Re: Mouse card --- Can EPROM go Bad?

[Rubywand <rubywand@swbell.net>]

Vernon Bartlett writes ...
> 
> On Sat, 21 Nov 1998 10:27:50 -0800, Wayne Stewart
> <waynes@intergate.bc.ca> wrote:
> 
> >Without allowing for me replacing them before they're toast,
> >the SCSI cards would outlive the computers since they also use
> >EPROMs of one sort or another. I haven't run across bad
> >EPROMs in a IIgs yet but I have replaced some in a few II+
> >and IIe machines.
> >
> 
> A good reason to maintain all eprom codes on a backup disk - maybe
> even CD. And to keep a programmer.

     For sure. Our II+'s custom EPROM code is backed up on diskette and
our prommer just got a checkout.  

 
> Question is how long do unused eproms last?

     No one really knows how long any of this stuff is going to last. We
can be fairly sure about some things-- e.g. that a 7404 IC produced
today will be good for at least 20 years of 'normal use'.

     Based upon experience, we have a pretty good idea about durability
for lots of IC's. The catch is that _we_ are the ones doing the real
world durability 'testing'.
  

> I've had to throw away a
> few that totally refused to erase/reprogram after sitting around in
> storage.

     One popular notion is that solid state integrated circuits in a
parts box do not change. They do; but, so far, over the time span of a
couple decades, the change for most IC's seems not to be ruinous.

     As level of integration increases and everything becomes smaller,
the effects of aging should become noticable more quickly-- there is
less stuff (silicon, etc.) per circuit on a die which needs to change.
On the other hand, costs are dropping and the trend is toward junking
whole boards/modules after just a few years. If some VVLSI chip on a
graphics card rots after 10 years, no one but a few experimenters are
likely to notice or care. 

     I've come across EPROMs like those you mention. Of course, there
are many good techy sounding explanations for the problem: maybe the
chip has been through too many erase/program cycles, maybe a Write
Enable circuit blew due to over-voltage during programming, maybe the
chip was zapped by static electricity, maybe the UV eraser cooked some
critical gates, maybe the EPROM gods are angry, ... . 

     That some are unprogrammable due to age is an explanation which did
not occur to me. Still, if there is a limit on number of erase/program
cycles, it is conceivable that age might have similar effects. (A couple
things to try are longer erase times and, on the programming side, a
longer Write strobe time and slightly higher Programming voltage.)     


> Will they still be making 27xxx 12 years from now?

     Maybe; but, it seems likely that EPROMs will be of the
electrically-eraseable sort rather than UV-eraseable. Small EPROMs (like
anything under 64kB) may disappear in favor of including some kind of
EEPROM memory on all micro-controllers.

     None of this will be a major problem for II users. If replacements
for our EPROMs cease to be available, we'll just cobble together
adaptors and use the new stuff.



Rubywand