How much light actually ends up where we want it? (Reflector idea)

[Flummo]

Something I've been thinking about for some time... Only a small part of the light from the lamp actually is reflected off the reflector, some of it hits the lamp socket, and nearly half (?) of it goes straight out the reflector glass in completely wrong direction. Now, that can not be very efficient, if only half the light hits the reflector that means that half the light ends up where we want it and the rest is spread all over the place.

My idea: Put a hemispheric reflector in front of the lamp too, reflecting the light that would have gone straight out back past the filament and to the ordinary reflector. That might also reflect some heat back to the filament as the Osram IRC, increasing lamp efficiency slightly.

What do you think, would it work? (If it does, I bet someone is already using it...) If you dont understand how I mean I may be able to draw a picture or something...

[Norm]

I'm sure that I have seen exactly that in some automotive lighting.

Originally Posted by Flummo

and nearly half (?) of it goes straight out the reflector glass in completely wrong direction.

Isn't this where we want the light to go?

[Flummo]

I have seen automotive lighting with something that looks like that (that's where the idea comes from), until you look closely: In fact it is not a reflector, but a matt black(!) shield to reduce glare for oncoming traffic. There may very well be lights with this kind of "double reflectors", but so far I haven't seen any. (But I want to!)

We want the light in the direction we point the reflector, not spread over 170 degrees or so as some of it is. "Straight out" may have been a bad choice of words, but anyway... English isn't my first language, you know.

[Norm]

Just thinking about your idea I think you would end with something similar to a reflex speaker (guess mode on) so you would be able to make a reflector more compact, a shorter reflector with similar performance to a deeper reflector (guess mode off).

[Norm]

I've been doing some googling and I this this is what it would be like the diagram is a telecope but if you were to put the light source where the eye piece is it would be what you are describing.

[Norm]

More of the same.

[Flummo]

Now I have used my incredible paintskillz to show exactly what I mean, to avoid all misunderstandings... Not a pretty picture but it shows what I mean, it seems to have the potential to put something like, say, 40% more light where we want it. That would be a big improvement from something as simple as putting a piece of shiny metal in the right place... If I had a suitable metal piece I would have tried it by now, but unfortunately I don't.

[ResQTech]

The problem with this is that you will have ZERO sidespill, which may be fine if you want a spotlight. However, the scattered, unreflected light is what helps to contribute to the nice soft corona around the hotspot that I feel makes a light more useful.

[greenlight]

there is already a flashlight like that, with the led mounted towards the user, and all of the light is bounced off the reflector. What's it called?

[Archangel]

Pelican's Recoil series

[Flummo]

No sidespill could be a problem in some cases, but in other it may be a good thing. If you want to put the light at something 50 or 100m away a dark night you do not want it to shine on the low white fence (or anything else near the lightbeam) 3 meters in front of you, blinding you so you can not see whatever you pointed the light at. A similar phenomenon can bee seen on some motorcycles with a white front fender, the headlight makes the fender alot brighter than the road ahead, making it difficult to see anything. (This happens to be exactly my intended guinea pig for this reflector, more light on the road and less on the fender, up in the sky etc.)

If the second reflector isn't perfectly placed or perfectly shaped, meaning that all the light isnt reflected exactly back to the filament, that should mean that you get some extra spread of the light. Additionaly, if there are areas where you want the light to go past the reflector it is a simple matter of making cutouts in it. In the motorcycle example that might mean that you want some light on the ground to the right and left near the bike, but no light on the fender or up in the sky.

[Norm]

Pelican's Recoil LED Technology™

Pelican's Recoil LED Technology™ works like a reflecting telescope in reverse. By pointing the LED backwards towards the reflector, 100% of the light is captured and projected forward in a super bright, white searchlight type beam that cuts through the thickest smoke, fog and dust. The square beam is proof that all available light is being used - square LED chip, square beam. An exclusive built-in voltage regulator maintains the LED's peak brighteness longer.

From here http://www.pelicanaustralia.com/ledlites.htm

[nerdgineer]

Lightwave Infiniton also works this way. Nice bright beam (about 1000 cd running out of a 1W Lux), zero sidespill. In practice, not all that useful.

[Archangel]

Depends what you need it for. Backscatter can kill you.

[carbine15]

same idea implimented a bit differently. I like the idea of the reflective shield around the emitter. It could be moved in relation to the emitter to create a variable flood / spot light.

[h_nu]

Or perhaps just a deeper parabolic reflector if you have room. Simpicity can be an advantage.

[Flummo]

Simplicity is always good, and I always try to follow "k.i.s.s." (keep it simple, stupid). But in this case, or at least in my intended application, it will be alot simpler to keep the original reflector and try to do anything possible to improve it. Adding a simple small extra reflector should be simple, finding a better reflector that fits relatively good is probably alot more difficult (and expensive!) and then I probably would have to modify it to make my bulbs fit too.

