We have recently experienced a failure in the field of the PMIC (TI P/N: TPS65217A) which powers a TI AM3352 Sitara CPU. The failure of the PMIC is not a complete failure. The part does operate but with very little heat (~35-40C) air temperature in our chassis the PMIC shuts down and turns off.
Over the past week we have been working to isolate and determine the root cause of the failure of the part. At this point we could use some assistance from TI as we believe we know what is causing the PMIC to shut down. LDO1, 1.8V, which is connected to the VRTC (real time clock) of the AM3352 goes out of spec and rises to almost 3.3V and the PMIC turns off.
In our testing we have monitored all the other output rails from the PMIC and have not observed any other voltage rail going out of spec. We have also measured the current draw from the PMIC voltage rails and all are within the operating specifications of the PMIC. Finally we measured the case temperature during our tests and the case temperature never exceeded the air temperature by more than 1 or 2 C.
|
PMIC Rail |
Min (mean) |
Max (mean) |
Mean (mean) |
PMIC Rec Operating Max |
|
VSYS |
187mA |
294mA |
225mA |
2500mA |
|
DCDC1 (1.8V) |
41.4mA |
267mA |
83.2mA |
1200mA |
|
DCDC2 (3.3V) |
24mA |
197mA |
108mA |
1200mA |
|
DCDC3 (1.1V) |
195mA |
431mA |
260mA |
1200mA |
We would like to know if TI can help us and provide some insight into what might cause LDO1 to be damaged or in this case partially crippled and temperature sensitive. Determining the root cause of the failure is of extreme importance to us in order know if there are any potentially design issues using the device. Current questions we have for TI are:
What stimulus to the PMIC could cause LDO1 to be damaged as we have observed. LDO1 regulates at room temperature but at 35-40C LDO1 goes out of regulation and the PMIC shuts down.
While we have reviewed the schematic page of the PMIC circuit, could TI also look at this circuit for accuracy. One discrepancy we noticed was that of C27 connected to pin 48 of TPS65217A, INT_LDO. Datasheet shows 100nF capacitor connected to this pin. We have a 10uF capacitor. Schematic sheet is attached to this email.
Can TI perform a root cause analysis of the device.
Is there any known errata for theTPS65217A? I did not find anything on TI’s website.
