Flex Memory Endurance calculator

[stefan_nagel]

Dear NXP-Team,

I have difficulties to understand the calculations of the endurance of EEPROM.

First we have the table from the data sheet:

stefan_nagel_0-1612278442404.png

Second I have the output of the calculator:

stefan_nagel_1-1612278498226.png

I configured the calculator as follows:

65536B of EEPROM Backup size, 4096B of EEPROM size

--> I assume this should lead to a FLEXRAM ratio of 16.

I further more configure 4096 records with 1 byte each. In my book this should lead to 100k write cycles as indicated in the table above. However ... it does not.

I dont understand the section from green to red at 12.500 flash cycles. The table above indicates a max of 1.000 flash cycles

I furthermore dont understand the relation between 100.000 update requirements lead to 25.000 flash cycles. In my understanding, it should lead to 100.000/16 (ratio) = 6.250 flash cycles.

Could you please clarify my wrong understanding?

Kind regards

Stefan

[SPaulus]

Hello Daniel,

an additional question to FlexNVM endurance: Does it matter in which arrangement data is stored in the FlexRAM / Eeprom?

To get a (logical) separation of different data areas, I created 1kByte blocks in the eeprom in the linker script and distributed the eeprom data accordingly. Since this data never has a size of 1kByte, free areas remain between the individual blocks. Does this affect the lifetime of the FlexNVM?
For a better unterstanding see attached file "eeprom_layout.png"

[danielmartynek]

Hello Sebastian,

It does not matter:

"EEE writes don’t have to be to sequential location, indeed user can write to any valid location and/or write to the same address more than once."

AN11983 Using the S32K1xx EEPROM Functionality

Section 3.6.1 EEE writes 

https://www.nxp.com/docs/en/application-note/AN11983.pdf

Regards,

Daniel

[SPaulus]

@danielmartynekThanks a lot!

[danielmartynek · Accepted solution]

Hello Stefan,

Please have a look at the footnotes in the table:

danielmartynek_0-1612369238675.png

The 100k specification is for 16-bit and 32-bit writes only.

This is because one 8-bit record take the same amount of the EEPROM backup flash as one 16-bit record.

Thus the 25000 flash cycles in you example.

It is 12500 in the example below that uses 32-bit writes only.

danielmartynek_1-1612369594203.png

Basically, the FlexNVM is more efficient in the Emulated EEPROM mode.

As the footnote 6 says, "For usage of any EEE driver other than the FlexMemory feature, the endurance spec will fall back to the specified
endurance value of the D-Flash specification (1K)."

Regards,

Daniel