Hi to all.
i'm using the flash programming in my project: from time to time
device adds some data to info memory. The algorithm of read/write
operation with segments is according to datasheet slau056e.pdf.
All devices were verifying out and there were no troubles with them.
But after a time some devices(for about 30) had fallen out.
Examination had shown that instead of the data there is just 0xff in
the flash(info memory). this problem occurs oftenly enough.
Supervisor is on therefore i'm not sure that supply causes it. During
all my experiments with the supply voltage not once error had
occured. Actually it seems very hard to initiate this error in all
the way.
besides supervisor, it was decided to use other hints in a project to
escape unexpected flash erasing. Each segment is written twice,
algorithm uses CRC to check the storing data that must be written,
each segment can be written only after the previous will be veryfied.
After all, the number of cases with errors was appreciably decreased,
but the problem isn't solved.

I'm using msp430f413.
surfing the blogs i saw that this problem is widespread. but i didn't
find no solving

does anyone can help me? Start a New Thread

Hi all,

Is not easy to detect the source of such kind of problems, and I'm not
sure if we can help you on this question, but we have some expertise in
this field, and I can confirm you that using similar techniques to store
data in flash we detected the same problem one year ago, but using other
msp430: series f149 and f148.

Some ideas that maybe can help you to trace the problem:

- First check that your flash algorithm is correct running in parallel
many msp devices with long erase/write flash cycles (100.000 cycles
could be enough).
- If there's no problem check the clock source select by your flash module
- Check if you have the interrupts disabled on the E/W process.
- Check the quality of your source of power. You say that you are using
a supervisor, but some picks are not detected.
- If your device is battery powered, check the max pulse current allowed.
- Another source of problems can be noise in the Vcc... try to use EMC
filters to protect the cpu.

Good night and good luck

Mario Comas
m...@gmail.com

goshakoshkin escribi
> Hi to all.
> i'm using the flash programming in my project: from time to time
> device adds some data to info memory. The algorithm of read/write
> operation with segments is according to datasheet slau056e.pdf.
> All devices were verifying out and there were no troubles with them.
> But after a time some devices(for about 30) had fallen out.
> Examination had shown that instead of the data there is just 0xff in
> the flash(info memory). this problem occurs oftenly enough.
> Supervisor is on therefore i'm not sure that supply causes it. During
> all my experiments with the supply voltage not once error had
> occured. Actually it seems very hard to initiate this error in all
> the way.
> besides supervisor, it was decided to use other hints in a project to
> escape unexpected flash erasing. Each segment is written twice,
> algorithm uses CRC to check the storing data that must be written,
> each segment can be written only after the previous will be veryfied.
> After all, the number of cases with errors was appreciably decreased,
> but the problem isn't solved.
>
> I'm using msp430f413.
> surfing the blogs i saw that this problem is widespread. but i didn't
> find no solving
>
> does anyone can help me?
>
>
> Yahoo! Groups Links
>
> Start a New Thread