Construction of an LED-light

[keks]

Hi, as i have found no proper LED-light available for purchase, I have decided to try making one by myself. Demands: good light quality, suited especially to light small objects, affordable.

I have bought some Yuji-LED's (BC Series High CRI COB LED - 135L, Daylight white (5600K).

I have found an aluminum tube (inner diameter about 10 mm) with a very reflective inner surface, that i want to place in front of the LED to direct the light the right way. The tube can have any length from zero to infinity.

I know there will be some more parts necessary to make that thing working, but I just would like to hear some opinions if that approach is meaningful or not.

[zzffnn]

Why not use a collimator lens, instead of an aluminum tube? You can use acrylic lens to save cost.

[Chris S.]

Keks,

What role do you see the tube serving? To me, it seems like a "snoot"--a tube that constrains the angular spillage of light. While snoots certainly have their place in photographic lighting, the much more common requirement is to spread the light out. (After which comes the next, vital, challenge: Make the light appear to have come from a broader, taller light source.)

As Fan suggests, substituting a collimating lens for the tube could increase transmission efficiency. But this still assumes that you want to limit the output angle of your LED. Do you, in fact?

A (too?) brief search indicates that your LED's have a 120-degree emission angle. If brightness is even across this angle, it's not difficult to imagine using these LED elements without snoots or collimating lenses, but with diffusing or reflecting elements, for macro lighting.

Since you've already purchased these LED elements, I'd suggest you wire them up and give them a try before buying additional components. The results you get will help determine what other components you may want for these LED's and your preferred style of lighting.

--Chris S.

[keks]

I agree the light source should be taller than the object, but i wanna avoid just loosing lot of light because its spread in all directions. Therefore I want to limit the angle of the light emission (emission angle of 120 ° for the LED is correct). The tube will have a diameter of about 12 mm (depending on the tube size finally used), object size would be about 4 mm or smaller for the intended use.

I consider the tube as a simple kind of reflector, maybe not perfect but easy to build. Due to its small dimension it can be placed very close to the gap between microscope lens and object. I am not aware of any other kind of reflector that fits to the square LED-element and doesn't increase the size of the whole lamp considerably.

The optimal tube length could be very short, maybe only 1 or 2 cm (less than 1 inch), I would have to test that.

Using a collimator lens instead sounds interesting and its probably the most appropiate technical solution, but I'm not familiar with that topic. A hint to a suited lens would be much appreciated. The diameter of the emitting area is about 11 m (I don't have the LED's ready to hand right now).

[enricosavazzi]

Look up "light guide" and similar expressions on Google. A glass or acrylic light guide should be more efficient than a polished aluminium tube. Light guides for teeth bleaching use are often a simple acrylic rod of circular section. High-efficiency light guides have tapered sides, rather than parallel ones (probably to reduce the number of internal reflections).

The good thing with LEDs, though, is that you can place them close to the subject. Rather than using a light guide, you should try to place the LED as close as practical to the subject (but not so close that it overheats the subject by radiated IR). Mount the LED on a solid copper rod that acts as a heat conductor to a heatsink, so the heatsink can be large enough and far enough from the subject to stay out of the way. Use a heat pipe instead of a rod if you need a higher heat conductivity.

Collimating lenses may also work, but unless they have a really high NA, they may produce a dimmer illumination than no collimator lens at all.

[Troels]

Keks, as I understand it you want to find a way of minimizing light wasted (and I would add: potential disturbing light) because your light source (the LED) is too big to be positioned close to the object and has a large emission angle.

As Chris S. points out in most cases you would benefit from using some kind of diffusor, but that does not necessaryly eliminate your problem because it could still be difficult to position your lamp correctly without spreading light all over the place. Remember: light from the wrong direction could result in serious flares in your optics.

Just read Enrico Savazzi's advise on using light guides instead of tubes. His advise is always worth listening to.

I will just add: If you for some reason prefer to try a solution with a tube, you should be aware of the fact that the light emission angle from the front opening of a perfect reflecting tube will be the same as the emission angle for the light souce inside. If the inner tube surface is not a perfect mirror (i.e. with some diffuse relections) it will be even greater.

If the end of the tube is easier to position near your objec/diffusor than the LED itself this might be a solution to your problem. But if you want a more narrow angled light beam (making it possible to maintain a longer distance) you have to either use a snooze with dead black inside walls, some kind of collimating lens or a light guide as proposed by Enrico.

Troels

[keks]

I already had an issue with a tungsten bulb that warmed up an object (very small flower) in a way that it languished faster than I could stack, thats why I'm quite interested to get a "cold" light.

I have operated 1 of the LED's at approx. nominal power on a little heat sink for some hours, that has worked well. The LED got hot but obviously didn't overheat. Heat emission on the front side was surprisingly little.

Light guides are interesting, but those for dental use on eBay are only 8 mm in diameter. Perfect for me would be about 12 mm diameter and a few cm in length.

Heat pipes: I have seen those for CPU-cooling on eBay, fascinating actually, but the dimensions are bit disturbing. I may come back to these in case I experience a problem with the heat sink solution.

[keks]

Troels, I believe the aluminum tube can safely be considered as a not perfect mirror. Actually I was expecting a somewhat narrower emission angle at the end of the tube, because light rays emitted at a larger angle would have to be reflected more often till reaching the end of the tube and thus be more dampened than other light rays who take a more direct way with less reflections. But I certainly may be wrong with that consideration.


However, sounds like directing the light with some proper means is advantageous. I will have some further look on that matter, any help is welcome!

