Multibody Systems modeling is a very systematic approach usually results in large sparse Jacbian matrix. I am working to model a system consisting of 11 bodies and 63 constraint equations as soon as i finished the assembly of the system`s Jacobian matrix I got a very ill conditioned matrix even I reviewed the model several times but still have this problem. Later I discovered that it is a common problem using such this method. I am now in mid of my research there is no time to go step back so i tried to deal with this matrix and compared results with other methods there were great differences shown in graphs below . I need an urgent help how to overcome this problem as I said no time to go step backward. I also tried to read about the saparsity there are dozens of research talking about that with a very complicated mathematics far away from my research field. So Any suggestions please to solve this problem.