Posted on December 20, 2011 at 13:05Hi,
We are using the STR911FAWX6 Rev H device on one of our products that has been in manufacture now for around 2 years. We have shipped hundreds of these without a problem.
However, on the latest build the CPUs were sourced separately and although they are clearly marked as rev H, they do have different a batch code on them from the previous ones we used. The problem is, the exact same software fails 100% of the time with this CPU.
I narrowed this down to the routine that writes to FLASH. We run the software from bank 0 and use bank 1 to store configuration settings and so on. Our firmware has the ability to erase and reprogram bank 1 as necessary. With the latest CPUs, if I prevent the software from writing to FLASH, it runs perfectly, but with the write included the CPU falls over in such a state that even with the debugger connected the only option to recover is a hard reset.
So my question is, has anyone else noticed any difference in FMI operation between different batches of the same Rev H device?
Many thanks
Michael
