I’m not sure where to start.
There has been some interest in light pipes and lens arrays. From some of my experiments and raytraces I’ve been able to develop a basic understanding of how they are supposed to work and some of their faults. From what I can tell they both do a similar thing, even the screen illumination by overlapping multiple light sources, reflected in the case of the light pipe or refracted from the lens array.
The light pipe is usually used with a diverging light source such as an elliptical. The rays are directed into the opening of the light pipe were they are reflected multiple times and then emerge from the other end with the same angle as it entered. The problem with this approach is that each reflection creates an image of the source ending up as a matrix of light sources. The more reflections there are inside the light pipe the greater the number of cells that make up the matrix and the greater the reflection loss there will be.

The first idea I came up with was to use a light pipe that had tapered sides. And instead of using a reflector to collect the light and direct it into the light pipe I thought to use the tapered light pipe as the reflector itself.
Click to view attachment

The sides of the pipe need to be very close the arc so as to keep the matrix small. The further they are away the larger the effective light source will appear to the fresnels
Click to view attachment

This still produces an image at the triplet that is too large because of the magnification from the fresnel combination. So to use this method the rear fresnel would need to have a longer FL than the front, the magnification now becomes negative and the arc image shrinks
Click to view attachment

So using a 600mm front fresnel means that the rear will soon become very long.

To be continued…

Obviously this projector would be too large. But this was using full size fresnels so shrink this whole optics assembly and use it in a similar way as you would an elliptical set-up.

Click to view attachment

The advantage of this approach is that it will collect allot of light and at the same time even and square the light onto the LCD.

I like that!! Where exactly is the arc in those pics? Can you add a spherical relfector behind it?

Here's a quick test reflector I made that shows the multiple arc reflections

The arc is at the apex of the light pipe.
The idea is to collect the light from axially around the arc, the ends emit very little, so a spherical may not work. I’ve had a chance over the last week or two to do some preliminary tests using small aspheric lenses that show promise. I just need to figure out what angles and FLs to choose. It may even need to use multiple lenses to get the correct FL.

DJ

If the concept does work then it would be a very efficient light engine. Reflecting nearly all of the light from 360 degrees around the lamp and evening it out in one swipe. Reflection loss would be limited to one or two reflections. AND no void to deal with like the elliptical and parabolic.

what lamp are you using?

For quick tests I’m using just a 5mm filament halogen

I use the same type of lamp for testing, works out well. One can power it on a lower voltage and not blind yourself.
What is that test reflector made from?

Just polished aluminum sheet. If it proves itself then I would use some cold mirror cut to the correct sizes.

That shows real promise.

Would the light chamber prior to the tunnel allow for multiple bulbs?

DAZZZ looking good !!


Btw did yo sleep last night ? I went to bye..byes about 1am as my eyes were starting to get as bad as yours.


Will the FS mirror handle the heat so close to the arc ?

Can you work out from the test setup and the the raytrace program the dimension for all components...

Not yet.
I'm still working out the effects of different components and what the limits are.

DAZ - how different/similar is your small light guide/reflector compared to one of these?:

http://www.bulbconnection.com/ViewItem/bcr...c/FXL/item.html

I have a couple of these little lamps. I've concluded that the tiny facets make it act like a larger parabolic to provide a more even light distribution - still falls somewhat short.

It looks like your design acts more like an elliptical reflector...so there's one big difference.

DAZZ could you use one mini diverging/negative fresnel instead of the first two mini positive fresnels...

Their similarities are that they both reflect light and the both create a large light source. The key to my idea is the combination of three things, collection, shaping and shrinking. It could be possible to use one of those lamps and just add the shrinking optics.

Lots of experiments to do yet.

Scrub that... hmm thinking that you could one small enough ..you would not have a focused arc from using a small diverging feresnel..

Wow, you've eliminated a lot of issues with this engine.

You might want to do a patent search on this idea.

Do you think that the result would improve if you were able to fabricate a conical light pipe as opposed to the pyramidal shape?

The original idea used a conical, then the pyramid shape was added to square the output. The conical shape does have the advantage of creating a small light source because all the surfaces are closer to the arc.

Do you think it is better to square the output? I'm trying to visualize the third dimension on this... Wouldn't a conical shape more accurately follow the model you have drawn in two dimensions?

You would end up with a ring of light rather than 4 arc reflections. That might even things out a bit, but then again, it might make things worse.

But since we are applying the output usually to a 4:3 rectangle ...

Yeah... but aren't we always just working with a cone that is cropped to a rectangle?

I would think that a fresnel would ideally "see" a round source as small as possible.

Maybe I'm just being a little thick today. I do see how a squaring the light source could be a good thing, but I wonder if the corners are going to be invisible and if any patterns are going to be introduced by the shape.

Interesting....what are you using for shrinking the light source?

I thought about using a condenser with the little FXL lamp - to gather more light. But that would enlarge the light source and likely require an even longer FL fresnel...

