To describe a circuit, you must provide a `netlist'. The netlist is simply a list of the components with their connections and values. The format is essentially the same as the standard SPICE format.
Before doing this, number the nodes on your schematic. (A node is a place where parts connect together.) Then, each part gets a line in the netlist (circuit description). In its simplest form, which you will use most of the time, it is just the type, such as `R' for resistor, or a label, like `R47', followed by the two nodes it connects to, then its value.
Example: `R29 6 8 22k' is a 22k resistor between nodes 6 and 8.
Node 0 is used as a reference for all calculations and is assumed to have a voltage of zero. (This is the ground, earth or common node.) Nodes must be nonnegative integers, but need not be numbered sequentially.
There should be a DC path through the circuit to node 0 from every node that is actually used. The circuit cannot contain a cutset of current sources and/or capacitors. If either of these cases occurs, it will be discovered during analysis. The program will attempt to correct the error, issue an `open circuit' error message and continue. This is rarely a problem with real circuits. Most circuits have such a path, however indirect.
Semiconductor devices require both a device statement, and a .model statement (or ``card''). The device statement, described in the Circuit description chapter, defines individual devices as variations from a prototype, as is required for different devices on the same substrate. The model statement, described in this chapter, defines process dependent parameters, which usually apply to all devices on a substrate.
The .model card syntax is:
.model mname type {args}
Mname is the model name, which elements will use to refer to this model. Type is one of several types of built-in models. Args is a list of the parameters, of the form name=value.