• State
  • Locked Locked
  • Replies 8 replies
  • Subscribers 35 subscribers
  • Views 1057 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.

Original question:

DP83TC811S-Q1: unlink issue of DP83TC811S-Q1

DP83TC811S-Q1: dp83tc811s High packet loss rate

kevin.huang_FJXM
Expert 2925 points

Part Number: DP83TC811S-Q1

A few days ago, we used EC20 and PHY dp83tc811s to communicate with each other through SGMII interface, but there was a problem that Ethernet Ping could not work.

According to the investigation, the high packet loss rate is caused by the wiring harness and the real vehicle environment. After the packet loss rate is high, the Ping will not work after 30 minutes to - 1 hour.

 Conclusion

Because the signal quality on the line is not good, it will cause some error packets,

When there is a continuous error packet (from the captured SNR, we can see that there is a continuous state of very poor signal-to-noise ratio), then 4G module (using Ti dp83tc811s) sends a pause packet to 5g, and then Ping will not work

 Now I want to know,

1. Is there any mechanism in the bottom layer, when the packet loss is high (the packet loss of the faulty line reaches 5% - 20%), it will directly lead to the Ping failure?

2. Why does the pause packet occur when the signal quality is poor? Is this the mechanism of this chip?

  • 0
    TI__Genius 11725 points

    Hi Kevin,

    We are looking into this issue and will provide feedback by Wednesday this week.

    Thank you,

    Nikhil

  • 0 in reply to Nikhil Menon
    Expert 2925 points

    There was an error in some values of the registers last sent.

    When Ping fails in the laboratory, the value of register 0x0198 is 0x0122.

    The register has not been read in the real vehicle, so the estimation is lower.

    From this result, it is true that the harness quality is poor, and the kernel does not print error messages.

    Excuse me. How does the value of this SNR register affect the network? Will it lead to the final Ping failure?

  • 0 in reply to kevin.huang_FJXM
    TI__Genius 11725 points

    Hi Kevin,

    A poor SQI will to poor link quality and signal quality. This can cause errors and lost packets, including ping failure. What we can do is try to locate the source of the poor signal quality. What type of cable and cable length is being used? Can you provide a schematic for review?

    Thank you,

    Nikhil

  • 0 in reply to Nikhil Menon
    Expert 2925 points

    Nikhil:

    Use common double Jiao shielded wire, 13.4 meters in length

    Using the wire in Figure 1, the packet loss will be serious when the length is 13.4 meters, and eventually the Ping will not work.

    There is no obvious packet loss in the laboratory when using the wire rod of specification Figure 2, and there is no obvious packet loss in the real vehicle, but the actual routing has not been carried out according to figure 1

    Question: if the packet loss is caused by wire, will the Ping fail after a long time? What is the principle?

  • 0 in reply to kevin.huang_FJXM
    TI__Mastermind 18425 points

    Hello,

    Thanks for the follow up. Replies on E2E are delayed due to end of year holidays. We will get back to you by Jan 7.

    -Regards

    Aniruddha

  • 0 in reply to Aniruddha Khadye
    TI__Genius 11725 points

    Hello,

    If there is packet loss or poor signaling due to cables or connectors, we do not think the effect on ping should be time-dependent. It sounds like you have both a working setup and non-working setup. Based on your statement "There is no obvious packet loss in the laboratory when using the wire rod of specification Figure 2, and there is no obvious packet loss in the real vehicle," are these two setups error free? If so, what are the changes in conditions fir the setups with ping failure?

    Can you provide an SQI reading for both working and non-working conditions. In the non-working condition, you mention ping fails after 30 minutes to an hour. Can you take an SQI reading in this setup both before and after ping fails?

    Thank you,

    Nikhil 

  • 0 in reply to Nikhil Menon
    Expert 2925 points

    Nikhil :

    Thank you for your reply

    1。 what are the changes in conditions fir the setups with ping failure?

    As shown in the figure, when it doesn't work properly, we find a lot of errors and dropped through ifconfig

    2、Can you provide an SQI reading for both working and non-working conditions. In the non-working condition, you mention ping fails after 30 minutes to an hour.

    In normal operation, the value of register 0x0198 is: 0x0231

    When it does not work normally, the value of register 0x0198 is: 0x0122

    Through the register manual, we can calculate that the SQI is 37 when it cannot work normally and 49 when it works normally

    So can we infer that the quality of wiring harness really affects Ping?

    3、Can you take an SQI reading in this setup both before and after ping fails?

    Yes, the value we read out through the register of 0x198 is: 0x0122

  • 0 in reply to kevin.huang_FJXM
    TI__Genius 11725 points

    Hello,

    It is possible that the harness is the cause of the poor SQI if SQI value is ok without the harness. 

    In regards to my point 1: I'm not clear on the setup difference between the working and non-working conditions. I understand when failing, there are a lot of errors. Is there a change in boards, cables, temperature, etc.?

    2. There is a noticeable signal degradation. Was there any change to the setup between normal and failing SQI readings? 

    3. ok

    Additionally, can you describe this setup in the picture below in more detail? Is it three lengths of the same cable model, separated by interconnects? If so, is SQI still poor for a short cable length with no interconnects?

    Thank you,

    Nikhil