• State TI Thinks Resolved
  • Locked Locked
  • Replies 9 replies
  • Answers 1 answer
  • Subscribers 66 subscribers
  • Views 897 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

TPS544B25: CML error found after power up

user3317283
Prodigy 100 points
Part Number: TPS544B25

we found TPS544B25 has CML error after power up, TPS544B25 drive the SMBus alert signal to low. The CML error shows "Invalid Cmd" and it can be cleared. I wonder whether there is same error on TPS544B25 demo board, then we can be sure our design is OK. Do you have TPS544B25 demo board to check the status? And have you ever met similar issue before?

Image preview

  • 0
    TI__Expert 4795 points

    Hi

    Let me check with my team member and feedback to you.

  • 0
    TI__Guru 54900 points

     

    Are you seeing this after launching the TI FUSION Digital Power Designer software?

    It is likely that you are seeing a manifestation of the FUSION's "Address Scan" process interacting with the TPS544B25.

    In order to identify connected devices that use different methods of digital discovery, the FUSION software "scans" the connected bus by reading from several I2C registers / PMBus command codes at each PMBus Address.  The TPS544B25 uses the standard IC_DEVICE_ID command code, but does not support the MFR_SPECIFIC_FD (Command Code FDh) and MFR_SPECIFIC_FE (Command Code FEh) commands that FUSION uses in the Scan.

    You can confirm this is the route of the issue by opening the FUSION software without the TPS544B25 connected.  In the "Device Not Detected" window, you can select "Change Device Scanning Options"

    This will bring up the "Scan Editor" which allows you to set the scanning options used for each device address.

    You can change all of the addresses to use IC_DEVICE_ID (Command Code ADh) by selecting "IC_DEVICE_ID" on the top line, or you can update the specific address you are using to only use IC_DEVICE_ID, by selecting the pull-down menu and picking the IC_DEVICE_ID option

    If that eliminates the Invalid Command CML error you are seeing initially, it is a valid warning where the device is informing the system that it rejected a transaction because access to an unsupported command was requested as part of the scan.

    If you are not seeing this from the FUSION interface tool during scan, can you explain the exact process you are following to see SMB_ALERT low and the invalid command bit set in STATUS_CML?

  • 0 in reply to Peter James Miller
    Prodigy 100 points

    Thanks for the information, it is helpful. I will check it.

  • 0 in reply to user3317283
    Prodigy 100 points

    We use FUSION Digital Power Designer software. Initially TPS544B25 have CML(Other Comms Fault, Invalid Cmd) fault. We follow your steps to set the setting and still have CML fault but only Other Comms Fault. The CML fault will repeat after press “Clear Faults”. The TPS544B25 is connected with BMC, not sure if BMC is using any command TPS544B25 doesn't support.

  • 0 in reply to user3317283
    TI__Expert 7500 points

    Hi,

    We are looking into this, will feedback to you soon.

    Thanks,

    Lishuang

  • 0 in reply to Lishuang Zhao
    Prodigy 100 points

    one update, if we don't use Fusion software, we use BMC command to read out the fault register 0x79 and 0x7e, they are 0, but the alert signal is low. Then I use BMC command to clear the fault register (command 0x03), the alert signal is high now. It is strange that fault register is 0 but TPS544B25 still drive the I2C alert to low. We need clear the fault register every power cycle. We are OK with this work around solution, we can close this case if TI don't want to continue debug.

  • 0 in reply to user3317283
    TI__Guru 54900 points

      

    0x79 is reporting 0000h (16-bit word, all 0s) but the ALERT line is pulled low until you send a CLEAR_FAULT?

  • 0 in reply to Peter James Miller
    Prodigy 100 points

    Yes. And 0x79 is reporting 0000h (16-bit word, all 0s).

  • 0 in reply to user3317283
    TI__Guru 54900 points

     

    Thank you for the detail.  I am going to look into that.

    Can you tell me how VSET is terminated?

    How us SYNC/RESET_B connected?

    Additionally, are you able to read the SMBALERT_MASK for STATUS_WORD?  This requires using the Write Block - Read Block process call to read out the MASK on the upper-byte of status WORD.  While the PGOOD_Z bit in STATUS_WORD is masked by default, if this MASK bit is cleared, a PGOOD low condition will assert SMB_ALERT.

    To read the ALERT_MASK for STATUS_WORD:

    Start // Device Address + Wr // Command Code for SMB_ALERT_MASK (B1h) // Command Code for STATUS_WORD (79h) // Repeated Start to change direction of transaction // Device Address + Rd // Block Size = 01h // Mask Data // Stop