Inconsistency in I2C based eeprom write operation

[GoogleCodeExporter]

It has been observed on few boards that the write operation on the AT24C256
(256kB Atmel Flash) is inconsistent. The read-operation seems to work fine.
The exact technical explanation for this behavior is still not known. We
are working on a solution for this issue and are hopeful that we will be
able to resolve it as early and possible and also suggest some work around
for the existing customers.

It has now been documented in the UserManualV1.2 for the Blueboard, while
we continue to look into this issue, it would be really nice if someone in
the group with sufficient I2C experience throws some light in this matter.

Problem details and other observations:

1. We have been using i2c_master_read(), function to check the presence of
I2C eeprom. The theory is that if this function returns '1', then it
implies that the eeprom is detected.

Observations:
-------------
When the SLAVE address is changed in the firmware; the i2c_master_read()
function returns error; i.e. if the address is changed from 0xA0 to 0xB0
the function returns error.

2. A proper test for I2C eeprom would be to write a known set of data and
then read and confirm if the data matches. This test seems to fail of few
random boards.

3. On some boards the read/write operation works fine.

4. Some observations have led us to conclude that the Atmel part might have
been faulty; but its not very conclusive.

5. The I2C line pull ups are 2.2K, changing them to 10K has had no positive
impact.

6. The voltages measured across all the pins on the eeprom seem to be fine.

-Ashwin

Original issue reported on code.google.com by Ashwin.A...@gmail.com on 2 Nov 2009 at 6:20

[GoogleCodeExporter]

After few sleepless nights, finally resolved this issue :).... Its a little 
strange
that none of the customers/members have reported this yet.

Update to the revision 42 for this fix

Original comment by Ashwin.A...@gmail.com on 5 Nov 2009 at 11:56

• Changed state: Fixed