Projector lenses etc.

Discuss about cameras, projectors, calibration panels, turntables etc.
User avatar
Micr0
Posts: 586
Joined: 15 Nov 2016, 15:20
Location: New York City

Projector lenses etc.

Post by Micr0 »

Has anyone given any though to changing the lens on you projector? specifically something Sven said a while back about most projector lenses being cheap and the week link in this whole system. I was thinking specifically of doing it to not only increase the throw of my projector to better match my camera lenses but to also change (actually eliminate) the vertical offset that all home theater projectors have. I have a small low rez projector that U don't use. Maybe I'll hack into it and have a look at whether this is possible.

Thoughts?
µ
User avatar
avogra
Site Admin
Posts: 116
Joined: 11 Nov 2016, 17:27
Location: Munich, Germany
Contact:

Re: Projector lenses etc.

Post by avogra »

When I opened the Acer K11, I immediately discarded the thought to replace the lens. But maybe there is a way. The main problems to overcome in my opinion:
  • flange back distance (german: auflagemaß): you have to find a lens that is designed to be mounted far enough from the sensor. C-mount has a FBD of ~17,5mm, which will be too close to the sensor. no idea about photolenses though.
  • lens shift: the entire construction is designed, so that the light is reflected at an angle from the dlp chip. To get rid of the lens shift, the lens would have to be translated parallel and still catch the light reflected from the chip. I'm not sure, but I suppose, you need a very high apperture for that.
  • Space: I think it will be difficult to fit a different lens into the housing and mounting it outside will further increase the flange back distance.
I have some 100mm concave mirrors with F=100 at home, that might work as a single element optic. No idea if that means distortion. i'm not as deep into optics as i would like. If I find the time, i will try it after christmas!

About the lensshift: Instead of removing it at the beamer side, one might as well add a similar lensshift to the cameras. at the same time one could also add a lensshift in horizontal direction. In my imagination, the result would be, that the cameras and the beamer are mounted in parallel, while the field of view and the sharp plane would be the same.
That idea is going through my head for some weeks, but i didn't manage to draw a sketch up to now. maybe i should really do that soon or just print and try such a lensshift adapter :P
User avatar
Micr0
Posts: 586
Joined: 15 Nov 2016, 15:20
Location: New York City

Re: Projector lenses etc.

Post by Micr0 »

avogra wrote:When I opened the Acer K11, I immediately discarded the thought to replace the lens. But maybe there is a way. The main problems to overcome in my opinion:
  • flange back distance (german: auflagemaß): you have to find a lens that is designed to be mounted far enough from the sensor. C-mount has a FBD of ~17,5mm, which will be too close to the sensor. no idea about photolenses though.
  • lens shift: the entire construction is designed, so that the light is reflected at an angle from the dlp chip. To get rid of the lens shift, the lens would have to be translated parallel and still catch the light reflected from the chip. I'm not sure, but I suppose, you need a very high apperture for that.
  • Space: I think it will be difficult to fit a different lens into the housing and mounting it outside will further increase the flange back distance.
I have some 100mm concave mirrors with F=100 at home, that might work as a single element optic. No idea if that means distortion. i'm not as deep into optics as i would like. If I find the time, i will try it after christmas!

About the lensshift: Instead of removing it at the beamer side, one might as well add a similar lensshift to the cameras. at the same time one could also add a lensshift in horizontal direction. In my imagination, the result would be, that the cameras and the beamer are mounted in parallel, while the field of view and the sharp plane would be the same.
That idea is going through my head for some weeks, but i didn't manage to draw a sketch up to now. maybe i should really do that soon or just print and try such a lensshift adapter :P
My motivation to change the projector lens is twofold. First is to change the focal length to better match my cameras. High resolution lenses are expensive and it's better to replace one projector lens than two camera lenses. The resolution of the projector is less critical. Second,regarding lens shift, Lens shift is a distortion of the image from the source. With projector optics being relatively cheap production components, who knows how much the image is changed? David is supposed to calibrate around that to some degree, but better to eliminate it all together than to have to fix it in the mix so to speak. Also, eliminating the lens shift would allow the projector to be parallel to the cameras in the vertical plane. This is how all the high end scanners with dedicated beamers do it.
µ
Curiousjeff
Posts: 121
Joined: 16 Nov 2016, 22:31

Re: Projector lenses etc.

Post by Curiousjeff »

I hope I am not to much off topic, but have a look at this article :

https://www.osapublishing.org/oe/viewme ... 4657&seq=0

I found it when I was looking for the best but low cost solution for perfect projection of patterns. I don't think it exists as a stand alone product, but it's interesting.

