ShopBot Visuals...

[geneb]

For quite a while now, a friend and I have been working on an "impossible" project. The site linked below shows our progress in building the world's first DIY collimated cross-cockpit display. This is a display system for commercial flight simulators - you can learn more about how they work here:
http://en.wikipedia.org/wiki/Full_flight_simulator

This is actually a pretty big deal for us and for the homebuilt flight simulator community in general because "common knowledge" says that these displays can't be built by DIYers. Reasons have included "too complex" and "too expensive". Well thanks to my PRT Alpha 96, "too complex" was whittled down to "mildly difficult". :)

Here's the project site: http:/www.diy-cockpits.org/coll

Thanks for reading!

g.

[myxpykalix]

although i didn't read everything, scanning thru the pics you get a sense of how the project is progressing. Looks like a lot of work and you certainly are documenting it well. Good job.

[geneb]

Thanks Jack. I tend to be obsessive when documenting projects - you should see some of my other stuff. *laughs*.

g.

[Acmeaviator]

Nice Gene - I've got a simpit in my den that I built last year. Nothing super fancy - Y E T:D

[br928]

Gene,

Very interesting project.

How many projectors will be used and how will you blend them together? I assume this is a front projection screen.

After looking at some of your vids, it seems like you could use an off the shelf PID controller to operate the bypass valve for the mirror draw-down. They can be bought for $100 or so.

[Brady Watson]

Very impressive, Gene! I love the way that mylar sucks down and how good the results are!

I know you have been working on this project for a while now...Nice to see this milestone!

-B

[geneb]

Stan: The system will use three Epson 705HD projectors @ 1280x800. They'll be driven by a Matrox Triple Head 2 Go. Distortion correction is being done by a program called NThusim Plus. It enables us to "wrap" the image to the spherical section formed by the screen.

The system IS front projection, but you're viewing a reflection of it, not directly at the screen. The projector fires out to a first surface fold mirror and that bounces the light to the screen. You view the reflection of the screen using the spherical section mirror formed by the Mylar.

The PID controller is based on an Arduino, which you can get from $9 to $30 depending on the vendor. $100 for a COTS PID controller is a pretty poor deal in the face of that. :)

Brady: Thanks! It's taken us so long because Wayne & I really only work on it one day a week. Had we been doing this full time we would have gone from the prototype to the full size mirror in about 30 days. :)

g.

[br928]

I have had the Arduino controller show up in a number of people's applications. I guess I am going to have to buy one to see what all it can do.

My background is the process control industry, and a standalone full featured PID controller w/ setpoint and process variable display and alarm output for less than $100 is cheap by industry standards anyway!

The way you are mixing the video is intriguing.

[geneb]

The Arduino is both a software & a hardware stack. The core is typically an ATMega328 microcontroller. The software stack provides a bootloader that's locked (so it can't be accidentally overwritten) that allows you to download new "sketches" to the board over a regular serial port. No special programmer is required. It uses avr-gcc as the back-end compiler, but it's piped through a bit of an abstraction layer to make it easier for non-programmers to get going with. http://www.arduino.cc has a ton of info. SparkFun is the place to go if you want to order one. (http://www.sparkfun.com)

If you want an Arduino you can plunk on to the end of a breadboard, you can't beat this: http://adafruit.com/products/72 (Boarduino)

When finished, the controller will be pumping data to a host machine that will display the PID data as ongoing graphs and will allow tweaking any of the parameters in real time. The client is written in Delphi and we'll be showing that off once it's a bit further along. Most of our time has been sunk into the mirror as the "hard bits" for the bypass gate controller were solved.

g.

[Brady Watson]

The Arduino is cheap enough to justify tinkering with one. Glad I don't have to program it in ++.

For El Cheapo PLCs - I like these: http://www.tri-plc.com/e10intro.htm

-B

[MogulTx]

Gene,

What is the approximate screen radius you chose to work with? It looks pretty short. I have some experience in collimation and would be curious about the primary and the seconday radii (the projection screen).

How did you get involved in this? Will you use it for entertainment? Is it a real budget hog?

Let me know. I am working on a single channel system that will be 30+ degrees high and nearly 60 degrees wide on a 37" radius. ( Too short to use as a cross cockpit. Must be a single eyepoint only.) I am not using my bot for much of it! My mirror will be a solid mirror. And if you ever had a need, I MIGHT be able to scrounge you up a solid mirror... if I know the parameters.

Best Regards,

Monty

[geneb]

Monty, the mirror has a radius of 4 feet and the screen is about 2 feet. The size was limited by the Mylar we could find. A 54" wide roll is just wide enough to cover the conic section that the mirror framework defines. We found out recently that you can purchase really wide rolls from the likes of CAE, Rockwell-Collins, etc., but you're looking at $5k for a roll. :)

With the size of mirror we built, we can do cross-cockpit for "typical" GA types like the Cessna 172 or Beechcraft Bonanza. We're looking into the possibility of building something along the lines of an Embraer of some type as well. (I'm also going to build a single-seat fighter type and a single-seat car rig for Need for Speed:Shift)

I've been deep into home-built flight simulators since about 1997 or so. It's gone wildly out of control since then. :) This project started when I wanted to develop a better display system for the smaller cockpit design I'd done. I had built the prototype framework, but was chasing down a blind alley. I was going to use "gores" of material to give myself a direct spherical section display screen. The collimated display came about when Wayne Ledzian contacted me via the Mycockpit.org forums. He knew the math that had to be conquered in order to pull this off. It was pure accident that the prototype framework I built just happened to be the right dimensions - three of them put together build the 8' max diameter the Mylar can cover.

Anyway, between the two of us we managed to get enough cats running in the same direction to pull off the prototype mirror. This was a pretty big deal too - it had never been done by hobbyists before. Not only that, it was considered impossible due to the cost/complexity of it. The bad news is that we had to use a patented feature in order to properly form the mirror. We're going to try to obtain a non-commercial license for it so we can publish plans so folks can build this themselves. Without that, we'll only be able to publish so much and then go, "look at patent so-and-so and figure that bit out yourself".

The only real complexity comes with the math - that's pretty well understood and Wayne has built this awesome spreadsheet tool that spits out the magic numbers for you.

The prototype mirror used 2mil Mylar, the full curve mirror uses 1mil Mylar. Wayne scored a 50' roll (54" wide) from a local vendor for about $40.

I've probably got about $500 in raw material for this project and not all of that was spent at one time. In the wide scheme of things it's actually a pretty in-expensive project to build. Well as long as you don't factor in the cost of the 'bot. :)

With a solid mirror, are you going for a beam-splitter type collimation system?

g.

[myxpykalix]

"The prototype mirror used 2mil Mylar, the full curve mirror uses 1mil Mylar. Wayne scored a 50' roll (54" wide) from a local vendor for about $40."

Gene,
Where is this local vendor? The prices i've seen were higher then that?
I am trying to make a solar water heater and need some mylar. I want to make a big curved trough with this mylar in the back, what in your opinion is the best thickness? I know you use 1,2 mil but you need it to form to your curve off and on.

[geneb]

Jack, the vendor is http://www.mirrorsheeting.com/ and they're local to us so we don't have to pay shipping. :) I was incorrect on the price though - the 50' 1mil is $31.

I would recommend the 2 mil or thicker. 1 mil is VERY delicate. It won't tear that easily, but it's very easy to mar the mirror coating.

g.

[myxpykalix]

thanks gene