In comp.arch.embedded,
Juergen <juergen.steger@bbraun.com> wrote:
>Hallo I'm using AT90S1200 at 3,3V
>The application read the content of the flash via the serial programming interface.
>Then the contents will be compared with an CRC BYTE .
>
>Sometimes we found that 3 Bytes of Flash Memory get lost.
>Address 00, 01, 200 HEX
>
>Is there any experience with corrupted F L A S H Memory ?
>
>We found no way to generate this behavior ?
>But it happens 1 per 100 units per month ???
We've had AT90S1200's completely loosing flash contents (blank check
OK). This seemed to happen after some time. We would produce and test
a batch and send them to our customer. Once there, they found some
dead units which where fine when they where shipped. As a test we
stocked a batch of units for a few weeks and did some tests every
week. I think we had about 1 in 100 with cleared flash after one week.
After re-programming they just worked fine. Never found the actual
problem and Atmel did not find any problems in our circuit as well.
Finally we switched to the AT90S2313 and never had the problem again.
All this was a few years ago and not knowing the actual reason still
bugs me. So if you come up with anything, please drop me an e-mail.
--
Stef (remove caps, dashes and .invalid from e-mail address to reply by mail)
Chemistry is applied theology.
-- Augustus Stanley Owsley III
Reply by Juergen●February 14, 20052005-02-14
Hallo I'm using AT90S1200 at 3,3V
The application read the content of the flash via the serial programming interface.
Then the contents will be compared with an CRC BYTE .
Sometimes we found that 3 Bytes of Flash Memory get lost.
Address 00, 01, 200 HEX
Is there any experience with corrupted F L A S H Memory ?
We found no way to generate this behavior ?
But it happens 1 per 100 units per month ???