Terminal Definition

Terminals are the assumed external contact points used when extracting the inductance matrix. Unlike capacitance extraction, inductance and resistance extraction requires terminal definitions, and the results will depend fundamentally on the terminal locations.

Terminals are specified by creating boxes or polygons and labels on special layers. The feature will define as equivalent all nodes that touch or are enclosed in the shape, for any Z coordinate (the features are in the X-Y plane).

Each terminal feature must have at least one overlapping text label on the same layer, that provides the terminal name. Terminals must resolve to pairs, where each pair is a ``port'', taking the inductance matrix as an N-port network. The pair is ordered, with one terminal being the ``plus'' terminal, the other the ``minus'' terminal.

This is all accomplished by adherence to the following rules.

1. The features (boxes or polygons) and labels which equivalence nodes and define terminals are created on special layers. The layer name is the same as the layer name of the conductor which provides the nodes. The purpose name is the special keyword ``fhterm''. Thus for example, for a metal layer named ``M1'' (which has the default ``drawing'' purpose) the corresponding special layer has the full layer-purpose pair (LPP) name ``M1:fhterm''. Such a LPP should be defined for each conducting layer in the technology file.

2. Terminal features must touch or enclose at least one node. Nodes can be found at the center of each edge of each tile. When connecting to the end of a metal strip, for example, the entire transverse width of the strip end should be enclosed in or touch the terminal feature, so that current flow is uniform.

3. Each terminal feature must have at least one overlapping text label on the same layer giving a terminal name.

4. Each terminal is one of a pair, the pair representing a port. The terminals of each pair must contact the same conductor group, i.e., be connected.

5. It is possible for a terminal to be used in more than one port, in which case the terminal will have more than one overlapping label.

6. Port-terminal association is by name. Name labels must follow these rules:
   1. Terminal names consist of a port name and a suffix. If the name string contains punctuation or white space, the first occurrence of such is stripped, and the port name is taken as the characters to the left, and the suffix is taken as the characters to the right, of where the punctuation or white space resided. If there is no punctuation or white space, the port name is the name string with the rightmost character stripped, and the suffix is this character. The port name and suffix must each contain at least one printable character or a fatal error results. The port name is arbitrary, but must be unique among the ports.

   2. Both terminals of a port must have the same port name, case sensitive. It is a fatal error if a terminal can not be paired.

   3. Both terminals of a port must have different suffixes. The suffix is used only to order the terminals in the port. This is done using lexicographic ordering of the suffix strings. Beyond ordering, the suffix is ignored. It is a fatal error if the suffixes are the same.

   4. Both terminals of a port must contact the same conductor group.

7. For each terminal feature, a list of intersecting nodes is created internally. The first node in the list is taken as the reference. If there are additional nodes, they are equivalenced to the reference node, using the FastHenry ``.equiv'' construct.

8. For each port, a FastHenry ``.extern'' construct is used to provide the reference nodes of the two terminals in order, followed by the port name.