Light guides for combining multiple LEDs to a point source?

[KreAture]

I have been playing with diff torches and reflectors using 1-9 LEDs.
My biggest issue with multiple led is they are harder to focus.

For immense throwing distance I have been replacing the glass on some torches with lenses colaminating the beam into a 0.5-1 degree spread max. This gives incredible throwing distance but works best on a point source. Using multiple leds tend to create multiple blooms or hotspots at the target.

So, I thought, why not use lightguides!
Some sort of christmastree of 8-10mm lexxan-like material curved and melded together from each led towards a common output. Kind of BIG optical fibre. The output of the fibre could then be focused by a twist-adjusted lense at the tip for custom focus like some maglites have.

Waddya think? Anyone have equipment to try this?

 

[saabluster]

I had the same idea. If you look hard enough you will find some information on the internet but not much. I've been working on this for about 5 months and have yet to get it to work very well. I know its possible but it will just take time to develop the right technique. If I can make it work it could be the one thing that could beat LPS for throw. It will be awhile though. Its all about surface brightness.

 

[Duglite]

I'm really interested in this too. I'd like to be able to combine several LEDs into a single point source, and I came to the conclusion that light guides as described by KreAture would be the best way to do this. Unfortunately, it's beyond my technical ability, and an off-the-shelf device is probably waaay expensive. I'd love to know if anyone makes any progress on this.

 

[KreAture]

OOOOh!
Look what I found!
http://spie.org/x8762.xml

Now this is exactly what I wanted to see.
It's possible, feasible and possibly not too hard to build/mold.

 

[saabluster]

OOOOh!
Look what I found!
http://spie.org/x8762.xml

Now this is exactly what I wanted to see.
It's possible, feasible and possibly not too hard to build/mold.

That is a GREAT find. Only one problem. Where do we buy it? This is the problem I had. I found some neat solutions out there and they are either WAY too expensive or not even able to be bought.

 

[LukeA]

You could put each LED at one focus of an elliptical (not parabolic) reflector, and angle the LEDs in reflectors so that the other focus of each ellipse is common and also the focal point of the lens. You will lose a good bit of light as spill from the reflectors, but you'll end up with much more light passing through a single point than by most other methods.

 

[Duglite]

OOOOh!
Look what I found!
http://spie.org/x8762.xml

I never bookmarked an article so fast in my life!:goodjob:

 

[DM51]

Very good find indeed.

This is obviously still pretty much at the research stage, but it isn't too hard to imagine that in only a very few years, this idea will have combined with increasing LED efficiency (= decreasing heat problems) and fiber-optic internal-reflection technology, to the stage where the output from many LEDs could be concentrated into a single tiny point.

This experiment has already worked with 3 LEDs. There is no reason to suppose that it could not in due course be done by 30 LEDs, or 300, as they won't need to be mounted side by side as in this experiment – they could be widely dispersed, with fiber-optics bringing the light in to the single point.

The point source is a very significant part of this – the smaller the point, the better it can be focused.

If this happened with 30 LEDs emitting even only 200 lm each, you already have a point source of 6,000 lm. With 300 LEDs, it would be 60,000 lm…

 

[KreAture]

Exactly DM51. There is no rules as to topology.
Injecting another light source into a fibre/lightguide is not too hard.

LukeA, using the technique you describe will let light pass through a point, but not originate from same point with any degree of colmination. The lightsource will still be big. Trying to focus it to a cominated tiny beam will be very hard.

Next time there is fog here I'll take some beamshots of my cominated 1w Cree. You'll love it. The beam is like a intense light-rod inside a dimmer tube, continuing on in the distance untill it is diffused by the fog.

I have access to some lexxan material in rods and will see if I can fuse them into some beam-combining contraption.

 

[DM51]

Next time there is fog here I'll take some beamshots of my cominated 1w Cree. You'll love it. The beam is like a intense light-rod inside a dimmer tube, continuing on in the distance untill it is diffused by the fog.

Sounds excellent. What lens are you using? Have you seen the other threads on aspheric lenses?

