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


I finished something! I FINALLY FINISHED SOMETHING!

*laughs*

g.

hi, i went to the link and seen a blank video nothing even played??

Awesome looking setup there Gene. I have been keeping up with this for a long time and a few questions keep coming back. With the background noise in the video I am assuming that is the vacuum to pull the display tight; so when you wear headphones or use the volume in game do they overcome the sound of the vacuum? Also I remember once that you were trying to accomplish this without the vacuum but it appears that you were unable to do so; so would any old vacuum work and have you considered moving the vacuum to any other area away from the computer but still attached via a longer hose to reduce the noise? I know that a vacuum works better the closer to the source it is pulling against but if you move it farther away would that effect the display any?

Looks great!

@The Nephilim - the video should play - it's just a regular YT video. You might try going to http://www.youtube.com/f15sim and see if that helps at all.

@Reschke - The vacuum is actually about 18' away and on the other side of a shelving partition. I still have a few things to do to completely silence it.

As long as you don't pinch the diameter of the feed line, you won't reduce the vacuum over a long distance. It just takes a bit longer to pull the mirror down because it would have to remove all the air from the feed line before it could start on the mirror plenum volume.

Thin film mirrors are really the only way to go with this - the solid form was a blind alley for me.

tnx.

g.

That is too cool! Congrats on success of the project so far.
Derek

Wow congrats Gene that is truly amazing! But will it fit around the F-15? LOL!

Nope, won't fit the F-15.

g.

Hi Gene,
Like Resckhe I've been following your progress with interest. Very nice to see all your hard work finally paying off in such a spectacular way.

I suspect, like most of us, you'll be tweaking until the day they put you in your box .......

Well done !!!

Andy

Been following this for a while. Congrats on getting it finished!

Looks awesome in the video but I bet that's nothing compared to being there...

Yeah, the video camera really can't do it justice. More map tweaking to do on the screen distortion, then comes the Need for Speed: Shift test!

g.

Nice job

Thanks! I've got a Need for Speed video I'll be posting this weekend.

g.

Hey Gene, congrats on a really impressive build. Look forward to seeing it in action, nicely done!!

sweet

Gene, that is looking very cool! I am very happy for you

I have a question about fs rendering or other games for that matter. I know when using triple screens the side monitors stretch the image. This isn't so bad for some stuff as your focus is at the center monitor and because I use trackir when I look to the side and am still mostly looking at the center monitor.

With your setup you are not using track ir so I wondered if you had one large eyefinity type display or were you rendering the front and each side separately.

On a similar note, the software you use to create a map to remove distortion, can that be used to compress the stretching that happens on games that do not support rendering 3 screens separately?

Anyway, looks great!

Any image generated as a single perspective window will have stretching towards the edges. This is an unavoidable consequence of a perspective projection. Another consequence of a single perspective projection is that it is geometrically impossible to reach 180°, and the distortion gets really bad as the FOV gets large.

To get around this limitation, we're using 3 separate windows in FSX. That can be used to eliminate the stretching, but cannot produce more image than is generated, as in a game that uses only a single window won't ever be much good over ~120°, if you're looking for a 'correct' image. With FSX, because we're using three separate windows, we are able to go beyond 180°. I emphasize 'correct', as Need for Speed (a single-window game) looks absolutely incredible, even when it's stretched to 210°. Since you're not paying much direct attention to the sides, it's just peripheral vision and adds tremendously to the immersion. FSX, on the other hand, would feel strange and make pattern-flying difficult if the view directly to the side were showing something that was significantly in front of me.

Keep in mind that we're using a trial license of Sol7 at the moment, which has the capability to apply the screen fitting warp as well as a secondary geometry-correcting warp. There's a good chance we'll go back to NThusim at some point, as Sol7 on 3 channels retails for ~$6000. NThusim is capable of removing the distortion, but is only configured with an alignment routine to fit a rectangular grid to your screen. If you want to correct the distortion further, you need to know what the grid should look like when the image has been corrected, and make it look like that on your screen.

Well I'm an avionics technician from Duncan Aviation. I just finished some training on the Cessna Sovereign at Flight Safety in Wichita.

Part of our training was hands on time with one of their real full motion collimating display cockpits set up with the Sovereign cockpit and Epic avionics. What a blast.

I have to say, I'm impressed by this project. I hope that you guys keep at this because I'm hoping to try to make one of these myself soon. I've been bitten by the simpit bug.

I'm curious how you came up with the dimensions for the Mylar support structure?

Anyway, great job guys.

Thanks! You've got a fun job I bet. The dimension was picked out of the air - it just happened to be the right size for the max Mylar size we could get. I had no clue about it until Wayne got involved - I was going to use a different system all together (that would have not worked).

g.

