Inductor

A real inductor, solved by the matrix: the one two-terminal passive whose state is a current rather than a voltage.

An inductor cannot be modeled as a voltage behind an impedance the way most parts here are, so it is analog-native from the start: it seeds an analog island and the matrix carries its branch current as an unknown to be solved for, rather than the part guessing at it.

Pins

Pin	What it does
1	Either end.
2	The other end.

Properties

value: inductance in millihenries. 100 mH is a typical small choke, 1 mH a filter inductor, 10 000 mH (10 H) a large audio choke.

What the model gets right

Because the branch current is a real unknown in the matrix, an LC or RL network here rings and settles the way the real components do, not the way a lookup table pretends they do.

What it does not model

No winding resistance and no mutual inductance. This is an ideal inductor. If a circuit needs the resistance a real coil's winding has, put a resistor in series with it, exactly as you would account for a real inductor's DCR on paper.

Nor is there a core: no saturation, so the inductance never falls away as the current rises the way a real ferrite's does; no core loss, no hysteresis, and no self-resonance from the winding's own stray capacitance, which on a real 100 mH choke puts a ceiling of a few tens of kilohertz on where it behaves as an inductor at all. There is no tolerance, no temperature coefficient and no current rating.

Common mistakes

Forgetting the series resistance a real inductor always has. An ideal inductor with nothing damping it rings forever in a simulation that has no losses to bleed the energy away, where a real coil's own resistance would settle it out in a few cycles.

See it in action

RL filter on the scope pairs an inductor with a resistor and a signal generator to show exactly this ringing and settling behavior. Open it from /templates.

The flyback diode page switches a 10 mH coil off with and without a diode across it and catches the spike on the scope.