peg@psuecl.bitnet (11/13/89)
In article <1308@rodan.acs.syr.edu>, pwasilko@rodan.acs.syr.edu (Wasilko) writes: > Query: What is the current state of the art in non-volatile solid > state memory technology? > > The most recent material in our library is of 1985 vintage with > references to 512K bubble memory chips, 128K removable bubble memory Hey! I certainly can't profess to being an expert on non-volatile memories, but I did do some looking for a recent project. It seems that the bubble memory craze sort of faded out. Apparently they found other ways of doing the trick that cost a lot less. I don't know if you would consider it "state of the art" but the EEPROM seems to be pretty well-liked right now. I have databooks from SEEQ and XICOR in front of me (1988) that list EEPROMS up to 128k bytes. The access time is quoted as 200ns. Write time is usually much, much longer--these say < 150 us. There are articles in these databooks comparing EEPROM to bubble memory, and it seems the EEPROM wins in the areas of read timing, write timing, power supply simplicity, operating/storage temp range, size, weight, etc. Probably cost, too! There are also articles on pushing EEPROM up to 1 Megabit, which may have been done by now. By the way, I'm not associated with either of these companies--they just have nice products! Contact: Xicor, Inc SEEQ Technology, Inc 851 Buckeye Ct 1849 Fortune Dr Milpitas, CA 95035 San Jose, CA 95131 408-432-8888 408-432-9550 Paul
sparks@corpane.UUCP (John Sparks) (11/13/89)
In article <65243@psuecl.bitnet> peg@psuecl.bitnet writes: >In article <1308@rodan.acs.syr.edu>, pwasilko@rodan.acs.syr.edu (Wasilko) writes: >> Query: What is the current state of the art in non-volatile solid >> state memory technology? >> >> The most recent material in our library is of 1985 vintage with >> references to 512K bubble memory chips, 128K removable bubble memory >I don't know if you would consider it "state of the art" but the EEPROM seems >to be pretty well-liked right now. I have databooks from SEEQ and Don't EEPROMS also have a problem with a limited number of writes? That is, after so many writes to the chip, it begins to degrade? They are great for storing data that needs to be changed and non-volitile (such as modems use to save parameters) but probably not so good for use in RAM replacement where there is a constant read-write situation. -- John Sparks | {rutgers|uunet}!ukma!corpane!sparks | D.I.S.K. 24hrs 1200bps ||||||||||||||| sparks@corpane.UUCP | 502/968-5401 thru -5406 The next sentence is true. The previous sentence is false.
kurt@tc.fluke.COM (Kurt Guntheroth) (11/15/89)
There are several non-volatile memory technologies.
1. (Most vanilla) battery backed CMOS RAM. There are very low standby
power CMOS RAMs kept alive by watch batteries.
2. EEPROMs. These are fast read access, slow write access devices.
3. (Most out-there) Ferroelectric RAM. These are the semiconductor version
of core memory. RAMTRON is the US company that makes these. These were
all over the ISSCC conference proceedings a couple of years ago, but
there are problems with the manufacturing process that are limiting their
commercial development. Ferroelectric RAM has the theoretical potential
to beat dynamic RAM in density (if you believe the press releases) is
intrinsically non-volatile, and has fast read and write access.yahoo@unix.cis.pitt.edu (Kenneth L Moore) (11/15/89)
In article <12459@fluke.COM> kurt@tc.fluke.COM (Kurt Guntheroth) writes: >There are several non-volatile memory technologies. ==>1. (Most vanilla) battery backed CMOS RAM. There are very low standby ==> power CMOS RAMs kept alive by watch batteries. I've always called these pseudo-non-volatile. ==>2. EEPROMs. These are fast read access, slow write access devices. ==>3. (Most out-there) Ferroelectric RAM. And what ever happened to mag bubble memory? -- yahoo@unix.cis.pitt.edu (Kenneth L Moore)
peg@psuecl.bitnet (11/16/89)
> Don't EEPROMS also have a problem with a limited number of writes? > That is, after so many writes to the chip, it begins to degrade? > > They are great for storing data that needs to be changed and non-volitile (such > as modems use to save parameters) but probably not so good for use in RAM > replacement where there is a constant read-write situation. You are absolutely right in that EEPROMs can only be written to so many times before they fail. I certainly would not recommend building a computer with EEPROM for RAM. Notice the letters ROM in EEPROM :) ! It depends very much on your application whether EEPROMs are useful. If you can put programs in ROM, run software in RAM, and only use EEPROM for data storage you may be in good shape. For example, many stage lighting controllers use EEPROMs to store preset light levels. Xicor claims about 500 k writes on average. It depends greatly on the ambient situation, apparently. Paul
ncas@cbnewsj.ATT.COM (clifford.a.stevens..jr) (11/16/89)
