Hello MSP430 Team,
I have a customer who is experiencing some issues during their testing for voltage transient tolerance. The application is a simple system monitor for a master micro controller. The MSP receives a SPI message from the master on a regular interval, and in return it feeds the master a 200 Hz safety clock to verify its own functionality. There is no external crystal on the MSP, it runs off of the internal DCO.
Currently their MSP comes back after a voltage dip in some indeterminate state. The output clock generated by the MSP is operating properly, but the MSP is unresponsive to messages over the SPI bus, and the WDT is not resetting the MSP after the expected timeout.
A temporary workaround has been tested where there is a slightly longer wait for the DCO to stabilize on startup, with the assumption being that the MSP core is being reset after a voltage transient, and re-initializing all the peripherals in the main loop. But there is still concern that they haven't addressed the root of the problem so the issue could come up again.
One of the questions that keeps coming up is what is the expected behavior of the MSP when VCC drops below the minimum required voltage for the system frequency (figure 1 in datasheet), and the actual brown out reset voltage (~1.4V)? Is there anything we can do to detect or help the MSP get out of this weird state after voltage has recovered?