When using the MSP430F5438A chip, the customer often experience program run-off. The following test is set to see if the flash is abnormal:
Testing:
Testing with a new chip, read the flash of the chip before burning the program, and read it again immediately after burning the program, and then when the program runs abnormally, read flash again, then save to 3 folders respectively.
When reading out immediately after burning a program, the contents are consistent, whether it is read immediately after burning or run for a while. The read is different until a problem is identified.
There is no actual flash in the program, neither read nor write.
Issue Analysis:
Comparing the files read before burning and the files read immediately after burning, the header and tail are identical, with only the program memory areas 0x5c00 to 0x2e800 not identical. Indicates that the burn process is normal and that there are no areas that should not be erased.
Comparing the files read after the fault, it was found that there was a lot of inconsistency from 0x1100 onwards, 0x1000 to 0x17ff should be the BSL flash region, and under the normal condition could not be overwritten. The user program area has been overwritten with a lot of blocks in the 0x5c00 to 0x45bff areas being erased as 0xff.
The customer would like to know the following 2 points:
1. what's the possible reason for the fault?
2. Is it necessary to see the file read out of Flash? If needed, could you please provide the route of transmission?
Thanks!
Best Regards,
Cherry Zhou
