The Luxeon Side-Emitter and reflectors

I've been pondering the reason it's so hard to get an extremely tight focus with a Luxeon Side-Emitter and a reflector. I can obtain pretty good focus, but nothing like I expected, given the SE's reflector-destined design.

The high-dome seems to focus to a much tighter and cleaner spot, but a spot that is not as bright as the total light thrown forward from an SE. An SE seems to exhibit a large range of "focus", ranging from a circular hot spot with a dark then bright ring, to a bright medium tight circle of light with a slightly brighter center (but smooth and not ringy).

I jammed an SE into a TurtleLight reflector, which focuses a single ground-down Nichia to a very tight hotspot. I performed some tests between the SE and a HD to see which I preferred. The HD did focus tighter, but threw out less light into the hotspot. The SE made an uglier beam, but was brighter and more useable.

As has been determined before, the ideal light source for a parabolic reflector is a single point of light. The HD seems to mimic this better than the SE - though the amount of light emitted to where it's usable for the reflector is MUCH less with a HD (at least with the 1W) than an SE.

However, the SE suffers in the reflector world as well. It actually presents 2 points of light source to the reflector - one from the "dome", and one from the "cone". Today, I applied some black tape to the different sections of the emitter lens to see what effect that had on the emitted light pattern, and the final beam.

This picture should explain a lot:


You can see the effects that each part of the optics have. I seem to have placed the emitter such that it is somewhere in between the optimal focus for each source (cone and dome).

I don't have any dimpled or textured reflectors, so I can't say if they negate the effects of the SE's emitting pattern.

Phew!

-Darin

Darin - I have more limited experience with the SE than some of the other modders here, but my take on the SE is that they prefer "deep" as opposed to "open" type designs.

I've found that really deep parabolic reflectors give tighter beams. This is most likely due to the dispertion angle of light coming from the SE.

I have a 5W "junk" emitter mounted on a 3/8th's diameter aluminum rod located centrally in a round finned heasink. In addition to "torture testing" overvoltage direct drive, this setup alows me to drop any reflector over it to get a beam sample at any feasible height within the "sweet-spot" of the reflector.

No one else has ever expanded on the idea of light patterns from two distinct areas of the SE being a potential cause of beam defects. Thanks for sharing your results!

Darin: Nice methodical investigation

Darin,

Good job! I too have been trying to harness the light from the SE's and have come to similar conclusions. I have found that the 1 W SE is more manageable as it would appear that the 2 sources of light are close enough to a single "sweet spot" to allow for effective positioning in a stippled reflector. I can get a good wide but dense beam in a E series reflector and in a turbo head, a very effective and concentrated spot is possible.

The 5W SE is another story though. In the E series reflector, the 5W SE gives you a decent but wide flood beam. In the turbo head, the best beam I have found is obtained by placing the rear source of light at the focal point and allowing the light from the cone to bounce off the reflector in such a manner that its light is concentrated in the center of the beam but spreads out into a wide overlaying flood. The total beam is a useful, center weighted spot with diminishing amounts of light as you radiate out from center. I think the beam is useful but if you are after maximum throw with cookie cutter spot coupled with even flood of low light concentration, the SE doesn't do it. The 5W HD's on the other hand give you a much brighter and tighter spot beam with a significant increase in penetration and throw.

If you apply about 2.5 volts to a 5W SE, The light is dim enough for you to view the LED from the side and you can see the two sources of light you have mentioned. My conclusion is that if you want to throw the light of a 5W, you need a HD in a large diameter and deep reflector.

Thanks all - this was bothering me all day yesterday, so I skipped some research work to test this out and write it up.

I think we've had a reflector conversation here before, though I couldn't find it. Therefore, much of this may be review and rehashing...

It seems that a "wide" reflector, like a SureFire turbohead, will compensate for a "wide" light source (such as a bulb filament). By the same token, a "deep" reflector will compensate for a "tall" light source (such as the two-point source of an LS SE). The high-dome may appear as a "wide" light source, but the larger reflectors that we use negate that. There is no need for "deep" reflectors with high-domes, as they emit light from a single plane.

