Hello again everyone!

It's been about 2.5 years since I last posted on the forum. Some of you may remember that I was messing around with the "black screen" paints. Since then, I bought a commercial projector (I know...shame on me) to help the screen research efforts but that project has pretty much died and been taken over by others on other parts of the internet. Also, the damn bulbs are too expensive to replace.

Now I'm back and looking to put together a kick-ass 1080p projector. I've been trolling around the threads looking at the various advancements...particularly with the light sources/engines. I've come up with an idea and want input from anyone interested or more knowledgable than I.

The idea is to extend the length of the box and use additional lenses to concentrate the usable light. Since most are watching movies/TV in widescreen formats, doesn't it make sense to squeeze the light source in the vertical dimension?

I've attached a basic sketch of the idea.

Pros: More usable light from the light engine.
Cons: More length in the light engine...longer box.

The con is a big one especially for straight shooting beamers...but I figured this might actually work out well for "behind the couch" vertical folded designs using the Pro lens equipment.

Anyways all, let me know what you think.

I think I've read (somewhere on the forum) that each lens or mirror between the light source and the triplet will reduce the amount of light that reaches the triplet by 10 percent, or so. You've got 9 layers (including the LCD) between your light source and triplet in your diagram.

I'm no expert, but I don't think it'll work.

Here's a quote by jonjandran where he tested and found that just Lexan reduced the amount of light by 12 percent.

Quasi,

I just did the calculation. Taking into consideration each new lens added to the engine, and the "squeeze" you get, the new light engine is only 10% better than the standard. Personally, I don't think 10% is worth all the extra effort.

I used 90% transmission per extra lens element as you suggested.

The only way this would improve is if the extra optical lenses used are greater than 90% transmission.

Oh well, was just an idea. Maybe someone can prove me wrong and make this better

Oh, another thought...

If you could find LINEAR fresnel lenses, one with positive focal length and one with negative focal length, then the number of additional lenses could be reduced to 3 instead of 6 as I show in the diagram.

Running the calculation for this situation (2 linear fresnel lenses +1 circular fresnel), you can get a 51% increase.

Where can we get linear fresnel lenses?

You could skip adding extra lenses and just use an anamorphic pre condenser lens. The problem is cost.

There is a lot of gain to be had if you can reshape the light cone. If you plan to only show 2.35 wide screen and have a round light cone you wast almost 40% of the light.

With a lower power lamp you might be able to use a standard anamorphic lens with a vary short FL condenser lens and it should work.

The problem is, for a bright PJ, you really want a 575 HMI lamp and they makes lots of heat. A plain glass anamorphic lens after the condenser will run much cooler so it just might live but you will have to be willing to risk a couple of big ones to see if it will crack.

Cylindrical lenses seem to be one of the few types that can be found surplus made from fused silica (one sixth the thermal expansion of pyrex). Surplus shed carries a few PCV cylindrical lenses for about $40, and this place carries a PCX cylindrical lens for $90 (getting way too pricey for my taste): http://www.sro-optics.com/cylin.html

A PCV would have to be used in a horizontal stretch, and a PCX in a vertical compression. I would prefer the PCX, as that would mean you could probably use it with the standard 4.5x6.5 PCX precon that is so popular, but a PCV would allow you to get the arc closer to the lens, so it may be better performing if you can find a good match for the second precon. sort of like dazzla's superprecon, but only in one dimension. Anyone want to trace it?

I think 5% is more typical for optical quality glass, with <2% loss for AR coated glass IIRC.

Ok. I've revised this concept. Take a look at the picture.



Again, assuming we are trying to squeeze the light only in the vertical for a 15.4" wide panel, this could be accomplished with only 2 additional fresnel lenses.

The first fresnel (a standard 220mm Pro lens) collimates the light from the point source.

The second fresnel (650mm linear Fresnel) focuses the light onto a line. Let it converge, then diverge.

The third fresnel (400mm linear Fresnel) re-collimates the light, now compressed in the vertical.

The 650mm linear Fresnel can be had at BHLENS.COM

The 400mm linear Fresnel is the required focal length to re-collimate the light. I don't know where to find this and need to search.

This adds even more length than the original concept.

Thoughts anyone?

I think AV is on the right track... an anamorphic lens (typically cylindrical) is the way to go... less light loss (brighter), maybe cheaper if you can find appropriate lenses surplus, no significant increase in PJ size. What's not to like?

Isn't it more than that? If I remember my trig, the coverage of the circle should be

A_lcd/A_cir = {sin(atan(AR)) + cos(atan(AR)}/PI (proof left as an exercise for the reader )

So, for a 16:10 panel, only 57% of the circle is covered, and for a 2.35 AR, less than 42% (yikes!)