JaydenDP Posted November 13, 2013 Share Posted November 13, 2013 Hi all! Is there a problem with using glue constraint network to create bigger non-voronoi shaped chunks, and how that result behaves with mass? I've done a setup with a constraint network, but if you watch the .mov (attached) you can see that the bigger chunks that maintain a strong glue bond, don't seem to have an accumulated mass based on the sum of the pieces that make up the chunk, resulting in an unrealistic sense of weight. Almost like polystyrene! Any workarounds/suggestions would be much appreciated! Thanks Jay Quote Link to comment Share on other sites More sharing options...
JaydenDP Posted November 13, 2013 Author Share Posted November 13, 2013 (edited) not sure my MOV is attaching : EDIT: Thats better GlueNetwork_VS_MASS_.mov Edited November 13, 2013 by JaydenDP Quote Link to comment Share on other sites More sharing options...
JaydenDP Posted November 14, 2013 Author Share Posted November 14, 2013 hmm, maybe what I'm asking doesn't make sense? Quote Link to comment Share on other sites More sharing options...
JaydenDP Posted November 15, 2013 Author Share Posted November 15, 2013 having a bit of a monologue here ;P Also, whenever I try using a spring constraint network instead of Glue network, it explodes! Only way to stop it is to have very low strength values, which defeats the point... Am i missing anything? Seems it might have something to do with the network complexity, if I make a spring constraint on a very simple (10 piece) box it works fine...any tips? Thanks Jay Quote Link to comment Share on other sites More sharing options...
JaydenDP Posted November 15, 2013 Author Share Posted November 15, 2013 having a bit of a monologue here ;P Also, whenever I try using a spring constraint network instead of Glue network, it explodes! Only way to stop it is to have very low strength values, which defeats the point... Am i missing anything? Seems it might have something to do with the network complexity, if I make a spring constraint on a very simple (10 piece) box it works fine...any tips? Thanks Jay .MOV attached Spring_network.mov Quote Link to comment Share on other sites More sharing options...
cwhite Posted November 15, 2013 Share Posted November 15, 2013 For springs, check what your rest lengths are - the rest length on the Spring Constraint Relationship DOP is multiplied by the 'restlength' primitive attribute on the constraint network's geometry to compute the rest length of each spring (same for other attributes like 'strength' and 'damping'). The gold release had a bad default value (zero) for Rest Length on the Spring Constraint Relationship DOP, which would cause every spring to get a rest length of zero. That's been changed to 1 in the latest daily builds. For your glue issue, there was a bug fixed in 13.0.207 that may be relevant: Fixed a bug in the Bullet Solver that caused the inertia tensor and angular velocity of a glued object to be incorrect in some cases, which resulted in incorrect rotations or unstable behaviour. Additionally, the solver now correctly handles the Rotational Stiffness setting when gluing objects. Quote Link to comment Share on other sites More sharing options...
JaydenDP Posted November 15, 2013 Author Share Posted November 15, 2013 For springs, check what your rest lengths are - the rest length on the Spring Constraint Relationship DOP is multiplied by the 'restlength' primitive attribute on the constraint network's geometry to compute the rest length of each spring (same for other attributes like 'strength' and 'damping'). The gold release had a bad default value (zero) for Rest Length on the Spring Constraint Relationship DOP, which would cause every spring to get a rest length of zero. That's been changed to 1 in the latest daily builds. For your glue issue, there was a bug fixed in 13.0.207 that may be relevant: Thanks, yeah hopefully updating the build next week will help with the glue stuff. As for the restlength, its completely normal in this case. set to 1 in the DOP. Can get it stable for few frames but still seems to eventually explode depending on complexity... I'll try some things, check my collisions etc Quote Link to comment Share on other sites More sharing options...
JaydenDP Posted November 18, 2013 Author Share Posted November 18, 2013 Nope, still no luck maintaining a stable spring network past a certain network complexity. Going to have to give up on it Quote Link to comment Share on other sites More sharing options...
cwhite Posted November 18, 2013 Share Posted November 18, 2013 What sort of spring stiffnesses and masses are in your simulation? Bullet's springs can be unstable if you have a small mass attached to a very stiff spring (this isn't anything particular to Houdini's Bullet solver - this is how the Bullet engine works). You could trying increasing the density or providing your own mass for the smaller pieces. Quote Link to comment Share on other sites More sharing options...
JaydenDP Posted November 18, 2013 Author Share Posted November 18, 2013 I have very large masses with very stiff springs. the problem is, i can get stable springs with high mass if the spring has a very low stiffness, but thats not what im after. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.