Now, reflectors from a Maglight, and TurtleLight (nearly the same size), are pretty wide and deep - apparently wide enough, but not deep enough for the SE.

From what I've seen, the Surefire turboheads are really wide, but not very deep. Anyone know of slightly wider, but MUCH deeper reflectors than that of a Maglight?

-Darin

While Maglite reflectors might not too ideal for 1W SE, it is perfect for 5W SE. It produces tight hot spot and smooth spill light. I was stunned by its performance when I received the Megaclop from Lambda.

Alan

Actually, since yours' Alan, I've even been Krylon sputtering the 5W mod reflectors, as the beam gets even better! A dimpled, or orange peel reflector makes all the world of difference with the 1W, and noticeable improvement for the 5Ws.

Lambda,

What's the exact procedure for your home-brewed textured reflector? I understand it involves a reflector, and krylon clearcoat, and sputtering - I'm still reluctant to try without a more detailed idea of how to proceed.

I made a cyclops myself, and it does produce a nice beam with the SE. The SE doesn't snap into a tight focus like a HD would (though there is more light in the SE beam).

I'm still inclined towards using the SE - even though there are more losses in the emitter's lens, more light is thrown in the direction that a reflector can use vs. a HD. (Plus, the SE is more interesting to talk about once someone sees that wierd little hat on the LED )

-Darin

Basically, to sputter the reflector, you want to give it a shitty coat of paint. By that I mean, you do not want an even layer of Krylon, as it will only blur the beam pattern. You do want a uniform, speckeled coat of Krylon droplets on the reflector surface.

I hold the nozzle about 14" from the reflector at about a 45 degree angle. Use very short (1/2 sec) burst of spray, step around 180 degrees to the other side of the reflector and spray again to get a uniform coat. Step around 90 degrees and repeat the process above to ensure a uniform coating. Immeadiately dry! Do not let it 'smooth' out. Wave it around in the air if you have to, or use a hair dryer. You want to dry the Krylon droplets in their place. For an ultra smooth beam, apply a second coat a few minutes later, but not too much, or you'll lose reflectivity.

You may want to pick up a cheap flashlight reflector to practice on, but it's really fairly easy to do with large reflectors. Small reflectors are much harder to get good results with.

Excellent! I may have to give it a go on my Dorcy $3 light I've been hoping to smooth out the reflector on my MagSE.

-Darin

Darin,

Great work!

The masking was a good idea, but I can't help feel that perhaps something may have been missed out by simply piecing the beams together this way.

By looking at the NX05 or NX01, one can surmise fairly well that a self-contained reflection microsystem plays a big role in any optical design for the luxeon. For example, the NX optics internally-reflect sidespill light from a low-dome forward using the conical part of their construction. I'm guessing that this reflected light is the secondary beam. The primary forward beam (or hotspot) comes from the interior mini-lens that is directly in front of the emitter itself. So, the NX themselves, too, project a composite beam.

I'm hazarding a guess that the same principles are in play too for the SE's cone optics. Except in this case, I'm guessing that external reflection has a part to play too.

To my eye, the dome appears to bounce a lot of its light off the external cone surface, hence projecting light sideways .ie. the dome and cone work in tandem. This is not the same as a composite beam, where you combine beams. This tandem system requires 2 parts to form its primary light path. Masking the cone would only reveal the radiation pattern of the dome in isolation, which may not be the effect the optical designers had intended.

Similarly, masking the dome only allows internally-channelled/reflected light to be emitted from the cone, which may be quite incidental only.

Also, by analysing the datasheets, you can work out the effective angles of incidence needed on a reflector to make it "ideal" for the SE. In other words, not any 'ol parabola will do. A reflector whose parabolic profile matches the radiation characteristics of the SE will yield the best beam, and all others will vary in mileage. To my knowledge, most of the parabolic reflectors out there aren't optimised for the relatively narrow effective radiation angle of the SE. You might get lucky now and then with some reflector you find in a another flashlight, but a purpose-built reflector destined for the SE is another thing altogether.

