Collimated display build thread...

Ok guys, I know it's been a long time without updates, so I decided to create an "official" build thread that will show our progress.

Wayne & I have both been pretty busy with the holidays, etc. and we've not made a great deal of progress that's very photogenic. :)

I finished the last sheet of parts a couple of weeks ago, so now the mirror framework is ready to be assembled.

The parts:

http://www.geneb.org/images/mirror_f...omponents1.jpg

http://www.geneb.org/images/mirror_f...omponents2.jpg

The parts marked "TOP PLATE" and "BOTTOM PLATE" are cutting jigs. In order to properly fit the "mask" that the mylar is attached to, the top & bottom of the mirror frame has to have beveled doubler plates attached. The jig was attached to my band saw and then I tilted the saw base 20 degrees and Wayne ran the parts through it.

The doubler plates are made of 2 or 3 layers of 3/4" plywood. Here are some in the glue-up stage:

http://www.geneb.org/images/doubler_glue_up.jpg

When cut, it looks like this:

http://www.geneb.org/images/bottom_doubler_arcs.jpg

The pic above shows the 2 layer bottom doubler. The top is made of 3 layers.

Here's what the cut-off from one of the top doublers looks like:

http://www.geneb.org/images/top_doubler_cutoff.jpg

The next step will be to cut out the parts that make up the table that the mirror framework will sit on, and the base platform that the table will rest on. Some of that work should happen tomorrow (2/27/11) providing things go as planned.

The really hard work has been done - it's all down to assembly now. :)

We've also come up with a neat way of managing the shape of the mirror that allows us to continue to use a shop vac as the vacuum source. Here's a short video that describes the Arduino controlled bleed-air valve:

http://www.youtube.com/watch?v=QC7H61palXs

The current iteration of the firmware works much better than what is shown in the video. I'm also in the process of writing a stand-alone application that will show data from the PID controller as it does it's job and will allow us to tweak the various controller parameters real-time.

Thanks for reading folks!

g.

Looking really good... how many of these sensors are you using? I'm guessing just one? I suppose you would use 1 per vacuum section. Would be possible to create a more complex one with multiple vacuum cleaners and subsequently more sensors but I suppose that would probably only be necessary for really large areas. Question is... how much larger do you think the setup can get before you are no longer able to get away with just 1 suction source?

There's just one big chamber behind the mirror, so there's no way to individually control separate parts of the mirror. A single sensor is sufficient; we don't need to know the exact position, only whether it's too close or too far from the sensor. The firmware does the rest.

There's a lot of leakage around the prototype frame, the final version will be much better sealed. Even so, we reached a point during testing where we were trying to eliminate as much leakage as possible, where the valve box full-open position was still too closed-down to keep the proper vacuum level. We don't anticipate any problems with the single suction source.

The prototype is using 2 mil mylar; the final version will use 1 mil mylar, requiring half the vacuum draw. If anything, the vacuum we're using now is extreme overkill.

Overkill? Wazzat? :D

g.

Excellent... thanks for that.

If anyone is curious, it takes four sheets of 3/4" plywood for the complete mirror chassis. It looks like the table may take two sheets. With any luck, we'll get some of that cut & assembled this coming Saturday.

g.

Quote:

---

Originally Posted by geneb

If anyone is curious.......

---

Anyone ? Anyone ?
We all... sit every day for our screens, waiting for more !

Thanks both so much for all this information.

Hessel

Thanks Hessel. I (and I'm sure Wayne does) appriciate your and the others on this forum for their enthusiasm for our project.
The good news is that we're closer to the end than we are the beginning at this point. :) With any luck, much sawdust will ensue tomorrow. (FYI, a 30 gallon garbage can full of sawdust is heavier than you'd think! - I have a "turbine extractor" (fancy name for a stupidly expensive special garbage can lid) that separates the heavier chips into the garbage can, keeping them out of the DC)

g.

I look you sistem of control mirror position, and i see oscillation in first time: forward, back, forward, back.
How you have solved this problem?
Or oscillations fade also position is stabilized?

