflector may outperform a light orange peel when used in tandem with a defocused clear optic.

I'm wondering what your Nichia high CRI emitter would do in a SureFire Invictus.

The significance of collimation for your design is that it may help medical professionals targeting the throat while wanting to keep the light out of the patient's eyes.

The significance of High CRI LED illumination can be explained by my new little buddy I met last weekend.

IMG_0590.jpg


I'm more likely to know what he is by the red/orange pattern on his skin than by the shape of his head or the depth of his eyes in their sockets. He's a man of few words and communicates with color and behavior.
 

yaesumofo

Well-known member
Joined
Jul 21, 2003
Messages
3,698
Don this question got missed elsewhere so I will try again here.
I was wondering if the emitter in the Light engine which you so generously sold me which contains an 083 emitter if that emitter is one of the hi CRI emitters which are to be found in the sundrop?
It is the best LE I have for use as a mule. I can't use it for anything but the mule. I am very happy with it. So is it a hi CRI 083?
Thank you.
Yaesumofo
 
Top