PDA

View Full Version : FXWars! Roller Coaster!: MConte05


MConte05
11-04-2004, 03:32 AM
I have done some little testing for what I hope to achieve....and it works, so I am in as well....

Just an image of my track....

ACantarel
11-04-2004, 06:57 AM
Which software do you use?
In which way do you want to solve the dynamics?


André

d3l
11-04-2004, 12:23 PM
ahhh, finally a B&M guy :)

good luck with that one.
i don't want to join the contest, but i hope to learn something about dynamics.
like already asked, which software do you want to use for the dynamics?

MConte05
11-04-2004, 03:56 PM
I am using 3dsmax6 and using reactor for the dynamics. What I have done is loft the track, assign that as a rigid-body. Then modelled the chassis and four road wheels (two for each side). Then assigned those to a toy-car constraint. I have also made 4 guide wheels, two for each side as well, and assigned those as a rigid-body with no friction. I have done several tests (which I am looking to upload somewhere) with up to 6 cars, all linked together using a point-to-point constraint. I still need to redo the trains with the upstop wheels, cause the main problem I am having right now is the cars coming off the track on rough spots and hills.

I am making an original B&M inspired floorless style coaster. Realistic design with 2 trains. I am also trying to make a way to have kicker-wheels in the station, brakes, and perhaps a lift hill. One thing I am thinking of doing is having a lauch similar to Hulk, but that goes into a diffferent element......

Once I find someone who can host my test videos, I will post em....

Thanks for the intrest!

MConte05
11-04-2004, 05:38 PM
Ok, I got two videos to show. One being the first test I did with just a single car on a U-shaped track.

http://tyrant.interfix.net/dynamics.WMV (http://tyrant.interfix.net/dynamics.WMV) (454k)

The second one has the train going over the hill and the back lifting up a bit due to the fact that I have no upstop wheels right now. I'm still not sure what causes the train to de-rail on the bottom of the drop though. Also, the ties were not part of the RB collection, so they aren't affected by the train, nor are the trains affected by them....

http://tyrant.interfix.net/dynamics3.WMV (647k)

Be sure to right-click and "Save Target As"

Thanks to the guy who hosted the videos: Bandwidth provided by InterFix Communications, Ltd. - Web Development and Consulting. http://www.interfix.net/

MConte05
11-04-2004, 07:54 PM
Just a quick material test....

The track is considerably low poly right now so that reactor doesn't have to calculate as much (wheels and track are concave mesh)....

I used an the gloss-paint architectual material....

Also, for those of you who aren't sure what a B&M Floorless coaster is, this is a picture of the cars from Kraken in Sea World Florida....

MConte05
11-05-2004, 06:50 AM
Ok, seems no one is interested in these parts of the forums :shrug:

Anyways, today I laid off the simulation part, and worked on the modeling of my cars......

This is just the seating part for now, I am working my way down instead of up....

http://img103.exs.cx/img103/4610/seatfront.jpg

http://img103.exs.cx/img103/2451/seatback.jpg

ACantarel
11-05-2004, 07:14 AM
"seems no one is interested in these parts of the forums"


Why do you think that? Ok, this part of the forum is not as high frequented as the the galleries for example.
But during working on the challenge, only the participants are looking here. The next big amount of people comes when the voting starts.

So don´t become demotivated!
Look at my thread. There´s also nearly no answer :)

And also it makes more sense to get the simulation working instead of modeling the whole rollercoaster. What, when the simulation doesn´t work? Then you modeled the whole thing for "nothing".

André
(ACantarel)

Sir_Vuclan
11-05-2004, 01:38 PM
Finally someone else with Max6 and reactor, everyone else seems to be using LW or Maya or whatnot. Anyways

How...How did you make the track. Not the rails but all the little support brackets. Oh, and what are you using to bank the track around the corners and loops and such? I am using a loft and then editing the twist curve, but thats a lot of work and very picky in some spots.

Looks like you have a good start. I have some what the same setup as you, though I didnt use the toy car, just the axle constraint. I noticed your cars aren't very long so four wheels (front and back of the car) shouldnt cause much problems. I could only put two wheels on mine or it would get stuck at bends or twist in the track.

I was thinking the same thing for launching the cars (the Hulk ride) I have it set up with about 20 wheels so far (you can see them in one of my tests) but they dont seems to push the cars very fast. (the friction for both wheels and cars is 1.0) I have found that it greatly depends on the weight that you are trying to push, so im gonna have to work on that more.

good luck,
Sir_Vulcan

MConte05
11-05-2004, 10:29 PM
ACantarel: I have already gotten my simulation stuff to work well, and I was getting somewhat bored with doing that, so I started on my modeling....