We're using an arduino-based PID controller to control the mirror position. Our controller is not yet tuned in that video; Gene is writing an interface that will let us view the sensor readings so we can tune it better. Once everything is tuned properly, there will be much less oscillation.

I just finished putting sanding sealer on the table parts. Flat black paint is next.

http://www.geneb.org/images/table_parts.jpg

The three primary legs and the two "ear" legs have been assembled.

g.

We got a ton of assembly work done today!

http://www.geneb.org/images/base_underside.jpg

The above photo shows the underside of the base plate that the mirror table & framework sits on. The base is made from 2x4s with OSB sheeting. I got the base assembled & painted a few weeks ago.

We got the mirror framework assembly completed today. Here's what that looks like:

http://www.geneb.org/images/mirror_framework.jpg

Here's a picture of me sitting in there to give you an idea of scale:

http://www.geneb.org/images/framework_gene.jpg

I'll try to get more frequent posts as we get more done.

ttyl!

g.

More pics! :) Wayne & I got some great work done this past Saturday and the framing parts for the new screen were cut out of Ultralight MDF. (it weighs roughly 60% of the standard MDF you get at your local BORG)

http://www.geneb.org/images/210_screen_frames.jpg

http://www.geneb.org/images/210_screen_frame_joint.jpg

http://www.geneb.org/images/210_screen_frame_joint2.jpg

http://www.geneb.org/images/210_frames_and_rings.jpg

The large half-circle assemblies are made up of a few different parts in order to get the max utilization out of the 4x8 sheet of ULMDF. One will hold the projectors and the other will form the "lid" of the screen. The small arcs are mounting rings that will become the upper and lower frames of the "blades" that will form the screen. It's hard to see now, but you'll get the idea really fast once it's assembled.

g.

Yesterday (5/14) we got the screen & support "spider" assembled from the parts that were cut last week.

Wayne's got a slick little camera that does time-lapse video, so he set it up to capture the screen assembly process:

http://www.youtube.com/watch?v=t7553ZkkPXY

I've got pictures of the whole thing that I'll get posted soon.

g.

I love it!!! Thanks :)

Matt Olieman

This update should cover the past couple of weeks worth of work... :)

The "spider" arms for the screen & projector support platform were cut out last weekend.

Here's what they looked like right after getting sanded & painted with shellac sanding sealer:

http://www.diy-cockpits.org/coll/ima...creen_back.jpg

In order to strengthen the material joints in the screen support parts, we needed to cut a set of doublers. Part of this was seeing if we could get the 'bot to do the countersinks for us. I had a 3/8" 82 degree countersink that I could fit to the 1/4" collet on the 'bot. Wayne created a file in VCarve that would have countersinks at various depths - turns out .210" is perfect for a #8 flat head screw:

http://www.diy-cockpits.org/coll/ima...untersinks.jpg

Having the 'bot do all these is a LOT easier than me doing it with a microstop!

http://www.diy-cockpits.org/coll/ima...ountersink.jpg

This is what the doubler parts looked like as they're being cut:

http://www.diy-cockpits.org/coll/ima...g_doublers.jpg

...and after being glued & screwed to the screen support panel:

http://www.diy-cockpits.org/coll/ima..._installed.jpg

After getting those in, we added the "spider" legs - you can see from the picture why I call them that:

http://www.diy-cockpits.org/coll/ima..._installed.jpg

Here's a picture of what it looked like after we got the top layer added.

http://www.diy-cockpits.org/coll/ima..._assembled.jpg

Like most spiders, this one doesn't like being on its back...

http://www.diy-cockpits.org/coll/ima...er_on_back.jpg

Here's a long shot with the screen support resting on the mirror framework.

http://www.diy-cockpits.org/coll/ima...side_wayne.jpg

Just for the heck of it, Wayne clamped in the first screen section we ever built - this is actually a small slice of the full size screen you saw being assembled in the time-lapse video.

http://www.diy-cockpits.org/coll/ima...est_screen.jpg

I figure I'll hang on to that ratty old thing - who knows, the Smithsonian might want it some day. *laughs*

http://www.diy-cockpits.org/coll/ima..._and_wayne.jpg

This picture is pretty amusing. Wayne is all, "Hurr! This is SO cool!", while my friend Rob is all, "Hrmm. I do believe we CAN take over the world with this!" :D

Please note the pink foam screen analog. :D It's about the right size for a 19" display.

