DIY anamorphic lens

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I thought I'd start an anamorphic lens topic for any AVS Forum members browsing around or anyone else interested in the topic.

First off, here's a link to Tor Arne's site describing in great detail his construction of an anamorphic lens (several actually):

http://home.c2i.net/ahustvedt/arnemorph/

For you German speakers, here's a link to the site that did the original R&D on the lens:
http://ww2.bepo.com/jochen/anamorph/

I think there's a link there to the German discussion forum where the original ideas were hashed out. Sorry, I don' t have it handy and don't speak German well enough to find it.

There's also some interesting info on prisms and anamorphic lenses here:

http://appft1.uspto.gov/netacgi/nph...p=1&u=/netahtml/PTO/srchnum.html&r=0&f=S&l=50

We've been moderated out of existence on the AVS Forums so I'm hoping some of the other 'rejects' from there will come over here and we can get a good discussion going.
 
Hi, I will check back to this forum daily and hope there we get a lot of replies from AVSforum members.


Some notes. :)

-The lens works upside-down too, but this gives more lens-shift and more uneven focus.

-I will try to compensate for the barrel distortion in only one direction instead of both like I have with my previous models. The compensation gives some focus-drift towards the corners that I want to avoid. Compensating in one direction only will give a better optical surface as it's possible to bend more accurately than with my previous methods. The left and right edges will have curving to them just as a non-corrected lens, but I don't see that as a problem. The problem area with non-corected lenses are the top and bottom edges IMO.

-I will try to make the corrective surface on the oilprism instead of the waterprism. I was worried that the lexan wouldn't be able to withstand the turpentine as it is a strong solvent, but after bathing a piece of lexan in red spirit for ten hours I'm confident it will work (I didn't have any turpentine with me at work to test with).

-Pure turpentine still works great as a substitute for turpentine oil. They are basically the same thing, turpentine contains 75%-100% turpentine oil addording to the data-sheet.

-The water does seem to attack the epoxy-glue somewhat and silicone sealant works better for the water. However silicone sealant isn't overpaintable, which is a drawback. The paint will form drops on the surface of the silicone after a few seconds. My final version will probably have silicone instead of epoxy for the waterprism. I'm not sure how the silicone sealant stands up against turpentine.

-I want to try making a horisontal expansion lens as well. It shouldn't be any more difficult, all I have to do is design it backwards and rotated 90°

-I wish I had a projector which didn't have such an extreme short-throw lens. It does however make my design very versatile and if it works with my projector then all other projectors with longer throw should be able to use a similar lens.

-If it's possible I might try to make a 2.35:1/1.78:1 adjustable lens, but I'm worried about the amount of chromatic aberrations and barrel distortions becoming too bad. If I compensate the geometry in only the vertical direction the problem will be reduced, however.


I hope I will have the time to experiment more in the future, I might have to send my projector in for repair (again) and my left B&W CDM7-NT speaker has started to make a fretting noise from the bass-driver so I will have to send that in as well. I can't possibly watch movies without the speaker or projector. :( I will be dead for the period they are away. ;)


Tor Arne
 
Hi Guys,

Interesting ideas!

I have previously been exploring the topic of anamorphic lenses myself, but in my case I was looking into using a pair of cylindrical lenses (one each PCV & PCX - both with the same focal length) to expand the beam horizontally, and possibly also using another pair to compress the beam vertically at the same time. (This would reduce any distortion in a single axis.)

Unfortunately, the lenses I have been able to get so far are very poor quality, and the results were unimpressive.

Your use of liquids to form the prism body is cool. Have you considered exploring those transparent epoxy packs available for use in embedding things within a plastic block? You might be able to get away with only needing one liquid device. (I'm thinking about leaks etc.)

Bill.
 
Is it somekind of molding material? The material would have to have the right refraction index, somewhere close to turneptine. If the refraction index is different the angles need to be changed.

