I seem to be up against a brick wall. I have a design using a TPS65262-1
Input 12 V
Out1 5V
Out 2 3.3V
Out 3 2.5 V
Lin 1 1.8 V
Lin 2 2.2 V
The 5 Volt is enabled immediately (if input is above a set minimum). The others come up after a derived 1.8 V as part of a power sequencing for a Zynq.
There are two first article PCBs in existence. Between the two, we have blown four of these ICs, and I don't have a clue as to why. The boards have been in use for a month or two in development labs, so its not like they always blow or have a design error. When the first one blew, I blamed the lab supply that was providing the 12 V. It is known to generate an overvoltage spike sometimes when the AC switch is shut off. But the next two failures weren't even using that supply, and the most recent failure occurred when the DC was first connected, the PS wasn't being turned on or off. As an added protection, a 12 V TVS was placed across the input before the most recent failure occurred.
The failures are all the same. When the 12 V power is applied, the circuit only draws about 20 MA. V7V has the expected 6.3 V on it, but all five of the outputs are 0 or only a few mv. Normal power up draws about 200 ma from the 12 V supply and generates the voltages mentioned above.
The exact extent and nature of the failure isn't known, because without the 5 V output (out 1) the secondary regulator has no input, and without it generating 1.8 V, the other outputs wouldn't be enabled. So it is possible Out 2 and 3 and the linear regulators are fine, but it is known for sure that Out 1 is not functioning.
I am at a total loss as to what is causing these occasional failures. They occur either during the previous power down, during unpowered handling, or at the moment of power up. The last one occurred with nothing else attached to the board, except an SMA cable delivering a 7500 Hz sine wave (1 V p-p) to another part of the board. I just don't have a model for overvoltage, stray voltage, an ESD issue or anything else. The input has 30 uf of X7R capacitance and a TVS diode across it, which should be robust against just about anything less than a connection to AC power.
Are there any known vulnerabilities I should be designing around or protecting against? Is there some way the secondary SMPS could spike the 5 V output and damage the FETS? (there is no other voltage source, and the 5V line has at least 75 uf of ceramic capacitors).
Here is the schematic of how the IC is used. Not shown is the 12 V input coming in the upper left corner, and an SMBJ12A TVS diode from there to ground.