1
$\begingroup$

[Blender 2.73a] I have a scene, where a helmet is falling on a surface.

enter image description here

It is quite complex: high-poly helmet, every link in chainmail is a separate element and all this stuff are rigid bodies. I want them to act physically correct, that is why so. All modifiers and transformations are applied: objects are clean meshes. I've already done simulation tests with this model earlier (https://www.youtube.com/watch?v=gszj0GUkigI), now I've updated the model and try to make a new simulation, but it fails:

enter image description here

Surface (was a passive body plane first, now an active non-dynamic cube — doesn't matter) first collides, but then chains and helmet pass through it. I've tried everything — no idea how to fix this.

Here are screenshots of settings:

Helmet:

enter image description here

Chainmail:

enter image description here

Surface:

enter image description here

Scene:

enter image description here

(all objects are real-size, high values of steps and iterations always well helped increase accuracy)

Any ideas, why is this?

UPD

Subdivided surface, increased steps to 1600 and solver iterations to 120. No luck

enter image description here

$\endgroup$
  • $\begingroup$ you can try to add more geometry to your ground. (subdivide it) $\endgroup$ – Bithur Feb 14 '15 at 19:09
  • $\begingroup$ this matters? O_o ok, I'll try $\endgroup$ – Terion Feb 14 '15 at 19:12
  • $\begingroup$ can help sometimes $\endgroup$ – Bithur Feb 14 '15 at 19:12
  • $\begingroup$ Have you applied the scale to all of the meshes? $\endgroup$ – GiantCowFilms Feb 14 '15 at 19:13
  • $\begingroup$ Yes. Everything is applied $\endgroup$ – Terion Feb 14 '15 at 19:14
1
$\begingroup$

Well, with some tries I've come to understanding, that problem is not in the surface, but in chains. In Blender's physics engine mass is corresponding to durability (sic!). So, realsize chains with a calculated mass near 1g each is too weak and they "break" during collisions. The only working workaround was to scale everything up several times and increase the mass of everything...

$\endgroup$
  • 1
    $\begingroup$ Yep, rigid body sims usually need to be scaled up quite a bit for the calculations to not wack up on small objects. Sometimes this will make the objects move slower (I am not talking about lag), if this happens just increase the overall sim speed, check out Andrew Price's rigid body tutorial. $\endgroup$ – PGmath Feb 15 '15 at 23:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.