Has anyone tried or know the theory of using a G12 Base 150w Phillips SA bulb with the following ellipsoidal reflector:

http://www.edmundoptics.com/onlinecatalog/...?productID=1412

It looks like the G12 150w bulb will fit the hole on these reflectors. Can we incorporate it into our fresnel/precondensor configuration? I did some measuring and the light chamber would have to be pretty long. Does anyone know how the focal lengths would interact between the ellip. reflector and collimating fresnel? Just curious because the look like REALLY nice reflectors and the don't collimate so maybe they are useful with our G12 ceramic bulbs. Please post your thoughts/experience.

Thanks

Hi Sinner7,

I'm trying it in my project, have not fired lamp yet. Small details to finish light engine before igniting.

Plog here http://www.lumenlab.com/forums/index.php?showtopic=20875

Planning on using 550mm first fresnel and 330mm secong fresnel.

Me2! has a plog.... http://www.lumenlab.com/forums/index.php?s...mp;hl=faceplate

I have the larger of the two Elliptical reflectors for sale if anyone is interested (PM me)

I’ve traced these reflectors before and they never seem to look like they would work well for a 15 or 17” LCD. But I just drew another and it appears that the smaller of the two looks close for a 10.6” LCD and a 317mm fresnel. Of coarse you’ll still have to deal with the void, maybe an integrator could be used.

DJ

EDIT:
You could construct an integrator with some mirror strips, light pipe.
I remember ROX playing around with a similar reflector and he said he got a spot size of 10mm, I’m a bit skeptical though. A very small arc maybe could be 10mm but not both ends of a larger 10mm arc. Perhaps he was only viewing one end of his arc, where the other was converging or diverging.

generally a elliptical reflector isn't working well in a projection system because you don't get even light.
Some kind of integrator is needed.

So that brings this concept to the integrator or light pipe. Edmund has some nice light pipes. Here's my scaled drawing of the design, I'm going to contact an Edmund engineer and get their opinion. Any ideas?

Click to view attachment

I know it will make for a pretty long light engine but it would be a bit more efficient to collect light from all sides of the lamp arch and homegize them into one dense beam.

My question is can we fit the focused reflector light into the end of a 5mm hexagon light pipe since it's much shorter than the 10mm light pipe? Anyone own the 150SA G12 bulb and the smaller refector to measure it?

Well I sent the diagram to Edmund and spoke with an engineer and he's going to take a look at the design and email me which light pipe/reflector/lens combo will work the best. I hope it's good news he said it would take a day and its free .

Interesting - I've been wondering about using a reflector like this since I saw inside the lamphouse at the theater.

Something interesting I learned there - if you take the lens off of the projector while the light is on, an image of the holder for the lamp end is projected onto the screen; you can see a large void in the center, and you can also see the 3 arms that hold the lamp. But when you put the lens back in, the light is perfectly even. I think that maybe if you put the light pipe at that point (where the film would be, just before the second focus of the reflector) then the light would come out even. I would like to run the theater lamp while the lamphouse is open, and take pictures of the light pattern at different distances, but I don't want to break anything

Well in theory the lightpipe integrator should homogenize and even out the brightness since that's exactly what it's designed for. The next challenge is using a fresnel to expand the condensed rod light. Maybe a post-rod precondensor will help expand the rod's output. We shall see

I haven't tried these reflectors in my setup but I'm guessing it looks like this without a homogenizer, the bottom pic show a perfect homogenized light pipe face ready for magnification :

Click to view attachment

There was an article about shaping the end of the lightpipe to a lens shape to spread out the light.
I will try placing a condenser lens at the end of the pipe in my light engine when I get there.


Just saw that the idea was used in your diagram.

Since you are going to try a light pipe, why not use a square or rectangular pipe?
I realize that Edmond say that a hex pipe is more efficient than a square but in this instance I think the square would be more appropriate for a square LCD.

DJ

Edit:
And if the arc image is too large to fit the pipe, it may be possible to stack more than one.

I was think of starting with a square entrance tapering to rectangle ( 10:16 ). WUXGA

I did try using a piece of square aluminum tubing, I did give even light. I used a ratio of 12:1 lenght/cross-section. Tubing needs to be polished more.

I also have the fiber optic faceplate to try, seems to lose to much light though.

Ok I tried to make a rough effiency estimation with this diagram:

Click to view attachment

I like the hexagon light pipes for a few reasons:
1) They are designed and made with precision. (I don't look forward to cutting precise kalidescope mirror sides and mounting them precisely.)
2) They have high quality coatings.
3) They collect an even amount of light from the entire reflector output.
4) It appears more efficient to start with as much light as possible (hexagon pipe) and trim it down as it travels into the system.

As to which orientation A or B I'm not sure. Figure A would send more light to the panel but may show more dimness at the corners. Figure B would send less light to the panel but it may be more homogenized. If these pipes are very evenly lit from corner to corner then Figure A is my guess for sure.

I know real life testing is the only way to find out, I'm just talking some theory here for fun. I should hear more from Edmunds tech support today. Any more thoughts fellow luminaires?

NOTE: This diagram puts the light pipe inside the reflector beam, in reality the light pipe shape would swallow the entire beam.

Oh boy - another chance to ask a stoopid question.

