Question about collimated display systems.

[geneb]

That's an interesting idea Jordan. The only issue I can think of is that the required stop point would change over time, but a differential feedback mechanism might solve that.

The pump I bought is small enough to fit in the palm of my hand. It's 110v AC and doesn't have much of a volume rating, but I'm hoping it can pull at least 6 PSI.

g.

[Atomic_Sheep]

I'm no expert in physics but based on my limited knowledge, I've got a feeling that no sonar based solution will work (but I'm basing that off the fact that mylar is quite thin. Maybe this is what you were referring to when you said sonar has inadequate resolution but I'm no exper in the area)... well at least no sonar based solution that is cheap. Pretty sure optical solutions are on the right track and I think it's easier to get right also or maybe I'm just more comfrotable with this tech. Either way, this vacuum revelation has just made me realise that this collimated screen will be much harder to make than I first anticipated.

The other question that I've thought of in light of this information is, is it possible/advantageous to have mutiple vacuum sources i.e. you have pumps in various locations? The way I see it, where the mylar is attached, it will not stretch at all, so there will be differences in tension in the mylar. The only simple solution that I see to this would be to approximate where the mylar should be stretched to from the begining and then attach it in a stretched state. Based on the CAE video, I cant help but feel that this is what is happening when the guy goes over the whole screen on the ladder. You can see distinct vertical lines in the mylar that are equally spaced which to me indicates that the material has been stretched at predefined intervals. So if this is the case, then I suppose basically one vacuum source will be enough but maybe distributed sources would allow greater control over the shape of the mylar?

[vyper883]

I've been thinking about this nasty problem for the last three weeks. So much so that I didn't do a thing with it today - I gave my poor IBM 5160 some quality time instead. *laughs*

First, you need one of these...



...but that's not all.

I'm going to try to get a frame that will fit into the one shown above to retain a solid skin - hopefully I'll get to it tomorrow.

One thing I noticed about the video from How It's Made - the mylar _crawls_. I've only seen that when a material is being drawn against a solid surface. I strongly suspect there's either a solid skin with a Metric Buttload(tm) of tiny holes, or there's a gas permeable solid like MDF there. Fortunately, I've got me a little gadget that's perfect for drilling holes in Metric Buttload quantities.

I should note that the frame in the picture above is a 60 degree wide slice of a 48" radius circle. It's 40 degrees high. The top is at zero degrees. This is a good size for a practical test. It also means that if it works, I only have to build two more to get 180 degrees.

g.

Hi Gene

Up until now, I've been lurking here for a long time, and I have followed the collimated topics from the very beginning. I'll introduce myself officially in the proper thread later. Great job on your prototype !

I've watched the how it's made video dozens of times, and that JUST what I was thinking too. It crawls as if it was being drawn onto a surface. However, I highly doubt that there are any permeable materials into play. IMO, even the slightest of pin holes would probably show up as some kind of distortion on the mylar surface. I think what is happening is:

The mylar has the greatest slack in the center. That would allow the center of the mylar to get drawn down onto the surface FIRST once vacuum is applied, and gradually get drawn down until it reaches the outer edges. That way, there would be no air pockets trapped underneath the mylar. Also, there may be a possibility, that the frame that holds the mylar, might be adjusting the rate of tension somehow, via an automated, or passive mechanism, during the vacuum stage, to allow said gradual contact of mylar, from the center to ouside edges. Of course, this is JUST my theory, but one that may fit the facts.

On the other hand, the mylar itself, may just be thick enough NOT to distort, if there IS some kind of perforated forming block with very tiny pinholes on it's surface. The only thing I TRULY believe after watching that video over and over, is that the mylar ultimately is being drawn onto some kind of surface. As you've said before, the properties of mylar itself, do form a great image, but in time as it is streched and released, it starts to develope more and more slack, that would cause undesired geometric changes in the image over time. To get a constant image every time, it would only make sense to me, to have a permanent forming block behind the mylar. Just my opinion though.

[wledzian]

The other question that I've thought of in light of this information is, is it possible/advantageous to have mutiple vacuum sources i.e. you have pumps in various locations? The way I see it, where the mylar is attached, it will not stretch at all, so there will be differences in tension in the mylar. The only simple solution that I see to this would be to approximate where the mylar should be stretched to from the begining and then attach it in a stretched state. Based on the CAE video, I cant help but feel that this is what is happening when the guy goes over the whole screen on the ladder. You can see distinct vertical lines in the mylar that are equally spaced which to me indicates that the material has been stretched at predefined intervals. So if this is the case, then I suppose basically one vacuum source will be enough but maybe distributed sources would allow greater control over the shape of the mylar?

