Too quick to celebrate. One of the reason why I thought things were working was because the system was small and it wasn't too easy to miss out branches. The row operations to get rid of the large resistances and isolate the stiff equations was correct. But the subsequent matrix manipulation to make the matrix symmetric was wrong.
The consider the diagonal element of a manipuled row to be the sum of the offdiagonal elements is no way right. Simply because the diaognal element may already contain a branch present in any of the off diagonal elements and so the effective resistance/inductance may be larger.
I was trying to do this because I didn't want to do another parameter read from the objects. But looks like that may be the only way. Or at least can't be completely avoided. Need to figure out how to minimize the object read.
No comments:
Post a Comment