RGB LED's for more incan-ish light?

[JRTJRT]

Is it possible to group red green and blue LED's together to produce fuller spectrum incan type lighting characteristics?

TV's and notebooks use RGB to produce white light...

How about a 3 led flashlight but with red green and blue led's. Or a monster 158 LED showerhead or RGB's? What kind of light would that produce?

 

[SemiMan]

The "quality" of the light it would produce would be no better and probably worse than using current generation white LED. You would be better off using a current generation white and adding amber and red to get the middle and long wavelengths back.

TVs look great using LED backlight because you are directly viewing the light source and which can "simulate" any color needed. However, once you bounce that light off something else, the conditions change completely and you need a broad band light source to be like incandescent.

Semiman

 

[wakibaki]

I don't know of any reliable reports on how this works in practise, but there is good reason to think that RGB leds will provide a more satisfactory colour rendition than the current blue+yellow 'white' ones.

The reason blue+yellow is so predominant is that green & red leds lag behind the efficiency of blue leds and even blue leds + a yellow phosphor. Plus the difficulty of making the outputs concentric. The output from the yellow phosphor is more broadband than that of an led, but the output from RGB leds can synthesize a wider colour gamut under the right conditions than blue+yellow under any conditions.

Given that there are a wide variety of entities in nature that light may impinge on (transparents, pigments, fluorescents, reflectors, interferers) the RBG approach is at least as likely (and in my opinion more likely) to ultimately produce 'good colour rendition' as blue+yellow.

When the RGB rebel stars become readily available in the UK I shall be doing some experiments with them...

What would be really good would be a bunch of leds with colours from blue to red with their centre frequencies spaced so that the high and low frequencies of each overlap permitting the synthesis of a flat or fairly flat continuous spectrum.

w

 

[sORe-EyEz]

if not changing d bulb can try using photographic filters to alter its tint. my Pelican M6 LED 2390 & 2330 tint did have a slight shift in color tint from slight bluish to a warmer tint.

hint: the higher d intensity of d LED the darker shade of the filter was used.

Cokin plastic (dunno wad kind of) photographic filter 81B on M6 2330, 81D on M6 2390.

not sure if i am of help...

 

[sysadmn]

Is it possible to group red green and blue LED's together to produce fuller spectrum incan type lighting characteristics?

TV's and notebooks use RGB to produce white light...

How about a 3 led flashlight but with red green and blue led's. Or a monster 158 LED showerhead or RGB's? What kind of light would that produce?

Yes, it's possible, and there are fixed lighting fixtures (MR-11, MR-16) that use this principle. The problem is that you can get 100+ lumen white leds cheaply, but R, G, and B leds run ~40-70 lumens or so. And since white is the high volume product, it tends to be cheaper than a colored led - much less three of them. THe electronics would be more complex also, since you'd need to tune drive currents to account for variation in individual led performance and different color led efficiency.

I'd also guess that an RGB solution would be more flood than throw, but that's mostly because the products I know of are for track-lighting type applications. For those, you want good light 8-10 feet away, not 50 yds away.

 

[JRTJRT]

Thanks guys. I'd like to tinker with this stuff but I don't have the cash right now to build and experiment as much as I'd like to. If you can't learn from experience, learn from others... :thumbsup:

 

[Luminescent]

Yes, it's possible, and there are fixed lighting fixtures (MR-11, MR-16) that use this principle. The problem is that you can get 100+ lumen white leds cheaply, but R, G, and B leds run ~40-70 lumens or so. And since white is the high volume product, it tends to be cheaper than a colored led - much less three of them. THe electronics would be more complex also, since you'd need to tune drive currents to account for variation in individual led performance and different color led efficiency.

I'd also guess that an RGB solution would be more flood than throw, but that's mostly because the products I know of are for track-lighting type applications. For those, you want good light 8-10 feet away, not 50 yds away.

I was surprised to find that all of the issues you mentioned have been resolved.

It is hard to combine the beams without screwing up the light pattern, and you do indeed need a sophisticated driver, but apparently some DLP projectors that used to use a spinning dichroic color wheel now use tri-color LED optics.

This company has some RGB combiner optics for the tiny Rebel LEDs

http://www.polymer-optics.co.uk/products LED optics luxeon.html

(It's a british company, so don't expect the documantation to be Intelligible)

Your comment about the need for a sophisticated driver is right on target so Linear Technology and others have custom drivers for this application.

http://www.linear.com/pc/productDetail.jsp?navId=H0,C1,C1003,C1094,P37747


Three 70 lumen outputs would total 210 lumens. It would probably be even less after you got the color balance correct, but this is quite respectable.

I agree that you could get much better color rendering with this type of light, and as a bonus you could have a full spectrum 'Rainbow' mode capability with the capability to generate fully saturated color light in any color desired.

Cool!

