We have an optics expert at my work, and he was telling me that one could get spectacular results with just using curved mirror, no lenses! He ran through the idea for me, and it does make sense. He said it helps to prevent color shift, and keeps light in phase. However to do this, wouldnt be no buy and try adventure. Someone would have to run some serious numbers to ensure results. Has anyone done this? or even thought about it?

I thought about it but i am in no shape mathmatically and financially to try it.

It seems quite feasibly mathematically, and optically, but not financially. I know some projectors do use mirrors, but the whole thing is quite small. See, you'd need a large mirror to collimate the light to go through the LCD. The only place I see to get mirrors that size within tolerance for optics is quite expensive. Check out prices for parabolic mirrors for telescopes. They're out of this world (pun very much intended).

Maybe if the LCD is small enough it may be financially viable, but not on a 15" screen.

One option that may be worth exploring is a mirror instead of the triplet. I know that Edmund Scientific (or maybe Anchor) sells parabolic mirrors on a 30/60/90 degree angle. That may work, but I don't know if a parabola is the appropriate shape or if some other conic would work better. I suppose if someone had an optics lab at their disposal they could try out a few things.

Actually the curved mirror is a lens... when I was a kid I used to take my sister's 'magnifying' make-up mirror and reflect an image onto a wall, don't know why I did it.. just did it.
And like AP said, there really isn't any reason to try this as the cost would outweigh any possible advantages.

They did announce an extreme short throw projector that used mirrors instead of lenses. http://www.shopcol.com/necwt610.html This one uses more than one mirror which sounds complicated.

I would like to try building one as a proof of concept experiment with a transparency in place of the LCD panel but I can't really find any detailed optical information on how mirror projection works. Best I have done is some pages on star dome type of projection but naturally that won't work on a standard flat screen unless you used a corrector lens which defeats the purpose of using no lenses.

You would need a concave mirror as you want to make the image larger for projection. Parabolic would likely be the best type for projection. I might experiment with a large makeup mirror just to see how it works with a bright light against a wall. There will be a problem with that simple approach in that you would want to put the panel in front of that mirror and shine a light through that panel and onto the mirror which then reflects back a larger image. The panel itself will block some of the projection. You would need to tilt the mirror and correct for keystone to avoid panel blockage. Reflecting telescopes have the same problem as well but because the little mirror is small compared to the larger primary mirror, it is a worthwhile trade off in light gathering power.

The advantage of this system would be the extreme short throw depending on the curvature of the mirror used. It might also be possible to run this at higher heat, resulting in fewer fans since the mirror itself wouldn't be bothered by heat.