[ODE] messing around with units

Ricardo Kirkner ricardokirkner at gmail.com
Fri Jun 2 19:59:18 MST 2006


Hi:

I know this has already been asked several times before, and I know
that ODE is somewhat independent of what units I am using.

However, my problem is. I have scaled all my units so that the lie
around 1.0, but I am getting some issues. Actually I need the
simulation to behave as realistic as possible, so that I cannot just
make the values up (or hand-tune them until it 'seems' right).

I am modelling an object that consists (more or less) of two boxes
joined together with a hinge joint.

Each box has a side length of 32mm and weights about 110g. The hinge
joint's motor can have a maximum torque of 23Kg/cm (the units are
alled in different scales, because that is how I got them).

I chose the following scale

1 Length Unit (LU): 32mm
1 Weight Unit (WU): 1Kg
1 Time Unit (TU): 1s

With those values, I have calculated that

Gravity = 9.81 * (1m / 0.032m) = 306.5625 LU*TU^2
Maximum Torque = 23 * (1Kg / (1cm / 3.2cm)) = 73.6 WU*LU

Are these values correct, or have I made a mistake?

The problems I am getting right now, (I think) don't really have to do
with the unit scaling (but I just wanted to check it out with you
guys).

What is happening is that the joined boxed seem  like bouncing of each
other (like colliding between them, although they are being held
together through the joint).

I have tried playing with CFM, ERP and number of steps (as I am using
the quickStep method). CFM and ERP didn't help much, and increasing
number of steps actually made the problem worse (which is very
strange, from what I have read).

Thanks for any kind of help you can give me


More information about the ODE mailing list