We are trying to use a GForce to model a nonlinear element in an assembly. When we define this spring inside the adm file, the model runs OK and converges quickly. However, when the spring is defined inside a dll created using c-code, the convergence takes longer. The internal and external definitions for the spring create similar forces (we overlapped the two).
1. The dll stops converging (for sim/stat) when we use stiffer springs, even for linear springs. Has anyone had this issue before that a spring converges when it defines inside the adm file and does not converge when defined inside dll? For sim/dyn, both springs work ok (dll needs more iterations).
2. Can it be due to Jacobian? Is it possible to help dll converge by providing the Jacobean for the function? We use GSE. However, in the simplified version, only GSEoutput has the linear spring equations.
