pre-collimator

[snimmer]

Hello,

I keep reading about lights with pre-collimators and think I understand the function....but I cannot seem to find any threads or other online content that is as descriptive or gives direction on how to choose one or install one. I have two 18650 torches with asphericals and I'm trying to decide if a pre-collimator would help .....and if so.....what should I get and how complicated is it to position it correctly between LED and aspherical. Can anyone point me to a thread or other site that discusses? Thank you.

 

[gcbryan]

I'd be interested in the answers as well. I guessing that the practical answer is going to be that you just have to experiment.

 

[bshanahan14rulz]

I'm curious too. It's probably something like picking a small diameter short focal length lens and pairing it with a larger diameter longer focal length lens.

Probably something like
small lens f1 on the emitter image
large lens f1 on small lens' f2

I think the chinese goods sites have little plastic aspherics too... or you could see what happens when you put an aspheric in front of a smaller aspheric light. I have a lens assembly from an LCD projector, but I know that once I disassemble it, I'll end up scratching or chipping the lenses, or I won't be able to put the thing back together.

 

[gcbryan]

I think the biggest problem with trying to make our aspheric lights more efficient is the fact that so much of the light from the emitter is going to the sides.

A small reflector that would focus it's output on the aspheric would be good. I think that's what the DEFT uses (not positive).

None of this would be necessary if the emitter beam angle was a point source or a lot narrower and straight out instead of to the sides.

I also imagine you lose a lot transmission wise with two optics. Maybe the lose is as much as the gain?

 

[Th232]

The DEFT uses a precollimator lens.

I remember from somewhere Saab said that the pre-collimator should only alter the LED output very gently compared to what the main lens does.

Here's an applet I found and have played around with a bit to get an idea of which lens combos might work well.

Most of my precollimator experiments were for a relatively small aspheric, it ended up that in every case that the precollimator interfered with either the LED or the main lens.

Now that I'm moving up a bit in size I'm starting to look into precollimators or a small reflector again, but nothing settled so far.

 

[gcbryan]

I'm not sure, but I think in one of the many versions of the DEFT a small reflector was used (or at least experimented with).

I'm just recalling one of Saab's (Michael or whatever he prefers to be called) posts.

 

[Th232]

I've followed the DEFT for some time, and while he definitely experimented with small reflectors, I don't recall him ever selling DEFT that used a small reflector. Now I'm curious, might as well ask him.

 

[gcbryan]

I've followed the DEFT for some time, and while he definitely experimented with small reflectors, I don't recall him ever selling DEFT that used a small reflector.

You're probably right. He did say that the place on the die of XP-G's were optimal for this sort of thing and he certainly doesn't use those in the DEFT so maybe the reflector was just an experiment.

My guess is that, unless you do it right (as he is able to do), you will find that the negatives will outweigh the positives.

If I try to just use whatever optics I can find I think I'll find that I would be better off just using one optic as far as throw is consider.

I'd like to be wrong however.

 

[Th232]

Yeah, it's a really interesting field to see how even two optical elements interact with each other. I'm hoping to use a reflector + aspheric combination to get good throw with some high CRI P4s I have. Would be a lot easier task if there were high CRI XP-Gs or XR-Es available.

 

[gcbryan]

It seems like (at least for any homemade solution that I can come up with) the primary benefit of an aspheric ends up being just the narrower beam rather than any greatly enhanced throw.

I read about someone using an aspheric on a XP-G (wouldn't be my first choice) so I went outside with my XP-G R5 based flashlight and just held a 28 mm glass aspheric lens in place to see how far it would shine.

I'm just basing this on which trees I can see and which I can not. Without the aspheric I could see the same trees that I could see with it. I just had more light without it.

The situation is pretty close to the same when I use my XR-E R2 with aspheric and then when I try it without the optic.

In the case of the XR-E R2 it of course throws further than the XP-G R5 but the main effect of the aspheric is just to limit the spill.

This is useful of course as it makes it easier to focus on the "subject" but the beam isn't really throwing much further than the same light without the optic but using a reflector.