Here's a couple of shots with the screen installed:

http://www.diy-cockpits.org/coll/ima...ng_heights.jpg

http://www.diy-cockpits.org/coll/ima...en_mounted.jpg

I would have taken a wider angle shot, but I couldn't back up any further while it was in the location it was. Stupid thing is so big you have to back WAY off to get it all in one frame!

During the week I'll be applying sealer to the screen frame components and the screen mounts so they can be painted. With any luck we'll be able to get the foam inserts installed into the screen soon!

g.

This is truly amazing... SO impressive!!!! NICE WORK!!!!

Matt Olieman

Update for the work done on Saturday, 21May11:

The screen support "spider" is completed and painted:

http://www.diy-cockpits.org/coll/ima...ns-support.jpg

The only remaining work that needs to be done is to have the projector mounts located & drilled.

We got the foam blocks cut for the screen. That stack of material used to be four 96" x 24" sheets. Wayne built a slick jig that allowed them to be bevel cut without getting your fingers too close to the saw blade.

http://www.diy-cockpits.org/coll/ima...for-screen.jpg

The blocks shown above were hot-wire cut in a special jig that used spare ribs from the screen in order to get the shape perfect.

Look carefully at this picture:

http://www.diy-cockpits.org/coll/ima...ab-attacks.jpg

Can you understand why I'd want to whack it with a crowbar as I went by? :D

Each mirror segment and the backs are sealed with a 1/32" rubber gasket to minimize the amount of air leaks in the system.

http://www.diy-cockpits.org/coll/ima...panel-seal.jpg

Applying the framework back panels...

http://www.diy-cockpits.org/coll/ima...ack-panels.jpg

The colors for the frame will be yellow for the "ear" boxes and peripherial framework and blue for the back panels. Gotta keep with the whole Simpits/Link Trainer colors you know. :)

We also test-fit the new goggle-mask.

http://www.diy-cockpits.org/coll/ima...k-test-fit.jpg

http://www.diy-cockpits.org/coll/ima...-test-fit2.jpg

It's quite likely that when the mask is attached "for real" to the framework, the mylar for the mirror will follow shortly behind. Right now we figure we're only one or two work days from performing the first draw-down of the mirror.

The screen was assembled and the first layer of filling compound was applied.

http://www.diy-cockpits.org/coll/ima...or-sanding.jpg

Once the sand/fill cycle is done and the shape is correct, I've got some Behr screen paint we're going to try with it.

Thanks for reading!


g.

We made more great progress yesterday. Once we were confident that we'd have the mylar in place, we decided to take advantage of the sunny day for a barbecue.

We worked late into the night, and got the mirror mounted, the screen painted and mounted, and one projector tested.
Drawdown of the 220° mirror

More to come!

I agree, it's insane ;)

FANTASTIC!!!!!!!

Matt Olieman

Thanks Matt. Wayne & I are very proud of that beastie. :D

Here's a walk-around I did today:
http://www.youtube.com/watch?v=wdr0-PVqjMk

g.

Thanks for the walk-around. I get goose bumps every time you show your project. :)

Matt Olieman

!!! Masterpiece !!!

Thanks again guys.

Hessel

it is cool! I wait when you switch on a projector!

The build is DONE! I wish I could have been there to see first light in person.

http://www.youtube.com/watch?v=Bx5eb5yTDQw

We still need to tweak the image warping, but that's just details at this point. It works!
Now that we've worked out the vast majority of the kinks during the design process for display #1, display #2 should come together much more quickly. I guess this means I should start cleaning up my office...

