Tool/software:
Hi,
We have two designs that use the TPS6594-Q1 to power Xilinx FPGAs. On both of these designs, we have seen periodic failures where the BUCK output pins become shorted to ground. They measure 0.1 ohms after failure. After the failure a SC interrupt is raised for the failed regulator.
The failed TPS6594-Q1 is removed from the board, the part measure 0.1 between the SW_Bx pins and ground. The part is replaced and the design will function for some time until it happens again.
Some theories we have had is: 1) too high of a voltage spike on the input PVIN_Bx pins, or 2) a spike on the switching nodes, the SW_Bx pins.
On our first design that uses this part, we have only seen 3 failures on 6 boards that have been built over a period of almost 3 months.
On our second design, we have seen 12 failures in just over a month.
The second design only uses buck1 in single phase mode and is lighly loaded. Failures are always on buck1 since that is the only switcher in use.
The first design uses buck1+2, buck4 and buck 5. We have seen two failures on the combined buck1+2, only 1 phase fails. And a third failure on buck4.
Since the issue appears to happen on power down or power up, we ran a test were the device under test was power cycled repeatably, with a wait in between to allow for the voltage rails to decay to zero. The interrupt registers as well as the current draw of the bench top power suppler were monitored to determine that they were in range. A scope was set up to monitor the input and output rails,
Has anyone seen similar issues?
