After looking at the Parabolic Reflector test and the great real life experiments done by "lucky me" and others I propose a similar and hopefully a very doable solution to making a more efficient light engine.

My conclusion from looking for an easy answer for the Parabolic Reflector dark spot is not to use a reflective dish but use a fresnel instead.
A Parabolic fresnel or a curved fresnel.

A perfect Parabolic frensnel would be a 3 dimensional one. Yielding the higest gain and the most even light. Have a 180 degree encasement of the lamp arc and have a spherical reflector on the other side. See the second drawing.

However while the efficiency would be great this solution would be almost impossible to make. Everything would have to be custom made. large molds and lots of money.

So what can be made that could approach the efficiency of a Parabolic fresnel?

A 2 dimensional curved or parabolic fresnel designed made for just that purpose.
This frensnel would have the following properties.

1. A focal point of 300mm [more on this later]

2. It would be made of PVC so it could be curved.

3. made to fit 17" and 19" LCD

4. groove pitch could be large .5mm or more

A focal point of 300mm might be optimal so a full 180 degree curve could be maintained on a 17" LCD plus the increased distance will allow for lower temps

PVC might be considered crude for a lens but it has several advantages for this use. It's bendable, this is a must, the lower transmissiveness is really very small and it is lower cost than acrylic,a plus. It does melt easer but this can be delt with.

A test lens could be 300mm x 550mm this would cover 17" and 19" LCD

The real advantage of this setup is that it could accept a longer arc lamp by using a slightly flattened out curve the fresnel will see a larger arc.

It's more efficient, maybe 2 or 3 times that of the current system so a lower wattage lamp could be used.

This lens could be made just like any other fresnel except it would only be in focus when curved.

Any curved shaped could be used. A 2D parabolic would give more even light and slightly higer gain but would a lot harder to frame in the proper shape.
A simple half round would be the easiest to form.

Any comments Welcome

Dave
Click to view attachmentClick to view attachment[atta chmentid=12398]Click to view attachment

It's interesting.

I've never seen a nonplanar fresnel, are they even possible???
You show a parabola and a semicircle but for this to work you would need a 3D surface i.e. a rotation about the central axis of what you have draw and it would need to be perfectly sized to work.

No 3d just 2d. It would be better to go 3d but then it could not be made the same way as the curent fresnels so we leave the ends open
What we would have is a curved fresnel that would still be in focus. flat it would be a mess just like if you curved a nomal fresnel
To the fresnel maker it just a strange lens. Just type the numbers into the computer and bang your done.

i wonder if there is other way than starting from a solid block (3d) and eat all the unnecessary material...

making a flat 2D fresnell, you can´t work out this 3d "object". if you push a fresnell from the center, it will bend a bit, but not as much as you want.

To rox: were using PVC it bends look at 3dlens and look at their PVC product.
If this is ever going to work it first has to be cheep. Cheep is better than perfect. perfect means it will never be made. what I am sugesting is just a PVC fresnel cut differently. the finished cost in production could be $10.00 after the start up.

Another way to look at it is take our current fresnel bend it 1 degree in only one plane toward the lamp. Now the lense is all whacked so re design it to be corect. If it can be "fixed" for one degree then it can be "fixed" for 180 degrees. That is what we would have.

If I remember correctly, the very first fresnels ever made, (designed by Mr. Augustine Fresnel himself) were used in lighthouses, and were actually a collection of prisms surrounding a lightsource. They were curved as you suggest.

HERE WE ARE! Found a perfect explanation of both the origin of the fresnel lens and a pic of the 3d fresnel lens as well: http://www.lanternroom.com/misc/freslens.htm

But really, we should definitly look into these round fresnels. Really they would make the most ideal light engine possible, completely surrounding the bulb, and with a good reflector, absolutely no light would be wasted. Also the edges will actually be recieving more light then the center which is good since most of that light is reflected away from our eyes so it would even things out.

But I bet round fresnels are expensive and difficult to make.

I estimate this would make about a 150% increase in brightness.

We would still lose the light at both ends top and bottom. I was thinking to have it made flat just like current fresnel then curve it later.
Great drawing. Thanks for the input

Might work but curving would probably result in no focal point.

Making a round fresnel I suppose would require an overturned bowl mounted on a spinner, cuts would have to be made in the side of it at the appropriate angles.

A fresnel lens has many small facets on the surface. On a normal fresnel each line is at a slighty different angle. On this lens their would be a large difference between each line almost as if the lens had a fl of 2mm then spread out untill the fl was infinity.

Another way to look at this. A fresnel is not realy a lens it is a series of small prisms that act like a lens. If you look at placing a series of small mirrors in a curve each can have a perfect focus on the arc. You can't do this with a normal lens it would have to be 20" deep. But a curved fresnel can. If a PVC fresnel can be curved then this can be done at a low cost.

A frenel is really a series of prisms ( reflraction not reflection), the fresnel may be flexable but this doen't mean that it maintains its optical properties in anything other than a planar form. The prisms are cut into the plastic to bend light, each one moving out from the center, a little more than the last. If you bend the plastic the prism, and thus the light bending charactareristics are the same but the prism isn't pointing in the same direction so the bending will not line up as it did when the fresnel was flat.