My humble 2 cents....

hotfoot,

Thanks for your thoughts!

The middle picture of combined beams *IS* the real combined beam (that is, no masking tape, completely new picture), I haven't tried combining the individual masked pictures to see what the results would be.

However, I was thinking the exact thing about external reflections off the cone. The dome of the SE does emit light forward near where it joins the cone, and external reflections off the cone may add to the light output. That means that instead of 2 distinct focus points, there might be a range points of emission, from the bottom of the dome, to the top of the cone, and all points in between. It's quite a complex emission pattern, but with a large enough reflector, the pattern is close enough to a point to not matter much.

This was all driven by my MagSE. After building it, it never snapped into a sharp focus. There seemed to be a range of "focused", from the first point where a coherent bright circle of light is made, right up until the "black hole" appears. I still haven't decided on what focus point is ideal for the most throw.

-Darin

Originally posted by lambda:
Actually, since yours' Alan, I've even been Krylon sputtering the 5W mod reflectors, as the beam gets even better! A dimpled, or orange peel reflector makes all the world of difference with the 1W, and noticeable improvement for the 5Ws.

Click to expand...

<font size="2" face="Verdana, Arial">I know that I'll get my 3rd or even more Megaclops soon I just can't help it:-(

Alan

evan9162,

My light meter sees the brightest spot when: the head is screwed in until the black hole appears (usually + shaped), then back the head out about 1 1/2 turns. In that general focus area lies the brightest beam on the ones I've put together.

Darin,

Hehe - yes, it does look like we're on the same track here.

Some time ago, I tried to quantify the datasheet graph in manner that would help me better identify or specify reflector curvatures that would work with the SE. I ended up with a re-interpreted SE dispersion chart below:



Basically, the redder the segment, the greater the output. A reflector that takes these hottest segments and has a curvature with the perfect counter-angles to the angles of incidence of the respective segments should theoretically cast a "perfect" spot, on an infinite-point-of-focus basis, ie. 0-degree beam.

I dunno if it makes sense, but when I drafted out a reflector based on the above chart, I ended up with a nice parabola! Turning that parabola into a real physical reflector is another story altogether and I'm still trying....

lambda, what if I sputter a 3d m*g reflector and use it with an incandescent bulb? Will the m*g now have a decent beam, and will the paint hold up to the heat of an over driven bulb?
Ted

Looking at the design of the side emitter, I conclude that essentially all of the light that appears to be coming from the side of the cone segment is externally reflected light, as has been suggested above.

Doug-
Interesting - if that's the case, I wonder if silverizing the inside of the cone would gain any advantage in side-emitted light.

hotfoot-
you've modeled the SE as if it were a single point source of light. In reality, it's more like there is a region in the vertical direction of source of emitted light. Thus, for the red portion, some may be coming from the point source as above (dome), and some from a point slightly above the point source (reflections from the cone). If your reflector is made for the dome, then the light emitted from the cone has a slightly different angle of incidence to the reflector surface, creating a different beam.

Good information everyone!

-Darin

Leo,

That's a cool graph you have there! Now for extra credit, draw in the volume of light origin as opposed to the point of origin of light.

If your graph is even close to accurate which I believe it is, I think it is clear that especially with the SE, the depth of the reflector is not as critical as the curvature of the reflector. I would assume that if you were to draw two parabolic curves on top of eachother, with each based on one of the two basic points of origin and then merged these two curves weighted by the relative light intensities of the two points of origin, that if nothing else, I have just typed a really long sentence.

I suspect that I can be shown the errors of my thinking without much effort but I think that the HD or lambertian is the LED that can really benifit from a deep reflector as the deeper it is, the more light the reflector will catch that projects in relativly shallower, off axis angles.

Mute point since most of us can't machine parabolic surfaces; best we can do is stick LED's in existing reflectors and move them around and see what we get.

I have engaged neither spell nor brain check in above post; beware!

For the rest of the day, I remain,

Don

Don, I am trying to remember the useful term that you coined awhile back that is relevant to this discussion. Was it LCV, Light Center Volume?