Yep. The frame dimensions were a lucky accident. I had done some design work independent of Gene, and used a 4' radius as the largest mirror I could fit in my office. I don't know exactly why Gene chose 40° from the equator down, but the math showed that we could work with it. It worked well when we built the prototype, so we stuck with it.

The screen dimensions, on the other hand, were not so random. I spent a good couple weekends programming a spreadsheet to calculate the required screen geometry when given an eyepoint and desired field of view. By picking a field of view and using it to locate the eyepoint, the program spits out the screen shape.

It was pure luck that:
- Gene had selected the same 4' radius that would fit in my office.
- The geometry that Gene had already used for his first frame worked out well.
- His existing racing pit fits under the mirror with the eyepoint in just the right spot
- the widest Mylar we could find is just barely big enough to cover the 210° FOV we wanted (three 60° frame segments plus ~15° usable in each 'ear')

Well today is my last day of training. I'm hoping to get back over to the sim center to see if I can get a look at how the projectors they have are set up.

At any rate I would love to see the dimensions you guys came up with.

The 40 degree vertical came from an FAA document I'd read on the Level D FFS certification requirements.

frans42000, In most commercial FFS systems, the display uses a translucent screen and the projectors are configured to "rear project" on that screen. We're having to do front projection because a) it's very expensive to manufacture a toroidal screen with a ~4ft diameter out of plastic and b) there's not enough room inside the screen to properly position three projectors. (well not easily anyway)

I'm hoping at some point we'll be able to release plans for this monster, but that all depends on whether or not we can obtain a free or reasonably priced license for the patented feature we have to use. (the ears)

g.

So is there any publications that a person can read about the collimating mirror to help them design their own?

There was significant discussion on how to design collimated displays on the MyCockpit site. See: http://www.mycockpit.org/forums/forumdisplay.php/322-Collimated-Display-Discussions

I suggest starting with the "Question about collimated display systems" thread. There are are few experiments that didn't pan out, but there is also a lot of really good material as well as a few pointers to reference material.

Thanks for bringing that up Mike. I'd completely forgotten about it.

g.

Well, I guess what I was really interested was if the radius of the top of the support frame is the same as the bottom.

From first glance, it appears that the display is not just a toroid, but also...uhh...cone shaped.

if its just a 210 degree semi circle, I had an idea to use a section of used truck or tractor tire as a frame. But I doubt it's rigid enough for a vacuum pressure.

Keep in mind that the framing is designed in such a way that when the Mylar is under vacuum it forms a spherical section, not just a conic section. Take a look at the pictures I posted here: http://www.diy-cockpits.org/coll/ and you'll see how it goes together much more clearly than I could describe it to you.

g.

Just plain awesome Gene. I've been lucky enough to see some collimated displays first hand, so I can appreciate the effect you've managed to create. I just can't imagine how awesome it looks when it's got a racing game or CLoD running on it though!

Brilliant , simply brilliant I have just spent the morning going through all of your posts, pics and vids Gene.Very nice work the two of you have accomplished. Look forward to seeing some CloD action on that setup

What's your thoughts on making it a commercial type venture? Has anyone shown any interest in it ? I would think you could install something like this at airports or game arcades and make a bit of a return on your investment?

Top notch stuff

Thanks Ajay. Because of a patent we used in the construction, we can't sell it or plans for it. We're working on coming to some kind of agreement to get around that though.

I'd never build something like this for use by the general public. It's not something you'd want John Q. Beerswiller anywhere near.

We found out last Saturday that at about 80 degrees ambient, the adhesive fails. There's a corner of the mirror that only had about a 1" band that was attached to the frame and that area wasn't enough to prevent the adhesive from "creeping" and letting go of the Mylar.

This means that this coming Saturday it all gets torn apart and we'll install new Mylar for a *third* time. The good news is, we're getting good at it, but the bad news is, we're getting practice at it!

I'm going to take this as an opportunity to put the system on a new, more flexible mounting table (no platform base this time) and to neaten up the wiring as best I can.

What sucked is I had the video cabling all nice and neat, and then found out that the Digital TH2Go won't work with the DVI to HDMI cables I bought, so I had to go analog VGA.

Here's a video that Wayne took the weekend prior after a marathon warp-map & edge-blending session:
http://www.youtube.com/watch?v=4R2bcZ-3eX8

We took the FSX Grumman Goose from Renton to Friday Harbor and from there to Arlington. It was pretty cool.

As to CLoD...well when it hits the $19.99 bin, I'll put it on my todo list. I'm not interested until it has DeviceLink support and then only if they haven't lobotomized it for multi-player like they did to the rest of the IL2 series.

g.

Hmm, a foam rubber backed bezel would reduce the chance of the adhesive slipping, and if painted matte black would get rid of stray reflection from the portion glued to the frame. the trouble is to find a way to fix it to the frame without altering the structure. Strong magnets, perhaps?

