Odd Lux V idea

Correct me if I'm wrong, but one Lux V die is essentially four Lux I dies crammed together, right? Could Lumileds cram together four dies of different colors, say; red, blue, green, and yellow, to get better color rendition? Or would the electronics be too tricky? Either way, discuss.

KeyGrip said:

Correct me if I'm wrong, but one Lux V die is essentially four Lux I dies crammed together, right? Could Lumileds cram together four dies of different colors, say; red, blue, green, and yellow, to get better color rendition? Or would the electronics be too tricky? Either way, discuss.

Click to expand...

Each LED color has slightly different Vf requirements -it'd be very tricky to pull it off in the current small packages.

plus each die is in a different place and not directly in the middle of a reflector....would the colours be mixed by it properly?

Stu

RustyKnee said:

plus each die is in a different place

Click to expand...

Not only that, but the die themselves are shaped differently.

I second what greenLED and RusyKnee said. The emitters would not line up in the reflector so you would have a multicolored spot instead of a white spot. To get white you would need a diffuser first; far enough away from the LED to get good mixing. Then you would need a much larger reflector because of the large source size.

You would need to match the colors electronically. If you are going to do this you might as well have a microcontroller so you could set the color temp or get special colors.

I think you would be better off with separate emitters for each color.

Okay, so as of now, it's not really feasable. How would you make it work?

They make tri-LED holders/lenses that are mainly intended for red/green/blue mixing. The whole thing is only about 2" across.

I am working on a two color (amber/cyan) light. If you look at the 1931 CIE color chart, the outside edge represents monochromatic lights. A 1-color LED has a pretty narrow range so it would be just a little inside the edge. Draw a line from one point on the edge (one color) to another point on the edge, and you can mix those colors to get any color along the line.

Usually they use 3 colors to make a triangle. Then you can mix them to get any color within the triangle. Since my goal is to basically follow the blackbody curve, I hope to do that with only 2 colors.

OhMyGosh: Thanks for the chart, that is some really cool stuff. Since so many "Incan vs LED" threads have been popping up, I've been curious about ways for LEDs to more effectively render colors. Also, what are "black bodies"? I've heard the term used several times, but I've never heard a good explanation of what they are.

IIRC, "black body" is used to describe a surface (or an object, I guess) that absorbs all incident radiation. Not sure how that ties in with OMG's plans, though.

By blackbody I just mean something thats puts out light by getting really really hot Like a tungsten wire. Or the sun. If you dim a 'wire' light it gets more yellow - orange - red. It is actually getting cooler but we call the color change warmer - go figure. It tends to follow a curve on the CIE chart as it cools. However 'blackbody' only truely applies to something black (doh) like carbon.

I finally found a graph of color temperature here.

I am really only interested in 'warming up' some white LEDs that are to cool (blue). From the chart it looks like ~585nm would be best for 3000k - 6000k. However I have read about blind tests vs halogen that say 640nm works best for improving the color of white LED's. Time to do some testing.

OhMyGosh said:

I am working on a two color (amber/cyan) light. If you look at the 1931 CIE color chart, the outside edge represents monochromatic lights. A 1-color LED has a pretty narrow range so it would be just a little inside the edge. Draw a line from one point on the edge (one color) to another point on the edge, and you can mix those colors to get any color along the line.

Usually they use 3 colors to make a triangle. Then you can mix them to get any color within the triangle. Since my goal is to basically follow the blackbody curve, I hope to do that with only 2 colors.

Click to expand...

I assume you aren't concerned with CRI then? A 2-color light can make "white", but the fact that there are only 2 colors effectively means that the chromatic range seen will approximate what a color-blind persons sees.

I've been trying to do the same thing myself, but the problem for a red-cyan setup is that the left side of the CIE curve is much more spaced out than the right side; the slightest variation in nanometers can turn a two-color red/cyan light from yellowish to magenta. I'm not aware of any LED retailer who can deliver a diode within one or two nm.

A blue-yellow might be easier, using a blue (470nm) or royal blue (450nm), but now you need a yellow emitter in the 570nm range or so; amber Lux I's and III's top out around 589nm. I've never seen any Lux or other such emitter in wavelengths like that.

Your luck might improve in 5mm, but I don't know where small quantities at specific wavelengths can be had. (I would love to know, though, as I've got a bunch of 1st gen blue Inova X1's waiting for mods.)

As for blackbody, it's not a straight line through the CIE space, so I doubt you'll be able to get it with two monochrome sources.

Ah, just read your last post. Yes, 640nm or so is probably what you need, if you are working from high color temp whites. Check The LED Museum site, he's got all kinds of spectrographs for various different emitters that might give you an idea of where to bolster the spectrum of your sources.

I think you're right and I'm wrong :laughing: A two-mono color light would look good on a white wall, but when you shine it on other real colors it would look ... weird. Since the white LEDs have a phosphor that puts out a 'blackbody type' range of colors, I am better off using white LED's for the bulk of the light with some RYO (red-yellow-orange) added to adjust the color temp. I am wondering if one color is sufficient or if I need two colors like 580nm + 640nm to get the 'incandescent simulator' effect I am after.

PS: one main reason for this is that the 'coolest color' white LED's also tend to be the brightest. I would like to clean up the color without adding a filter that reduces the brightness.

This is not only cool, this is the direction I hoped this thread would follow. Please, continue your discussion of testing here!