dualrail

Dualrail refers to a data encoding technique in digital electronics in which each logical bit is represented by a pair of wires rather than a single line. In the common 1-of-2 scheme, one wire carries the “true” value while the other carries the “false” value. A separate spacer or null state may be defined where both wires are inactive, and a state where both are active is typically considered invalid. This encoding supports certain asynchronous and self-timed designs by making data transitions explicit on both lines.

In a typical dualrail scheme, valid data are indicated by exactly one wire being asserted for a given bit. For example, a pair of signals (A, B) might encode 0 as (1, 0) and 1 as (0, 1), with (0, 0) used as a spacer between data tokens. The requirement that the two rails are not both asserted at the same time helps to avoid some timing hazards. There are several variants and terminology, including different conventions for the spacer state and for how control information like handshaking is conveyed.

Advantages of dualrail encoding include reduced sensitivity to gate and wire delay variations, potential delay-insensitive operation in properly designed circuits, and improved hazard resistance in asynchronous logic. Disadvantages include the need for roughly twice as many wires and gates, higher circuit area and power consumption, and greater design complexity when integrating with conventional synchronous logic. Consequently, dualrail is most commonly explored in asynchronous, self-timed, or research-oriented circuits rather than mainstream synchronous designs.

See also: 1-of-2 coding, asynchronous circuits, delay-insensitive data paths, self-timed logic.