Accessing Core Memory

Core memory is read by setting the bits of a machine word to zero and measuring the output of the sense lines of the cores written to. This means that reading core memory is destructive: the information stored in the cores being read is lost. Hence the bits read must be re-written after reading them.

Diagram D3 illustrates how a byte is read from core memory. Red color denotes "one" polarity and current, blue color denotes "zero" polarity and current. The X-driver and Y-driver lines of the cores storing the byte are driven in such a way that the polarity of all the cores is flipped to zero. When flipping a one bit to zero, a pulse (red) is generated in sense line. When flipping a zero bit to zero, no pulse is generated. The pulses of the sense lines are decoded by the sense amplifier, resulting in the bit pattern stored in the cores. After reading the byte, all its bits have "zero" polarity.

To write a value to core memory, the byte to be written to must be read first in order to set all of its bits to zero. After clearing the byte, the X-driver and Y-driver lines are driven in such a way that all the bits of the byte are set to one. When writing to core, the sense/inhibit lines are used to inhibit the change of polarity in specific toroids. This process is illustrated in diagram D4.

The inhibit lines of the bits not to set to one are driven in the direction opposite to the line they are paired with. Driven inhibit lines are displayed in blue color. By driving a driver and inhibit line in opposite directions, the magnetic field of the driver line is neutralized and so the cores attached to these lines are not flipped to one.

[Larger image 620x350, 45K bytes]

The logic to drive inhibit lines when writing to core and to decode the sense lines when reading core is quite complex, so it is frequently placed on a separate board. The picture shows a DEC H-217C sense/inhibit driver module.

Nils M Holm 2004, 2008