Color Mixing With LEDs?

I made a first generation multi color LED flashlight. ("Torch" for our viewers abroad) It will light up a tent in any color you want. The various LEDs are clumped in groups across the face of the light. This means that when two or more colors are on they are a little offset from each other, so you get psychedelic colored fringes around the edges of things because each side will have a little more of one color than the other.

I'm thinking of various ways to make a second gen. unit and I'd like to try for smoother color mixing. The ways to do this that occur to me are:

1.) have the LEDs shine sideways or backwards onto a reflector (possibly with a pebbled finish) and try to get the reflector to mix the light.

2.) Have a hazed or frosted or faceted lens in front of the LEDs and try to get it to diffuse the light more evenly.

3.) have each color LED scattered evenly across the face of the unit (instead of each color in its own cluster)

What have people here seen or tried? What works well? Comments?

Ah, you seek the holy grail RGB design. GOOD LUCK. As far as for back lighting with multiple LED's, arrange them in a pattern, like a CRT. A light frosted cover will help blend the colors. I have a Prolight 3W RGB (3 1w dies in 1 package), it works GREAT for flood lighting, but when focused, it leaves alot to be desired. Even with a C m@g lens painted white with a dimpled finish it is still blotchy.

a setup like this would work well

R-G-B-R-G-B
G-B-R-G-B-R
B-R-G-B-R-G

although it will be a nightmare to wire up

Jeff

Quite a challenge. I think Jeff is correct and the LEDs have to be clustered to mimic the pixels in an LED display or the cells in a plasma display.

You did not say how big your first light was, or how big you are thinking to make your next one.

One of the big obstacles is that by the time you insert even 3 clusters of 3 or 6 LEDs, you have used up a lot of area and seriously compromised the reflector.

This effect is noticeable in the SF A2 or SL twintask with only 3 single color LEDs.

Would it be possible to use one group of 1W RGB LEDs? How about mounting them on a small post through the center of the reflector and fire them "backwards". A polished reflector would maintain its integrity and an LDF lens might smooth the beam out. My assumption here is that you are after smooth color(s) flood with throw being less important.

Also this would avoid the difficulty of trying to wire in groups many small LEDs.

Mark

[ QUOTE ]
Ken_McE said: when two or more colors are on they are a little offset from each other, so you get psychedelic colored fringes around the edges of things because each side will have a little more of one color than the other.

[/ QUOTE ]
I play around with multi-colored LED lights - mainly a multi-color LED pen and the fun Sauce color changing Mini-LightWand.

I did a sort of ad-hoc and no claims to be scientific experiment to see the effects of mixing LED colors.

LED Colors and Vision (pics)

The main problem was that the LEDs were not emittling their light from the same point - therefore color fringing and bias could easily be detected.

A true RGB LED probably would be "better" from the sheer fact that the emitters are physically closer together - but I think they still will suffer from color fringing and bias problems.

>You did not say how big your first light was, or how
>big you are thinking to make your next one.

The case is steel, 4.5 inches by 4.5 inches, by about
2 inches deep. It was designed around the batteries.
This size is just right for three "D" cells and a circuit
board. I haven't picked a size for the next although I'm
considering a two "AA" battery model. Two AA batteries
fit nicely in an Altoids tin.

>One of the big obstacles is that by the time you insert
>even 3 clusters of 3 or 6 LEDs, you have used up a lot
>of area and seriously compromised the reflector.

On the first one there is no reflector. It is permanently
set on "flood"

>Would it be possible to use one group of 1W RGB LEDs?

I've been looking at using 5 or 10 mm Leds in an Altoids
tin with this arrangement. Turns out I need to order
more 5 and 10 mm leds.

>How about mounting them on a small post through
>the center of the reflector and fire them "backwards".

The problem with doing this with one watt leds is that
the one watters are supposed to have a heatsink in
back of each of them. I don't know if I could house
them in baby oil or vodka or something and let this
carry the heat away from them.

>A polished reflector would maintain its integrity and an
>LDF lens might smooth the beam out.

What's LDF?

>My assumption here is that you are after smooth color(s)
>flood with throw being less important.

Actually I'm after entertainment.

>Also this would avoid the difficulty of trying to wire
>in groups many small LEDs.

I wound up simplifying my first light because it was
getting to be a nightmare of little wires.

Hi Ken,
Wildrices' comment on the holy grail is quite correct.....you've embarked on a great quest, whose path is scattered with many a collection of failed RGB clusters...good luck and please share any innovative ideas that get good results.