A combination of the two would probably be the best. Circular at the apex and squared at the exit. This should make use of the otherwise wasted light.

Makes it a bit more work to fabricate...it's doable and would be more than worth the effort..

I intend to use lens magnification to shrink the light source. Magnification of the image occurs when the object distance is less than the image distance. The opposite is also true. When the object distance is longer than the image distance the object will appear smaller. I won’t be using fresnels to do this because of their facet inefficiencies.

Here’s the latest reflector I’m testing.
Click to view attachment
Hmm a telescope



And the ray trace for it.
Click to view attachment

what motorbike did you steal the mega from .. heheheheh

This one

The second lens in the ray trace above is used to adjust the ray angles presented the real fresnel. If it is moved to far forward it increases the magnification too much. So choosing the correct fl of the first lens and the distance to the source is important.

So far I’ve been able to illuminate a 17” fresnel and collimate the light. I’ve only held everything by hand so I’ll need to make a prototype to get some accurate tests. That’s if I go with this design.

DJ

What's the actual length of that megaphone?
Hmmm..... I wonder how it would work if you replaced the MH lamp with high wattage spiral CFL.
Lux measurements?

DAZZZ

What the !!

A horn of plenty ?! plenty of light....

Raytrace looks great...where the...would one get one of them there um...horns !?

That’s an interesting question. My gut feeling is that it would create an image that was to long so the rays after the first lens would be harder to control. I’m also thinking about trying a 100mm parabolic and converging the light slightly. To many test, not enough time.

It's about 300mm long.

DJ

I was just thinking about that. I’m getting some strange rings that I’m almost certain are from the surface irregularities inside the reflector. If it works then perhaps it could be turned on a lathe.

DJ

It's not as long as one imagines it is in the pic....
perhaps a sphere on the end or like with a high lumen compact fluoro..

It is amazing how quickly it flattens out those rays....

It's not to big so a lathe could do it..and it wouldnt have to be worked on the outside ..(taper only on inside)

If you mean how close it converges the rays, that’s because I’m using very short FL aspheric lenses. It would be good if I could find larger diameter aspherics as it would mean that there would be fewer reflections inside the relector.

DAZZZ

what is the FL you are after and I will have a search tomora.. Im guessing the diameter is about 2inches what you are using..

Cheers and good night !

I just tried the compact fluoro in a ray trace. It didn’t look very good. The further the source is from the apex (vertex?) then more it magnifies the image. The fluoro is just too long.

Just to get a perspective of what is happening. The filament is 5mm axial. With the reflector it becomes a 30-40mm diameter ring. Then the lens magnification reduces it back down to 5-10mm

DJ

The best one I have ATM is about F0.7 (50mm diam, 35mmFL). If it were 100mm diam then the cone wouldn’t need to be as step.

Dazz,

Could you do a easy test for me please? Remove the lens from the end of the megaphone. Measure the diameter of the
large end of the megaphone. Now power up the light and measure the beam diameter at about 12 inches away from the
same end of the megaphone. I'm trying to figure out how parallel the exiting beam is - thanks.

Yeah, I was wondering the same thing.

As far as elongated cones, it's a good year to look for some. Seems like there around Christmas you would see more chrome trumpet shapes. For that matter, if you just find the right shape, you could have it chrome-plated for high reflectivity.

I wouldn't throw out the possibility of making a mylar cone/megaphone structure, though heat might be a problem. Or aluminum flashing for that matter.

Or perhaps a cone curled from a flexible mirror.

DAZ - love your originality and willingness to experiment and try new ideas.

Your raytrace shows most of the light getting reflected 3-4 times before passing through the 2 "condenser" lenses. Wouldn't this significantly affect the efficiency - assuming typical stainless steel reflector?

How about taking a regular small parabolic reflector - like in a commercial PJ lamp - and then somehow condensing/shrinking down that resulting arc image?

This is a real stinker of a math problem. Oh Rox - where are you?

Given a megaphone built at a certain angle, light will bounce how many times before it exits if sourced from a given angle
of the lamp? And what is the final bounce angle?

60mm opening, 220mm image, 300mm throw.

Yes it does. The first square reflector only has 1 or 2 reflections. The narrower the cone angle the more reflections occur. Here’ a pic of the light pattern that shows each reflection and successive reduction in light. You can also see how deformed my reflector is.
Click to view attachment


I would use aluminum rather than stainless steal

I really think that to make the void in a parabolic blend, the reflector shold be much larger than the lamp.

DJ

Keep up the good work Dazz. I always like to see what you're coming up with.. it's funny to see a thread that will stop and wait until you weigh in on the topic
Nice pic- experimental setups are just that... I wouldn't expect perfection just yet... (unless it was made of MDF by pun15her )

My goal is to use a 150W powerball and achieve around 150 ANSI. Using this approach, which may turn out to be a flop, or some other method.

DJ