 

[KreAture]

I haver this huuuge 4" dia lense that I use to see what effects I can get with lensing the torch. Then I look to see what f number I will need. If I am lucky I can just pick up a pair of cheap circular glasses at bestbuy or similar for $2 to fit instead of the torch original glass. (plastic lenses, easy to trim...)

I also found my local ee shop has lightguides!
http://www.elfa.se/images/highres/h10859.jpg
They come in strips of 10 and cost $1 a strip. 3mm round tips, 15, 30 or 45mm long.

 

[DM51]

If you have a very small point source of light, the best result will be achieved by putting the lens at a distance from it equal to its focal length.

You can get a good idea of the focal length of a lens on a sunny day by measuring the distance from the lens at which the sun focuses into a point having passed through it.

I'm very much looking forward to hearing about your further experiments with light-guides.

 

[KreAture]

Jusr remembered, I havce some coctail sticks here!
They are crystal clear plastic with polished edges and triangular.

One end has sides 5x5x5mm while the other has 1x1x1mm and they are 90mm long. I bet those could be made to collect light from multiple LEDs and simply merge to a point.

 

[saabluster]

Jusr remembered, I havce some coctail sticks here!
They are crystal clear plastic with polished edges and triangular.

One end has sides 5x5x5mm while the other has 1x1x1mm and they are 90mm long. I bet those could be made to collect light from multiple LEDs and simply merge to a point.

I highly doubt the plastic used would be of a high enough quality to work for light transmission. The plastics used for optics is not your everyday clear plastic.

 

[DM51]

The main problem you will have is heat - it will melt the plastic. LEDs get very hot, so you will need glass.

Another point is that the ideal fiber-optic configuration is a thin cylindrical tube. A tube with a triangular cross-section will be only half as efficient at transmitting light, as there are 4 escape surfaces rather than just two. The amount of light that will escape from each surface is equal. Only a quarter of the light will reach the far end of the stick, rather than half of it.

 

[frenzee]

This is a fascinating subject. I looked at LPI's patent copy and it looks like the key to their design is an "SMS" or Simultaneous Multiple Surface optic which looks like a combination fresnel lens on one side and an asymetrical refractor type lens on the other side. Very interesting, but I still haven't figured out how it's supposed to work, but if it does work as they claim, which sounds like it does, this would be holy grail of RGB lighting and flashlight optics, not to mention automotive lighting, solar lighting, etc. I wonder why we haven't heard more about them. (I wish Google's Sci/Tech section would cover more real science like this instead of game consoles and social networking sites...grrrr). Thanks KreAture for digging this stuff up.

 

[saabluster]

The main problem you will have is heat - it will melt the plastic. LEDs get very hot, so you will need glass.

You don't need glass. There are tons of plastic optics and plastic reflectors used with LEDs.

Another point is that the ideal fiber-optic configuration is a thin cylindrical tube. A tube with a triangular cross-section will be only half as efficient at transmitting light, as there are 4 escape surfaces rather than just two. The amount of light that will escape from each surface is equal. Only a quarter of the light will reach the far end of the stick, rather than half of it.

"escape surfaces" Where did you get this idea? What we are discussing would be TIR optics.

 

[DM51]

You don't need glass. There are tons of plastic optics and plastic reflectors used with LEDs.

In this case the material would need to be in very close proximity to the LED surface, if not in actual contact with it. I don't think there is a plastic material that could withstand the heat involved.

"escape surfaces" Where did you get this idea?

I think you may need to read up on how fiber-optics work.

What we are discussing would be TIR optics.

In fact we were talking about fiber-optics. Not quite the same thing.

 

[65535]

I don't know about glass, but most plastics soften and warp around 80C so majority of them are out of the question.

 

[saabluster]

In this case the material would need to be in very close proximity to the LED surface, if not in actual contact with it. I don't think there is a plastic material that could withstand the heat involved.

Close yes. In contact no. You would ruin the led. You would need to use an optical gel in between the LED and optic. I do know there are plastics that can handle the heat. I work with plastics all the time prototyping and molding my own parts.

I think you may need to read up on how fiber-optics work.

I know exactly how they work. You still did not answer the question.

In fact we were talking about fiber-optics. Not quite the same thing.

Fiber optics use TIR. The 3 LED combiner we were talking about is a TIR optic. It is made from plastic.