In article <1230@corpane.UUCP>, sparks@corpane.UUCP (John Sparks) writes: > In article <65243@psuecl.bitnet> peg@psuecl.bitnet writes: > >In article <1308@rodan.acs.syr.edu>, pwasilko@rodan.acs.syr.edu (Wasilko) > writes: > >> Query: What is the current state of the art in non-volatile solid > >> state memory technology? > > >I don't know if you would consider it "state of the art" but EEPROM seems > >to be pretty well-liked right now. I have databooks from SEEQ and > > Don't EEPROMS also have a problem with a limited number of writes? > That is, after so many writes to the chip, it begins to degrade? > > are great for storing data that needs to be changed and non-volitile (such > as modems use to save parameters) but probably not so good for use in RAM > replacement where there is a constant read-write situation. Exactly! We use a EEPROM in our *MODEMS* to hold options and other such info, but they do have a write limitation. They are fairly slow on writes and have a limited number of writes, but operate at the speed of the bus, forever (As forever as any RAM!) on reads! ------------ Who, me worry?!? Cliff Stevens MT1E228 att!cbnewsj!ncas (201)957-3902
davidb@braf.inmos.co.uk (David Boreham) (11/16/89)
Try looking at "Flash EPROM" devices. These are either bulk or sector erasable (most require 12v Vpp but some from ATMEL are 5v only) and can be programmed byte-by-byte. The best have 10,000 program-erase cycles and the cheapest cost about $160/Megabyte. Much cheaper than full-featured EEPROM. Some new handheld computers (for instance the new one from Scion in England) use Flash-EPROMS for simulated disk. Manufacturers: Intel, Seeq, Atmel. Future manufacturers: TI, Toshiba, NEC, SGS-Thomson Microelectronics. David Boreham, INMOS Limited | mail(uk): davidb@inmos.co.uk or ukc!inmos!davidb Bristol, England | (us): uunet!inmos.com!davidb +44 454 616616 ex 547 | Internet: davidb@inmos.com
nemeth@gpu.utcs.utoronto.ca (Gabe Nemeth) (11/17/89)
Dallas semi makes a number of really slick non-volatile parts by adding a lithium power source directly to the die. They have non-volatile 8031's, and lots of rams. Don't know prices though. /leonard
mmm@cup.portal.com (Mark Robert Thorson) (11/17/89)
kurt@tc.fluke.COM (Kurt Guntheroth) says: > 3. (Most out-there) Ferroelectric RAM. These are the semiconductor version > of core memory. RAMTRON is the US company that makes these. These were > all over the ISSCC conference proceedings a couple of years ago, but > there are problems with the manufacturing process that are limiting their > commercial development. Ferroelectric RAM has the theoretical potential > to beat dynamic RAM in density (if you believe the press releases) is > intrinsically non-volatile, and has fast read and write access. Not quite true. Ferroelectric RAM is only like magnetic core in the sense that it stores data and is non-volatile. It is not at all a magnetic technology. Ferroelectric technology got its unfortunate name because of the similarity of ferroelectric materials to ferromagnetic materials. Both can be used to create bistable devices. In the case of ferromagnetic materials, an imposed magnetic field can be used to flip the magnetism of a core memory cell. In the case of ferroelectric materials, an imposed electric field can be used to flip the charge on a capacitor. A common ferroelectric material is lead zirconate titanate, which you'll notice contains no iron. Sources of ferroelectric RAM's are Ramtron (Colorado Springs, CO) and Krysalis (Sunnyvale, CA).
wbrown@beva.bev.lbl.gov (Bill Brown) (11/18/89)
In article <1989Nov16.174857.22800@gpu.utcs.utoronto.ca> nemeth@gpu.utcs.UUCP (Gabe Nemeth) writes: >Dallas semi makes a number of really slick non-volatile parts by adding a >lithium power source directly to the die. They have non-volatile 8031's, >and lots of rams. Don't know prices though. >/leonard Last time I checked (a few months ago) 8K by 8 were in the neighborhood of $25.00 in small quantities (1). I don't think they have distributers which means ordering direct. -bill wlbrown@lbl.gov Disclaimer: These opinions are my own and have nothing to do with the official policy or management of L.B.L, who probably couldn't care less about employees who play with trains.
wtm@neoucom.UUCP (Bill Mayhew) (11/18/89)
I noticed when I looked inside my AT&T CEO 2224 modem that it used a RAM chip with internal battery for holding the setup information. The nonvolitile RAM was about double the height of a typical J type 24 pin JDEC package. It apparently has its own internal lithium cell (which appears to be non replacable). It has been a while since I had the modem apart, but I believe the part was manufactured by Dallas Semiconductor. Seems like a saw a write-up on the aforementioned parts in EDN or a similar magazine a couple of years ago. The battery RAM contains a powerdown circuit to prevent blowing away the data when Vcc goes down. I recall from using a 5313 2K * 8 EEPROM from Seeq (old technology now) that you had to be careful holding the write enable line while power was applied so as to avoid zeroing the chip. The number of writes on an EEPROM is several hundred thousand, so one needn't worry too much unless considering using it for scratchpad memory durring program execution. One wouldn't want to do that anyway, as the write time is pretty long. The older parts like the 5313 had write times on the order of a couple of mS! Bill