Newton's Cradle

[i-rex-i]

Hi

I seem to be having trouble when trying to get the right settings for the rigid bodies for the balls.

They just don't seem to be bouncing right. Anyone had any experience with making one of these before?

Here's the file if you want to have a look: Newton's Cradle

[joojaa]

Dont try to do a newtons cradle with dynamics. Because in essence it requires instant contact which isn't easy to do in the kind of simulation environment that depends on discrete steps for collisions.

This said, keying it manually is a job that takes less than one minute to key so theres not umuch point doing it with dynamics either.

[duncanbrinsmead]

I just tried a quick newton's cradle with Maya rigid bodies, and it worked OK. The key was to scale the spheres just slightly smaller than touching( .99) and to make the collision tolerance low (0.001). Other things worth noting: I used standin spheres on the rigidBody nodes and also made the bounce = 1. For threads I used two pin constraints in a v shaped configuration for each ball.

Duncan

[joojaa]

Hmm yes that works. But thats hardly the point, theres not much point in using dynamics to something one can key with perfect accuracy and waste computation time on inprecicion.

I mean keying is as much simulation than using a dynamics engine. Except it allows for a solve and since the solve is easy here its better to make for the exact solve thats independent on steps than actually making the dynamics do it. So there are times when one wants to do things differently.

So the problem is that initial tought of people is that simulation equates to full inner workings of the world but while that could be done. Many times thst kind of detail is a bit exagarating the fact that quite often it not needed. Certainly not needed for engineering purposes and even less for media graphics.

There are ties when keying does the job better this is one of them.