Sir_Vuclan: I modeled the ties using splines, then extruding a couple images (looking at reference pics). I then used the Spacing Tool to line it up with the original path of the lofted rails. Suprisingly, if you have your axis orientation the same as the track, the ties will follow the banking of the twist modifier in the loft deformation.

As for the wheels, I am following real-life examples. The B&M floorless cars are kept balenced by two road-wheels on the top of the rails on both of the cars. Then a long connection to the other car extends out. That long extension is what keeps the center-of-gravity balenced on the cars. Also, the upstop wheels (which were not applied in the test videos) is what keeps the cars from coming off the track. In my test simulations, I didn't have those on, so thats why it would come off the track......

Also, for your launch system, are you key-framing the wheels, then hoping for them to push it? Or are you using reactor-based action? If not, I would try using a motor action, with a high angular speed and a high gain. I haven't tried it yet, but just by looking at the reactor manual, seems that it would work.....

This is an image of the B&M floorless cars with the axles showing:

Zazu Yen
11-05-2004, 10:40 PM
Did you figure out why the car derailed? You may have to shrink your stepsize...

Sir_Vuclan
11-06-2004, 01:38 AM
The motor contstraint was giving me problems, as the wheels would just fall to the ground and start rolling around. So I used the same thing i used on the coster wheels. the axle constraint and then just applied the angular speed and gain in its properties, that way I could have the coster framing or world as the parent and the wheels would stay put and push the cars.

The problem i have is that, even though I have a high angular speed and friction, the wheels still slide across the cars. so it pushes it nicely but it doesnt shoot it out like on the hulk ride. Still working on that... One thing that helps to get more boost is to up the steps per frame. I doubled it and the cars took off a lot faster.

MConte05
11-06-2004, 01:39 AM
Zazu Yen:The cars derailed because I didn't have any upstop wheels on my tests. My final ones will have upstop wheels....

Sir_Vulcan: Very interesting to know, also, what are the typicall substeps you use? When I preview and make final simulations, I typically use 50-75.

Zazu Yen
11-06-2004, 01:54 AM
Ah, yes a coaster without upstops will tend to do that :)

It's looking good though, I'm looking forward to seeing your next animation.

MConte05
11-06-2004, 06:18 AM
Two new renders:

http://img130.exs.cx/img130/4034/frame1.jpg (http://img130.exs.cx/img130/4034/frame1.jpg)

Shows the overall so far....

http://img130.exs.cx/img130/5599/wheel.jpg (http://img130.exs.cx/img130/5599/wheel.jpg)

Showing the detail of the wheels.

For the simulation I will be using a dummy box and dummy cylinders that will have the main mesh linked to it. That way reactor won't have spend probably hours figuring out the collisions of the high-res meshes.....

MConte05
11-10-2004, 01:07 AM
Well, I am proud to say I finally got everything in my simulation to work correctly. This is just a small test, the coaster is coming out of the station, heads to the lift.....and then rolls back....

http://tyrant.interfix.net/Lancertest5.mov (http://tyrant.interfix.net/Lancertest5.mov)

Tonight I hope to send the coaster over my lift hill and see how it runs.....

MConte05
12-06-2004, 09:55 PM
After pretty much giving up on this project, I had a realization last night on how to get it work, got everything assembeled, and sent it around the track, and to my suprise it actually worked. Then realizing I only had a couple hours left to get this contest entry done, I have been rendering in low detail with horrible lighting, but it does the job of showing the track. Its a little rough because there is some space between the object and the rails, which causes shuffling. I didn't have the time to remake the object, re-sim it, then render....

Also, in the video it seems that the rails intersect the train. I only simulated the train and the rails, and when I went to un-hide the spine and supports I found that I had deleted those. So I ended up just taking my model and trying to line it up as best as I could. Keep in mind I was pushed for time, so its sorta sloppy :blush:

Anyways, here is my entry.....

POV: http://tyrant.interfix.net/MCONTE05_FXWARS(320X240).mov
Exterior Shots: http://tyrant.interfix.net/MCONTE05_FXWARS(320X240)EXT.mov
WireFrame: http://tyrant.interfix.net/MCONTE05_FXWARS(320X240)WIRE.mov

Sir_Vuclan
12-08-2004, 11:16 PM
Nice simulation, few spot where it shakes a little, but looks like it works pretty nicely.

The coaster is really nice looking to. Too bad you didnt have time for an enviroment.

CGTalk Moderation
01-19-2006, 06:00 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.