Saving to EEPROM.....

[JAM2014]

Hi All,

I have a project that among other things derives latitude and longitude from a connected GPS. I'd like to store these values in EEPROM to speed up the booting process and/or eliminate the GPS requirement once an initial fix has been determined. I parse the latitude & longitude from the NMEA string, and then convert to floats. I'm wondering if there is a 'clean' way to store these values to EEPROM during this process?

Here is an example of how I get the latitude:

[Ttelmah]

If you look in the 'drivers' section of the compiler, there is a standard bit of
code, 'floatee.c', which shows how to write a float value to EEPROM. Basically
you store the four bytes that 'make up' the float. In the library for the
internal EEPROM (internal_eeprom.c), there are similar functions for the
internal EEPROM, and ones for the various sized integers as well.

However big caveat. What you should do, is read the value that is currently
stored, and only write a new value if the value has changed. Thing to
understand is that EEPROM has a limited 'write' life. Every time you perform
a write it uses a 'life' for the cell. This happens even if you write exactly the
same value. Hence adding a test, and not actually writing if the value
hasn't changed is worthwhile. The EEPROM has to be thought of rather like
a stone carving. Quite a lot of work to do and a lot of work to change. Not
something to be written frequently like the RAM.

So make your code perhaps look at the reading, and wait till it has been
stable for several minutes, and only then write it to the EEPROM. When it
actually writes check the already stored value and only perform the actual
write if this is different. This way you won't write when the unit is moving,
or if the value has yet to fully settle. Result very few actual 'writes' to the
EEPROM.

[JAM2014]

Hello Ttelmah,

Awesome! That is exactly what I need!

Yes, understand about writing too often to the EEPROM! BTW, the folks at Dangerous Prototypes did a project a few years ago to determine the ultimate life of the internal EEPROM on a PIC device. My recollection is that the real world experimental result was quite a bit more (2x or 3x) the manufacturers stated life!

Thanks,

Jack