Please refer to our application note that details on the design considerations for EFT.
Let us know if anything is not clear or you need further help!
Thanks for the information.
I have placed 2 ferrite beads in the communication lines of an external device, and passes the test successfully.
Even so, I can not understand there is a situation in which the microprocessor is properly powered and stays "frozen" by a noise and do not "awake" via internal watchdog or similar...
There could be two reasons why this happens:
1. Latch up: Latch-Up is a situation where the transient-induced noise does not actually do the damage. It just sets things up so that the power supply can destroy the part or the circuit becomes nonfunctional without a power cycle reset. The ground bounce or shifting of the ground reference due to transient-induced noise can drive CMOS circuits into latch up. Specifically, it is the creation of a low-impedance path between the power supply rails of a CMOS circuit, triggering a parasitic current carrying path which disrupts proper functioning of the device. A power cycle is required to correct this situation. Latch-up can lead to the destruction of the device due to over current.
2. Memory Corruption: Transient-induced noise can damage memories such as flash or RAM due to them interfering with system clocks or flash write voltages. When the memory gets corrupted, the system can fail to start due to a flash checksum error or can lose the functionality due to corrupted data or code in flash or RAM. Flash corruption might be permanent or might require power cycle or reprogramming to recover the normal state. On the other hand, RAM corruption might require a power cycle or any other reset to resume normal operation.