If you’ve heard of core rope memory, it will probably be in the context of vintage computing equipment such as Apollo-era NASA hardware. A string of magnetic cores and sense wires form a simple ROM arrangement, which though long-ago-superceded by semiconductor memory remains possible to recreate by the experimenter. It’s a path [Nicola Cimmino] has trodden, as he’s not only made a few nibbles of core rope memory, but incorporated it with an Arduino as part of one of the most unusual LED flashers we’ve ever seen. The memory holds a known sequence of bits which is retrieved in sequence by the Arduino, and the LED is kept flashing as long as the read values conform to those expected.
The memory itself is simple enough (and not to be confused with magnetic core memory). The cores are ferrite rings that form a sequence of small transformers that become the bits of the memory. Individual bits are set high or low by either passing a sense wire through a core to create a primary, or bypassing it. Multiple sense wires can be used for separate nibbles in the same cores, so for example his four nibbles all share the same four cores. Pulses are sent down the wires, either passing through a core or not, and equivalently picked up or not on sense lines.
In this case the sense wire is driven directly to ground by Arduino pins which means that the circuit is relying upon the current limiting of the ATmega328 to avoid destroying itself, it’s possible we’d add a driver transistor. The bits are read meanwhile from the secondary windings through a diode rectifier and capacitor to an Arduino analogue pin.
Core memory has been paired with an Arduino before on these pages, though of the RAM variety.
Source: CORE ROPE MEMORY MAKES ONE OF THE ODDEST LED FLASHERS WE’VE EVER SEEN
French 
English