Jeff
User avatar
avogra
Site Admin
Posts: 116
Joined: 11 Nov 2016, 17:27
Location: Munich, Germany
Contact:

Re: Projector lenses etc.

Post by avogra »

Micr0 wrote: My motivation to change the projector lens is twofold. First is to change the focal length to better match my cameras. High resolution lenses are expensive and it's better to replace one projector lens than two camera lenses. The resolution of the projector is less critical. Second,regarding lens shift, Lens shift is a distortion of the image from the source. With projector optics being relatively cheap production components, who knows how much the image is changed? David is supposed to calibrate around that to some degree, but better to eliminate it all together than to have to fix it in the mix so to speak. Also, eliminating the lens shift would allow the projector to be parallel to the cameras in the vertical plane. This is how all the high end scanners with dedicated beamers do it.
I don't think, that lens shift means distortion in the first place. But if optics are of medium quality, lens shift will certainly render it even worse. I didn't think about that, so your idea is probably the best solution. I just can't imagine, how one could remove it if you don't want to build a new housing with different geometry. I will draw a picture of how the optics in the K11 are arranged. I suppose it is very similar in the K132. Maybe if you can mount high quality optics, the effect of lens shift is then acceptable?
User avatar
avogra
Site Admin
Posts: 116
Joined: 11 Nov 2016, 17:27
Location: Munich, Germany
Contact:

Re: Projector lenses etc.

Post by avogra »

I tried to sketch the geometry inside the K11 from memory, please no comments on my drawing skills :P
The left is when you look from top, the right is when you look from the right side. I had the Lens removed and could see the mirror and the DMD chip.
CCF14122016.jpg
The problem is, that when you look at the right picture, there is this mirror, a bit below the DMD chip. It reflects the light onto the DMD chip at an angle, so that the reflection from the DMD goes over the mirror into the lens. So the rays come from the DMD already tilted to avoid the mirror. I think to remove the lens shift, the lens has to be in line with the center of the DMD. No idea, if there is a way to accomplish this.
Curiousjeff
Posts: 121
Joined: 16 Nov 2016, 22:31

Re: Projector lenses etc.

Post by Curiousjeff »

Texas Instrument has a family of DLP products especially for structured light:

http://www.ti.com/lsds/ti/dlp/advanced- ... tions.page
Attachments
Capture.PNG
User avatar
Micr0
Posts: 586
Joined: 15 Nov 2016, 15:20
Location: New York City

Re: Projector lenses etc.

Post by Micr0 »

Curiousjeff wrote:Texas Instrument has a family of DLP products especially for structured light:

http://www.ti.com/lsds/ti/dlp/advanced- ... tions.page
This is just the DLP with no light source or lens or power supply. but I can see where the dedicated beamer design starts. I found a couple of high end blue light scanners that have swap-able projector lenses. (we are talking $60K systems) The lenses look like the type of projector lens that I see online for about $1K. I assume this is to allow for the greater versatility to scanning large and small objects. FWIW the ability of my system to capture fine detail is excellent. Now I want to improve things like surface quality and overall accuracy.
µ
Curiousjeff
Posts: 121
Joined: 16 Nov 2016, 22:31

Re: Projector lenses etc.

Post by Curiousjeff »

I had also looked at this back in the days when you were removing the fan in the Optima 8-)

http://www.opto-engineering.com/product ... -projector

This would be great. I don't know the cost.

Frequency, pixel orientation, resolution, lens quality, distortion...many things that cause problems.

Today. while testing the PointGrey in a lower resolution, I scanned the panel. Strange that the left panel is smooth and the left has waves. This is with a single camera on the left side of the projector.
Attachments
2016-12-14_16-11-41.png
Curiousjeff
Posts: 121
Joined: 16 Nov 2016, 22:31

Re: Projector lenses etc.

Post by Curiousjeff »

My previous message should have read:

"The left panel is smooth and the right one has waves. This is with a single camera on the left side of the projector."

I wonder if items that are parallel to the plane of the camera chip (right panel is parallel to camera chip) are more prone to surface artifacts and deformation ?

Or is it due to the stronger light reflecting towards the camera from the right panel ?

I have a photo filter: Variable ND. I'll give it a try to see if by reducing light on the right side, I get a better scan.

If the problem is in relation to parallelism, that could be useful when deciding how to scan a part.

Jeff
Post Reply