DP83822I: Inquiry about intermittent connection failure during initial EtherCAT connection.

Part Number: DP83822I

Tool/software:

Hello,

I am using Ethernet products (DP83822IHRBR) to operate EtherCAT.

Out of the two EtherCAT devices, one intermittently fails to establish an initial connection when repeatedly turned off and on.

(There are no communication disruptions once connected.)

I wonder about the initial connection sequence and the reset timing during initial connection. 

Could you also provide guidance on what to check regarding the intermittent connection issues?

Please provide an answer.

Thank you.

  • Hi,

    I wanted to clarify some details about your setup. 

    • What speeds are you running?
    • What is the link partner in the test setup?
    • What exactly do you mean by "repeatedly turned off and on"? Is this referring to power-cycling the PHY multiple times in a short span? Why is this being done?
    • Is this referring to the link status which is intermittently failing to establish?

    Best,

    Vivaan

  • Hi,

    The EtherCAT is set to 100Mbps.

    The link partner communicates with the EtherCAT master (another integrated controller).

    One day, after turning off and on the controller to download a software update, the EtherCAT link failed to connect. Since then, I have been repeatedly turning the controller on and off to test the link connection (the controller contains an EtherCAT communication circuit). The controller has two EtherCATs, EtherCAT0 and EtherCAT1. One connects the link well without any issues when turned on and off.

    Thank you,

  • Hi,

    Does the link partner also use 822 PHY?

    Were both PHYs working correctly before the software update? What was the software update for? 

    Can you share the register dump of both PHYs in this state?

    Best,

    Vivaan

  • Hi,

    I am not sure if the link partner uses 822 PHY. It might be a different PHY.

    The software is for the operation of our controller. Even before the update, one PHY occasionally had connection issues.

    Could you explain the method for a register dump? It might be difficult to share it now, but I will do so if possible.

    When the initial PHY HW reset time was slightly extended, the intermittent disconnection seemed to decrease somewhat.

    Is there a minimum or maximum time for the initial reset time?

    Thank you.

  • Hi,

    Could you explain the method for a register dump?

    This depends on how you are accessing the registers for the PHY. Are you using a linux kernel to access the memory, it can be done using the ethtool command ethtool -d <ethernet interface>, where <ethernet interface> is the interface where the PHY is connected, such as eth0. 

    Another thing I would like to try is simply switching the PHYs by swapping the cable connections from the link partner. Then we can see if this issue follows the link partner or the PHY itself

    Is there a minimum or maximum time for the initial reset time?

    I wanted to clarify this point. Is this referring to a hardware reset performed after power-up? For a regular reset, the minimum time is 10us. We must also make sure that a stable clock is present before we initiate a reset. 

    Could you describe the power-up sequence for the PHY?

    Best,

    Vivaan

  • Hi.

    It seems that I have replied too late. I apologize for that. Also, Happy New Year to you.

    Currently, while testing, I found that when an external communication connection is made, the Clock does not work, causing links to fail to connect.

    I am continuously testing this issue.

    Thank you for your responses so far.

    Best regards,

  • Hello,

    Thank you for your query. Vivaan is OoO and will be expected to respond to your query early next week.

    Sincerely,

    Gerome

  • Hi,

    Happy new year to you as well!

    Which clock are you referring to in this case? An inactive clock does cause link issues. 

    Best,

    Vivaan