Skip to Main Content
Insuring safety of flight in high-performance aircraft requires great advancement in the reliability of flight-control computers. While such improvements are not attainable by component improvement alone, they can be effectively achieved by redundancy. This paper describes a combination of redundancy techniques developed to meet flight-control requirements, using majority logic supplemented by self-combinational networks and parity-check error detection. Alternative redundancy techniques are outlined and simple approximate expressions for reliability are derived, showing the reasons leading to selection of the desired combination. Theory of majority logic is explained, and expressions for optimum voter arrangement and resultant reliability are derived. Implementation of a model flight-control computer, built for the Air Force, is explained. Judicious computer organization enhances reliability of the basic nonredundant layout, achieving a mean lifetime of 1,500 hours (failure probability of 0.005 for an 8-hour mission). For an 8-hour mission, the redundant computer has an equivalent lifetime of 110 years (failure probability of 10Â¿5).