Bad wording on my part[ will fix] but the example was to show how it could work not
how is does work. Work backwards pretend to cut the prisms to the correct shape when the fresnel is curved then flatten it out that is the shape we need to make. It may be that it doesn't fit but I bet we could get close.

to Mikau: Great find now I can't wait to start hunting ebay for used light house fresnels! few minunts with a glass cuter and bam instant PJ
I know that this is not a first. Curved fresnels are used in police lights to make the beem focus foward.

To pagercam: good point about the prisms. It is probable that the prisms in the horizontal plane will distort alot this is bad. The top will be pulled down thus changing their focus. The vertical prisms should have very little distortion
The actual shape of the cuts that make the prisms looks to be very robust and the deformation should be very uniform in the horizontal plane therefor it should be easy to correct for.
[edit] Many small cuts could be made in the vertical plane to almost totally eliminate the stress on the prisms when bending the Fresnel. This will leave many small voids in the light pattern but this effect should mostly be filled in with stray light and could be reduced if necessary by using vary thin shallow cuts. It looks like the fresnel could realy be curved and not distort.
A 160 degree might work almost as well as 180 degree coverage. There is no reason to focus on the lamp base.

Actually I think police car fresnels, and many other lighthouse fresnels, are just cyllindrical fresnels rather then spheres. But see what you can find.

I put the frsenel in a curved fixture in hopes to lengthen the effective focal length. The result is an oval light pattern at the triplet. This got rid of the dark corners.

The test was with a 330 mm EFL and the 135 mm triplet. The curve was cut with a 1/4" router bit then I filled half of that with a glued in wood strip.

Click to view attachment

Click to view attachment

Close up.

Did you curve all four sides of the fresnel or just two?

Just 2 sides. The light that hits the triplet is oval in shape. Closer to the format of the LCD.

It does get rid of the dark corners. I did not notice an improvement in brightness.

rlwoodjr: You are making DIY projector history. Awesome.

arizona: A machinist might be able to make a jig to suck PVC or whatever plastic fresnel lenses are made from into a semicircular shape and then a CNC lathe might be able to carve the grooves into it. I don't see machining these things from solid blocks.


This whole DIY LCD projector movement, this site, and the enthusiastic amateur support going into developing engineering solutions to problems that no commercial venture would be willing to consider is really amazing.

My best freind has been making plastic molds for 20 years it's sooooo much to have one made here but if we could come up with something that looks like it could realy work then I or we pay the startup cost for them to make thousands.
The guy from 3dlenes made a pro 550mm fresnel on his own before even a real lens to go with it is out. I have the money if I had to but there has to be a real need an a 99% proof that it would work. Brain might not like comp here so if he won't sell them and make it worth it for me then it would be a lot harder to sell them but first it has to be proven and pass the ROX test and if everyone scream for it then we will see.
It may be that just the polar mods might make enough of a diffence that this will not be needed. looks like they may get 20% to35% more gain.

Once the polarbear issue is nailled, we'll be looking for extra lumens elsewhere, I can guarantee that! I like the direction you are heading with this, keep it up. The link that mikau posted showed the fresnels wrapping around the source. It looks like whatever fresnel you come up with, it should be machined on both sides, not just one like the current ones we use.

Just my 2 cents
I once saw a guy's website who had made his own handheld dreamcast. He custom built his own vacuum forming machine to create the casing for the handheld console.
He used a household/industrial vacuum cleaner mounted below a chassis.
In the chassis, he sat a mould he had prepared (from foam/polystyrene).
Above the chassis, was suspended a coil similar to an electric bar heater.
Between the heater and the chassis was a mechanism for raising and lowering the material he wanted to work with.
The idea was to heat the material until it was flexible enough, then lower the material onto the chassis and apply suction from the vacuum cleaner to draw the material into the mould.
The whole contraption looked very simple and cheap to build and his results were great.
The handheld dreamcast resembled a professional product.
Would it be possible to get a flat fresnel of the correct dimensions, and then create a mould for the fresnel to be drawn into?
I'm out of my league here as to whether this would make the fersnel useless from our optical point of view, but the vacuum forming method looked so simple and effective.
Maybe some googling would uncover the site, as I'm sure it would explain it better than I have here.

Well, I found it.
Here you go, anyone that's interested http://www.halloweenfear.com/vacuumformintro.html

Actually, these are easy to make and work well. I still have an oven and table in the garage from when I use to do RC helicopter bodies.

The vacuum table is a wooden frame with a pegboard top.

My oven is a sheet metal box in a bigger wood box with insulation in between. I used electric heating elements and a thermostat to control temperature. I also added gravel in the bottom to help maintain a steady temperature.
I made my own oven so that the door could be on top (also it kept the wife happy) You can use an electric house oven.

Then you need a metal frame(mine are aluminum) to fasten the plastic sheets to. The plastic can be held with office supply binder clips.

You can form plastic up to 1/4" (6.4 mm) thick without using a 2 stage table.

Its real easy to do. The pre-teen guys in our church made primitive soccer playing robots and vacumm formed the bodies

It's the master thats the problem. if you could cut a master you might be able to use casting resin.

SMART.

edit

some fresnels are stamped a machine shop could make the mold on a cnc no problem.
found this online
"One of the advantages of such a lens is our ability to stamp it out of a clear piece of plastic, much like making a record (disc). A mold in the shape of the prism surfaces is pressed into the plastic, leaving us with a flat lens which works the much the same as the thicker, heavier lens.