• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 4 answers
  • Subscribers 33 subscribers
  • Views 841 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:

DS125BR820: DS125BR820 application issue.

DS125BR820: DS125BR820 application issue.

Hugo Liu
Genius 5095 points
Part Number: DS125BR820

Hi Sir,

Please refer to picture below.

  • 0
    Genius 5095 points

    Hi Sir,

    I just update detail signal loss information(all value@4GHz) and diagram as below.

    Attached file is our current EEPROM setting of repeater.
    So far, we guess the repeater driving strength is too strong on the TX side that goes to retimer(mark in red circle).

    We are trying to lower it.
    If you have other suggestion, let me know.

  • 0 in reply to Hugo Liu
    TI__Mastermind 43795 points

    Hi Hugo,

    I did not see the EEPROM configuration file attachment.

    Based on the diagram here are the settings I would suggest.

    Upstream: (towards CPU)

    Function       Register value 

    EQ                02'h  (Datasheet - Level 3)

    VOD             100'b  (Datasheet - Level 4) 

    VOD_DB      000'b

    Downstream: (towards PEX)

    EQ                01'h  (Datasheet - Level 2)

    VOD             100'b  (Datasheet - Level 4) 

    VOD_DB      000'b

    Regards,

    Lee

  • 0 in reply to Lee Sledjeski
    Genius 5095 points

    Hi Sir,

    Add the EEPROM configuration file

    JBOF_20171225(VOD_DB_000_Max_VOD_010_0.71).zip

  • 0 in reply to Hugo Liu
    TI__Mastermind 43795 points

    Hi Hugo,

    I adjusted the file to decrease the equalization gain.

    Fullscreen JBOF_EQ01_VOD02_VOD_DB00.txt Download 
    :200000006700100033000033000033000033000033000033000033000033000000000000D1
:2000200000000000000000000000000000000000000000000004070001AA00001AA000014F
:20004000AA00001AA009800354000035400003540000354000005454000000000000000073
:20006000000000000000000000000000000000000000000000000000000000000000000080
:20008000000000000000000000000000000000000000000000000000000000000000000060
:2000A000000000000000000000000000000000000000000000000000000000000000000040
:2000C000000000000000000000000000000000000000000000000000000000000000000020
:2000E000000000000000000000000000000000000000000000000000000000000000000000
:200100000000000000000000000000000000000000000000000000000000000000000000DF
:200120000000000000000000000000000000000000000000000000000000000000000000BF
:2001400000000000000000000000000000000000000000000000000000000000000000009F
:2001600000000000000000000000000000000000000000000000000000000000000000007F
:2001800000000000000000000000000000000000000000000000000000000000000000005F
:2001A00000000000000000000000000000000000000000000000000000000000000000003F
:2001C00000000000000000000000000000000000000000000000000000000000000000001F
:2001E0000000000000000000000000000000000000000000000000000000000000000002FD
:00000001FF

    Regards,

    Lee

  • 0 in reply to Lee Sledjeski
    Genius 5095 points

    Hi Sir,

    Thank you very much,

    We will pass to customer try it.

    Hugo

  • 0 in reply to Hugo Liu
    Genius 5095 points
    Hi Sir,

    This issue improve.

    our Customer have other question,
    If the trace is too Long over one DS125BR820 capability,
    Do you have any suggest for this application?

    Hugo
  • 0 in reply to Hugo Liu
    TI__Mastermind 43795 points

    Hi Hugo,

    We like to avoid using multiple redrivers whenever possible.  Without direct knowlege of vendor link training algorithms there is more risk to the link adapting to a non-optimal configuration.  Typically in PCIe applications at Gen3 we can utilize a single redriver for links up to ~30dB.  The actual maximum loss is highly dependent on the layout and system components.

    Regards,

    Lee