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.

DRA821U: SGMII compliance with 100BASE-T

Part Number: DRA821U
Other Parts Discussed in Thread: DRA821

Hi TI

Can you please clarifiy if DRA821U SGMII interface and Ethernet FW are able to support 100BASE-T compliance.

Customer uses the Marvell 100BASE-T Ethernet-PHY (Marvell 88Q1111) via DRA821U SGMII. Their board produces the packet loss and ping error.

I am unable to find the specific information about 100BASE-T from below.

Note
The SGMII interfaces electrical characteristics are compliant with 1000BASE-KX per IEEE802.3
Clause 70.
Note
The SGMII 2.5G / XAUI interfaces electrical characteristics are compliant with IEEE802.3 Clause 47.

  • Jack,

    Yes, the SGMII has been functionally and electrically validated to support 1000Mbps as you have noted from the datasheet. I do not believe that there should be a concern with supporting the 100Mbps speeds as it can be considered a subset to 1G speeds but I have not verified (electrical compatibility at 100M speeds). The device is capable to support 5G/10G speeds and hence 1G speed support is considered on the slower side.

  • Hi ShreyasRao

    Thanks for your kind reply.

    Can you please provide us with the trouble shoot guidance regarding the packet loss problem when using SGMII?

    Even though DRA821 was not validated wifh SGMII-100BASE-T, Customer tried to interface with 100BASE-T of Marvell. then the packet loss problem is reproducible from their custom board. 

    They still don't have any idea how to deal with it as of now. We need your helpful comment on this topic.

    Regards, 

    Jack

  • Hi Jack,

    Yes, I have received it over email, and is being looked into.

    The SGMII communicates at 1G link speed even during 10/100M mode. We are waiting on the CPSW stats before and after the test to determine the errors.

    The same 1G SGMII configurations must be used in 100M case.

    10/17:

    As we have been debugging, we have found out that the preamble bytes are not compatible with the Marvel PHY and hence lead to the rejection of the packets.

    The workaround is to either use another PHY which will ignore the preamble bytes or to use 1000M mode which doesn't have this issue.