Plated-wire memory is a variation ofmagnetic-core memory developed byBell Laboratories in 1957.[1] Its primary advantage was that it could be assembled by machine, which potentially led to lower prices than magnetic core, which was almost always assembled by hand.
Instead of threading individualferrite cores on wires, plated-wire memory used a grid of wires coated with a thin layer ofiron–nickel alloy (permalloy).[2] The magnetic field normally stored in the ferrite core was instead stored on the wire itself. Operation was generally similar to core memory, with the wire itself acting as the data line, and the magnetic domains providing the individual bit locations defined by address (word) lines running on either side of (and perpendicular to) the data wire.
Early versions operated in a destructive read mode,[citation needed] requiring a write after read to restore data. Non-destructive read mode was possible, but this required much greater uniformity of the magnetic coating.
Improvements in semiconductor RAM chips provided the higher storage densities and higher speeds needed for large-scale application such asmainframe computers, replacing previous types of memory, including both core and plated-wire memory.
Plated-wire memory has been used in a number of applications, typically inaerospace. It was used in theUNIVAC 1110 andUNIVAC 9000 series computers, theViking program that sent landers toMars, theVoyager space probes,[3] a prototype guidance computer for theMinuteman-III, theSpace Shuttle Main Engine controllers,[4] theKH-9 Hexagon reconnaissance satellite,[5] and in theHubble Space Telescope.