[ODE] mass of collisions - imagine a brush
Hendrix_ at gmx.net
Fri May 11 07:44:58 MST 2007
this is a little movie I just made. It is a 4x10 array of bristles,
each bristle is made of 5 capsules. There is no restoring force that
makes the bristles go straight, so they just fly around. Looks awful
The settings are:
capsule 1.2 x 0.2 cm, 0.0348 grams
the anchor is in the center of the spherical part of the capsule, the
capsules overlap in the sperical parts
i used universal links
surface friction 10, bouncyness 10%
gravity -9.81 downwards
delta_t = 200 fps
SOR solver, 100 iterations
max correction velocity 1
The capsules are handled as trimeshes, I havent imported the capsule-
stuff from the trunk yet.
On 9. May 2007, at 15:43 Uhr, Tobias Zimmer wrote:
> As Remi Ricard wrote, the Universal joint does exactly that! I used it
> to connect the bristles to the head as well.
> But the idea of a restoring torque is really good, I haven't really
> thought about this point yet, but that is very
> important! I guess, this force has to be built into the joint between
> bristle and head especially.
> Thank you for your ideas and fast answers already,
> Patrick Enoch schrieb:
>> - have all bristles as single geoms, even worse, subdivide the
>> bristles into several parts connected by universal joints.
>> - the bristles are connected to the head by a ball joint;
>> preferably a joint that does allow rotation "to the sides", but
>> not "twisting rotation". I have posted the "free6" joint to the
>> list (that is capable of this), but the email has not been
>> approved by a moderator.
>> - the bristles can collide among themselves, ...
>> - but not a single bristle with itself (lets hope they dont bend
>> too far so self-intersection is possible)
>> - the joints stiffness should be possible via a motor with
>> rotational velocity of zero and a maxforce (however, this wont
>> make the bristle return to its original "straight" state)
>> - you need to determine the relative rotation of the subdivided
>> parts of the bristle (a relative quat), and add a restoring
>> torque. maybe this should be built into the joint itself.
> ODE mailing list
> ODE at ode.org
More information about the ODE