2764 EPROM data retention

P

Peter Bennett

Guest
I am looking for information on how long we can expect a 2764 (or
similar UV-erasable EPROM) to retain its data. We have a number of
microprocessor systems on site using 2764s which were installed as
early as 1980, and we are now wondering if we should be concerned
about them losing their programs.

A quick search on Google turned up a page at ST about this, and an
enquiry to Intel's "museum" gave me a datasheet which didn't address
the question. (The parts we used at the time were from Intel)


--
Peter Bennett VE7CEI
email: peterbb (at) interchange.ubc.ca
GPS and NMEA info and programs: http://vancouver-webpages.com/peter/index.html
Newsgroup new user info: http://vancouver-webpages.com/nnq
 
Hiya!

I repair old video game PCBs, from 1975 onwards, and for the most part,
EPROMs seem to hold their data very well, I find very few boards that
have roms with corrupted data, it's more likely for the chip to
electically fail completely.

Yours, Mark.

Peter Bennett wrote:

I am looking for information on how long we can expect a 2764 (or
similar UV-erasable EPROM) to retain its data. We have a number of
microprocessor systems on site using 2764s which were installed as
early as 1980, and we are now wondering if we should be concerned
about them losing their programs.

A quick search on Google turned up a page at ST about this, and an
enquiry to Intel's "museum" gave me a datasheet which didn't address
the question. (The parts we used at the time were from Intel)
Click to expand...
 
Dear Peter,

Peter Bennett wrote:
I am looking for information on how long we can expect a 2764 (or
similar UV-erasable EPROM) to retain its data. We have a number of
microprocessor systems on site using 2764s which were installed as
early as 1980, and we are now wondering if we should be concerned
about them losing their programs.
Click to expand...
Good question. It seems strange, but the data retention heavily depends
on the programmer device used. If it has programmed the bytes using good
voltage margin, you won't have any problems with corrupted memory contents.
IIRC (but can't state a source) Intel guaranteed 10 years if programmed
according to the specs.
Of course, these 10 years are meant for the full temperature range.
If your 2764's are normally running at room temp or so, you can safely
assume 30 years.

HTH
Wolfgang


--
From-address is Spam trap
Use: wolfgang (dot) mahringer (at) sbg (dot) at
 
Hi, yes you should be concerned because you cant replace them if they fail. The
smart move would be to copy the contents to disk.
 
Back them up as quickly as you can...then play around with making
spares to leave in the systems in case of failure and you can't find
an Eprom programmer handy and time is not on your side. Also keep some
spare CPUs as I am finding that older microprocessors are starting to
fail (Z80, 6808, 6502, etc..) nearing the end of their life I fear...

John :-#)#

On Thu, 05 Feb 2004 13:20:20 -0800, Peter Bennett
<peterbb@nowhere.invalid> wrote:

I am looking for information on how long we can expect a 2764 (or
similar UV-erasable EPROM) to retain its data. We have a number of
microprocessor systems on site using 2764s which were installed as
early as 1980, and we are now wondering if we should be concerned
about them losing their programs.

A quick search on Google turned up a page at ST about this, and an
enquiry to Intel's "museum" gave me a datasheet which didn't address
the question. (The parts we used at the time were from Intel)
Click to expand...
(Please post followups or tech enquires to the newsgroup)
John's Jukes Ltd. 2343 Main St., Vancouver, BC, Canada V5T 3C9
Call (604)872-5757 or Fax 872-2010 (Pinballs, Jukes, Video Games)
www.flippers.com
"Old pinballers never die, they just flip out."
 
Hiya!

The 6808, 6502 are going to be a problem in the future I agree, as they
are getting harder and more expensive to find.

Don't worry about the Z80. Zilog still make the 10MHz CMOS version,
which is pin compatible with all the previous versions. I'm going to
have a go at protoing an adaptor board to plug a Z80 into an 8080
position. I beleive it should work as the Z80 contains all the 8080
instructions and according to the datasheet, is 100% software compatible
with it. Got to be worth a try, unless someones tried already and has
either failed, or got one onto an adaptor board they sell - LMK either
way on that to save me the effort!!

