Hacker News Reader

Comment by userbinator | Hacker News Reader

Core eventually miniaturised into https://en.wikipedia.org/wiki/Ferroelectric_RAM , which remains in limited use (TI has a series of MCUs with FRAM instead of EEPROM) --- and one curious application is in the toner cartridge DRM chip of a Lexmark laser printer, for which there is this ongoing reverse-engineering effort:

https://www.eevblog.com/forum/projects/lexmark-toner-chip-ti...

(Normally those use EEPROMs, so the choice of FRAM is a bit odd.)

View on HN

> Core eventually miniaturised

Well, sorta. The physics behind the device is the same, but the manufacturing is distinct. Core was produced by threading tiny ferroelectric donuts cut from an extruded tube onto copper wires. There's no "miniturization" continuum between that and FRAM, which is produced with planar lithography.

You make it sound as tho FRAM is old technology... its actually just starting to be used in volume (in microcontroller based systems). FRAM has some very desirable properties compared to current technology uC that use both SRAM & FLASH. Its still slightly expensive at this point in time but that will solve itself in time.

> FRAM has some very desirable properties compared to current technology uC that use both SRAM & FLASH.

Could you please elaborate?

it's as fast as sram but persists like flash

There seems to be some confusion in this thread between ferroelectric vs ferromagnetic.

These are two distinct physical phenomena. FRAM is based on the former whereas magnetic core memory is based on the latter.

With this in mind, I’m not sure it is really accurate to regard FRAM as a successor to magnetic core memory.

It's a very cool but strangely marginal technology; it's not going to have the density of Flash any time soon, sadly.

The interesting property of FRAM is that it behaves in same way as core. Random access and non-volatile (and without significant write/erase wear). End result is that for applications that will fit into practically sized FRAM array you essentially do not have to thing about restoring the state after power loss, because the state simply is there as it was on core-based minicomputers (one thing that I find funny is that MSP430FRxxx in fact is PDP-11 on single chip with slightly different instruction encoding).

Reply on news.ycombinator.com