 

[JRTJRT]

"Full spectrum 'Rainbow' mode"


Ok now I just might sell my car for one of those.

:twothumbs:huh::thumbsup::bow:oo:

 

[brickbat]

Is it possible to group red green and blue LED's together to produce fuller spectrum incan type lighting characteristics?

It certainly can be done. The big advantage is you can control the color temperature of the source. But, the spectrum isn't 'fuller'. In fact, it'll be composed of three peaks; red, green, and blue; with very little between. I don't know what the best CRI might be, but have a hunch it's less than that available from 'white' LEDs. I've experimented it with a while back (with some Lamina Ceramics arrays), and the CRI was, to my eye, pretty horrible.

 

[brickbat]

...TVs look great using LED backlight because you are directly viewing the light source and which can "simulate" any color needed. However, once you bounce that light off something else, the conditions change completely and you need a broad band light source to be like incandescent...

Very well put, and that corelates exactly with my findings. I was able to easily dial the color temp of the LED array to whatever color temp I wanted. And when viewing the source directly, its color was undistinguishable from a high CRI 4100K fluorescent, or incandescent source. But when I used the array to illuminate objects, the colors were off. Skin looked magenta...

 

[shakeylegs]

Is it possible to group red green and blue LED's together to produce fuller spectrum incan type lighting characteristics?

TV's and notebooks use RGB to produce white light...

How about a 3 led flashlight but with red green and blue led's. Or a monster 158 LED showerhead or RGB's? What kind of light would that produce?

Perhaps LITEMania can answer some of these questions. Look what he has here:
http://www.litemania.com/front/php/product.php?product_no=579&main_cate_no=1&display_group=2

 

[Luminescent]

Very well put, and that corelates exactly with my findings. I was able to easily dial the color temp of the LED array to whatever color temp I wanted. And when viewing the source directly, its color was undistinguishable from a high CRI 4100K fluorescent, or incandescent source. But when I used the array to illuminate objects, the colors were off. Skin looked magenta...

Good points. Natural objects do reflect very complex spectrums.

I am not sure that I agree that the situation is as bad as your quick testing may have led you to believe, because by your logic, 3 chip color cameras which use 3 separate CCD's and narrow band dichroic filters would all render totally unacceptable color scenes, but they seem to do an outstanding job in the real world.

The eye has it's own limitations in this regard, due to the the tri-stimulus characteristics of human color vision, so I suspect you could improve things by choosing LED's with optimal center wavelengths and spectral widths.

Did I miss something? Is there some difference between three narrow band LED's bouncing off a surface, vs. a broadband white light source being bounced off the same surface, then narrow band filtered by equivalently narrow band dichroic filters in a 3 chip cameras?

In a pinch you could always use a variation of the CYMK trick used in printing, by including a high color rendering index white LED along with the Red, Blue, and Green chips.

In this configuration, highly saturated colors would be rendered by the R, G, and B drivers alone, and less saturated colors would be mixed with more and more 'white' light from the white LED as the color saturation decreased.

So there would be no compromise, even if the issues you raised do cause problems.

 

[wakibaki]

color cameras use 3 separate CCD's.

Nicely reasoned.

w

 

[sysadmn]

I was surprised to find that all of the issues you mentioned have been resolved.

Cool!

I wasn't
I tried to find an earlier thread were this was discussed. I had one of those links in it, and another supplier out of LA. It's not that the problems are insurmountable, just that the solutions are not cheap, and not ready for flashlight use.

 

[hank]

The Litemania chip looks like it might be very nice for a flashlight.

If I could read the language I'd do better -- anyone got a pointer to that in English? Sorry.

 

[brickbat]

...Is there some difference between three narrow band LED's bouncing off a surface, vs. a broadband white light source being bounced off the same surface, then narrow band filtered by equivalently narrow band dichroic filters in a 3 chip cameras?...

If the transfer function of the dichroic filter matched the LED's spectral power density function, then I see no difference. I'm no expert on filters used in imagers, but I do know that efficient, commercially available LEDs are not available with arbitrary wavelengths and spectral half-widths.

 

[Luminescent]

If the transfer function of the dichroic filter matched the LED's spectral power density function, then I see no difference. I'm no expert on filters used in imagers, but I do know that efficient, commercially available LEDs are not available with arbitrary wavelengths and spectral half-widths.

Amazing the kinds of things that can happen when driven by a mass-market application.

No one these days even thinks about the man-years of development that went into developing efficient properly color balanced phosphors for color televisions 40 odd years ago, but it happened.

It's pretty clear that the RGB LED drivers and optics that have already been developed are for portable DLP projectors and other high quality RGB display applications, so more than likely those optimizations are already ongoing.

With the same mass-market requirements driving development, I am sure that matched RGB LED's with great color balance and efficiency will be along shortly, if they are not here already.