Page 1 of 2 12 LastLast
Results 1 to 10 of 15

Thread: two projection lenses for each horizontal side of a DLP chip, possible?

  1. #1
    Join Date
    May 2018
    Posts
    56

    Default two projection lenses for each horizontal side of a DLP chip, possible?

    DLP chips these days are made in the standard 16:9 ratio. Imagine trying to split a 1920x1080 pixel DLP chip between two projection lenses. One should receive the 1080x1080 pixels from one side of the chip, the other the 840x1080 from the other side of the chip While we can have an ordinary beamsplitter after the chip , is there a known suggested way to block the pixels not needed for each projection lens? Simple absorber slit positioned to block half of the beam will suffer from diffraction and soft transition rather than sharp cut, from my limited optics knowledge. Any ideas and known solutions to this kind of problem?

    Last edited by piydadorto; 07-05-2018 at 11:16.

  2. #2
    Join Date
    Jan 2006
    Location
    Charleston, SC
    Posts
    2,147,489,446

    Default

    You might be able to live with the diffraction effects if you don't mind sacrificing a narrow band of pixels...

    I would try a two-step approach. Mask each beam coming out of the beam splitter (before the projection lenses) with a slot that is slightly larger than needed, and then apply a second mask after the projection lenses that is tighter to get just the part you want.

    If you're willing to accept some loss in that transition area, I think this would be the easiest solution (and one you could set up and test pretty quickly).

    Adam

  3. #3
    Join Date
    Jun 2009
    Location
    St. Louis, MO
    Posts
    1,199

    Default

    Zeiss has a DLP projector (called Velvet) where they focus the light from one DLP chip pixel for pixel onto another DLP chip to achieve 2.5 million native contrast - so yes it's possible. But taking a high resolution projector and converting it into two low resolution projectors doesn't sound like it would ever be a cost effective design decision.
    "There are painters who transform the sun into a yellow spot, but there are others who, with the help of their art and their intelligence, transform a yellow spot into the sun." Pablo Picasso

  4. #4
    Join Date
    May 2018
    Posts
    56

    Default

    Quote Originally Posted by laserist View Post
    Zeiss has a DLP projector (called Velvet) where they focus the light from one DLP chip pixel for pixel onto another DLP chip to achieve 2.5 million native contrast
    I tried to find info how it actually works but couldn't. Any ideas how they did it? Something like an expensive optical fiber wafer is not something I think I could source and use but if it uses something easier to source I could try it out.
    Quote Originally Posted by buffo View Post
    I would try a two-step approach. Mask each beam coming out of the beam splitter (before the projection lenses) with a slot that is slightly larger than needed, and then apply a second mask after the projection lenses that is tighter to get just the part you want. If you're willing to accept some loss in that transition area, I think this would be the easiest solution (and one you could set up and test pretty quickly). Adam
    Interesting idea. I will need to test and see how much I can reduce the transition area with two slits. Maybe you can also suggest what to use for the slit to improve it even more?

  5. #5
    Join Date
    Jan 2006
    Location
    Charleston, SC
    Posts
    2,147,489,446

    Default

    Several folks here have had excellent results using the sharpened edge of a razor blade for masking. (Granted, that was for a spacial filter to clean up the beam from a multi-mode diode, so not quite the same as your application, but the concept is similar.)

    Depending on how much room you have to work with, this might be a good first try - at least for the initial, smaller mask between the beam splitter and the projection lens. You can use small neodymium magnets on a flat vertical post to hold the blade in place, and you can nudge it around with a flat-bladed jeweler's screwdriver to adjust it.

    Eric Meyer (Planters) posted a good video on his solution here: https://www.youtube.com/watch?v=A5sjyEGFOhE Skip to around the 13 min mark to see him applying and adjusting the razor blades. Since you only need to mask one side of the beam, you only need a single upright and a single blade, not the large square mount with 4 blades that he used.

    Adam

  6. #6
    Join Date
    May 2018
    Posts
    56

    Default

    Thank you. I wouldn't want to put neodymium magnets too close to the DLP micromirror chip but I'll see what I can do.
    Last edited by piydadorto; 07-02-2018 at 04:49.

  7. #7
    Join Date
    Jun 2009
    Location
    St. Louis, MO
    Posts
    1,199

    Default

    The Velvet uses a folded optical path, and I believe a parabolic mirror at the fold. Someone told me they used two free form optics to get the resolution they were after. The simplest way I can think of the accomplish your task is to use a 1:1 conjugate ratio lens to get a real image of the DLP chip and use a mirror where the edge of the mirror is at the focal plane of the real image to grab the pixels that you want. Then you use lenses to project the two real images. If the lens has the resolution and the mirror is positioned accurately enough the edge of the mirror will be in the gutter between the DLP pixels, and while the pixels away from the edge of the mirror will be progressively more out of focus due to their distance to the focal plane it should be possible to design the optical layout to make sure the out of focus portion of the diverted image never gets past the edge of the mirror.
    "There are painters who transform the sun into a yellow spot, but there are others who, with the help of their art and their intelligence, transform a yellow spot into the sun." Pablo Picasso

  8. #8
    Join Date
    May 2018
    Posts
    56

    Default

    The simplest way I can think of the accomplish your task is to use a 1:1 conjugate ratio lens to get a real image of the DLP chip and use a mirror where the edge of the mirror is at the focal plane of the real image to grab the pixels that you want. Then you use lenses to project the two real images. If the lens has the resolution and the mirror is positioned accurately enough the edge of the mirror will be in the gutter between the DLP pixels, and while the pixels away from the edge of the mirror will be progressively more out of focus due to their distance to the focal plane it should be possible to design the optical layout to make sure the out of focus portion of the diverted image never gets past the edge of the mirror.
    Thank you. I will pass this info to an optical engineer and see what he thinks and if he can explain it to me in simpler terms. I assumed we wouldn't need lens after the DLP chip to get projection the size of the chip if the illumination beam was made parallel before reaching the micromirror chip but it appears to be more complex than that. Does that also have to do with diffraction and micromirror edges, or impossibility to get perfectly collimated LED or laser illumination beam?

  9. #9
    Join Date
    May 2018
    Posts
    56

    Default

    Quote Originally Posted by laserist View Post
    The simplest way I can think of the accomplish your task is to use a 1:1 conjugate ratio lens to get a real image of the DLP chip
    Please tell me if I understand this correctly. Basically create another a focal plane like how it's done for making reticles in focus in rifle scopes? If yes, it is possible to skip this step of needing to create a focal plane by having the illumination LED beam going trough an LCD matrix or DLP chip collimated?

    and while the pixels away from the edge of the mirror will be progressively more out of focus due to their distance to the focal plane
    I'm trying to understand why this is an issue. While on the mirror itself they will be out of focus, if it is 45 degrees before reaching the projection lens all the beams should have passed the same length and should be focused the same way by the final projection lens, correct?

  10. #10
    Join Date
    Jun 2009
    Location
    St. Louis, MO
    Posts
    1,199

    Default

    Having the image infinitely collimated and using a beam splitter would allow you to use all or part of the imager in either image at the expense of intensity. In a infinity collimated beam the pixels are 'relatively' huge - a knife edge would produce images with a significant vignette.

    In the layout I described the intensity for each sub image is 100% as long as the out of focus pixels don't bypass the mirror edge and therefore reducing the intensity of those pixels. (I doubt that would be a problem, but it should be considered.)
    "There are painters who transform the sun into a yellow spot, but there are others who, with the help of their art and their intelligence, transform a yellow spot into the sun." Pablo Picasso

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •