On Thu, Oct 31, 2002 at 09:02:53AM +0100, David Brown wrote:
> I think flash write errors are going to be a very rare case, and if you *do*
> get an error, it doesn't really matter at what time you see the error
> message, since you will probably have to replace the chip anyway. So I
> would think the priority should be to improve the speed while everything is
> working correctly.
Yes, I think this is the way to go. Do you really have so many FLASH-write
failures? Then there must be another problem (bug) that should be
fixed. I think FLASH write failures only occur with low supply voltages
or heavy used FLASH (more than 100000 writes). The latter a very rare case.
And for the first - you can start with a short write (to detect that FLASH-
write is possible at all) and then write with large block sizes.
(Newer MSP-parts with brown out circuit can detect that the supply level
is ok for flashing).
Matthias
