EEPROM & Flash lifespans

[Dr_Colloid]

We are working on a new version of our instrument that will use a Freescale MC9S12DP512 MCU. In the Device User Guide, there is a table indicating that over the operating range of -40C to 125C, the NVM Reliability Characteristics, which we aren't sure if we're even interpreting correctly are:

Flash: Data Retention of 15 years if the memory is cycled 10 times.
EEPROM: Data Retention of 15 years if only cycled 10 times, dropping to 5 years if the EEPROM is cycled 10000 times.

Out instruments have had useful lives of 20-25 years, so a 15 year life is worrying. Further, we are saving use settings in EEPROM that can change daily, and a 5 year life there is even more of a problem. On the other hand, the instrument is used in a laboratory setting over comfortable room temperatures, not at the extremes of 125C and -40C. Is there more detailed information on this subject available somewhere?

Thanks.

[Nabla69]

Dear Dr. C,

You have special Engineering Bulletins describing how Freescale came up with these numbers:

EB618-TypicalDataRetentionForNVMem.pdf

EB619-TypicalEnduranceOfNVMem.pdf

If you work at room temperature, the cycling before failure was measured but NOT guaranteed at several millions program/erase cycles.

I imagine your instruments don't spend their time modifying the EEPROM.

In this case I am totally confident in a 25 years life.

Freescale Flash/EEPROM specifications might sound worrying compared to competitors... But Freescale does guarantee the values across the FULL temperature RANGE. Not all of them do !

Cheers,

Alvin.