A different kind of LED light

[VaughnH]

I have been thinking, which can be dangerous, and it occurs to me that one could make an inexpensive aquarium light using a fresnel lens plus a Luxeon Rebel Endor configuration LED, which can produce about 500-700 lumens. This would have the LED mounted at the focal point of a short focus fresnel lens, so the light would exit the fixture as a 10 inch diameter, for example, beam. My crude calculation indicates that such a fixture might give 50 micromols of PAR, which, because the beam would be close to a parallel beam of light, would not lose much intensity at a two foot distance. Just two of these, side by side, would light a 10 gallon tank.

Needless to say, I haven't done any experimenting with this. One problem is that fresnel lenses of that size and short focal length are expensive, perhaps as much as $150 each. But, the quality of the lens would not have to be very high, so relatively cheap stamped plastic ones could be used. And, it is possible to make such lenses photographically too - a transparent sheet with black line circles spaced just right also acts as a diffraction fresnel lens. That could be very cheap.

Why wouldn't this work?

 

[SuperColey1]

Not sure if this helps or not but this is 3 x 3W LEDs underpowered at 7.35W over a 33USG tall = 0.22WPG. No optics, no lenses just the plain stars (inc shimmer):

AC

 

[VaughnH]

That is a very impressive demonstration!

After thinking some more - bad habit? - I realized that my "new" idea can best be done with a parabolic reflector, not a fresnel lens. By pointing the LED star back at the reflector, with it mounted at the focus of the reflector, you should get a reasonably uniform large diameter beam of light, and with a super LED, that might let just one provide good uniform light for a square foot of aquarium. In effect, just reverse the optics of a reflector telescope.

The only good reasons to even look at such an idea are to get around the inverse square relationship of light and distance, and to get away from needing dozens of LEDs and their drivers. Even the newest, most powerful LEDs aren't a great deal more expensive than the older one watt ones, and not needing multiple drivers reduces the cost substantially.

Of course someone will soon recognize that I'm describing a big flashlight. (I believe that is a "hand torch" in real English?)

 

[SuperColey1]

There are 100W+ LEDs available you know!!! Would require huge heatsinking but what you are suggesting is very feasible.

LED Lighting - Technology & Application (LEDs, OLEDs) - Edison Opto has introduced a 100 Watt single-package LED

Its distributing the light to spread that would be the hardest part. Getting the light via a single super high power LED is easy


This 10W one is probs more feasible
4x 10W Warm White High Power 600LM LED Lamp Light on eBay (end time 09-Aug-09 04:51:24 BST)

The world of lighting is moving pretty quickly

Soon the box (luminaire/reflectors etc) will still be the same size so as to distribute well but the actual light source will be tiny.

AC

 

[Philosophos]

Essentially what you're doing is setting up something like the diffusers that professional photographers use, but with a lens? Sounds good. I don't think I'd limit the application to LED either. I was considering something similar (minus the lens) for the new setup.

While the distribution would be better, wouldn't more small LED's be preferable? If overlapping the spread is more desirable, more points overlapping more than the light beside them would be superior for even diffusion, would it not?

Also, SuperColey, how's that in-line heater working out for you? I've just put in an order for one, hoping it won't turn to buyers remorse.

-Philosophos

 

[SuperColey1]

While the distribution would be better, wouldn't more small LED's be preferable? If overlapping the spread is more desirable, more points overlapping more than the light beside them would be superior for even diffusion, would it not?

I think we talk about overlapping to simplify things. What we are really talking about is an even spread at a decent Par. Therfore if the light is not directed at the tank and a reflector can distribute the light equally across the tank it should give the sameish result.

Also, SuperColey, how's that in-line heater working out for you? I've just put in an order for one, hoping it won't turn to buyers remorse.

It is working fine. Been on there for 2 years or so. It was the third one though. First was second hand off ebay and was cracked so leaked. The second one was brand new from ebay and was part of a recalled bunch. Kept buzzing and staying on. The third one was from a trusted retailer and has been on there ever since

AC

 

[Philosophos]

I can't see it giving the same result. Mutliple points should cause lighting from multiple angles, making it harder for higher growth to overshadow the lower areas.

-Philosophos

 

[SuperColey1]

I can't see it giving the same result. Mutliple points should cause lighting from multiple angles, making it harder for higher growth to overshadow the lower areas.

But the 'reflector' or diffuser can be used to turn the single LED into a similar effect of having multiple point sources.

AC

 

[Philosophos]

If that were the case, diffusion should be able to eliminate shadows just as well as two light sources. Hypothetically, this would mean that diffusion should be able to completely eliminate some of the shadow in behind an object aligned with the bulbs center. After all, any light not on center should be able to get some direct light below an object to the side of it; basic line of sight stuff. If this is possible, I'd like to see it.

-Philosophos

 

[SuperColey1]

There will always be some element of shadow. You could have 1000000 little LEDs above the tank and still will have the effect of shadow.

i.e. 1 area of substrate receives light from all 1,000,000 LEDs and an area under a low leaf receives light from only 400,000. The area receiving light from all 1,000,000 will obviously be brighter than the other although they are both receiving some direct light.

The 'shadow' wouldn't be a perfect silhouette of the leaf as we see in nature. More a gradient to the darkest area

Maybe I am totally wrong but that is what I am assuming.

AC

 

[Philosophos]

Sure, there will pretty much always be a shadow from any decent sized object under overhead light. Exceptions get out of the range of dimensions within an aquarium.