You are right, there are differences in tension along the edges. There is reasonably even stretch perpendicular to the edge, but almost none at the edge. In fact, this is the main cause of the edge distortion which forms the unusable "dead band".

The vertical lines are just a result of being slightly unevenly stretched during mounting. It's got nothing to do with intentionally stretching at predefined intervals. He's going over the whole mirror with a ladder bit-by-bit to get it as even as he can. If you just attach one edge first, it is very difficult to get the film to lay right, Gene and I ran into that issue if one of us got even a couple inches ahead of the other when attaching the film. If you look at the video of the first application of vacuum on the prototype, you'll see a similar pattern of wrinkles. Distributed sources of vacuum would require separate sealed zones, and would produce more distortion.

[Atomic_Sheep]

Thanks, very helpful. I'm seriously contemplating purchasing some mylar in the very near future to start experimenting with it a little bit.

[SU-GGG Dispatcher]

I was doing some mathematical calculations (Pythagoras) yestarday to design a frame suitable for my cockpit , and i discovered a big barrier , it's the "huge Space" .. but i think it would be a great way for the design if we all recognized the validity of it ..

I'm not sure if those calculations are correct or not , but they seems so logic .. correct me if I'm wrong ..

well , my Cockpit width is 165 cm ( Sorry for the non-US unit ) , so this is the idea of the calculations :-

1- Forward- Projection screen's radius
- I began by the base of the Forward-Projection screen , I planned to be as width as the cockpit .. plus say 5 cm , so the light rays would be free of hitting the cockpit , because it's the first task to guarantee that all light rays from the Forward-Projection screen would hit the whole of the mirror totally , and after i guarantee that task , i can plan the mirror radius / widths / etc according to those calculations .. See Picture below !




2 - Mirror's Radius
- After i calculated the Forward-Projection screen's radius , I then calculated the Radius of the Mirror .. The idea is the Screen would be in a distance from the Mirror less than the focal length , So i first assumed the screen would be 20 cm in front of the focal point , so the focal length would be [ screen's radius - 20 cm = 249 - 20 = 229 cm ] , Now I could calculate the Mirror's radius by multiplying the focal length by 2 [ 229 x 2 = 485 cm ] , See Picture below !



3- Mirror's Lower ( Bottom ) width
- The triangle of the lower width , it's base would be the Mirror's lower width that we aim to calculate , as we know the total sum of any triangle = 180 degrees , so as the base angle = 40* , so both the other angles = [ 180 - 40 / 2 ] = 70* .... & see the Picture below !!




4 - Mirror's Upper width
- see the picture , i guess it's now clear !!




Remarks
- The first 30* angle , to draw a latitude of 30 South ( If we looked to the screen as the earth )
- The 2nd 40* angle , because our Field of View ( FoV ) = 40* .. drawing latitude of 70 South .
- The opposite side is the same as above .
- As the half circle = 180 degrees , so the remain angle ( Mirror's Lower width's angle ) would equal [ 180 - {(30* x 2 ) + ( 40* x 2 )} ] ...

Looking forward for your thoughts

Mohammed Sayed

[castle]

For those who have been following this discussion or anyone for that matter here is a link to an excellent set of tutorials on how image warping works and techniques for projecting flat images onto curved screen surfaces

http://local.wasp.uwa.edu.au/~pbourk...hics/cylinder/

For the more adventurous, there is source code on the OpenSceneGraph website (http://www.openscenegraph.org/) detailing the mathematics.

JW

[Hessel Oosten]

And here ....

On this site:

http://www.mod-gmbh.com/Collimated-Display.48.0.html

is an interesting paper about the collimated display principles:

http://www.mod-gmbh.com/fileadmin/us...Sale-Paper.pdf

Hessel Oosten

[vyper883]

Hi Guys

Forget mylar.....
Get a load of this: http://www.youtube.com/watch?v=HIqxd...eature=related


And THIS: http://www.anishkapoor.com/works/gal...urve/index.htm

If you want even more info, just do a google search for "Anish Kapoor C-curve" This guy is a sculptor, so how did he make this? If we find out, then things get a bit simpler.

[Matt Olieman]

Hessel, Thanks for the links. The last link is VERY interesting.

Matt Olieman