Re: [ng-spice-devel] convergence


To ng-spice-devel@ieee.ing.uniroma1.it
From Steve Hamm <Steve.Hamm@motorola.com>
Date Tue, 1 May 2001 14:57:43 -0500 (CDT)
Delivered-To mailing list ng-spice-devel@ieee.ing.uniroma1.it
In-Reply-To <01050102085802.26584@hobbes >
Mailing-List contact ng-spice-devel-help@ieee.ing.uniroma1.it; run by ezmlm
References <CF654D964573D311A1CA0090278A36FF4CBE4C@EDIN_EXM1 ><0104291404120G.10628@hobbes ><15085.31090.944666.981605@pearls ><01050102085802.26584@hobbes >
Reply-To ng-spice-devel@ieee.ing.uniroma1.it

---"Al" == Al Davis <aldavis@ieee.org> writes:

Al> For consistency, take form 1 and add a last step to reevaluate F
Al> using the formula F(x_(n+1)) = F(x_n) + J(x_n) * (x_(n+1) - x_n)
Al> Now it is consistent and KCL is satisfied.  Note that J is not
Al> re-evaluated.

Uh, we have a different definition of "consistent". Your formula just
back-substitutes and doesn't really relate to KCL: Your F, above,
won't be the same as KCL, which would be F evaluated at x_(n+1); the
formula acts to check the linearization, which is silly since we
generally solve nonlinear systems. And it's still inconsistent, since
the last real function evaluation was F(x_n) while the accepted
solution is x_(n+1).

When I mention consistent, I'm meaning that when the iterations stop,
I want some solution x, such that F(x) is close to zero, and F and J
have been evaluated at x so that conductances, capacitances, currents,
etc. have been evaluated at x.

--Steve

Partial thread listing: