Since it seems we're not going to find a practical light source more powerfull then a 400 watt halide, I was thinking of ways to capture more of the light the halide is producing, I got an idea for one but I don't know how much of an improvement it would make.

The idea is using a larger collector with a longer focal length and rear mounting it.

The pic below explains. This uses 16x12 fresnels. Using similar triangles the focal length would have to be roughly 400 mm, give or take depending on what your scaling by.

Anyways, a 16x12 inch fresnel has a surface area of 192 square inches. A 12x9 lcd only has 108 square inches. So we'd be getting about a 77% more light.

I'm sure it would work but first I don't know if we can find a large 400 mm fresnel, and second I've often noticed that decreasing the projection size to make a brighter image, doesn't make a dramatic differance. Then again, this often involves moving the triplet far out of the collectors focal point so theirs probably a lot of wasted light.

Opinnions?

That is almost what we can get if you use a395mm x395mm frensnel and a 15 inch LCD. If you go to a longer FL you lose just as much as you gain by having a larger frensnel. but the idea with jyst a larger frensnel will be brighter that is what I plan to to but for different reasons.

Yes, that will work. You lose keystoning ability with the unsplit fresnels, but it'll work.

For "moving out of the focal point" -- that's true, to a degree. You can adjust the lamp position to compensate, however. It does work...

how will having a fresnel bigger than the lcd help exactly? in an unsplit wouldn't the "extra" light be coming in at a fairly steep angle and be mmmmm... transmitted thru the lcd at a lower efficiency? am I missing something? you'd still have the 220 collimating fresnel right? even if that was larger also, since its purpose is toi pass parallel light thru the lcd, if its bigger than the lcd it won't...

That's the point. A larger 220mm collimating fresnel is a larger surface area. The larger surface area "captures" more light from the lamp, which in turn results in more light hitting the projection lens.

but if its bigger than the lcd how does it get into there? with the collector fresnel the same size, the extra light will be above the top and bottom/sides of the lcd won't it? the only way to make it pass thru the lcd is to place the unsplit pair further behind the lcd so the collector will aim them at the lcd, in which case they will come into the lcd at a sharper angle and I thought the lcd "viewing" angle was not that great...

The angle at which the light is passing through the lcd would be the same if we got the correct focal length. A 16x12 fresnel with roughly a 400 mm focal length will do the trick. See my drawing. The green line represents the lcd.

Ok here's another light engine idea.

I was thinking about how we could convert a non point source to a point source, and to get all of the lamp arc into the triplet and not just the center.

This is just an idea for how the light needs to be moved, maybe you guys know enough about optics to figure out what kind of lenses we need.

First we need to collimate the light, placing the lamps arc, tangent to the outside rays as my pic demonstrates. We then place the reflector centered on the blue X I drew. All the light in the lamp arc is now in the collimators field of vision, but theres a problem. The outside edges are only capturing a very small portion of the edge of the lamp arc. Thus the collimated light will be stronger towards the middle, and weaker the farther you move to the edge. I illustrated this by drawing the rays towards the edge, spaced farther apart. To reverse this effect, we'd need compres the outside rays into a smaller area so the light is more concentrated, but compress less and less as we move closer to the center. Once the light has been evened out, we just need to redirect it to a point.

Now either this arrangement could be done by glass lenses, and the light criscrosses at the focal point and enters the collimator, shown below. Or if we had fresnels to do this stuff we could just place the lcd after the last lens.

I know what you're all thinking, due to all the additional lenses the light would be filtered out resulting in little or no imrpovement. True but this is just a drawing to represent how the light needs to move. We may be able to achieve this light flow pattern with one or two glass lenses, or the right fresnels or whatever. We might be able to combine lenses 1 and 2 into one lens.

Just an idea.

hmmm.... ok I think I can see that... but I don't think it would be 77% more, given that the farther away from the "point source" the weaker the light... how would the distribution.eveness of the light be affected? I know you can't move the resnel too much closer to the light because the differential between the corners and center will become more pronounced(was that you or rox that said that?) would the same thing happen here?

When the light flow is being controlled, (if its moving in a straight line) the light does not decrease much at all. Thats why a laser pointer can illuminate an object from so far away. The light is moving straight out so its not spreading out and weakening. When light is moving out of a bulb in all directions, the light spreads out the farther you get from the bulb, I believe thats when the light decreases at the inverse of the distance squared.

thats what I was referring to.... the extra light you're capturing is from the "border " of the larger fresnel which is farther away than the "border" of the current fresnels and if you think about the commonest complaint(dim corners,edges) then I think while the larger fresnel might help a little it would be minimal, especially considering the increased box size and cost of the larger fresnels... but I could be wrong, it happens a lot....

Keep in mind the pro lenses will have a 220mm collimator, so really its no differant.

yeah, I thought about it some more and decided that the differntial would be about the same and yes, if you're going the pro lens route along with a bigger lcd it would be beneficial.... which makes sense (optics is not my strong point) since when I went from a 5" lcd to a 15" the brightness for the same size projection increased ....

Yeah, that's exactly the idea. Higher brightness. The square law doesn't apply after the collimator fresnel until after the triplet, because at those points, the light is no longer spreading out, but being collected back towards a point.

Its nice to be right now and then.

So what of my wierd lens idea?

Well, I'm not quite sure what you're trying to accomplish with the first one that the simpler example didn't accomplish, and in the second one, it appears to capture LESS light than a standard setup.

huh?

I didn't get any of that. Its basicly an effort to place the lamp arc slightly closer then the collimators focal length, thus to put the whole arc in the collimators vision. Only problem with this concept is the light at the edges would be pretty dim since its only getting a small piece of the edge of the lamp arc. So we'd need to compress the light at the edges closer to the middle to compensate for the dimming. Once we have this 100% collimated light, we will be able to focus it to a very small point and get the most ouf of the bulb.