TPS65911: VIO IS WAY TOO HIGH AND CAN'T CHANGE IT

Part Number: TPS65911

I am using the TPS6591133 to power a DM385. I am only using the VIO to power the DVDD_C, VDDA_1P8, VDDA_PCIE_1P8, VDDS_OSC0_1P8, VDDS_OSC1_1P8 of the DM385 and the DVDD of a TLV320AIC3101 audio CODEC. I have verified before getting the board booted that the VIO was 1.8V

I have had 2 boards now have the VIO go up to 3.5V even though the register setting was for 1.8V. When we tried to change it the only setting that works is 3.3V or off. Any of the other settings do nothing an the output stays at 3.5V.

I removed the output EMI filter to isolate it and see if it was the PMIC or something driving it. The output went to 4V with no load so I put a 100 OHM resistor in for a load and it still was 4V. All the other voltages are correct and can be changed.

What would cause this to happen? Any help or ideas would be great.

35 Replies

  • Brain,

    It looks like the feedback trace is not connected correctly or some external rail is driving the output high. It would be helpful if you can provide schematic and switch node and output ripple waveform before and after bootup to help debug the issue.

    Jay
  • In reply to Jayanth Kruttiventi:

    The VIO outputs the correct voltage at first but somehow along the way it goes up to the 3.5V. I have confirmed the feedback is correct on the board layout. For the schematic I copied the DM388 Camera Start Kit eval board that uses the TPS6591133. Literature number SPRUIC7. I will post the picture of the startup later.

    After I made this post I checked the startup sequence and found that it does not startup in the correct order even though I have the Boot pin tied to VRTC. Here is the boot order:

    Note: Each boot sequence is 2ms from each other unless noted on the right

    VRTC
    VIO 4ms after VRTC
    LDO8
    LDO4
    LDO3&6 4ms after LDO4
    GPIO0
    LDO5&7
    GPIO7 1.75ms after LDO5&7
    LDO1&2 2ms after LDO5&7
    VDDCRTL 2.5ms after LDO1&2
    VDD1 1.5ms after VDDCTRL
    VDD2
    NRESPWRON 4ms after VDD2
  • In reply to Brian Forster:

    I got the document wrong. Its the DM388 CSK Document Number MS_TI_TMDXCSK388_SCH_REVA. That is the schematic I used to design my board.
  • In reply to Brian Forster:

    Can you please send me scope shot of VIO switch node and vout ripple. Also can you disconnect the load from VOUT and check if VOUT voltage changes. Also please confirm if VIO register value does not change when output voltage change from 1.8V to 3.5V.

    Regards,
    Jay
  • In reply to Jayanth Kruttiventi:

    I have the screenshots but I can't seem to find a way to attach them to this post. Do you have an email I could send them to?

    When I disconnect it from the load it goes to about 4V regardless of register setting except the 3.3V setting.

    The register in the PMIC is set for 1.8V. In fact the only register setting that works is the 3.3V setting. All others just output the 3.5V
  • In reply to Brian Forster:

    Brian, if you select "Use rich formatting", it will provide a mechanism to attach files.

    Brandon

     

  • In reply to Brian Forster:

    Brain,

    Can you please check when VCC6 voltage comes up. In the sequence you have sent, it looks like VCC6(3.3V) is high only after GPIO7 is high which is wrong. This would cause LDO1 and LDO2 to power up wrong.

    Also can you disconnect the resistor between TPS_VDDA_1V8 net and VDDA_1V8 to check if the votlage is still shoots to 3.5V. Also, I would try to change voltages on other rails to check for feedback path between rails.

    Regards,

    Jay

  • In reply to Jayanth Kruttiventi:

    Jayanth,

    Good call on the VCC6. I am feeding it with the 3.3V that GPIO7 is turning on which explains the sequence mix up. We are trying to hack in a fix for this to see what else it might fix but I think the raised VIO voltage is a separate issue. I will post the results if we can get the board to work after our fix.

    I have an EMI filter that is directly on the output of VIO that I have removed to check VIO when disconnected from the circuit. VIO actually rises to 4V when removed from the circuit. 

    We have successfully changed VDDctrl to 1.35V. Do you need me to try a different one? 

    Thank you,