EEPROM Array Refresh?

[Steve H]

A previous post mentioned the 18F452 data sheet section 6.8. I looked at that and it seems to me to be very confusing.

From data sheet section 6.8
================================

Frequently changing values will typically be updated more often
than specification D124. If this is not the case, an array
refresh must be performed. For this reason, variables
that change infrequently (such as constants, IDs, cali-bration,
etc.) should be stored in FLASH program memory.


Note: If data EEPROM is only used to store constants
and/or data that changes rarely, an
array refresh is likely not required. See
specification D124.
=======================================================

The first para says an array refresh is needed if the variables are not written frequently?

The second para says the refesh is not needed?

Does anyone have a link to Microchip which talks about this refresh method - I'm confused ;-)

Steve H.
___________________________
This message was ported from CCS's old forum
Original Post ID: 13401

[PCM programmer]

<font face="Courier New" size=-1>:=A previous post mentioned the 18F452 data sheet section 6.8. I looked at that and it seems to me to be very confusing.
:=
--------------------------------------------------------

I agree.

I did a web search on this, as well as usenet archives,
and Microchip's forum archives. Here is a comment from
the forum:

"The bottom line is that if you write to the EEPROM constantly
in your application (like 1,000,000 times or more @85C) that
you should consider refreshing all stored data every 1,000,000th
write. This is NOT a data retention issue - if you write to the
EE, the data will be retained. The issue exists on nearly all
sub micron Flash / EE processes where excessive writes to the
array after an enormous number of write cycles(1,000,000 or
more) can potentially alter the state of a neighboring cell."


The key is that last sentence. They are worried about data
corruption of neighboring cells, if you write to one cell
a large number of times.

-----------------
There is also a separate issue of eeprom endurance. The eeprom
will lose its data after many years, unless it's refreshed by
being read and re-written. Go to the link below and read
the articles on NonVolatile Memory.

</font>
___________________________
This message was ported from CCS's old forum
Original Post ID: 13412

[Steve H]

___________________________
This message was ported from CCS's old forum
Original Post ID: 13416

[Steve H]

It also occurs to me that with everything using Flash now and that these endurance equations are all extrapolated that none of our equipment will last > 15 years! We’ll be lucky to get to the extrapolated value.

I have lot’s of equipment that I want to survive for a long time – I guess that won’t happen.

Is this the 'real' Y2K problem?

Steve H.
___________________________
This message was ported from CCS's old forum
Original Post ID: 13417

[Sherpa Doug]

:=It also occurs to me that with everything using Flash now and that these endurance equations are all extrapolated that none of our equipment will last > 15 years! We’ll be lucky to get to the extrapolated value.
:=
:=I have lot’s of equipment that I want to survive for a long time – I guess that won’t happen.
:=
:=Is this the 'real' Y2K problem?
:=
:=Steve H.

I worry about this a lot. On my bench at the moment is my trusty HP1631A logic analyzer which is about 20 years old. Does it have EPROMs in it, and if so how much longer will they last? Should I pull out all the EPROMs and make archive copies?

My analog scope is even older but it has no ROM. It is "fully transistorized" though ;-)

___________________________
This message was ported from CCS's old forum
Original Post ID: 13419