« first day (1 day earlier)      last day (32 days later) » 

andre
andre
00:01
Tomorrow, I think I will reactivate my spice (ngspice more precisely)
 
2 hours later…
xzczd
xzczd
02:04
@andre It's not that "magic" :) . I add the "eps vdiode'[t]" term just to modify the original equation set (which is a DAE system, it doesn't involve derivative of vdiode ) to an ODE set. eps is a properly selected small number to make the influence of the modification small enough.
 
9 hours later…
andre
andre
10:59
@xzczd Illuminating explanation ! Thanks. I suppose that the idea to add a artificial derivative to obtain a ODE comes from the Advanced NDSolve Tutorial about PDEs, specially the text concerning how Dirichlet boundary conditions are implemented
 
2 hours later…
xzczd
xzczd
12:30
@andre I believe it's one of the source of inspiration. (When writing the solution above, that tutorial isn't in my mind, but I should have been unconsciously influenced :) ) Another possible inspiration is Stabilization of Convection-Dominated Equations part of this tutorial.
It's these materials that tell me, it's possible to circumvent DE solving issue by slightly modifying the equation, without changing the solution significantly.
 
10 hours later…
andre
andre
22:07
@Tschibi I have reinstalled NGSpice on my machine. I have full control on it from mathematica. I retrieve the results and plot them automatically in mathematica. The inputs are *.CIR files (no schematic) :)

« first day (1 day earlier)      last day (32 days later) »