This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

LM5064: 48V after hotswap on/off frequently when 12V power on

Part Number: LM5064
Other Parts Discussed in Thread: LM5034, , UCD9090

Hello, TI experts:

we are from Nokia China team, and encountered an power circuit issue related LM5064 and LM5034.

I provide the detail information about this issue in attaced document, would you pls help us review the schematic in attachment and answer below questions:

how can we do to disable LM5064 detecting the voltage after MOSFET, and make LM5064 always turn on MOSFET?

Is there any tool could monitor LM5064 inside registors or its status online? Just like UCD9090’s emulator.

Thanks in advance

12V0_PowerRail_issue.docx

MOHA_X21_0209.pdf

  • Will look into this tomorrow, US time.

    Brian
  • Let us add something of the issue:

    No matter whether there is any fault, The value of register, E1h is always 0x0888, which should be default 0x08e0 according to the datasheet.

    When we write 0xff  in D7h to disable all protections,  LM5064 can also be closed by some fault.

    We think the status of these registers are not correct.  Would you like to check  whether they are the flaw in the chip?

  • Suggest you work with the EVM stand alone. Not sure what is happening or if it is tied to the system itself. load PI Commander from the product page.

    Brian
  • Hello, TI experts:
    We found that Vauxh is very sensitive and seems to need to be highly consistent with Vcc. There is a problem
    1. Is there a range between Vauxh and Vcc in LM5064?
    VAUXH is not normal, especially if VAUXH and VCC levels are somewhat inconsistent.
    1. Whether it will affect the stability of VDD.
    2. Will affect the fault inside registers.
  • Vauxh measures Vin for the ADC vor Vin telemetry. It isn't part of Vdd generation. Vin provides bias to the Vdd regulator. Normally Vauxh is tied to Vcc directly so there would be no voltage difference. I believe the reason for the separte pin is so that an RC can be placed on Vauxh that could not be done on Vcc due to bias current. This RC could be used to filter Vauxh if Vin is noisy. Voltage difference between Vauxh and Vcc should not trigger faults. Noise injection (not difference of voltage) may possibly trip a Vin uv or ov warn in E1h or 79h.

    Brian
  • When we read the register E1h of LM5064, it is always 0x0888 even no fault occur. But the default value should be 0x08e0 according to LM5064 datasheet.

    Would you like to check the status of this register E1h on your Demo board, and explain this phenomenon? Thanks

     

  • It is important that you work with the EVM using PI Commander GUI you can download from the product page. You can order one through your TI sales team. Recommend you review with the IC installed on it and then install the suspect IC to compare. The difference between 08E0 and 0888 = Vout UV warn is set, Input OC/PFET OP fault + Device off and Vin UV undervoltage faults are clear. Have your TI FAE involved.

    Brian
  • Brian,

     

    1, The difference between 08E0 and 0888 = Vout UV warn and Input OC/PFET OP fault are set,  Device off and Vin UV undervoltage faults are clear. Is it correct? And does it mean input OC/FET OP fault detected by LM5064 in our system board?

    Where to get the detailed description of E1 register bit?

    To my understanding, IIN OC/PFET OP FAULT = 1, we should see LM5064 to retry when RETRY# is tied to VEE, but actually, LM5064 works normally.

    2, Do you have internal test report on the 'Averaged VIN/IIN/PIN reading via PMBus commands'? We can't get a stable reading value even the power consumption of board is stable.

    Regards,

    Dawei