PDA

View Full Version : conveyorBelt/TankTreads in toolbox


HornBerger
07-20-2010, 02:06 PM
conveyorBelt/TankTreads in toolbox

hi!
playing with toolbox this weekend to create a conveyor belt (not the one in the visual guide on
the orbaz website) i wanted my belt to be suspended between 2 rollers, so when i dropped
an object on the belt it would dip inside a little due to the weight of the object. i got the effect
[attached max file]

(right click and select view image to enlarge)
http://img121.imageshack.us/img121/9339/conveyor.jpg

now i want to convert this conveyor belt into a tank tread,so as the rollers spin against the belt, the belt propels the whole system forward... easier said than done.. i am having problems keeping the distance between the rollers also the belt slips of the rollers when they move on a noisy terrain..

some basic issues i would like to point out..
a1) i cannot get a cylinder to move forward by making it spin like a wheel against the ground plane... (what i did..) created a physX flow, added a birth group operator and picked a cylinder lying on the ground plane of the physX world next added a spin operator followed by a physX switch operator with the spin checkbox ticked but the cylinder only moves forward by a inch. I want to make it accelerate forward ( i tried animating the spin rate spinner but it barely has any effect)
a2) used a motor force to achieve the above, and oriented the motor so that it makes the cylinder spin on its axis but its giving really weird results (the cylinder goes haywire after a couple of frames)

i have a couple more questions but i guess i will save it for later

(right click and select view image to enlarge)
http://img25.imageshack.us/img25/5945/failedtanktread.jpg

thanks!

[attached 2 max file (2011)
1) conveyor belt
2) failed attempt at tank treads]

JohnnyRandom
07-21-2010, 12:10 AM
Cool conveyor :)

I think a better approach for the tank threads would be more like using madcar, rig, hand animate,.or similar the drive wheels over terrain and use them as pflow collision objects and then just glue the threads together around the collision objects.

I did a simple test, it isn't perfect by any means, it does work (it could work better, my timing with rotation and ground is off a tad)

Max2010

http://4rand.com/TEST/ParticleFlow/Box2/TankThread/TankThread.gif

HornBerger
07-21-2010, 12:39 AM
thanks for the post

its amazing! thanks a lot! :thumbsup: i will have a deeper look at the setup and let you know... i was just doing this as a pet project i too prefer (yet) using a rig for tank treads rather than simulating them with physX engine (but its just a pet project) in havok 3 you can make a wheel spin by using the motor reactor helper but when i try to use the motor force or (match spin with physX switch) i cant get the wheel to accelerate forward! i was trying to do a similar setup to make the rollers produce torque and move the setup (treads and the rollers) over the terrain instead of hand animating them (maybe its easier to just animate the rollers with constraints and stuff but as i said its just for academic purpose i wanted the rollers to be particles and produce a torque to move the system forward :) ) ... thanks again :)

cheers! :)

PsychoSilence
07-21-2010, 12:48 AM
Cool conveyor :)

I think a better approach for the tank threads would be more like using madcar, rig, hand animate,.or similar the drive wheels over terrain and use them as pflow collision objects and then just glue the threads together around the collision objects.

I did a simple test, it isn't perfect by any means, it does work (it could work better, my timing with rotation and ground is off a tad)

Max2010

http://4rand.com/TEST/ParticleFlow/Box2/TankThread/TankThread.gif

hah! that's great! :D

JohnnyRandom
07-21-2010, 02:00 AM
Thanks Ansi :)


I gotcha, I totally understand where your coming from, I have tried a few times with limited success in making wheel driven objects, the biggest hurdle to overcome is binding the actual wheels to the body as currently the binds are axial locked, meaning they don't spin.

It is possible and not at all practical (yet) to build up some concave objects that will hold the axles in place, but keeping them attached is difficult. Binding gets a little insane as you are literally building nuts and bolts to fasten things together (much as you have discovered building the belt on you conveyor)

I have suggested Axial XYZ lock/unlock for binds, whether or not PhysX allows it I have no idea. Sure could do some cool stuff if rigid binds could spin :)

PsychoSilence
07-21-2010, 02:27 AM
interesting setting in the glue! quite some spring coef. to keep it all contained :D

JohnnyRandom
07-21-2010, 05:50 AM
It works without it, but since the bind gets stress when it goes around the wheel setting the min and max length plus elasticity gives it a little bit smoother ride. I am kinda curious how it would hold up on a large scale.

HornBerger
07-21-2010, 09:38 AM
I gotcha, I totally understand where your coming from, I have tried a few times with limited success in making wheel driven objects, the biggest hurdle to overcome is binding the actual wheels to the body as currently the binds are axial locked, meaning they don't spin.

exactly.. that was the question i was saving for later the binds wont support spin the moment the axial moves forward the spin stops and the object falls towards the axial.. also no matter what amount of spin i apply to my cylinder (wheel) it only moves forward a few inches before its dragged to a stop
i tried to play with dynamic friction without any luck (at 0 it just keeps spinning at its place and any value between 0 and 1 it only moves forward by an inch :(

It is possible and not at all practical (yet) to build up some concave objects that will hold the axles in place, but keeping them attached is difficult. Binding gets a little insane as you are literally building nuts and bolts to fasten things together (much as you have discovered building the belt on you conveyor)

i tried this too but its difficult to set it up without making the wheels go haywire. heh i guess at the end of the day rigging wins :( i did a rig by using path constraint on the treads so as the wheels move the script calculates the path constrain % for the treads and they move correctly (that or a path deform but i find using path constrain more flexible)

thanks!
cheers! :)

CGTalk Moderation
07-21-2010, 09:38 AM
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.