[ChrisR]

What is the subject going to "see"?
It'll still see a very bright, very narrow angle spot, with a disc around it of quite bright light.
The brightness of the spot is about the same as without the tube, the disc is new.

What I think would be more use, would be to make the light spread around the "sky" that the subject sees. So the subject sees light coming from a large angle, more or less evenly. The first requirement then would be a diffusing disc placed at the end of the tube.

A concave surface could achieve a greater angle, assuming that the subject were very close, relative to the curvature.

Even if the surface were not concave, it would still need to be diffusing.

[enricosavazzi]

(...)
Light guides are interesting, but those for dental use on eBay are only 8 mm in diameter. Perfect for me would be about 12 mm diameter and a few cm in length.
(...)

Retail sellers of acrylic stock can certainly provide thicker rods, probably 50 mm or more without problems. They may also be able to polish the ends of a rod cut to custom length.

Polishing the ends may not be necessary, though, since a coarse/sanded end will work as a diffuser. You need a diffuser anyway at the subject end, as Chris pointed out, and it might be good to have one as well at the LED side (in general, two diffusers spaced apart from each other have been found to work better, in many different situations, than a single diffusing sheet or surface). The sides of the rod, however, work better polished.

If you could find a rather acute cone, rather than a cylindric rod, and cut away the tip, it would work more efficiently as a light guide (with the wider end toward the subject) than a cylindric one. If you have access to acrylic machining equipment, you may also make a highly efficient light guide as a truncated pyramid with a square base machined out of a solid block of acrylic stock. This will have to be the shape of a truncated pyramid with the wider end only 1.2-1.5 times wider than the narrow end and a height 4-6 times the base width, not something as squat and low as an Egyptian pyramid.

Acrylic items made for use in jewelry (or other small merchandise) display come in a variety of shapes. I would not be surprised if they have a ready-made truncated cone the diameter of a finger, for example for displaying jewelry rings.

[keks]

Enrico, do I understand correctly that these dental light guides are just having acryl inside?
Getting cylindric material and cutting it (even sky-shaped) wouldnt be a problem, and I also can do some coarse polishing at inlet and outlet. I dont see a way to do the pyramid shape for now. Having a few (?) % less efficiency won't hurt that much.

By the way I am planning to build 4 of these lamps, and I have already prepared a box with 4 LED-drivers for adjusting the LED power independently from 0-100%. I guess such a large illuminating area should ease diffusing.

[siliconGary]

Acrylic/poly carbonate collimating lens are very inexpensive. Carclo makes several. You can match them to your LED and the light pattern you want at LEDSupply http://www.ledsupply.com/led-optics
I used one on a microscope illuminator it was ~$2 (They also come clear or frosted to serve as diffuser)

[enricosavazzi]

Enrico, do I understand correctly that these dental light guides are just having acryl inside?

That is correct, just a solid cylindric acrylic rod. No mirrored jacket, no optical fibers. They just rely on total internal reflection of light exceeding the critical incident angle. Usually heat-formed with a bend near the end that goes on the teeth, to make it easier for the dentist holding the tool.

[keks]

Ok as I understood so far:

Arcryl rods are a good way to direct the light to the location I desire, but they dont change the quality of the light (emission angle). I may try this because its easy to make.


With a collimating lens the light can be bundled to parallel rays, which appears to be the more advantageous kind of light.

So the desired lens would bundle the 120° light of the LED to parallel (0°) rays. And it should be large enough to cover about 12 mm diameter of the (my) LED. Those from Carclo are too small, I have found several models on eBay, but I have no clue which is the right one, the sellers don't provide much detailed information about what exactly their lenses are actually doing.

What would be correct designation/technical data for such a lens?

[enricosavazzi]

Arcryl rods are a good way to direct the light to the location I desire, but they dont change the quality of the light (emission angle). I may try this because its easy to make.

Essentially true for the light that enters cylindric rods, although there is some de-collimation at the exit, and light that hits the entrance surface at a high angle from the normal is largely reflected away and does not enter the guide. Correctly designed tapered rods do collimate the light.

With a collimating lens the light can be bundled to parallel rays, which appears to be the more advantageous kind of light.

Not parallel, except in the ideal case of a point-sized LED, but less divergent than before entering the lens. With a large cob, you can probably do no better than 30-50°.

So the desired lens would bundle the 120° light of the LED to parallel (0°) rays. And it should be large enough to cover about 12 mm diameter of the (my) LED.

The acceptance angle depends on the NA of the lens rather than on its diameter alone. You need to put the lens at a distance from the LED equal to the lens focal length (in practice, you should be able to see a magnified portion of the cob surface approximately in focus when looking through the collimator from a distance). Only then you get maximum collimation. I doubt you can find a collimator lens that accepts a cone of light wider than 90° or so, the rest is "spillover" that does not enter the collimator.

Old slide projectors of reasonably good quality used to have an aspheric pyrex condenser lens between the halogen lamp and the slide. This condenser can be used as a collimator for a not-too-large LED cob. It should provide a relatively narrow collimation. Low-end slide projectors often had a collimator with two planoconvex lenses facing each other, not as good as the aspheric collimator.

Superficially similar aspheric lenses are sold on eBay for mounting in front of LED cobs for interior illumination. The LED cob is usually surrounded by a plastic chrome-plated cup that helps to collect the spillover and reflect in into the collimator. You could start by testing one of these reflector + collimator assemblies and see how well it works, but I think they are meant to provide a rather wide angle of illumination. Wider than the aspheric collimators for slide projectors.