Dummy said:
The EEPROM will be programmed in factory before shipping out to
customer.
Everytime when radio is turned on, checksum will be verified. Checksum
error will occur when any bytes are corrupted in the EEPROM. If data
corrupted during radio ON, any checksum error won't be detected until
the next radio turned OFF and ON cycle.
The corrupted bytes are at random EEPROM address.
Some of the parts could be recovered after re-programming while some
could not. For those parts which damaged permanently, failure analysis
showed cell overwritten. Trying to inject some noises to EEPROM data
or supply line while performing write operation could cause checksum
error.
You say they are preprogrammed, but this implies that you are writing
them during normal operation. Which is it?
But all the voltages supplied to EEPROM are clean when in
normal use.
^ Famous last words.
The filter and regulator have taken care of the noises. So
it's not right to point to the noise as the culprit.
Most of the radios failed after being in the field from 6 months to 2
years.
If the eeproms aren't being reprogrammed in the field during normal use,
then a software error is unlikely, unless the magic write sequence is
stumbled upon during a freak crash. If they *are* being reprogrammed
(ie, you are saving some value when the user retunes the radio) then
I'll again say software. I'm telling you, lock out those interrupts!
The other possibility is a bad batch of eeproms. This is fairly
unlikely, but not without precedent*. Attempt to correlate the bad ones
with some lot. Talk to the manufacturer, and ensure that they don't have
a 'known' problem. Also, I wouldn't reuse the corrupted ones just
because you managed to program them. I'd swap them out as soon as practical.
* A company I used to work for decided to save 10 cents a ram chip and
forgo individual testing of the chips by the manufacturer. Sadly, it
turned out that those chips were bad 5 to 10 percent of the time. They
were selling high availability purple ethernet switches for hundreds of
thousands of dollars each. The engineer responsible was of course
promoted to VP, and given vast new responsibilites.
--
Regards,
Robert Monsen
"Your Highness, I have no need of this hypothesis."
- Pierre Laplace (1749-1827), to Napoleon,
on why his works on celestial mechanics make no mention of God.