DIY 3D printable optics tools

[Arturo]

Hi all.

I am in the proccess of building a homemade 3D printer (printrbot in my case, but any prusa-mendel-makerbot would do), and came across a set of 3d printable optical components on thingiverse http://www.thingiverse.com/jpearce/things/ which also point to an Open Source Optics page (http://www.appropedia.org/Open_source_optics). I will surely try some of these and let you know how this fares...

If anyone else has experience in this, I'd love to hear that!

Yet another 'thing': interferometer (http://www.thingiverse.com/thing:33020)

Cheers!

[Ed Wesly]

What would be totally hip would be an open source spatial filter thingy!

[favalora]

Neat, I will visit those links.
Have you seen Disney Research / CMU's work on printed light pipes?
http://www.disneyresearch.com/project/printed-optics/

g

[Dinesh]

Interesting work by Disney. I've been thinking along those lines myself. I started last year by making an edge lit HOE (http://holoforum.org/forum/viewtopic.php?f=10&t=278 ). The next task is to try and link up several of these edge lit HOEs. However, my interest is holographic neural networks, as opposed to just focusing and deviating optics. Of course, Disney is far more advanced, but then they're better funded than the two of us in this small corner of San Marcos!

[walschuler]

Dinesh,

This is by now old hat, but I have some questions. Am I correct that in the example photo the edge light reference beam
was diverging after entry, and the object beam was a beam focused by a lens or mirror through the emulsion? What emulsion did you use, what laser, what power, what exposure time? Have the experiments continued? Have you done any more theoretical work? Is there a difference between what you did and other edge-lit techniques?

After all that, Happy New year to you and Joy!

Bill

[Dinesh]

Am I correct that in the example photo the edge light reference beam was diverging after entry, and the object beam was a beam focused by a lens or mirror through the emulsion?

Not quite. It was recorded by one collimated beam sent into the edge of the emulsion, and the other beam was diverging from an objective and hit the face of the emulsion head-on, ie at zero angle to the normal. Thus, a collimated beam launched into the edge of the emulsion should focus from the face of the emulsion. In order to test the system, we created a U-shaped channel, about 3 inches long, and placed an orange led in the middle of the channel. The flanges on the channel slipped over the edge of the hologram to ensure that the hologram was being reconstructed by edge light only (as opposed to a very-high-angle beam). However, in this geometry, the hologram was being reconstructed by a divergent quasi-monochromatic source, which gives rise to all the aberrations you see. We did launch a thin, collimated laser beam - created by a cylindrical lens - into the edge and observed a focus from the face, but this was almost impossible to photograph.

What emulsion did you use,

I wanted to avoid the usual method of launching via a large glass block or prism, since this is not my idea of "edge lit". I think to be truly "edge lit", it needs to be reconstructed by an edge alone with no ancillary optics. We normally use dcg on 1/8th thickness window glass from a local hardware supplier and Joy cuts them down to whatever size we need (currently, we've made 8x10 and are going to be coating 12 x 16 soon). However, this glass has very rough edges and it was difficult to launch a clean beam into the edge. So, we ordered 1/4" glass and asked the glass supplier to polish one edge. We then coated this with dcg. One problem I foresaw was the problem of creating a latent image from both sides of the emulsion simultaneously. As you're probably aware, the latent image planes act to limit the actinic light from entering deep into the emulsion since the latent image planes are perfectly Bragg selective for the wavelength you're recording with, thus limiting the depth into which you can record. However, in this case, the beams were launched from the back of the emulsion and the front simultaneously, which, because of the high absorption of dcg to the laser light, would affect both the latent image and the ratio. The latent planes would be forming somewhere in the depth of the emulsion and it was impossible to predict how the latent image planes would affect the recording. My initial reasoning led me to believe that, therefore, a thin emulsion was required. However, on further thought, it occurred to me that this required a very Bragg selective set of planes, since the light is being redirected at 90 degrees, and so a thick emulsion is required. I went with the thick emulsion and coated an extra thick emulsion.

what laser,

A Lexel Argon set to 488. I have no idea which Lexel, since Joy does all the maintenance and tuning on the laser itself. I barely know how to turn it on and off by following a sequence that she's put up on the wall.

what exposure time?

I don't remember. I do have it in my notes, which I can dig up, but it's buried somewhere in a thick stack of notes on all the holograms made since. My memory is that it was fairly long because there was a lot of light loss. The laser itself was emitting at about 400 mW at 488.

Have the experiments continued? Have you done any more theoretical work?

No more experimental work, some theoretical work. The problem is that the customers get in the way of any real research, so I fit in any research between customers and don't really have a lot of time. I really do envy those companies that have investors so that they can afford to indulge in research without worrying about where the next set of beer vouchers is coming from! At any rate, when I did my paper for ISDH on aberrations, mostly it was a re-hash of Meyer's and Champagne's work, so I thought to put in some original stuff. I did get in a theory on holographic resolution and tried to get in some work on aberrations in edge lit holograms, but it was still incomplete and we had no time to fit it into the paper.

Is there a difference between what you did and other edge-lit techniques?

As far as I'm aware, all the edge lit holograms launch via a large glass block or prism. I wanted to avoid this, so I launched directly into the edge. This way, it can be reconstructed with an appropriate line of light anywhere with no ancillary optics. Also, as far as I'm aware, no one has made a HOE with edge lit light. I initially saw this as an opportunity for edge lit complex lenses with almost no volume requirement, similar to the idea that Gregg's link showed to Disney's light controlling optics. Then it occurred to me that I could stack these edge lit lenses in a variety of geometries, with a variety of input light shapes and a variety of output light shapes so as to deliberately create aberrations. For example, I can record a collimated edge lit lens and reconstruct it with a diverging source from the previous edge lit lens in a system of two edge lit lenses. This system would be similar to a node with an input, an output and a control signal (the form of the recon beam) which is ideal for a neural network (since this is basically one node in a neural network). So, I thought that perhaps one possibility was a holographic neural network. However, as I said, with little time and less money, it's probably going to take a little while before anyone sees a holographic R2D2 from Triple Take's lab!

After all that, Happy New year to you and Joy!

Thanks. And a Happy New Year to you too!