From my own observations, it's really hard. I think the things to note are:
- individual R,G and B sources should be as close as possible together
- alternate RGB sequence around the edges in particular. I've done a bit of simulation with ray tracing to come to that conclusion. In any beam, it is the edges with the greatest angular divergence, hence it is the most in need of RGB sequence. Wildrices setup looks good.
- Single RGB modules, particularly with a tight beam will cause more headache than good. The lens will diverge, not converge from a point. You might try slicing off the lens, putting a light diffuse surface (eg sandpaper) and putting a seperate lens (or reflector) from the surface of the diffuser.
- Diffusers eat up light.....it's an unavoidable tradeoff. However may be the only practical solution. Check out http://www.lrc.rpi.edu/programs/solidstate/completedProjects.asp?ID=48 as a start. You'll probably end up as you've predicted, some combination of diffuser and focussing optics....and balancing the problems of both.
- I'd be interested in your results with a dappled reflector. My guess is that you'll end up with a big collection of psychadelic colours, pretty but may not be what you want.
- have a look at the lamina or other modules with the seperate RGB dies already placed close together. I've had some correspondence with them on the problem, they are positioning their dies as best as possible...but there will always be some fringing....and again, its a wide beam solution only.
- try SMD LEDs....you can cluster them pretty close together.
My wish list:
- OLEDs with RGB layers stacked vertically, obviating the problem....may be a few years before we get stable high power OLEDS
- some sort of frequency dependant holographic/diffraction film that bends and mixes nicely....I suppose there'll be a light reduction hit there as well

In short, tight beam RGB mixing is going to prove very hard, unless you're willing to sacrifice some light output and focus your optics on a diffuser. Wide beam is less of a problem (I just use 120degree superflux (spider) LEDs).


Phil

Light Shaping Diffuser:
http://www.poc.com/

Read more about them:
http://www.ledsmagazine.com/articles/features/1/10/2

Much lower cost holographic and other options:
http://www.wft.bz

Other products:
http://www.astraproducts.com/index.htm

3M Vikuiti

Two things that come to mind in regards to the above:

1.) It sounds like I would get better results using smaller LEDs.

2.) it might be interesting to have fiber optic light pipes collect up the light from multiple leds and lead it all to a common outlet. This way you could spread the emitters out for cooling and easier soldering.

Two sub-questions that spring up from this are how effeciently can you "funnel" light into a fiber optic line, and will the light mix evenly in the course of traveling along the fiber?

well regarding using optical fibers...check the link
http://www.lrc.rpi.edu/programs/solidstate/completedProjects.asp?ID=48
mentioned above, their research albeit on a light pipe ( a fat fiber) sort of shows no real benefit.
...Those links you put up there Newbie look sort of interesting.....I'll have to have a read.

Incidentally, all this talk of diffuser films and such got me to disecting an old Palm Zire 71 that had a rough time in a toolbox, cracking the screen. It consisted of a diffuse film (like tracing paper, nothing special) bonded to a very nice diffraction grating which appeared to act as a reflector. This was behind a thin plastic layer that had the backlight LEDs glued on it (4 of them). This was interesting, between the LEDs was a 'spray' of little dots, presumably to even out the light. Then going to the top of the plastic layer was another diffraction grating, that gradually increased in line density. Over this was the cracked LCD/polerizer. (Note:I may have got the order of the diffraction grating behind the plastic LED thingy wrong...I was in a bit of a rush). When I reassembled the thing, minus the LCD/polerizer, it gives a nice bright even square of light from only four small LEDs mounted at the bottom.
Playing around with a laser and diffraction gratings is fun as well....though I do miss the Zire71.
Anyway, I think there is something in the use of diffraction gratings/holographic films that should be looked into. I'm hoping to get some samples from www.poc.com and will write on the results if/when they arrive.
phil

Are you certain they were diffraction gratings, or were they 3M BEF films?

rgbphil, Send me your addy by PM, I'll send you some BEF and other stuff...

well the thin sheet bonded to diffuser I'm pretty sure is a diffraction grating....I put a slit of sunlight in, and out pops a rainbow. It only works in one direction as well. Also a laser gives a nice seperation. On closer inspection, the plastic with LEDs bonded to it looks more like a dot grid though, with dots getting closer further away from the LEDs...
I've PM'd my email....I'd be interested in the BEM stuff.

ok....looked a little closer at the BEM datasheets and it looks like it might be a Brightness Enhancement Film...or in otherwords a film of little prisms. From the datasheets, it also appears the correct usage is the diffuser/film on top of the LED backlight, not behind it. Still gives a nice rainbow effect though. I wonder how you'd be able to tell the difference between a true diffraction grating and a prismatic film?