Controller Redundancy

In the “olden days” of single loop control hardware, it was rare for a process loop to actually be backed up with another. Usually, the fallback position was to switch to manual. On critical processes, however, the expense of another controller was justified for a one-to-one control loop backup. There is no doubt that there can be opportunities for redundant control for a single process loop. The battle cry for “single loop integrity” started in the early days of distributed systems, when shared loop control was the only economical way to use microprocessors. When central (mainframe) computers became available it was very difficult to make them redundant, but it was at least possible. In the controller design architecture that uses different card sets for each controller “file,” redundancy is a rarity. In those “single card” controller racks that are made by several vendors, only “multifunction cards” can back up their identical twin.