> If you need help, I might be able to offer some. Back in my last year of my > Mathematics degree (~23 years ago), I developed my own predictor-corrector method for > solving differential equations in much the same way that Runge-Kutta, Milne-Simpson, > etc do, and then wrote Fortran and Pascal programs to test the hell out of it and > compare the results against 6 or so other methods. Alas, like I say, this was 23 > years ago, so I am now very rusty. But, if you need help solving any differential > equations, let me know and I will do my best.
Wow. Your method/code would certainly help, once I knew how to transform a netlist into differential equations. That is your contribution would be step #2 and I'm already clueless about step #1 ;-)
I have the (very) vague idea that every element of a circuit would have to be described with its resistance, capacitive reactance and inductive reactance, i.e. 3 characteristics, and then each knot in a network would have at least two such characteristics influencing each other. I may be totally wrong here, though.
Edited by pullmoll (02/06/10 05:26 AM)
|