• State
  • Locked Locked
  • Replies 6 replies
  • Subscribers 34 subscribers
  • Views 107 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.

DP83826E: TI PHY DP83826E

noam n
Prodigy 20 points
Part Number: DP83826E

Tool/software:

Hello,

I would appreciate your help regarding the following issue.

 We use TI PHY DP83826E in 2 projects:

  1. First project which is based on Infineon XMC4800 worked so far with TI’s PHY: DP83826E.
  2. Second project which is based on TI F28P650DK worked so far with TI’s PHY: DP83826E.

In both projects, we recently got new boards with same PHY of type DP83826E, but different date codes.

With the new PHYs, we experience the following issue:

  1. Port Link acting as active high and not active low (as opposed to our previous boards).  
  2. We cannot establish ETHERCAT connection.

We can't establish a connection with an ECAT master such as Beckhoff controller.

 

In both projects, the only thing which is changed in the new boards, is the PHY date code.

When we replace the new PHY with old PHY – ECAT communication is ok.

 

The "old" PHY which is working properly has these details printed on it:

  • 826E

            TI 198

            ALSE G4

  

The "new" PHY which is not working properly for us:

  • 826E

            TI 218

            A2VN G4

 

  • 826E

            TI 168

            AEXF G4

 

 

I'd appreciate your assistance with this issue.

Are you aware with any functional modification between these PHYs?

  • 0
    TI__Genius 11786 points

    Hi Noam, 

    I also got the FAE's post on this issue. Having looked at this design, it looks like the LED1 is strapped to GND which means that the LED1 is meant to be configured as active high in enhanced mode. 

    When the new PHY is in old board, could you confirm that if the LED1 is active high or active low?
    Could you also see if the old PHY is active high or active low in the new board and can communicate with the ECAT controller?
    Lastly, could you check if the modeselect pin is pulled high or stays floating on both old and the new PHY in the old board and the new board if possible?

    Please let me know. 

    Best,
    J

  • 0
    Prodigy 20 points

    Hello, J.

    As Noam, my colleague, wrote we have two projects. 

    In the first project (let's call it "The First") we have NO LED connected to LED1 pin and despite this, we got different active levels of Link signal in Old and in New boards. As we wrote in the Old board the active level was "0", but in the New board the active level come be "1" without any hardware changes.

    In the second project (let's call it "The Second"), the situation is more strange. Here we have LED connected to the LED1 pin. The LED is connected by Anode to the pin 31 and Cathode to GND, so the expected active level is "1", correct?. In Old board it was as expected, but in the new assembled board with another date coded PHY, the active level come be "0". Again: without any other hardware changes.

    Modeselect pin in First project not connected. In the Second -  is pulled high.

  • 0 in reply to Sergey Kruglikov
    TI__Genius 11786 points

    Hi Sergey, 
    I understand the polarity changes. As far as I know, there has been no functional changes in the PHY, but I will internally check to ensure that there are no changes that I may not be aware of. 
    As I asked before, could you check on the following that I asked of Noam?

    J said:
    When the new PHY is in old board, could you confirm that if the LED1 is active high or active low?
    Could you also see if the old PHY is active high or active low in the new board and can communicate with the ECAT controller?
    Lastly, could you check if the modeselect pin is pulled high or stays floating on both old and the new PHY in the old board and the new board if possible?


    Based on your information, I have additional questions:
    The First:
    1. Is the PHY LED1 pin pulled to GND or VDD? If VDD, it is expected to be active low so this will be a PHY issue again. 
    2. Have you tried putting the old PHY in the new board and verify that the LED signals are coming out as expected?

    The Second:
    1. Have you tried putting the old PHY in the new board and verify that the LED signals are coming out as expected?
    2. Have you tried putting the new PHY in the old board and verify that the issue persists?

    For both projects:
    Could you get the register content of register 18h? This register holds the information of the link LED polarity. 

    Best,
    J

  • 0 in reply to J
    Prodigy 20 points

    Hi, J.

    Thenks for yuor effort and help.

    Here are my answers.

    1. When the new PHY is in old board, the LED1 is active high. The old PHY in the new board is active low and can properly comunication with ECAT controller. The "Modeselect" pin is float. We also try to connect it to VCCIO, but no any affect.
    2. The First:
      1. Pin LED1 pulled to GND by Strap resistor 1.5kOhm in both boards. LEDs no connected to the pin.
      2. Old PHY in new board is working properly.
    3. The Second:
      1. We didn't yet try to exchage PHYs.
    4.  Register 18h. I'll ask Noam to anwser.
  • 0 in reply to Sergey Kruglikov
    Prodigy 20 points

    Hi J.,

    For the 'first' project, which involves Infineon XMC4800 DSP, register 18h has the default value of 0x480 (bit 7 is high). 

    * I tried to set it also to 0x400 (bit 7 is low) - the register is changed, but it had no affect.

    * PHYSTS Register (Offset = 10h) value is 0x315 when power on.

    * If relevant, what should be values of MLEDCR (register 25h) and LEDCFG (register 460h)?

    Thanks again for your help.

  • 0 in reply to noam n
    TI__Genius 11786 points

    Hi Noam and Sergey, 

    To answer Noam's questions:
    LEDCFG should be the reset value of 5665h and MLEDCR does not matter for LED1. 

    I discussed this internally and seems like we had a rev change of the die a while ago that may impact the functionality of the PHY. I checked the schematic again, but it seems like both the old and the new PHY are the new revision since the strap configurations are up to the most recent configuration. However, could you do a read of register 3h of all PHYs to ensure that all the PHYs are the most recent one? If they are a different Rev, the revision ID will be different if the old and the new PHYs are different revision. 

    Also, does the new PHY cannot establish any link at all in the old board in the first project, or just LED is the issue? 

    Please let me know. 

    Best,
    J