But at the same time:
(excuse the crude doodle)

The shadow its self should be smaller under multiple sources, in each case offering a greater depth of penetration per watt than a single source because the effects of obscuring objects will be minimized. To get the equivalent in diffusion, the light would have to be so diffuse that it would actually cast a shadow that lights beyond the line sugested by the edge of its reflector on a nearby wall, or the floor.

-Philosophos

 

[SuperColey1]

the diagram is what I was saying but I'm lost with what you are saying r.e. the diffusion. I don't really understand this aspect.

AC

 

[Philosophos]

I'm saying diffusion can't come close to equaling that spread. Light doesn't just bend around objects to light as well as direct light without something like gravitational lensing.

-Philosophos

 

[SuperColey1]

I would say neither does the sun though. Although I accept the sun lights up the sky and the light comes in from all angles of the sky, however it hits the water surface and therefore only enters at the top just as we are trying to do with our above lighting.

I think what we are trying to do is just get light to come down from the whole hood rather than just some of it. Therefore I can't see why some sort of reflector could reflect light in a way to make it enter from many pointof the hood. Would be one hell of a complicated setup but I'm sure it could be done somehow. Beyond my abilities though. I would guess someone with a good knowledge of physics and mathematics could calculate all the angles much like a 'mirror maze' deflecting light from reflector to reflector and then some that reflect downward in the same positions as you would have each LED

Here is a silly example which I would guess is nothing like Vaughan is suggesting. This is a little like the mirrorball effect in a disco without being a ball and without spinning

One LED in the centre. point upward to a reflector which has many faced. comes to a point closest to the LED and then tapers out. like a cone but with flat sides.

So then all the light from the LED hits the 'cone' and is redirected outward, then once redirected outward there are reflectors positioned at different places (where the infdividual LEDs would be to capture the light and redirect it downward. I suppose you could put lenses in the path leaving the cone to focus the light at each outer reflector.

As I say this would take lots of mathematics to calculate and make sure that all the 'rays' went where they were supposed to and I would say is excessive in the least when you could get the effect from using seperate LEDS. the only plus would be hardly any soldering or wiring

AC

 

[VaughnH]

That complex reflector idea will work at spreading the light to all corners of the tank, but the light paths will still vary a great deal in length, so the intensity will not be uniform - due to inverse square loss of intensity.

But, a near point source of light, with a parabolic reflector, with the light source at the focus of the reflector, should result in an almost uniform intensity all over the area being lit. However, since no real life light is a point source, other than than from a distant star, and no parabolic reflector is perfect, there will still be less light around the periphery of the beam than in the middle, but the difference should be much less - I think. And, a single die LED is very close to being a point source. I haven't tried to calculate what a 500 lumen LED, for example, should give in PAR over a 10 inch diameter, using this method. When I get bored with my other projects I will try to do so.

Once you wire up your first multiple LED light, I think you have a great incentive to find a way to do the job with far fewer of them.

 

[lanvin]

Not sure if this helps or not but this is 3 x 3W LEDs underpowered at 7.35W over a 33USG tall = 0.22WPG. No optics, no lenses just the plain stars (inc shimmer):

AC

Hie,

can i know what do you mean by a 3W LED? how does it look like? Anyone can lend a hand and show me a hyperlink how it looks like?

 

[SuperColey1]

These are the LEDs from the tank picture you see above. Each one potentially 3W making a potential 9W total. They are underpowerd to 2.45W making 7.35W total

 

[shoggoth43]

I think one problem you may run into is a slight dead spot in the center of your photonic cannon. The LED is sitting there, but more importantly you'll need your spider AND you'll need someway to heatsink that LED. You might be able to do something with the heatsink if you put it on top of the reflector and run a heatpipe to it, in which case the spider might not be too big. I suppose you could also try water cooling through the spider arms as well to the LED. What kind of wattage are you looking at?

Also, at what distance do you need to be off the water for the multiple LEDs to have a decent shimmer vs. the strobing effect you otherwise get? My cube has 4 LEDs maybe a couple inches off the water surface and it's rather tough to look in for any length of time due to the stroboscopic insanity they generate. I'm not sure if more LEDs are the answer or if perhaps cutting down to one of the tri LEDs on a single star are the better idea here. As it is the unit is just sitting around unused due to the migraine inducing nature it has.

Any thoughts on that? I think Vaughn's photonic cannon will get around much of the strobing as it would more closely aproximate the MH pendants. I would guess three of the LEDs on a star would work well on the spider. Depending on the design you could do 7 tri LED stars radially arranged around a center one. Just aim them all up at a reflector instead of down.

-
S

That complex reflector idea will work at spreading the light to all corners of the tank, but the light paths will still vary a great deal in length, so the intensity will not be uniform - due to inverse square loss of intensity.

But, a near point source of light, with a parabolic reflector, with the light source at the focus of the reflector, should result in an almost uniform intensity all over the area being lit. However, since no real life light is a point source, other than than from a distant star, and no parabolic reflector is perfect, there will still be less light around the periphery of the beam than in the middle, but the difference should be much less - I think. And, a single die LED is very close to being a point source. I haven't tried to calculate what a 500 lumen LED, for example, should give in PAR over a 10 inch diameter, using this method. When I get bored with my other projects I will try to do so.

Once you wire up your first multiple LED light, I think you have a great incentive to find a way to do the job with far fewer of them.