I've been thinking about how ngspice and other circuit simulators select the 'state variables' for a network. I know that this will involve some graph theory but was unable to find any references that describe just what it is that the solvers do. Any references would help.
Asked
Active
Viewed 53 times
0
-
2Go to Laurence Nagel's UCB/ERL M520 on SPICE2. It's all covered in there and freely available to read. (Take note of the PDF link on that Berkeley page.) Skip to 2.5 ff for a quick dash. – jonk Jul 01 '22 at 03:56
1 Answers
0
Pretty much the only thing they can do is use capacitor voltage and inductor current*, or possibly capacitor charge and inductor flux. Those are the relevant states in an electronic simulation.
* leaving out the simulation of any transmission lines and direct simulation of linear transfer functions.
TimWescott
- 46,144
- 1
- 42
- 107