Heh. Yeah, you're gonna need every spare square inch you can find. :D

There is actually one more thing to make - the cap on the top of the whole thing. I'm thinking some 3mm black Coroplast would do the job quite handily. We can attach it using velcro. The display is quite bright even with the shop lights on, but it looks nicer with them off. We'll need a "blackout skirt" around the outside perimeter of the table too - black canvas would work there.

g.

All I can say is "WOW" and then again "WOW" :) :) :)

Matt Olieman

Gene and Wayne,

You guys have been doing great work. Very impressive!
Best,
Mike

Thanks folks! The nice thing is that mine ate up all the growing pains so Wayne's build shouldn't take more than a few work days.

g.

Just a few tweaks left. I'm going to be cutting parts for new projector mounts in a few minutes. The mounts I purchased are great for ceiling mounts. However, they're crap for what I'm using them for. You look at them wrong and they move.

This is VERY bad, especially when it takes 2+ hours to align the pre-warp maps properly.

here's the new mount:

http://www.geneb.org/images/pj-mount-small.png

It takes up the same space as the original (3-1/4") but can be locked into place much more firmly. The top bolt will hold the mount to the support frame and some sandpaper glued to the top will ensure it never moves once the wing nut is tightened down. I'll add a metal spacer to the pitch axis adjustment so it won't break the mount if that gets tightened down too much.

g.

FSX, fully aligned and edge-blended. Hopefully, we only have to do that once; I now understand why the visual techs get paid so much to do a projector alignment.

Personal note: There's no such thing as "leaning in to get a closer view" on a collimated display.

<iframe width="560" height="349" src="http://www.youtube.com/embed/4R2bcZ-3eX8" frameborder="0" allowfullscreen></iframe>

\o/

For those curious, that's the default FSX Grumman Goose and I was sitting in Friday Harbor at the time.

Wayne flew from Boeing Field/Renton (?) to Friday Harbor and I flew from there to Arlington. The flight really brought home the fact that I've forgotten most of the stuff I learned about navigation. Time to drag out the textbook again. *sigh*

"Hey look! It's a golf course!" *laughs*

g.

The most amazing DIY project I have ever seen. Absolutely amazing.

Andy

spent last week sailing the san juans with my son.

-cringes-
I just remembered I'll have to do at least one more alignment and blend on that setup to handle programs which use only a single window.
Then I'll have to do it all over again once mine is built.

I'll have to give NFS a shot and see how it does. You'd think because you blended the projector outputs that it wouldn't be different on a per-application basis.

I'm going to stop off at KPLU on the way home to pick up an updated sectional. :)

g.

That is f a n t a s t i c !!!!!

I must admit that this is a very most difficult Project I have ever seen in mc.org

The edge-blending pulls image from the area adjacent to the 'channel' area as source for the blend region. Since we're using multiple FSX windows, the edge-blending would wind up exposing the window borders. To get around that, I made use of the countermap to pull image only from further within each window, so that the edge-blend never gets to the border. The result is that each source window has a larger FOV than normal, but there's no vertical black line in the blended image. Additionally, the countermap is intended for three individual windows, and won't look anywhere near correct with a single source image.


With a single source image, I'll need to re-align and re-blend with the standard grid, and apply a countermap designed for whatever distortion is present in the source. If you take the existing map and turn off the countermapping (make a copy of the config first!), it should be a good start.

If I recall correctly, you mentioned that none of the suppliers you contacted would sell you rolls wider than 56" in affordable quantities. May I ask what the smallest roll of >100" wide would cost, and which supplier?

From what I've been told, you can buy from L3, Rockwell-Collins and CAE. Expect to pay in the high five to low six figures. I don't know what length rolls they sell are.

g.

Those are Simulator/Visual manufacturers.
I happen to know you can get a professional mirror re-skinned incl. material and labour for under 50000$.
There must be mylar suppliers that can offer rolls wide enough (10' or so?) for reasonable (4 digit or low 5-digit) prices.