I asked this question before, with no answer. I will try and reformulate it more clearly and with more details.
We are experiencing a strange and so far unexplained behaviour with TPS564208.
Our customer uses it to supply a Raspberry PI 3 and a 7" LCD with a 5V, 2.5A. Input voltage is either 12V from external supply, or 7.4V from a battery.
As an emergency measure, in case the RP3 hangs, he put a "reset" microswitch in series with the supply, to disconnect/reconnect both devices (I know... using the ENABLE pin would be more logical).
In several cases, this "reset" destroys the TPS and not always in the same way: short circuit between SW and VIN, or between SW and GND, or SW floating near 0V.
We performed many tests attempting to reproduce the problem, with no success. Among the tests we simulated the "reset" procedure cutting an external 2 Ohm load with a relay, but after 500 cycles the TPS was still very alive.
We also tried with inductive and capacitive loads, or short-circuiting the output... and TPS was still alive.
The external supply is (in my opinion) underpowered. When the loads are connected, VIN may drop a few Volt. The input voltage is still enough to produce 5V output, yet I wonder if such a suddenly drop can cause problems.
Can you suggest something?