--- In BoschDME@yahoogroups.com, "Weston Pawlowski" <WestonP22@...> wrote:
>
> The list at ElecAuto and other sites are based on the entire bin file's
checksum, not that 16-bit value written at the end of the chip.
>
> It would seem that the last two bytes of the chip are some sort of a checksum,
but I can't get them to match up to anything either. For example, I have one 4KB
program with a file checksum of 0x8D5A, but the value at the end of it is
0x8A00. When I subtract 0x8A and 0x00 from 0x8D5A, that's only 0x8CD0, so
evidently doing a checksum on the whole bin file is counting an area that's not
included in the 0x8A00 checksum. And I've had even bigger differences with other
4KB and 8KB chips, even if I do my checksum calculation based on only the top
4KB. I wonder if they're doing something other than a standard checksum when
they write that value at the very end of the chip. The ML1.2 chips appear to be
the same way.
>
>

That's interesting. I could not find a way to use the two last bytes as a
checksum. I guess it could be just a version number since it increases during
the development of the code. It is not used anyway.

The first 4k do have the number 0x0042. That is always the same in all versions.
The 4k eproms use the first 4k in the 8051 rom.

Your binary with 0x008A for the second 4k block points to version 1267355191 of
a 944.618.121.04 box.