I'm sure if you optimize everything perfectly the aspheric option must throw somewhat further.

I think it's just inefficient however you do it with an aspheric since the led isn't designed to emit most of it's light where it will be focused by the aspheric which is why it doesn't throw as far as you would think. It's having to work with such a small portion of the total output.

 

[Al Combs]

One possible use for a pre-collimator would be to reduce the virtual size of the die (i.e. smaller target, higher lux). This is a bit off topic, but comes back at the end. I did some experimenting with a P7 in hopes of eliminating the donut. I never posted the results because it didn't work. I still had the donut but the LED lost quite a bit of output, despite having a tighter beam. Both Luminus and Osram sell LED's without domes that are spec'd lower in lumen output than their domed counterparts. I believe the reflection of the dome causes a secondary fluorescence that both increases output and effects color. Color because the secondary reflection has no Royal Blue in it. When I cut off the LED's head, it made the beam both dimmer and more yellow/green. I had twin P7's to make the comparison.



Using the pixel count of a selection box in a paint program as an estimate, there was a 22% reduction in the size of the 4 LED cluster (not the whole phosphor patch). PhotoBucket resampled the pictures down to 1024*768. That picture shows only a 20% reduction, but essentially the same. Taking the lower 20% as an inverse square relationship should have netted a 56% increase in Lux output but didn't. If anything, a side by side comparison with a "normal" P7 (both Mag's) seemed to have lower surface brightness. Estimated by eye as I don't own a lux meter.

So back on target, I believe a plano-concave lens might achieve the virtual size reduction without loss of output that decapitating an LED causes. I should mention I never actually tried it. But it seemed at the time, the idea had potential. If the concave surface was both uncoated and of a radius that kept all parts of the glass equidistant from the center of the LED, it might actually increase the LED's lumen output by contributing to the secondary fluorescence of the Lambertian dome. The plano surface would be coated to reduce unwanted reflections at that surface. The sides of the lens would have to be taper ground and blackened, also to reduce reflections. The plano surface would have to be much larger to avoid vignetting the main aspheric lens. For a simple experiment, it wouldn't need to be so exotic. I have no idea what effects such a lens might have on chromatic or spherical aberration.

 

[007adan]

Great subject here..

I used a 28mm aspheric as a pre-collimator offset from my xre-r2 by a few mm to reduce reflection. My main lens was a 3" Aspheric. I was able to resolve the beam into a nice tight spot but didn't notice much difference. I do not believe another aspheric is the answer for a pre-collimator, I think you need a lens that collects and tightens up the light before it hits the main aspheric but not to the point that its too tight of a beam. I have done research and many say that a anti-reflective coating is necessary for a pre-collimator to be totally effective. I ordered a few of these fellas from Dx http://www.dealextreme.com/details.dx/sku.13570 to try as a pre-collimator. Saabluster has indicated that a carefully selected PCX lens is what he uses as his pre-collimator but will not go into detail because of all his R&D. Which I find Ironic because most of his R&D was done on this site and another CPF member was the one that turned him on to using a pre-collimator in the first place.

 

[Walterk]

The small aspheric has a small focal length, and therefore a wide viewingangle at the Led. In that way the 1x1mm2 surface of the Led is then multiplied with the mm2 of the surface of the lens. That is how much Lux is collected.

Lenses with small focal length, have a wider beam then the same size lens with a longer focal length. Thats why the light from the Led (collected by the first lens) is used; to use a lens with larger focal length and thus more narrow beam. The beam gets more concentrated in that way, so more throw.

The combination of two lenses in this way, is called a beam expander.We used it reversed then it is collimator. (same concept is used for laser and telescopes).
The practical explanation is that it gets more light to a larger emitting surface.

I posted some pictures in the thread Aspheric lens tricks., and Why do throwers have to utilize Asperics.

 

[saabluster]

I'm not sure, but I think in one of the many versions of the DEFT a small reflector was used (or at least experimented with).

I'm just recalling one of Saab's (Michael or whatever he prefers to be called) posts.