[Ra]

Hi Flummo,

Your drawings are very clear!! But this is a very old problem: Remember the reflective coatings on top of normal household-bulbs.. Same thing: Reflect the lumens going out the front back to the reflector, and they are reflected into or at least closeby the main beam!

What happens with torchlights if you do that: Indeed more lumens are reflected there where you want them. If the litlle spherical reflector is not exactly in focus you'll have more usable sidespill..: You're creating a ghost-image of the filament around the filament itself.

You don't increase throw this way tho..: The aparent surface brightness of the reflector stays the same!

It is not done often because each lamptype and reflectortype needs another approach ! And you have to buckle your mind over another reflector-mount.

Better use a deep reflector with a short focus: Maxabeam ! To cover most of the angle in which the light is produced by the lamp..


Ra.

[Ra]

And as for the Recoil LED: Yes.. 100% is captured BUT NOT REFLECTED !!!

Those reflectors have about 88% surface reflectivity! And a part of the light is obstructed by the LED-mount...

No, the best way with leds are the widely used 100% internal reflective optics made of lexan or whatever material...:



And..



They reflect the entire angle of light for nearly 100% by internal optical total reflection.. The only light-loss is about 4% at the front of the reflector.

Ra.

[eebowler]

Hi Ra. Something I've noticed about collimators. If you rest one on a Luxeon LED and power it up, the side of the collimator (vewing it horizontally)apprears bright. Doesn't that mean that light is escaping from the sides of the collimator and not 100% of it going through the front? (I"m refering to the normal 15 deg and 6 deg collimators that can fit in a minimaglite.)

Thanks.

[Flummo]

Originally Posted by Ra

You don't increase throw this way tho..: The aparent surface brightness of the reflector stays the same!

I'm afraid you lost me there... Thanks for your (very interesting) reply, but this part requires a bit more clarification.

[Ra]

Ok Flummo,

It will be a dificult job, but I'll try to explain:

A reflector reflects a magnified image of the light producing part of the lamp or led. The image reflected cannot and will not have a higher surface brightness than the source ! That is one of the rules of light..
In fact, the surf br. of the image is lower the the source due to reflection losses in the reflective material. (Al: 88% reflective)

Example: Take a smooth parabolic reflector, fit a white luxeon in the focus area and power it with 0.5 milliamps: Glowing very dimm. Look in the reflector and see the much bigger reflection: Also glowing very dimm, it looks brighter because it has a larger surface, but with about 88% of the sources surface brightness.

If you use a little mirror to reflect the light going out the front back to the reflector, you don't increase the surface brightness of the source! For that the extra light you reflect should come exactly from the direction of the source. BUT THAT IS IMPOSSIBLE: THE SOURCE IS BLOCKING THE WAY!

So you need to place the little mirror just out of focus to generate an image just around the source (filament of the lamp).

The result: Yes, more lumens are reflected forward where you want them. This results in a wider beampattern with the same throw as before, because you increased the dimensions of the surface, not the surface brightness!

I hope you undestand this...


EEBOWLER:


There are more types of collimators, exactly matched to the shape of the emmitter-dome, if there is no match, the collimator can leak light. The internal 100% reflection is optical only: there is no reflective material like Al.
Any deviation from the perfect form can kill the 100% reflection in certain areas, light-leak is the result!

But things like sckratches, dirt, crease and stuff also definitely cause them to leak light from the side: THEY HAVE TO BE UNDAMAGED AND ULTRACLEAN to maintain the 100% internal reflection! Only then they give most torchlumens out of bulblumens!

Edit feb 11 2007: As any optical element, the acrylic optic has losses: with 100% internal reflection, I do not say it has 100% efficieny! These optics, if very clean, can have up to 90% overall efficiency. This is much higher then the about 65% efficiency of conventional reflectors!


Regards,

Ra.

[eebowler]

Ra, thank you for the very clear respons.

[Flummo]

Okay, I think I understand.

Now I'm trying to find a reflector for some testing, I think I have one somewhere. I like what I've heard so far, so it's worth doing some testing I think.

[uk_caver]

Posibly it's a little off-topic, but I use reflectors with LEDs in a bit of a non-standard and possibly niche way.
I make light units for caving, to replace the original reflectors (~55mm diameter, 25mm deep) in mining headlamps. With a regular halogen bulb in the reflector, though there is spill outside the spot beam, there's often not much light down at the feet.
I mount Lambertian emitters facing sideways into half the reflector (with the reflector in use pointing forwards, the LED points down). Depending on alignment when mounting, this can actually give a pretty good spot beam which appears quite adequately even and circular in actual use, even though if firing at a smooth painted wall it can look a little more irregular. There is very little spill close to the spot beam, and the bulk of the light that misses the reflector goes downwards and sideways close to the user, where it is actually pretty useful.
One bonus for me is that I actually have nearly half the original reflector volume free to mount a heat spreading plate and driving electronics and still have the overall unit fit as a fairly simple replacement.
Finally, another feature is that even if the LED spot beam is running fairly bright, there is hardly any light above the horizontal axis, so if interacting with other people underground, as long as you keep your light pointed at or below their chin, you don't risk messing up their vision.

