The rest of the toolbox

Five things the app does that are not a circuit solve. Four of them are documented nowhere else.

By-Hand Equations

Tick it and a card appears with the system a first course would have you write: nodal, with supernodes, or mesh, with supermeshes. It names the method it used and counts the supernodes or supermeshes it needed.

It is written independently of the classic solve and then compared with it. The classic solve stays the authority and is never fed by the by-hand one — the point is that two different derivations agree, which is worth something when you are checking your own work against the machine's.

Use it to find where your hand derivation went wrong, not to avoid doing one.

The Numerical Solver

A property of its own, at /eqsheet/, reached from the app's Explore Numerically card. It is a TK!Solver-style equation sheet rather than a circuit solver: a list of equations, each variable marked Known or Unknown, and a numerical solve over the lot.

Two ways in:

• From a solve. The card hands over the system and the results from a DC or numeric-ω AC run, already loaded.
• Empty. Open it with nothing and type your own equations. It does not need a circuit.

Every unknown carries a Restriction — Unrestricted, Positive, Negative, or Range with a from/to pair. This is not cosmetic: a restricted solve runs a bounded least-squares rather than the unrestricted root finder, and a system whose root lies outside its restriction says so instead of handing back a boundary value that looks like an answer.

Any identifier works as a variable name, Python keywords included — is, the natural name for a source current, is accepted.

It needs SciPy, which the offline builds bundle.

The SPICE Translator

Both directions, in a card of its own.

Symbulator to SPICE returns a netlist with a title line and .end, ready to paste into ngspice or LTspice. Anything it cannot translate comes out as a * comment and is reported — it never quietly drops an element. Dependent sources translate whenever their value is affine in node voltages.

SPICE to Symbulator takes the linear subset of a netlist. Elements and directives outside that subset are dropped and reported, never guessed at.

It is a prototype, and says so on the card. Read the warnings both ways.

The schematic

Symbulator draws the circuit you described, from the description alone. It is worth a glance before trusting any answer: the fastest way to catch a node typed wrong is to see the picture disagree with the circuit in your head.

It is a drawing of what you wrote, not of what you meant — which is exactly what makes it useful.

The package

The solver is on PyPI and does not need the app:

pip install symbulator

from symbulator import dc
r = dc("e,1,0,10:r1,1,2,1'k:r2,2,0,2'k")
r.v2

Every answer is a SymPy expression, so it substitutes, differentiates and lambdifies like any other. In a notebook there are %%dc, %%ac, %%fd and %%tr cell magics that take a circuit written one element per line, the way the app's input card does.