Would a round cross-section truncated cone (or not conical?) light pipe (or even oval in cross-section?) work better/worse/same than one made up of planar surfaces?

Here is an article regarding lightpipes for projection.

http://www.patentstorm.us/patents/5829858-description.html

Yowzer! That pretty much covers it. Lunchtime read.

That's a good question, I believe the reason they use flat sides is to shape the "reflection population" as put in that posted article. So maybe matching the light pipe to fit the LCD aspect ratio is better, though it doesn't collect as much reflector light as the hexagon. Somone could make a round or oval light pipe easy with polished alluminum. I don't know the proper taper angle to achieve. One source I have says the average comercial lightpipe reflects each ray about 3 times before output. So that wouldn't be a very long light pipe, I have a feeling ours would have to be longer since we are dealing with larger lamp archs.

Awww, shucks. No pictures.

Good read though.

For a light pipe to work efficiently the arc image should fit into its entrance. It doesn’t have to fit exactly. If the light entering is conical it doesn’t need to fit into a circular pipe. The pipe could be square, circular, hexagonal, whatever shape. The shaping and evening takes place inside the pipe from reflections

If you conical beam from the ellipsoid reflector doesn’t fit the hex pipe, arc image is to big, simply stack three together so that the beam isn’t wasted.

DJ

Doesn't a ellipsoidal reflector with a focal point change the arch image into a donut shape? My understanding is the the light pipe/ellip reflector combo really transforms the arch image and homogenizes it.

In other words, the light pipe output becomes the new arch image so getting the most reflector output through the pipe creates a better pipe output. The eliip reflector reflects a 3d image of the lamp arch and it throws it into a donut shape that needs to be mixed and condensed. It just seems to be a different beast than our traditional concave reflector/lamp arch/precon.

I’ll draw a pic to explain what I’m trying to say.

The light converging from the reflector should fit inside the pipe. There is no need for the pipe to fit the shape of the beam, like when we try to optimize the circular shape of pre-con to the shape of the LCD.

Here's a crude pic of the arc image, in real life it would be no where near as defined as this:
Click to view attachment


Now when an arc image that is too large for the pipe:
Click to view attachment


And the way is should be:
Click to view attachment


Or a cross section:
Click to view attachment

I understand what you are saying, now if you beam is round and the pipe is square do you think the voids at the corners of the light pipe entrance affect the homogenized result? Doesn't the hexogon pipe have less voids and more angles to homegenized more efficiently. I know any shape pipe will collect all the light if the beam fits inside it's mouth. Why wouldn't a hexagon homogenize better than a rectangle? I just having fun with this.

There is a void in the middle of the beam. What the light pipe does is to create multiple arc images inside the pipe, the same as looking into a kaleidoscope. This multi arc image effect is similar to the way a lens array creates multiple images of the arc.

It’s really had to explain this so here’s a little experiment for you to try. Make a crude light pipe from a square section of aluminum or mirrored card board or what ever you can find. Put a small light of LED into one end (the arc image) and look through the other end. It should make this allot clearer to understand.

DJ

Edit:
You should see an array of arc images created by multiple reflections. All the images serve to even the void out, the same as multiple ceiling lights in a room stop shadows.

The longer the pipe the more reflections there will be so using the ceiling light analogy, less shadows because of more lights. But there is a catch. Each reflection looses light.

So a hexagon will create even more virtual arch images. Wouldn't this translate into less variation because the dark areas are smaller (but also more numerous )? With a 10mm lightpipe face, we have to magnify the face image about 26 times to fill a 10.6 lcd panel. Given the large magnification requirement, it seems more homogenization is better. Does the hex homogenize more than a rectangle?

I'm also aware of lens arrays, I just hope to find a way to use a lens array or light pipe in our designs. I don't thinks it's impossible, the trick is making it compatible on our larger scale and compatible with the collimating fresnel.

I believe it would. Going even further a circular pipe would be better again. But you will be loosing one of the light pipes features in that shaping the beam to better suit a rectangle LCD.

So carry it out to the extreme. If you morphed a circular pipe into a rectangular pipe and had the right taper is that the end-all light pipe of the DIY'ers dreams?

Perhaps

I'm email corresponding with Edmund right now and they are suggesting a holographic diffuser which is apparently 90% efficient:

http://www.edmundoptics.com/onlinecatalog/...?productID=1363

I'm asking how it will work with the elliptical reflector. These diffusers come in different angle profiles, I need to find out how to match one with their ellip reflector. Interesting...

What about fiber optics like this one. At 2mm it might be too small but if bundled into a large straw or tube could it work? Or possibly find a larger diameter?

For a lightpipe could a set of long, thin front side mirrors be used?

If so; then if you got a bathroom mirror or something, how could you cut the mirror down to the right size pieces to make the lightpipe walls?

I built a rough test bed over the weekend with a diy light pipe/ ellip reflector and fiber optic faceplate. I'll post a report about it soon. I did get a nice homogenized image but I'm not sure how much brighter it was. I'll explain it further when I have more time.

Sorry guys been a while..found the patent with pics go here and register

http://www.freepatentsonline.com

then put 5829858 in search..then click on the result and then open the PDF to view...

I tried to upload but the size of pdf is above the limit..pdf is 293kb

Wow ! there is even a twin lamp light pipe reflector.

The designs appear to be for smaller panels however.