Interesting idea about compressing and expansion working together. :)


Check out this patent with lots of ideas but unfortunately no calculations:

http://appft1.uspto.gov/netacgi/nph...75'.PGNR.&OS=DN/20020063975&RS=DN/20020063975

It has many auggestions for how to make a lens and how to correct different artifacts.


Tor Arne
 
Hi,

Of course I can't find a link when I want one, but to give you some ideas of the properties, and see if the stuff is in the ballpark, try these links:

http://www.loctite.com/pdf/doming.pdf
http://www.loctite.com/pdf/ComparPropertiesOpticallyClearEncaps.pdf

The stuff I was actually looking at is generally available from Home Depot in the US. It is clear, transparent, and designed for encapsulating things into table-tops, trays etc.

I was thinking of using it to make large PCX field lenses for the LCD panels instead of fresnels, but when used correctly, the fresnels weren't too bad.

You put whatever you wish to encapsulate on the surface, pour over the epoxy as if it were water, and let it set. (Of course, you must build a waterproof frame around the edges...)

I have seen this used with thicknesses up to an inch.

Bill.
 
It would be possible to make a mould from glass. The problem would be getting the prism out of the mould, even if the mould was expendable.


My apporach to the lenses is a bit silly as I'm trying to make it small. By making it a bit larger it will work better but it will be ugly. I just mention it so you guys know why I'm making so many different versions. :)


Tor Arne
 
For the loctite materials, the problem is that the current DIY lens design is based on the the optical properties of turpentine oil and water. It seems that the refraction and dispersion properties of both materials are important to get the desired bending of the light while reducing chromatic abberation. This is what I've read, anyway.

However, it seems that "commercial lenses" of this sort (liquid filled prisms) just some sort of mineral oil. I'm not sure how the correction for chromatic abberation is handled there.

There's a patent on handling chromatic abberation in prism based lens systems...

http://patft.uspto.gov/netacgi/nph-...704,008'.WKU.&OS=PN/4,704,008&RS=PN/4,704,008

...but it's over my head at the moment.
 
I made my first attempt at working on the lens yesterday. It didn't go too well.

I had a very hard time getting clean, straight cuts in the glass. It didn't break well and when I did manage to get something of a straight snap, the edge would be very "wavy".

After finally getting a few pieces that were close to desirable I attempted to glue them up. That didn't go well at all. I just couldn't get it to come together in the right shape. I ended up with epoxy everywhere and a shape that would never work. The result went into the trash.

I know I am clumsy but this was just ridiculous.

I think I'll try something other than the all glass and epoxy method of construction. Since I don't have access to aluminum bending gear, I'm going to search for some kind of small 'L' bracket material to make a frame, then glue the glass into the frame.
 
I made some formers from cardboard which I put between the surfaces of glass while glueing. I also use a lot of clear plastic tape and align the pieces flat on the table on the tape before I glue so that I can just swing the parts together using the tape as hinges.

It should be possible to make a wood frame for the prisms and painting it and then coat it with epoxy-glue thinned with red-spirit.

I started assembly on the failure-prism again today. I doscivered that the glue had released the aluminium frame where the frame wasn't painted, but where the edge of the frame was covered in paint the glue stuck wery well. So I painted the edges today and will glue it together tomorrow.


Tor Arne
 
Tor,

Great work so far!

I made a prototype completely out of lexan for the sides and surfaces of both prisms. This was about a week ago and no leaks yet. For someone who does not have acsess to metal and plate bending press i think the lexan is the way to go for the side pieces. However it does not work well for the light surfaces. Way to reflective. I am going to get some high quality non reflective glass for the finished version.

My observations so far:

My prototype works fairly well. I also have a short throw projector (panasonic 711xu) and the barrell distortion is significant. I haven't tried your outward curvature of one lens surface yet(not sure how to get the proper curvature drawn out and cut)
I also noticed different areas of the projected image were slightly out of focus. will have to see if this improves using better glass and tighter tolerances when mounting the two prisms.

