Where is the rest?
I'm pretty sure there's more story to publish there, or are you holding the other half back for another slow day?
A new memory technology which harks back to the 1950s is to be launched by the Ferroelectric Memory Company (FMC), a commercial venture being spun out of the Technical University of Dresden’s nano and micro-laboratory (NaMLab). NaMLab specialises in dielectric materials for semiconductor devices and focuses on the application …
*Two* Ferranti Argus telemetry computers with 1KB ferrite core storage each. The techs were on permanent toilet watch, as there were no spares in the world, and the project to move to a VAX was delayed. The machine was barricaded with furniture, as they were terrified a chair bump would do for.
Looks like a teenytiny replica of the old ferrite rod/coil arrangement in transistor radios strung on a grid of wires running up/down and diagonally. The XY wires were used to energize the coil (or not) for a single bit and magnetize the ferrite rod. To read it back you "wrote" back to the bit and "listened" for a momentary pulse on the detector wire that indicated a state change. If you got a pulse the bit was the opposite binary sense to whatever you wrote.
The ones I saw and used were mounted on 6 or 8x4 foot* "barn doors" on the corner of one of the cabinets that comprised the computer itself which could be fanned open to allow air into the matrices because they got hot.
They don't make 'em like that any more (thank Offler).
[EDIT] And I just realized you meant the cleverer modern version. So change all that to "Haven't the foggiest. Whose round is it?"[/EDIT]
* sorry, 60 year old meat memory can't remember clearly which it was.
If their choosing to go by the moniker "FMC Corporation" is any indication, this crew isn't too bright. If they're too dim to type type "FMC" into google and find out there's already a global, multi-billion dollar company with that name, then their prospects probably aren't good...
Exactly! My son works for FMC. Maybe they want to rethink the name because I'm pretty sure that it's already trademarked. Any good lawyer would have found that in the first 5 minutes.
Might want to rethink investing in a company that does so little due diligence
>If they're too dim to type type "FMC" into google and find out there's already several big companies with that name,
Fixed it for you! See:
Ford Motor Corporation
FMC Fairbanks Morse and Company - heavy plant
FMC Technologies - oil services
FMC.co.uk Dental industry publishing
It is acceptable to trade under an already used name if you are in a different sector. Do you remember a time when Apple Computer wasn't a player in the music industry, but Apple Corp was?
Do you remember a time when Apple Computer wasn't a player in the music industry, but Apple Corp was?
If memory serves, that was a sort of détente achieved after a long court 'n lawyers session, in which each party agreed never to tread on the others' turf in perpetuity on pain of draconian penalties.
The nibbled fruit mob had apparently forgotten all about this when they launched iTunes, until the other lot went all Smaug the Magnificent on their arses.
I'm sure it was reported here when the Apple Corp dragons gleefully pointed out in court otherapple's blatant breach of their long standing and legally binding agreement.
We had the GEC 4020 assemblies with thermite charge markings (no actual thermite charges though), these used core store so an initiated thermite charge was used to take them above the Curie point and effectively zero them. Would have been in the Nimrod AEW3 if it had been continued with, although the computer in the real systems would have been a 4060 to make it a bit faster.
If memory serves the whole shebang was contained in a pretty thick walled unit to try to contain the 3,000 C temperatures for long enough to comprehensively knacker things, meanwhile the Nimrod crew would have been doing whatever they did when a forced landing behind enemy lines was on the cards.
Airshows aren't the same now the Cold War is over...
When my father was posted to Vietnam during the late 1960's, his computer room contained a cabinet full of thermite grenades. If the base was likely to be overrun by the enemy, the computing staff was instructed put grenades on top of the computers and disk drive packs, pull the pins, and run like hell.
Indeed. And the safes containing codebooks and misc. hardware spares had a cupholder affair cut into it. Anytime the base was under attack or in attack threat status, a thermite grenade was to be placed in each holder. When the word came down, pull the pins, grab your rifle and hightail it for the assembly point.
10^5 cycles? Last I checked, that's only 100,000, which isn't really all that terrific versus better flash technologies out there. I think one of the higher bars set for any flash successor is a minimum longevity of 10^7 if not 10^9.
The rest of the specs also don't hold much promise, either.
Core memory is ferromagnetic--it stores with a permanent magnetic field which happens when molecules or domains align with each other and can remain aligned without thermal energy disrupting them.
Ferroelectric is when a material is made of polar molecules which have a built in electric field, can be made to align with each other and can remain aligned without thermal energy disrupting them.
If you make a capacitor (basically parallel plates that store charge) with a dielectric (fancy term for insulator) that is ferroelectric, the charge does not leak off. DRAM uses a capacitor dielectric which is not ferroelectric to store a bit, and the charge slowly leaks off, that's why it is "dynamic" and needs to be refreshed. When a ferroelectric material is used the permanent electric field of the dielectric keeps the charge present, so it is non-volatile.
In a capacitor, charge stored is proportional to the voltage between the two parallel plates. A DRAM is read by pulling off the charge as a current and reading the current. Then it has to be rewritten after the read. It is not necessary to pull the charge off a ferroelectric storage cell, you can simply detect the voltage, therefor no need to rewrite.
The "ferro" in ferromagnetic refers to iron (and its magnetic properties). The "ferro" in ferroelectric does not refer to iron, it is just there because the effects of the material are analogous but with electric fields instead of magnetic fields, so they made up an analogous word.
Companies have been making ferroelectric memory (FRAM) for years. The main one I can think of was Ramtron International, recently taken over by Cyprus Semiconductor.
Companies have been making ferroelectric memory (FRAM) for years. The main one I can think of was Ramtron International, recently taken over by Cyprus Semiconductor.
Have one of these in mind for an upcoming project:
Both devices seem to be Fujitsu parts.
Hwyl!
M.
"in the 1950s, early computers used a matrix of magnetic cores, which stored bits by either being magnetised or not magnetised"
How did that work then?
I was taught that the direction of the magnetism was changed, one way for 'one' and the other for 'zero'. Still in use in the Shuttle, hence being able to read the contents of the Challenger's memory after it exploded and crashed into the sea.
I still use a Data General Nova in an olde world lighting desk the OS is on one core and the backup decorates the dining room wall - all of 8k x 16 istr in 1971 when it was made it would have cost thousands. Problems can arise with the drive transistors, recirculate logic chips and the diode arrays apparently. They're a pig to debug / replace. The core is 6" x 9" and is hand made.
On the upside the core is radiation hard and you can now get a DG emulator for your PC (ReNOVAte - Wild Hare) RISC these babies aren't :)
LC
Not sure why you would consider the NOVA to be "not a RISC". Decently orthogonal instruction set, fixed (well, two possible lengths IIRC, but could be thinking of later members of the family) instruction length, predictable (and deducible) instruction execution times, at most one operand changed per instruction...
As for Ferroeletric memory, I recall seeing a photo of a prototype from Bell Labs, back in the 1960s. It was only the substrate and metalization, with off-board drivers, but had either 64 (8x8) or 256 (16x16) sites. Again, I don't recall. But, yeah it has been possible to buy FRAM for quite a while. Just as it's still possible to find "bubble memory", although IBM calls the on-chip version "racetrack", and NCR's Thin-film memory has a new on-chip incarnation whose name and sponsor I can also not recall at the moment . Perhaps I need a new memory.
You only needed to rewrite the bit if the (read) write caused a state change. The state change caused a change in magnetic field which in turn generated a current in the coil wound around the core which could be detected. Current made= state changed. No current = same state as written information - no rewrite necessary.
Grumblegrumblewheresmydinner?