However, unless someone wanted to make a really unusual-looking torch (like one with a semicircular head?), I guess it's maybe only practical for head-mounted lights - a torch that had sideways beam asymmtery might be somewhat less use than a headlight with some deliberate downthrow.

[Ra]

You're right uk_caver.. Its a little off topic ! This (quite old..) thread is about "how much of the emitted light do we actually grab and use with our reflectors"

So setting up a reflector-emitter configuration to enlighten a certain area of interest is not discussed here.

You can try this in the "general light discussion" area of these forums.

Edit: Ohh This is the general light discussion forum !!

OK, maybe I meant to say: Start a new thread about this !! and post pictures of what you have made.. I think there are members that would be interested !


Regards,

Ra.

[uk_caver]

Thanks for your reply.
Particularly as a new member, I was a bit unsure about starting a new thread, partly wondering what to say, and partly wondering if anything I said might be at all relevant, or would be ignored and/or seen as self-promotion.
Also, I do sell stuff, and though it's pretty unlikely anyone here would be a potential customer, I'd hate for people to get the wrong idea about why I was here (to buy stuff, learn things, and keep up to date with developments) so I thought I'd be safer in an existing technical thread.

Regarding posting pictures, what I make tends to be functional, but cosmetically maybe not on the level that people round here seem to reach. However, I'll see if I can get some shots that look reasonable, and build up a bit more posting confidence.

[2xTrinity]

Hi Ra. Something I've noticed about collimators. If you rest one on a Luxeon LED and power it up, the side of the collimator (vewing it horizontally)apprears bright. Doesn't that mean that light is escaping from the sides of the collimator and not 100% of it going through the front? (I"m refering to the normal 15 deg and 6 deg collimators that can fit in a minimaglite.)

This is why I think a nice combo is to have a collimator inside a typical reflector. This way, the light that would otherwise go "forward" out the front is collimated, and any light that spills out the side is at least reflected by the reflector in the genreal direction as you want. This gives you some side spill, but still concentrates most of the light into the beam -- IMO the most useful combination. Usually I don't like lights that are 100% throw with no spill. On the other hand, if I want flood, I have other lights that deliver mostly flood, such a photon clone using a completely sanded down LED for reading without any hotspots, and a couple of Cree retrofits that I use for lighting things up at close range.

[Ra]

Hi 2xTrinity,

I think you miss the point of the quote you posted:

Its not about the lightloss at the front, its about the lightloss at the side!

And I already posted that with 100% internal reflection, I did not mean 100% efficient, those are two different things.
Ofcource they have lightloss, every setup has ! But the lightloss in those optics is WAY, WAY less then the loss you have with conventional reflectors!

The collimators I wrote about alredy have a collimating lens inside, thats why they are 90% efficient ! With these emitters (Luxeon, Cree and Seoul) most of the light is emitted at the front, within a angle of about 40 degrees.. Conventional reflectors only grab what is emitted at the side, more suited for 'normal' lamps (hotwire, short-arc and HID)

Members that use conventional shaped reflectors with these emitters are simply throwing away alot of lumens! Because the light that is emitted within 30-40degrees at the front is not grabbed by the reflector and, on top of that, the reflecting surface mostly has a reflection efficiency of 88% tops (mirror smooth aluminium)!

No reflective material reaches the 100% of optical total internal reflection!

And as for throw: The high efficient acrylic optic is designed especially for use with these emitters: They have a very tiny focal spot, the actual emitting surface of the led is much bigger, so it already has the max throw it can have, and the emitting surface thats not in the focus generates sidespill.

Many members forget, or do not even know what huge difference in actual torch-lumens (and throw as well!) you encounter when comparing an acrylic optic with a conventional reflector with the same diameter.

Beleve me.. ITS HUGE !!

Maybe a few things I said are hard to understand, they are hard to explain as well ! But if you have specific questions, I will answer then the best I can..

I'm thinking of starting a thread about this soon, however, with my busy life I don't know how soon.


Regards,

Ra.

[uk_caver]

And as for throw: The high efficient acrylic optic is designed especially for use with these emitters: They have a very tiny focal spot, the actual emitting surface of the led is much bigger, so it already has the max throw it can have, and the emitting surface thats not in the focus generates sidespill.

I'm really a layman when it comes to optics.
Could rather narrower-beam acrylic optics be made if they were permitted to be somewhat larger? Is there some relatively simple relationship between physical size, efficiency and minimum beam divergence?

I imagine scaling everything up except the emitter would give the same geometry as having a smaller emitter in terms of the divergence of a beam (albeit with the beam 'starting off' from a larger frontal area), but I'm not sure if I'm imagining correctly.