• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Subscribers 62 subscribers
  • Views 794 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

UCD90320: alert setting

user6158403
Intellectual 280 points
Part Number: UCD90320


Hi

In my project, I use UCD90320.

Because BMC comes access to UCD at a frequency, only once per second. But a voltage into UCD change may be 1kHz.

Is UCD have a way to record which profile was triggered at that time and the value of its voltage, so that they save some in the register?

And then clear the register when the BMC read?

  • 0
    TI__Guru 67330 points

    Hi

    I assumed that you are talking about the profile feature here. each rail can have up to 4 profiles which are selected by two GPIs.

    The voltage change shall not cause the profile change. the profile changes only if the assigned GPIs are changed.

    The only thing I can think from UCD point view is to use GPI fault to log the GPI event. so that your host can know the given GPIs are changed

    But this only works when the GPI is changed from ASSERTED to DE_ASSERTED.

    Since these GPI are from external. do you have a way to read the GPI outside UCD?

    Regards

    Yihe

  • 0 in reply to yihe
    Intellectual 280 points

    Hi

    Sorry, maybe you misunderstood me.

    I use one rail and two GPI so rail has four profiles, then I will use LGPO to BMC when over warm or under warm and use BMC to read rail voltage.

    Because BMC comes to access to UCD at a frequency, only once per second. But a voltage into UCD change maybe 1kHz.

    Therefore, when the BMC reads the rail voltage, it will be inaccurate. 

    Is UCD have a way to record which profile was over warm or under warm at that time, so that rail voltage save in the register?

    And then clear the register when the BMC read?

  • 0 in reply to user6158403
    TI__Guru 67330 points

    Hello

    The profile index is not saved by device.

    If there is a fault, device will record the fault value into the eeprom and your host can read later. But it may not tell which profile.

    Why is profile so important ? If so, your host can query device by  PMBUs alert signal instead of regular timer.

    Regards

    Yihe

  • 0 in reply to yihe
    Intellectual 280 points

    Hi

    Fault value are mean over Fault and under fault? So over warm or under warm is not store eeprom?

    by the way, how can I use PMBUs to read fault value of eeprom and how do I distinguish over fault and under fault?

    How can I use PMBUs to read num profile?

    Thanks.

  • 0 in reply to user6158403
    TI__Guru 67330 points

    Hello

    Yes, only fault values are stored, warnings are not.

    Please refer below document for reading fault.http://www.ti.com/lit/pdf/slvaeg5

    Please read section 9.4 http://www.ti.com/lit/ug/slvuaw9b/slvuaw9b.pdf to read profile.

    Regards

    Yihe

  • 0 in reply to yihe
    Intellectual 280 points

    Hi 

    Your information is very helpful for me.

    I want to check  "Fault #1" is mean num profile #1?

    If so, is "Fault #1" include your return data ??? 

    At last, is the return data overwrite every time? or just record first data and wait for clear then continue to write?

  • 0 in reply to user6158403
    TI__Guru 67330 points

    Hello

    Fault #1 means this is the first fault device detect.  It has nothing to do with profile. you shall decode the fault to understand where the fault from.

    The fault log could be overwritten if you enable the FIFO mode.

    Regards

    Yihe

  • 0 in reply to yihe
    Intellectual 280 points

    Hi Yihe,

    Thanks, is very helpful for me.