Historical Computer Engineering - Memory
Mode: Visitor // Expert
Language: Deutsch English

Memory

1976: magnetic-core memory

Aufbau und Funktionalität

The magnetic-core memory is a non-volatile memory, that consists of a big number of magnetically hard cores, which are arranged on copper-wires. Magnetically hard means, that each core stores one state of a bit. Two mutually isolated wires, that serve the function of reading and writing information, run through each core. Accordingly one core, which are formed like a ring, is situated at one cross-way.

The storage occures with the aid of the direction of the core-remanence (remaining magnetism). To store data on a magnetic-core, a electric current of a specific extent on the writing-wire is needed, which creates a magnetic field. When the electric current is switched off, magnetism remains, that stands for '0' or '1'. To read out an informaton, a voltage impulse must be induced on the reading-wire, that finds out the direction of the magnetism. But this procedure destroys the core-data and that's why it is instantly necessary to store the information again, so that the data doesn't get lost.

The first computer systems needed 20 μs for one read-write-cycle. Until the 1970s, the systems reached a time of 0,3 μs. The fastes access times for data were reached by a diameter of 1 mm and 0,25 mm of the ring-cores.

Three of the exhebits are magnetic-coe memories: magnetic-core memory from 1976, magnetic-core memory of the IBM/360 and the data storage-board of the SM4. An own article elaborates on the last mentioned exhebit.