I have never offered a DEFT for sale with a reflector instead of the small pcx. There is more in this thread to comment on than I have the time right now so I will have to check back in later. Saab, Michael, whatever. It's all good. Just don't call me lusty.

 

[007adan]

-bump-

 

[wechnivag]

Try using a positive meniscus (both sides of the lens curves forward), maybe in the diameter range of 12mm-15mm or even 20 mm, if your main aspheric is a large 50mm lens. The reason for using a positive meniscus butted right up to the XR-E dome is that you can collect a bit more light as it curves round the dome contour (you can only butt right up to the dome with a plano convex or plano-aspheric pre-collimator)

The focal length need not be too short, 20-30 mm is enough to have a significant pre-collimation of the beam. Too short a focal length will result in a very large die image out of your main aspheric, which defeats the purpose.

It is a fine balance between collecting more light (pre-collimating) without enlarging the die image too much.

This is what i gathered from my experiments. I got my positive meniscus from Surplus Shed.

At first i went nuts and got a huge 20mm lens with like a 20mm focal length. It was almost like a bowl and could cup right over the XR-E and rest against the star with clearance to spare. It definitely collect and collimated all the light from the LED (apart from reflective losses) but the end result was unusable because the die image was huge.

 

[wquiles]

Try using a positive meniscus (both sides of the lens curves forward), maybe in the diameter range of 12mm-15mm or even 20 mm, if your main aspheric is a large 50mm lens. The reason for using a positive meniscus butted right up to the XR-E dome is that you can collect a bit more light as it curves round the dome contour (you can only butt right up to the dome with a plano convex or plano-aspheric pre-collimator)

The focal length need not be too short, 20-30 mm is enough to have a significant pre-collimation of the beam. Too short a focal length will result in a very large die image out of your main aspheric, which defeats the purpose.

It is a fine balance between collecting more light (pre-collimating) without enlarging the die image too much.

This is what i gathered from my experiments. I got my positive meniscus from Surplus Shed.

At first i went nuts and got a huge 20mm lens with like a 20mm focal length. It was almost like a bowl and could cup right over the XR-E and rest against the star with clearance to spare. It definitely collect and collimated all the light from the LED (apart from reflective losses) but the end result was unusable because the die image was huge.

Very cool - can you please share which parts worked that you got from Surplus Shed?

 

[wechnivag]

I can't remember exactly. About 15mm diameter and 30mm focal length.

Search under PMN with those stats. Most are coated which I nice. 4 dollars each so not too expensive to experiment.

Dia. Fl. Coated.
L4160 PMN 9.5 20 Y $4.00
L4163 PMN 12.1 29 Y $4.00
L4358 PMN 11.0 30 Y $4.00
L4476 PMN 9.9 20 Y $4.00
L4478 PMN 12.1 29 Y $4.00
L4480 PMN 13.0 26 N $4.00
L4563 PMN 9.0 19 Y $4.00
L5040 PMN 18.2 26 Y $4.00
L5125 PMN 16.1 24 Y $4.00
L5137 PMN 14.0 20 Y $4.00
L5139 PMN 14.0 20 Y $4.00
L5154 PMN 18.0 23 Y $4.00
L5599 PMN 15.8 21 Y $4.00
L5610 PMN 12.7 31 Y $4.00
L5976 PMN 12.0 31 Y $4.00
L6286 PMN 12.8 19 Y $4.00
L6341 PMN 17.5 24 Y $4.00
L6656 PMN 9.1 20 Y $4.00
L7160 PMN 12.2 29 Y $4.00
L8026 PMN 15.5 30 N $4.00
L6575 PMN 12.4 25 Y $4.00
L6592 PMN 14.7 22 N $4.00
L8166 PMN 15.0 30 Y $4.00
L11194 PMN 19.0 22 Y $4.00
L11957 PMN 19.0 24 Y $4.00

 

[Walterk]

I found again the thread that is the first pre-collimator led-light at CPF:

Long Johns; HID-Killer: Adaptor for 1xCree light

It uses a 9cm reading-glass.