Collision between Sphere and Box not working

[tometherton]

So I've implemented the Bullet SDK into my project, following the instructions in the User Manual. I've also created a wrapper for the Bullet Physics system.

Now I'm creating a sphere with btSphereShape as one collision shape and a box with btBoxShape as another collision shape. I then set all the properties of these and create btRigidBody for both of them and then add them to the world. When I run my program I see the ball falling, falling and falling...it doesn't collide with the box at all. Is anyone able to suggest why this might be? Am I doing something wrong?

Thanks in advance,

Tom

[CHAOS-THEORY]

make it collide with the box? lol i mean can you bring down the code dude cuz no one will help u this way...

[tometherton]

Adding a sphere:

Code: Select all

//Function for adding a sphere to the physics system
void
BulletPhysics::VAddSphere(float radius, Actor* actor, float specificGravity, PhysicsMaterial mat)
{
	//Create the collision body, specifying the shape
	//btSphereShape* const collisionShape = new btSphereShape(btScalar(radius));
	btSphereShape* const collisionShape = new btSphereShape(1.);

	//Calculate mass from specific gravity
	float const volume = (4.f / 3.f) * PI * radius * radius * radius;
	btScalar const mass = volume * specificGravity;

	AddShape(actor, collisionShape, mass, mat);
}

Adding a box:

Code: Select all

//Function for adding a box to the physics system
void
BulletPhysics::VAddBox(const Vector3D &dimensions, Actor* actor, float specificGravity, PhysicsMaterial mat)
{
	//Create the collision body, specifying the shape
	btBoxShape* const collisionShape = new btBoxShape(GetBulletVector3(dimensions));

	//Calculate mass from specific gravity
	float const volume = dimensions.x * dimensions.y * dimensions.z;
	btScalar const mass = volume * specificGravity;
	
	AddShape(actor, collisionShape, mass, mat);
}

The add shape function:

Code: Select all

//General function for adding shapes to the physics system
void
BulletPhysics::AddShape(Actor* actor, btCollisionShape* shape, btScalar mass, PhysicsMaterial mat)
{
	//Set up the local inertia
	btVector3 localInertia(0.f, 0.f, 0.f);
	if(mass > 0.f)
		shape->calculateLocalInertia(mass, localInertia);

	btTransform trans;
	trans.setIdentity();
	trans.setOrigin(GetBulletVector3(actor->VGetPosition()));

	btDefaultMotionState * const myMotionState = new btDefaultMotionState(trans);

	//Set up the body's info
	btRigidBody::btRigidBodyConstructionInfo rbInfo(mass, myMotionState, shape, localInertia);

	//Create the rigidBody
	btRigidBody * const body = new btRigidBody(rbInfo);
	//body->setWorldTransform(trans);

	//body->setFriction(15.0f);

	m_dynamicsWorld->addRigidBody(body);
	

	//Calls back the actor class to set the body up with this actor
	actor->VSetActorBody(body);

	//Set up an ID for this actor to be associated with
	ActorId nextId = 0;
	for(ActorIdToActorMap::const_iterator it = m_actors.begin();
		it != m_actors.end();
		++it)
	{
		++nextId;
	}

	//Sets the Actor's value of ActorId
	actor->VSetActorId(nextId);

	//Sets up and inserts the Actor into the list
	m_actors.insert(std::pair<ActorId, Actor*>(nextId, actor));
}

And then the call to Bullet to update the physics comes from my main loop with a call to:

Code: Select all

//Function for updating the physics system
void
BulletPhysics::VOnUpdate(float elapsedTime)
{
	//Pass in 4 so BulletPhysics carries out 4 substeps and will run simulation 
	//in terms of elapsedTime
	m_dynamicsWorld->stepSimulation(elapsedTime, 4);
}

I think that the problem is in there somewhere, maybe with how I'm creating the shapes but I'm not 100% sure. The box is getting created with 0 mass and the sphere with around 19 depending on the material I use. If anyone needs any more code shout, I just didn't want to post 000's of lines of it!

T

[tometherton]

I seem to have a new problem now - the btSphereShape falls down and goes slightly through the btBoxShape before stopping. So I can then move it about (through applying force to it) but it is stuck inside the box whereas I want it to be on top of it - any ideas why this would be?

T

[Nickert]

Perhaps the size is a bit off?

[CHAOS-THEORY]

lol i feel sleepy n i read ur script half eye opened so this is what i found i dnt even know what it is http://www.bulletphysics.org/Bullet/php ... 72&p=27984