Page 1 of 1

Low Torque on edge

Posted: Mon Apr 19, 2010 10:47 pm
by razer
When some object like cube fells on flow. It's starts to rotate but speed of rotation is very low and unrealistic. Same thing when I move some object to edge and make it fell.

I just compared what I seen on screen with cell phone felling from edge on table. When central mass of cell phone reaches the edge it rotates and fells but when I model similar thing in bullet it is much slow.

Why is that?

My box size is much larger then cell phone

It is 4x3x0.5

It still looks unrealistic.

Please, advise, Thanks

Re: Low Torque on edge

Posted: Tue Apr 20, 2010 10:45 pm
by razer
I set gravity 20 instead of 10. Slightly better but still not realistic.

Re: Low Torque on edge

Posted: Wed Apr 21, 2010 1:37 pm
by bone
What's the mass and inertia of the object?

Re: Low Torque on edge

Posted: Wed Apr 21, 2010 11:01 pm
by razer
mass is 1000, size 4x3x0.5 and initial inertia is (0, 0, 0)

I use localCreateRigidBody from demo to create most rigidBodies

Thanks

Re: Low Torque on edge

Posted: Thu Apr 22, 2010 9:02 am
by DannyChapman
razer wrote:I set gravity 20 instead of 10. Slightly better but still not realistic.
If you want to do a direct comparison with your cell phone then you would need to set gravity to around 500. This should result in relative motions being the same (i.e. the same amount of world time to rotate one revolution, or to translate by its own size).

Big things move like small things but on longer timescales...

Re: Low Torque on edge

Posted: Thu Apr 22, 2010 1:24 pm
by bone
razer wrote:mass is 1000, size 4x3x0.5 and initial inertia is (0, 0, 0)

I use localCreateRigidBody from demo to create most rigidBodies

Thanks
What you've got there is very roughly the size and mass of a car (a gravity of 10 indicates metric units, so 1000kg and the dimensions in meters), which will certainly act a bit different than a cellphone.

Re: Low Torque on edge

Posted: Thu Apr 22, 2010 4:56 pm
by razer
Things that have small mass fall as fast as heavy.
I set mass to 1 and there is no difference in bullet as well.
May be size make thing fall slower.
But it is unrealistically slow. when gravity is 10 and I set gravity 20 now.

May be bullet has some tuning or I should change some constant.

Re: Low Torque on edge

Posted: Thu Apr 22, 2010 5:49 pm
by bone
razer wrote:Things that have small mass fall as fast as heavy.
In free-falling due to gravity, yes you are correct. But in your case you have to rotate the object in order to fall off the edge of something. It is more difficult to rotate a massive object, not to mention that the moments of inertia go up with the *square* of the dimensions.

Re: Low Torque on edge

Posted: Thu Apr 22, 2010 11:05 pm
by razer
bone wrote:
razer wrote:Things that have small mass fall as fast as heavy.
In free-falling due to gravity, yes you are correct. But in your case you have to rotate the object in order to fall off the edge of something. It is more difficult to rotate a massive object, not to mention that the moments of inertia go up with the *square* of the dimensions.

But force is higher also. Size is important mass is not so important.
Cell phone is bad example.
It is too slow and not realistic.

Re: Low Torque on edge

Posted: Fri Apr 23, 2010 1:32 pm
by bone
razer wrote:
bone wrote:
razer wrote:Things that have small mass fall as fast as heavy.
In free-falling due to gravity, yes you are correct. But in your case you have to rotate the object in order to fall off the edge of something. It is more difficult to rotate a massive object, not to mention that the moments of inertia go up with the *square* of the dimensions.

But force is higher also. Size is important mass is not so important.
Correct. The increase in mass is directly offset by the fact that the force increases with mass.

So what you've got is the inertia vs. the torque (and torque arm, which at the initial state of this example is the lateral difference between the body CG and edge of the table). Removing the offsetting mass from both sides, you are left with the body dimension squared vs. the torque arm. So bigger objects, even if moved proportionally further off the edge, will still rotate slower because the squared term dominates.

Regardless, all we have established so far is that the cell phone vs. the car-sized object isn't a great comparison, which you've agreed.

If you provide the CG position vs. the edge of the table, and the full inertia properties (and current gravity coeff), we can determine exactly how fast it should start rotating.