softbodies problem

ok, so I know that alot of these topics seem to be poppig up since the new release, but I’ve looked to see if this has been covered yet and i couldn’t find it(if it has, I’m sorry).

Here’s my problem: I set up an object to act as a softbody - no problem. But i try to set up another object to act as a deflector (all you have to do it press deflection and that should be enough, right?) I recalculate the softbodies, but the “deflector” doesn’t work. THe softbody just passes right though. I baked the softbody(because i heard the simulation isn’t always 100% correct) but it still doesn’t deflect the softobdy. Any ideas what I’m doing wrong? thanks…

softbody colision only calculates the verts of the soft body and the faces of the deflection mesh. if you have a plane that is 4 wide falling on a cube that is 2 wide, none f the verts in the plane will touch the cube, and no deflection is calculated.

to fix:
subdevide plane mesh

Basically the deflector area of the new release isn’t very powerful. You need a lot of sub divisions which kills the CPU like a civil war reinactment.

For the deflector, you could set the ‘Outer Face Thickness’ up, this makes the soft body land above the deflector a bit. I assume you tried this and it still doesn’t work though, I’m not much help…

I just make my deflector and my soft body have wicked amounts of verticies and throw them in a seperate Layer and don’t bring them out until Christmas.

thanks for the replies guys. My delfector already is fairly subdivided, but I’ll try adding a little more. I actually didn’t think of upping the outer face thickness, so I’ll try that as well. Thanks!

no, not the deflector, that should be low poly (for speed issues)
the softbody only calculates verts, so IT should be subdevided

Right, well, they’re both sort of highpoly, though the deflector is less so than the softbody. And I just tried upping the our face thickness and that didn’t help…

EDIT: Figured something out, thanks for your help everyone!

I noticed that if the dampining is all the way down on the solid object the soft body will not interact with the deflector.

I’m just a noob, so take what I have to say with a grain of salt…

C911CA

im having the same problem;i read thou you should uncheck goals…

anybody know who created softbodies …maybe we can ask them whats up

I don’t suppose the soft object and the deflector are in different layers are they?

Just a thought!

no i checked that…
i get inconsistant deflection behavior;should none be selected or force
edge dependant behavior does this mean a highly subdivided shape is best for the softbody…

what about the deflection object is there a min num of vertices needed;i have a sphere which passes thru a plane

goals is off
force is 0

I would not say i created softbodies, but i coded physics simulation.

I

  1. If you complain, it would be nice to have a ‘.blend’ file to verify the “malfunction” ( please add information, if you’re using an official release like 2.37 or a home grown CVS build or … [or … ] )

  2. The SB collision system ‘as is’ was ‘borrowed’ from particle stuff ( with all the limitations there ) i’m not really happy with the performace, BUT it works fine. ( most of the ‘particle’ users got the concept of SB collision in a glance )

  3. Softbodies ‘per se’ was not designed to do crash test simulations ( though it could be blown up to do ) but do some nice jiggle stuff like env did.
    See :
    https://blenderartists.org/forum/viewtopic.php?t=45189&start=0

II
Stop ranting about things you did not think about.

Messages like :" i’ve heard … blah" spread the opinion that blender coders don’t do their job ( which in most of the cases is not their job to earn their living ) carefully enough.

That statements may help your ‘ego’, but are quite demotivating to code anything for you in my spare time.

Recently i met a friend of our family. I didn’t know he cares for blender at all. Anyway, he was asking me why SB objects may get bumped in on collision and i told him. The anwser i got was : 'ah, right its like a table tennis ball … ’

III
phew i should be old enough to ingnore post like that

but i could not resist

BM

Well …
i blew my distorted mood in the answer above …
but still
as a rule
make the collision target ( deflector you call it ) lowest poly as possible
make the SB to collide with target as high poly as urgently needed
will give best performance
… still it needs to check every vertex of the SB possibly colliding with any face in the set of collider objects
… got it ?
BM

Error limit can also affect the result. Depending on what scale your object are at. Try putting it down low.

But i think your problem is Goal. Turn off goal… Get your object to move with either gravity or wind. Use real low values first. Then speed things up.

Any --even enginnering ones-- physics simulations are not easy to code or operate. Theres a lot of detail to understand why things do what thay do. Remeber that you mite be better off faking the physics for some situations (ie use IPO and keyframes).

I have found softbodies great fun. And i will find them usefull as well. Thanks to all those who put effort into the codebase.

Delt0r