How Would You Create the NARROWEST Beam?

[Duglite]

What combination of LED and optics would you use if you wanted the narrowest beam possible from a given LED? I know it wouldn't be everybody's idea of an optimum flashlight, but I want a tightly collimated beam to squeeze every last drop of light into a small bundle of optical fibers (or a single large fiber ~6mm diameter.) Little or no spill. A lousy light for trying to find your keys on a dark beach, but maybe a spot your cat could chase.

-It's not about throw, but it probably would throw pretty well.

-I'm not worried about how pretty the spot is. Should be small though, let's say, at between zero and 2cm distance, a spot size of 3mm-6mm in diameter. The spot quality on a ceiling won't matter, because I'm grabbing it millimeters from its source.

-I'm partial to Luxeon because of the color selection, like red-orange, blue, green, cyan, etc., but I know others are making colored LEDS. Does another brand have more bang for my buck?

I've read interesting things about acrylic balls on this forum. Don't have any in my possession yet. Could these help me?

Battery life irrelevant. Will be powered by AC/DC converter. I'm not constrained by a flashlight-ish form factor.

I've tried some polycarbonate lenses, such as a 3-degree fiber-coupling lens from Quickar, and a funny-shaped little Fraen Fiber Light Injector model FFLI-07-LL-0. I'm looking for even more performance than these give. I've also put the fiber right up against the LED lens, but needless to say, all of the light does not get directed into it. But at least it gets points for less falloff.

So... brightest LED when battery life/efficiency is not an issue, and tightest beam when spot quality doesn't matter - what does the think tank think?

 

[enLIGHTenment]

Use an XR-E LED and two very low F number aspheric lenses in a condenser formation. If you can find lenses with an F of around 0.5 then you should be able to capture virtually all the light from the LED into a spot about the size of the LED die itself.

 

[monkeyboy]

I don't think you will do much better than a purpose built coupler but you could try this:

Start off with a domeless LED diameter h.

Single lens solution

To maximise light entry, you need to match image of the LED die, to the diameter of the cable, c. Required Magnification m = - c/h (negative since image is inverted)

Now m = f / (f-d) where f = focal length and d = distance from LED to lens. Rearranging gives d = f(1 + h/c)

So find a wide aperture lens and place this a distance d from the LED. Place the entry to the cable where the image is formed. (Alignment is very tricky).

To keep the die and image the same size like Enlightenment suggests use image and object distances of 2f.

Two lens solution (Simpler but probably less efficient)

Choose two lenses where f1/f2 = h/c. Choose widest aperture lenses possible. From left to right; Place lens 1 a distance f1 from LED and lens 2 close to lens 1. image will be formed distance f2 from lens 2.

 

[abvidledUK]

How's about a laser torch ?

 

[StandaT]

like this http://www.hbeamtech.com/prototypes.html? )

 

[Duglite]

like this http://www.hbeamtech.com/prototypes.html? )

I saw that thing... amazing, but certainly bigger, more $$ than I have and more light than I need.

 

[Duglite]

How's about a laser torch ?

That would be bright and narrow, but I'm expecting extended runtimes of up to 8 hrs. or more.

 

[Duglite]

Use an XR-E LED and two very low F number aspheric lenses in a condenser formation. If you can find lenses with an F of around 0.5 then you should be able to capture virtually all the light from the LED into a spot about the size of the LED die itself.

Thanks to this suggestion I started looking into the Cree products, and I see that they have the colors I need, and a narrower angle than the Lux products. So I guess I'm going to have to learn how to solder these puppies. I saw in some other threads that they can be found on stars, but there was mention that the thermal properties weren't that reliable. A star would help me, though.

Monkeyboy, I'm saving all of your useful info to study it more and make sure I know what I'm doing. I'm going to jump in at first with some aspherics I just ordered from Surplus Shed based on hints from this "Cyclops" thread (http://candlepowerforums.com/vb/showthread.php?t=148821) and then when I'm at the point where I'm ready to hire someone (probably from this forum) to machine a housing for me, I'll see about applying your formulas for a more accurate lens.

 

[monkeyboy]

I think the 4 die domeless osram ostar would be ideal. Take the "diameter", h, to be the diagonal of the LED die. (About 3mm for the 4 chip ostar)

A picture would explain things better.

Do = object distance
Di = image distance
h = LED diameter
c = cable diameter
f = lens focal length

Use the following distances:

Do = f(1+ h/c)

Di = 1/(1/f - 1/Do)

 

[Duglite]

I just looked at it on the Osram site. I'll bet that's a blindingly bright chip. I wish they had a color assortment.

Has anyone ever ground down the lens on a Luxeon Star? I wonder what a flat spot would do to its natural throw pattern...

 

[TMorita]

It sounds like you should be using a laser diode, not an LED.

Toshi

 

[monkeyboy]

Now that I think about it, you could try a domed LED. The dome acts as a lens, but since it's in contact with the die it shouldn't affect the distances much.

The smaller the die the better, as this would allow the LED to be closer to the lens, hence more light capture. You could use a Cree XR-E or SSC P4 LED.

Of course a laser diode would be better, but this option is far more expensive especially if you require a few different colours. Not to mention safety issues.

 

[Duglite]

Yes, my main concerns about a laser: runtime & heating, lifetime, colors, and speckle. This is going to be used in a decorative display, essentially an art piece. I should point out that my early prototypes use Luxeon IIIs and a 3-degree "fiber collimator" lens, and the results have been satisfactory. That being said, I'd really like to increase the apparent light output by some reasonable percentage, without getting too expensive or complicated. I'm hoping that a different LED choice and/or a better optic could help. for simplicity, things seem to be pointing to a narrower-angle LED such as a Cree, and maybe one of those ~50mm aspherics. I doubt I can afford a truly customized lens solution. Where I'd consider spending money would be on some kind custom milled of LED mounting/heatsink/lens mount/fiber coupling, which is probably well within the capability of the fellows doing mods that require them to turn aluminum or copper. I could see the thing looking like a stubby flashlight with a closed fiber fitting rather than an exposed lens. This will help me achieve these distance relationships between the die, lens and fiber tip that you have been telling me about. (Ideally it will have some threading that allows slight adjustments.) So someone is likely going to have a chance to make some money with their lathe before summer is over! BTW, I ordered some free XR samples from Cree. I have a dissecting microscope, so I'm not too intimidated about soldering the things. Well, a little, maybe.

 

[monkeyboy]

The positioning actually wouldn't be too much trouble with such a large cable.

For example, aligning a laser with a single mode fibre optic cable (few microns in diameter) is extremely difficult, even with a purpose built coupler. I once heard about a custom built precision piezo-electric coupler that cost around $20,000

 

[Duglite]

Yeah, my rough-built coupler solution right now allows about 1/8" adjustment all around, and seems good enough. I'm more concerned about getting repeatable focal distance setups that won't slip with the most minor impact or handling.