We plan to attach a black fabric surround to cover the portion of the mirror in the 'dead band' (image distortion due to non-spherical curvature). Under vacuum, there is roughly 9 lb/inch of tension applied along the edges. We've tried applying a second layer of mylar over the slipping area to double the adhesive contact patch, but it did nothing. A rubber-backed bezel would similarly do nothing to stop the slippage. The issue isn't that the glue isn't sticky, it's that it is no longer 'solid'. It becomes viscous above a certain shear load, allowing the mylar to slowly slide off. Apparently at ~80°F, the adhesive can hold 3 psi shear, but will not hold 4.5 psi shear.

The only real solution here is to re-skin the frame, being extra-careful to preposition the mylar so we're not left with a narrow border.

At least we're getting more practiced at it. By the time we skin mine, it'll be a no-brainer.

9lb/in. of tension? That works out to be just about 3300lbs. It's no wonder the mirror plenum moves a little bit under draw. *laughs*

g.

Lol Glad you guys are having so much fun with your work.I am struggling at the moment just trying to set up a triple screen deal at home...can you say ..missus is not impressed with my notion of needing more real estate...to fly your silly aeroplanes, i suppose if it does crysis it will be ok,i assured her it will do crysis .

On the il2 lobotomy front..if there wasn't an FMB the thing would have been doomed years ago on the single player front.And as you would realise, CloD hasn't done that reputation any favours

Ajay, you can always try the, "Well dear, it's Flight Simulators or Strip Clubs. You pick." ploy.

g.

Gene, your work continues to amaze me! It really is impressive.

I was so impressed I approached my wife with the idea of taking a part of the garage and turning it into a 'big screen' pit. She was almost sold on it until I told her I would need to cannibalize parts of her new vacuumn cleaner to draw the reflector tight...... The doctor says the stiches will come out in a week, but they haven't figured out how to get the joystick out of my ^@@...

Keep at it and I will live vicariously through you!

Dud

Heheheheheh

g.

Gene, if you were to put a valve between the box and the vacuum hose, how quickly or slowly would the screen loose it's focus? Just wondering how well you guys have managed to seal all the micro gaps. Would it be ready for the next stage, when a pressure detector notices a threshold being reached, the vacuum kicks in to decrease pressure? (With accompanying electronic valve)

There is enough loss in the system that it can't be just gated like that - even commercial mirrors use active vacuum management.

What the current system does is use a IR emitter/detector pair to "know" the position of the Mylar.

We configure the set point by adjusting the Mylar position until it just starts to touch a temporary marking block inside the mirror plenum. The set-point is adjusted with a potentiometer - the bigger the number, the farther away the Mylar is from the detector and vice-versa.

The controller is an Arduino Duemilanove that drives a servo that operates a bypass gate via a big plywood gear. This regulates the amount of vacuum that is being applied to the Mylar. Right now the system "breathes" a bit because we're still in the process of getting the controller response curves tuned in. I use the term "breathes" because the Mylar can be seen going in and out, similar to what your chest does when you breathe. It's not much, but it's enough to notice. Once the response curves in the controller are tuned, this effect will be either totally eliminated or greatly minimized.

We didn't get to replace the mirror this weekend, but that worked out because we're going to use the "extra" time we got to tune the controller a bit better using a live graphic application I wrote. It shows graphs of the PID values as well as the current mirror position in real time. Parameters can be tweaked and then saved to the Arduino's EEPROM.

Once the controller is tuned, the mirror will provide one last bit of data before it gets replaced - if it all goes horribly wrong, can it be popped with a total blockage of the bypass gate?

Here's a screen shot of the software I wrote:



There's some "debug" features on the screen (like the X= and "Memo1" areas) that won't be there in the final version, but you get the general idea of how it works.

For those curious, the application was written in Delphi XE.


g.

Nice job. And "Duemilanove" sounds like "two thousand nine" in Italian.

JAMF, It actually means exactly that. The Arduino was originally developed in Ivrea, Italy (see http://en.wikipedia.org/wiki/Arduino for more info).
Gene, if the mirror can't handle a completely blocked valve without popping or stretching out of shape, you could always put in a sacrificial window made out of an offcut, for example on one of the hidden sides of the plenum. Just rig a thumbtack or some other cutting device behind it with a way to fine tune the distance before a rupture is induced, and this ought to be enough to keep the main mirror safe.

Brandano, because the bleed-air gate controller has "cheater holes" in the side of it, we'll be testing a physically impossible condition - a direct draw on the mirror plenum. This is more to see how bad it _could_ get in a worse-case scenario situation.

g.

Gene, that looks fantastic. Great work.

Thanks!

g.