6
\$\begingroup\$

I would like to be able to electrically switch between horizontal and vertical light polarization. This polarization filter is supposed to be portable attachment running on a battery in front of a DSLR camera lens thus the filter should have the size of ~3x3cm minimum to cover the whole view. Light transmission through the filter should be the same in both polarization stages and as high as possible (i.e. ideally close to 50%). I would also prefer not to use any mechanical components, such as rotating the filter with a motor, to enable fast switching.

While I was looking for a solution I realized that LCD screens use two cross polarized filters and liquid crystal layer in between the filters to control what light passes through the display by twist the polarization 90 degrees. So I would basically need something like the LCD screen (without the need for controlling individual pixels) without the other polarizing layer and that's also fully transparent in both stages (twister & untwisted).

Another thing I was thinking is if I could possibly repurpose 3D shutter glasses, but not quite sure if this would work.

\$\endgroup\$
  • \$\begingroup\$ LCD screens, at least the liquid crystal part, ARE transparent. These only polarize the light or not. The light/dark that we see on a complete display is caused by the polarizer in front of the screen. Problem for you will be getting an LCD that is just "one giant pixel". \$\endgroup\$ – Bimpelrekkie Nov 2 '15 at 15:01
  • \$\begingroup\$ I just read that some older LCD screens used liquid crystal layer that scattered the light instead of twisting, thus wanted to bring that up. \$\endgroup\$ – JarkkoL Nov 2 '15 at 15:07
  • 1
    \$\begingroup\$ See if a "Kerr Cell" would work for your application. \$\endgroup\$ – Optionparty Nov 2 '15 at 16:12
  • \$\begingroup\$ Another approach would be a beam splitter, then through separate polarized filters. As early color TV split an image, then passed it through separate RGB filters. \$\endgroup\$ – Optionparty Nov 2 '15 at 16:28
  • \$\begingroup\$ @Optionparty: Most beam splitters have a (partially) polarizing effect, so I think this isn't a good idea. \$\endgroup\$ – sweber Nov 2 '15 at 20:34
1
\$\begingroup\$

In the simplest form just use two digital cameras, one with a vertical polarizer and one with a horizontal polarizer. Use a beam splitter if you need an exact duplicate image on both cameras. Then run both video signals to the inputs of an analog switch/multiplexor (or a reversed biased diode switch*) and then digitally select the needed polarity signal as the single output.

*https://www.google.com/search?q=diode+switch+for+RF&biw=1549&bih=730&tbm=isch&tbo=u&source=univ&sa=X&ved=0CCgQsARqFQoTCN2i6YGY88gCFYNZPgodb1AEkQ

\$\endgroup\$
  • \$\begingroup\$ This is for a portable attachment for a camera so can't really use two cameras. Also the additional 50% reduction of light due to splitting isn't good (would need to compensate with twice the exposure) \$\endgroup\$ – JarkkoL Nov 3 '15 at 18:42
  • 1
    \$\begingroup\$ @JarkkoL, you can use a polarizing beam splitting cube or Wollaston prism that splits an image into two polarization's with no (or only a little) loss in intensity. \$\endgroup\$ – George Herold Nov 6 '15 at 13:22
  • \$\begingroup\$ @GeorgeHerold Ah, using polarizing beam splitter could be a good idea, if you could direct both beams to the same camera (with additional mirror possibly?) It should also result in better effective transmission than polarizing filter. While this setup would halve the resolution (not necessarily an issue) it would also have the capture time, I need to look into this. \$\endgroup\$ – JarkkoL Nov 8 '15 at 3:59
0
\$\begingroup\$

I was doing some further research and an AC controlled filter called "Fast Polarization Modulator" does this kind of polarization switching that uses LC layer and polarizing filter like in LCD panels. For example a company in Sweden called LC-Tec (www.lc-tec.se) sells such modulators of proper size (2x2") for my application, which has ~44% transmittance. The basic FPM is able to do reasonably fast transition between the polarization stages (150us to non-twisted stage, 35ms to 90 degree twisted stage), while X-FOS of LC-Tec does faster 50us vs 1.6ms switching for applications requiring higher speed switching. The main issue for small-budget projects is the price however as the basic FPM retails for about 300USD.

As some people mentioned in comments, another potential option is to use polarizing beamsplitter that splits the light to P (transmitted) and S (90 degree reflected) polarized light. Then a first/front surface mirror (to avoid ghosting vs second surface mirror and further polarization) could be used to direct reflected S-polarized light back towards the camera lens. This would effectively result as a single image being split into two, where one half of the image is S-polarized version and the other half P-polarized version of the same image. The advantages are that this is completely passive system requiring no active control mechanism and thus doesn't have any limit for switching speed either. Also polarizing beamsplitter are able to retain around 90% of the transmitted & reflected light vs 44% of FPM. The disadvantage is that the resolution of the image is halved and there may also be more constraint regarding angle of incidence (i.e. at the edges of the image there might be some polarization discrepancies). Broadband Polarizing Plate Beamsplitter with AR coating (to reduce ghosting) of size 25x25mm retails for 199USD at Edmund Optics (www.edmundoptics.com) though, so it's quite expensive piece for projects with small budgets. This size might be too small as well for this particular application. FS-mirrors of proper size for this application with ~97% reflectance for visible light range can be bought for about 20USD from AliExpress (www.aliexpress.com).

If the cost of polarizing beamsplitter is too high, another alternative is to use non-polarizing 50/50 beamsplitter and adding super high contrast polarizing filters both behind the beamsplitter and between camera lens and FS-mirror. The downside with this cheaper solution is that only ~44% of the light gets transmitted through the system (vs ~90% of using polarizing beamsplitter). Non-polarizing 50/50 BS of proper size for this application can be bought for around 25USD at AliExpress. Polarization.com sells a 25x15cm sheet of "Super High Contrast Polarizer" for 22USD.

\$\endgroup\$
0
\$\begingroup\$

Check out this project: http://www.diyphysics.com/2015/07/27/dolpi-a-low-cost-raspi-based-polarization-camera/

They use modified cheap "Auto Darkening Welding Helmet Lens" which can polarize light from 0 to 90 degrees depending on an input voltage.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.