AM5728: Power Supply Issue

As stated in the data manual, the recommended maximum voltage is 1.89V.  The absolute max is 2.0V, so at this point damage could occur.  I do not recommend operating the device above the recommended max voltage (1.89V). 

I suggest trying to locate the cause of the problem.  If the PMIC is correctly programmed, it should NOT be generating 2.0V.  Double check the PMIC part number to ensure the correct device is installed.  (There should be a ID and/or Revision that can be read out of the device).  I believe the other E2E thread is suggesting possibly another power rail on the AM5728 device is not correct, and that voltage is 'leaking' from one power domain onto VDDA domain (source by LDO4).  This leaking would likely be inside the AM5728 device - so I would check to make sure all other processor power rails are correct and sourced in the correct sequence.   Also not the AM5728's IO are not fail-safe, meaning voltage should not be applied to an IO before power is applied to the VDDSHV power rail.  This too can cause 'leakage' - but typically goes away once power is applied to the device.

I am confused now,  is it the PMIC's issue or the processor's issue? can i use TPS659037 and AM5728?  

i did not programme it

Yes - TPS659037 can be used with AM5728, but TPS659037 is not a complete part number.  The PMIC has different configurations for supporting different power solutions, and each should have a documented complete part number.  The few TPS659037x PMICs I am aware all support 1.8V on LDO4, but there may be some configurations that support something else.  I recommended double checking the PMIC complete part number and confirming it is setup for 1.8v on LDO4.

As mentioned, it is possible the voltage is from leakage through the AM57 device.  I believe that is what was suggested in the other E2E thread.  If another rail is powered incorrectly or at the wrong time...or if voltage is applied to and IO before the rail is powered, it could cause an incorrect voltage to show up on other rails.

Hi Robert,

1.  the complete part number is TPS6590379ZWSR,  LDO4 is 1.8V output.

2. "If another rail is powered incorrectly or at the wrong time...or if voltage is applied to and IO before the rail is powered, it could cause an incorrect voltage to show up on other rails."---------there is only a TPS6590379ZWSR in the board.

With the correct PMIC part number, then it is generating the correct voltage (1.8V).  I understand there is only one PMIC on the design - but there are other power supplies, yes?  Are those supplies providing power to the processor's VDDSHV rails?  are those supplies powering other peripherals - and those peripherals are driving their outputs into processor prior its VDDSHV IO rails being enabled?   Specifically - look a the IO that are suppled by VDDSHV10 (since that is the rails with the issue).  What peripherals do those IO connect?  Are those peripherals all use only 1.8V IO, or are some at other voltages?

1. I understand there is only one PMIC on the design----->  Yes, only one PMIC on the design.

2 but there are other power supplies, yes?----> No, there is only one PMIC on the design,  no other power supplies.

3. Are those supplies providing power to the processor's VDDSHV rails?---->  No, there is only one PMIC on the design,  no other power supplies.

4.What peripherals do those IO connect?---->there are floating now,  they will be connected with FPGA

5.Are those peripherals all use only 1.8V IO, or are some at other voltages?---> they are expected to be 1.8V, but now they are 2V in fact

No other power supplies - then what supply/voltage is sourcing the PMIC?  Is that voltage (source to the PMIC) directly connected to the processor?  The fact that the IO is also 2V (and not 1.8V) indicates there is likely a incorrect power supply connection somewhere.

Hi Robert,  Perhaps you get me wrong or I did not express it clearly.

The design's power supply follows the attached Picture absolutely.

The question  is that , LDO4_OUT is 2V(it is expected to be 1.8V),  SMPS1_SMPS2 is about 1.18V（a bit larger than 1.15V）,SMPS4_SMPS5 is about 1.09V（a bit larger than 1.06V）................

The power net VDDA_1V8_PHYB is supposed to be 1.8V, but rather is measuring 2V.  You say the design is absolutely same as the block diagram, but that is NOT correct as VDDA_1V8_PHYB also connects to VDDSHV10 - and that connection does not exist in the block diagram.  

I don't believe the issue is between the PMIC and the AM57 device, but possibly between the AM57 device and the rest of the system.