Has anyone tried using mineral oil instead of turpentine?


Adam
 
tahustvedt said:

I was worried that the lexan wouldn't be able to withstand the turpentine as it is a strong solvent, but after bathing a piece of lexan in red spirit for ten hours I'm confident it will work (I didn't have any turpentine with me at work to test with).

Be careful with the lexan and turpentine. I was thinking of doing that too (oil prism of lexan) and just did a few searches on lexan/polycarbonate/turpentine with somewhat mixed results. One datasheet said that polycarbonate was resistant to turpentine. Several others listed turpentine as possibly destructive. I think I'll stick with glass.
 
Mineral oil is the key. It should not attack the lexan. I am using Edmond Scientific Glass which was quite expansive. It seems to have more reflection than I was expecting. I have the basic prisms bonded together and am now tweaking things in. Thanks to Tor for getting things going! Hopefully more AVS people will help contribute as well.

PP
 
I found a page stating that care should be taken with using turpentine with lexan. I guess soaking the lexan isn't a good idea, then. :)

When I ask about mineral oil at the pharmacy they have many different kind of oils available. Does anyone know what the appropriate oil is called?

I'm glueing the prism right now and will fill it later this evening. I will make a new prism soon with compensation in only one direction.

I will make some scetches describing why the compensation can give problems.


Tor Arne
 
I took some pictures of the prism while glueing. I forgot to take a picture of the top brace. Sorry! :)

An externally hosted image should be here but it was not working when we last tested it.

Here the flat side has ben glued to the frame. The brace is next.

An externally hosted image should be here but it was not working when we last tested it.

The brace has been glued to the frame and the correction-surface is setting.

An externally hosted image should be here but it was not working when we last tested it.

The finished prism prior to filling and painting the outside of the frame. Notice the brace which shapes the peak-side of the corrective surface.


Tor Arne
 
I'll try to explain to you the main reason why I want to try correcting just the top and the bottom of the picture.

In the first picture you see the prism from the side. The red lines describe how the lexan would be bent if it was corrective. This compensates for the barrel-distortion but it can introduce another problem. Since we won't be able to position the projection beam accurately on the surface of the prism the beam might be closer to the top or bottom of the prism. This would actually change the peak angle of the prism, if the projection beam exits closer to the peak the prism angle will be smaller than it should be and vice versa. The blue and yellow lines shows how the angles closer to each edge are different from ideal (black edges).

An externally hosted image should be here but it was not working when we last tested it.

Looking at the prism from the side.

In the second picture we see the prism from the top. Correcting just the top and bottom of the picture by curving the lexan according to the red lines in Figure 2 will not be a problem since the lens doesn't change the horisontal width of the picture. It will be easy to center the lens well enough to get an even focus.

An externally hosted image should be here but it was not working when we last tested it.

Looking at the prism from the top.

If one wants to correct in both directions it would be best to correct in one direction for each surface, and not like I have done with my prisms up untill now. For example bending the front surface by the vertical axis and bending the aft surface by the horisontal axis. I have been correcting the same surface in both directions all the time. :)

I don't think the focus drift will be as bad when correcting in only one direction, which might be another benefit from going with this design.


Tor Arne
 
I'm still working to get a basic prototype built... no fancy curved, geometry correcting prisms for me yet. :)

I learned the value of oiling my glass cutting tool yesterday. That trick finally allowed me to make clean, straight cuts.

I also picked up some thin sheet metal and made 'end caps' for the oil prism. I measured out the 24 degree angle, added in a bit of room for a flap to bend down around the edges, and cut the metal out with snips. Then I bent the edge down with a vice and hammer to get a nice sharp 90 degree bend. What I ended up with are top and bottom wedge shaped 'cups' to glue the glass into. It made the assembly process much easier.

I'll work on the water prism today.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.