The B-I-G problem in the future is going to be BiPolar PROMs, as they've
stopped making the smaller ones, only Cypress do EPROM equivs to the
82s181, 82s191 and some of the registered versions. The 82s135 (Zaxxon
amogst others) is already near impossible to find. The older PLDs as
well - PLS153 (82S153) is VERY expensive when you can find it, and I've
not come across a single blank CK2605 in three years!

Just a thought - if there was a problem with older CPUs in the future,
could we not either build them from CPLDs, or use a fast CPU to emulate
them? I know people have done amazing things with CPLDs already!!

Yours, Mark.

John Robertson wrote:
Back them up as quickly as you can...then play around with making
spares to leave in the systems in case of failure and you can't find
an Eprom programmer handy and time is not on your side. Also keep some
spare CPUs as I am finding that older microprocessors are starting to
fail (Z80, 6808, 6502, etc..) nearing the end of their life I fear...

John :-#)#

On Thu, 05 Feb 2004 13:20:20 -0800, Peter Bennett
peterbb@nowhere.invalid> wrote:


I am looking for information on how long we can expect a 2764 (or
similar UV-erasable EPROM) to retain its data. We have a number of
microprocessor systems on site using 2764s which were installed as
early as 1980, and we are now wondering if we should be concerned
about them losing their programs.

A quick search on Google turned up a page at ST about this, and an
enquiry to Intel's "museum" gave me a datasheet which didn't address
the question. (The parts we used at the time were from Intel)


(Please post followups or tech enquires to the newsgroup)
John's Jukes Ltd. 2343 Main St., Vancouver, BC, Canada V5T 3C9
Call (604)872-5757 or Fax 872-2010 (Pinballs, Jukes, Video Games)
www.flippers.com
"Old pinballers never die, they just flip out."
Click to expand...
 
"Mark (UK)" <jumbos.bazzar@btopenworld.com> schreef in bericht
news:c03l4n$si5$1@hercules.btinternet.com...
Hiya!

The 6808, 6502 are going to be a problem in the future I agree, as they
are getting harder and more expensive to find.

Don't worry about the Z80. Zilog still make the 10MHz CMOS version,
which is pin compatible with all the previous versions. I'm going to
have a go at protoing an adaptor board to plug a Z80 into an 8080
position. I beleive it should work as the Z80 contains all the 8080
instructions and according to the datasheet, is 100% software compatible
with it. Got to be worth a try, unless someones tried already and has
either failed, or got one onto an adaptor board they sell - LMK either
way on that to save me the effort!!

The B-I-G problem in the future is going to be BiPolar PROMs, as they've
stopped making the smaller ones, only Cypress do EPROM equivs to the
82s181, 82s191 and some of the registered versions. The 82s135 (Zaxxon
amogst others) is already near impossible to find. The older PLDs as
well - PLS153 (82S153) is VERY expensive when you can find it, and I've
not come across a single blank CK2605 in three years!

Just a thought - if there was a problem with older CPUs in the future,
could we not either build them from CPLDs, or use a fast CPU to emulate
them? I know people have done amazing things with CPLDs already!!

Yours, Mark.
Click to expand...
Hooi,

You can buy programs - say kernels - of the ancient micros in VHDL. You can
also write them yourself as well. You can put them into any programmable
hardware device you like (assuming it's large enough.) As for the old
memories, you can find various types that has the same and most of the time
more content. Problem may be the fan-in/fan-out and other electrical specs
of the oldies. Some kind of buffering might be necessary. Speed should not
be the problem. AFAIK the old things did 50ns, the newer ones <25nS.

petrus



---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.572 / Virus Database: 362 - Release Date: 27-1-2004
 
Hi, yes you could do those things but you end up with the most expensive slow
processor in history.
 
You must log in or register to reply here.

Welcome to EDABoard.com

Sponsor

Back
Top