(03-30-2009 07:22 PM)femnoob Wrote: [ -> ]thanks. you are the man. 
Hi Femnoob
been away from the forum, but I'm back now...
In response to your need to simulate a kinematic-dynamic assembly ( I believe this is what you want, right ...? define a mechanism topology (what connects to what, and the restrained degrees of freedom in the connection), define passive elements (masses moments and products of inertia), as well as active elements ( forces, moments, rotational springs and dampers, linear springs and dampers ) driving boundary conditions and external fixities )
Well what you want is called a MBD analysis package :
Each body is modelled as infinitely rigid, and the nonlinear ( in cartesian coordinates, but in Lagrangian coordinates it is linear ) equations of motion are solved by a newton raphson technique to yeld an Initial time viable solution; Starting from there, there is a time marching strategy that will integrate the equations of dynamics of the mechanism using the initial values you prescribe for velocities and the found values for positions and the driving values for accelearations and momenta ( they are non linear due to kinematic linkage coupling and second order in time ) with a robust integration scheme Runge Kutta 4th order ate least, to yeld a stable value for the position of the next time step.
After N integrations of this fixes size timestep, the program shifts to a more efficient time stepping algorythm, such as Gear algorithm, I prefer this one because it can integrate "stiff" systems without an arbitrary reduction in step size to acheive stability.
Note that the system becomes stiff because of :
1. A large specral radius of the reduced stiffness matrix [M]*[K]-1 due to the fact that the eigenvalues of the "spring" part of the above quotient may scatter along several decades span..
2. The inclusion of the algebric equations of reduction of degrees of freedom between links ( a revolute joint introduces three restrictions to movement in the links that must be accounted for when solving the equations of motion in the linkage, for instance...), giving the equation system a mixed nonlinear algebric and differential character...
In other words a numerical PAIN IN THE ASS.... :-D
I designed a code like this for 3d analysis of mechanical linkages several years ago...
made it in fortran 95... after reading a very good book on the suject by Parviz Nikravesh, Computer aided Analysis of Mechanical Systems., I recomend it to you....
The math can be a bit awkward to the undergraduate level, especially the part of quaternion transformation of the local rotational coordinates to global frames... but it is highly advisable.
IF you do not want to delve into the realms of Sir Isaac Newton, Joseph Louis de Lagrange or William Rowan Hamilton ( a long, harsh but very rewarding journey... ), you can just Download a code called MBDyn... from here..
http://www.aero.polimi.it/mbdyn/, that will do exactly what I described, and possibly more... ( I believe it does some elementary FEA analysis besides MBD, didn't tried it yet anyway... :-D ) or buy MSC.Adams, that does exactly the same with a lot of bells and whistles, a nice gui,... etc.
Hope this info was useful...
Enjoy
Alex
...