04-06-2020, 09:24 AM
(This post was last modified: 04-06-2020, 09:28 AM by josemendez.)
(04-06-2020, 08:46 AM)relevantname Wrote: Hmm, but how do I setup a rope like this? If I use a really long rope, it won't be tight? And it does not stick to the out of the tube.
Thank you so much
Edit: I think I need to use a real tube with the radius of around the rope's thickness
And in the rope around, if I exaggerate the move, the rope ends. So, it is using a single very long rope like you said.
Hi,
To prevent stretching (= slow convergence) you can just crank up the amount of substeps up. That will yield near unstretchable rope. Since there's no regular (Unity) physics involved, you don't pay the price of collision detection for each substep.
On the subject of convergence/substeps/iterations (which is physics simulation 101), I highly recommend reading the following link. Might take a while to sink in, but it is really useful:
http://obi.virtualmethodstudio.com/tutor...gence.html
Substeps basically divide each physics step into multiple smaller steps. Advancing the simulation using smaller timesteps prevents the rope from getting into "very" invalid state after each step, so the solver has to do less work to bring it back to a valid state (fewer iterations).
Think of constraint iterations as a way to fix the errors introduced by the fact that physics in games are advanced in "stop-motion", and think of substeps as a way to reduce the amount of error that is introduced in the first place.