Lpc2000

Leon,

don't forget to erase the cell between 2 write cycles ;-)  Seriously,
you could assign e.g. 512 bytes as EEPROM simulation cells and write
sequentially. Unfortunately I've been told that you need to erase the
cells between several consecutive writes into the same 512 byte page
of the LPC2106. To be more exact, writing 2-3 times seems OK but doing
more than 5 write cycles without erase generates reliability problems.
As I understand, the erase is like a re-calibration of the flash cells.

Hope this helps, Bob

--- In lpc2100@yahoogroups.com, "leon_heller" <leon_heller@h...> wrote:
> --- In lpc2100@yahoogroups.com, "lpc2100_fan" <lpc2100_fan@y...> wrote:
> > Hi,
> > 
> > as usual, Philips specification is rather conservative. Under very
> > simple lab conditions in an automated test we did not see any failures
> > within 500,000 cycles. This was just done as a sanity check. So under
> > "normal" conditions, which is that not all reprogramming cycles will
> > happen at maximum temperature, the number of reprogramming cycles will
> > be 1 or 2 orders of magnitude higher than the specified 10,000 cycles.
> > I guess this is where the difference between the marketing statement
> > "our device can do 1 Mio cycles" and the worst case specification "we
> > guarantee (under worst case conditions) 10k cycles" comes into the
> game. 
> 
> 
> Something I've thought of doing so that I can continuously update
> values in non-volatile memory like Flash or EEPROM with a limited
> lifetime is to keep writing to the same location and testing until
> that location fails, then moving on to the next location. I'm not sure
> how reliable this technique would be, though.
> 
> Leon