• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 2 answers
  • Subscribers 33 subscribers
  • Views 1553 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.

DS100BR111: SATA repeaters and autonegotiation problem

Romain Borlet
Prodigy 90 points
Part Number: DS100BR111
Other Parts Discussed in Thread: DS125BR111, SN75LVCP600

Dear,

I made an electronic design for embedded computer needs, that integrates SATA III link for which I add DS100BR111 repeaters. The SATA links are issued from Intel 6th generation Core i5/i7 processor on COM Express type 6 module.

I found the best pin configuration for the two repeaters I use, but the operation of the link is very unstable. It looks like the repeater causes a dysfonctionnement of the autonegociation of SATA link because sometimes when it well boot on the OS, the both SSD used are in 6Gb/s with good performances, but more often they are not in full speed or viewed in BIOS but not in Windows7 and 10...

Bypassing the repeater permit a best autonegociation process and the disk are always viewed in BIOS and OS, but performances are under.

Is there any "normal" explanation of this problem and any way to correct it?

Thanks for your help.

Romain 

  • 0
    TI__Mastermind 43795 points

    Hi Romain,

    What are the current settings used by the DS100BR111 device?  Pin values? Register values?

    What do the SATA channel segments between the CPU and COM Express Module look like in terms of topology and attenuation?

    Thanks and Regards,

    Lee

  • 0 in reply to Lee Sledjeski
    Prodigy 90 points

    Hi Lee,

    The design is based on COM Express Type6 with PICMG COM0 recommandations (SATA is routed as Z 85 Ohm diff). Our product uses COM module cExpress-SL from Adlink and we develop a carrier, a backplane and a SSD array. 

    The topology is in the following Excel table.

    We use two SATA lanes, one named SATA0 connected to a mSATA on the carrier board through a DS100BR111 redriver and a switch. The SATA0 lanes could also be switch to SSD array through two redrivers and a backplane. SATA1 lanes are connected to the second disk of disk array through two retrievers and the backplane.

    SATA_Layout.xlsx

    The DS100BR111, are configuring in pin mode.

    The best configuration is actually :

    • mSATA redriver: DEMA F , EQA1 0, EQA0 0, DEMB F, EQ1B 0, EQB0 R and VOD_SEL 0
    • SSD1 1st redriver (on carrier): DEMA 0 , EQA1 0, EQA0 0, DEMB 0, EQ1B 0, EQB0 R and VOD_SEL 0
    • SSD1 2nd redriver (backplane): DEMA 0 , EQA1 0, EQA0 0, DEMB 0, EQ1B 0, EQB0 0 and VOD_SEL 0
    • SSD2 1st redriver (on carrier): DEMA 0 , EQA1 0, EQA0 0, DEMB 0, EQ1B 0, EQB0 0 and VOD_SEL 0
    • SSD2 2nd redriver (backplane): DEMA 0 , EQA1 0, EQA0 0, DEMB F, EQ1B 0, EQB0 0 and VOD_SEL 0

    I made a test replacing the both redrivers on SATA1 lanes to the SSD2 by two small cables, and with the good BIOS parameters it works.

    I added the redrivers to have the better signal integrity as possible but I don't know why it looks to be worst for the SATA communication between the both communicators.

    Regards,

    Romain

  • 0 in reply to Romain Borlet
    TI__Mastermind 43795 points

    Hi Romain,

    Two redrivers in this relatively short span is too much.  This tends to over equalize the signal and most importantly makes OOB signal transmission more difficult.

    What setting are you using with SD_TH?

    What setting are you using with MODE?

    We may find that using the DS125BR111 will give a better response - it is just a linear equalizer and helps to re-transmit the OOB signaling without distortion.  The system problem you are facing is because the OOB distortion causes SSD configuration issues.  This OOB problem is mostly seen a "no drive detected" or a 6 Gbps drive operating at just 1.5 Gbps.

    Regards,

    Lee

  • 0 in reply to Lee Sledjeski
    Prodigy 90 points

    Hi Lee,

    SD_TH is F and MODE is 0.

    I also done a test removing one of both redriver, on first link at backplane side and on other link on carrier side. The problem was the same besides the fact that I had to grow a little up the settings.

    I finally remove all the redriver and disks are viewed all the times, but one of them never had 6Gb performances.

    I made a first design of all boards with SN75LVCP600, with which I had quite the same results. I think it could be interesting I work on it using what I know now.

    What do you think about that? Will I lost my time?

    Thanks for your help.

    Regards,

    Romain

  • 0 in reply to Romain Borlet
    TI__Mastermind 43795 points

    Hi Romain,

    I would try the settings below.

    MODE = High

    VOD_SEL = High (uses a non limiting output style which can help OOB noise transmission)

    DEM = Float

    SD_TH = Low

    You can also try the above settings with MODE = Low to see if that can help in your application.

    Regards,

    Lee

  • 0 in reply to Lee Sledjeski
    Prodigy 90 points

    Hi Lee,

    I maybe will try the previous settings you recommend me after the 14 st of may because I will be out of office next week.
    For informations, we tested a systems on which we short all the redriver and it works well at Gen3 with one SSD (Innodisk  3MG2-P 1T) and as Gen2 with the second (Innodisk 3MR3-P 128GB). I'm in contact with Innodisk to view if it possible to improve the second SSD functioning.
    On other hand Adlink sent me a special BIOS in which they modified the SATA lane setting as you can see folowing. But the results are not good with the actual settings of redrivers.


    Here are the changes for the SATA Settings:

     Original BIOS:

                 Sata 1 (ca 2")                     Sata 2 ( ca 2")

                    GEN1   GEN2   GEN3                 GEN1    GEN2   GEN3

    TX Voltage :    100%   100%   100%                 100%    100%   100%

    TX De-emph : -36.1dB -3.9dB -6.0dB              -36.1dB  -3.9dB -6.0dB

    RX DTLE    :   0.8dB  1.6dB  3.2dB                0.8dB   1.6dB  3.2dB

     

     

    TEST BIOS

                 Sata 1 (ca 2")                     Sata 2 ( ca 2")

                    GEN1   GEN2   GEN3                 GEN1    GEN2   GEN3

    TX Voltage :   64.1%  75.0%   100%                 64.1%  75.0%   100%

    TX De-emph :  -3.5dB -5.8dB -6.0dB               -3.5dB  -5.8dB -6.0dB

    RX DTLE    :   0.8dB  1.6dB  0.8dB                0.8dB   1.6dB  0.8dB

    Best regards,
    Romain
  • 0 in reply to Romain Borlet
    TI__Mastermind 43795 points

    Hi Romain,

    Thanks for the information.  I would remove all the Tx De-emph from the SATA card waveforms.  The individual repeaters will have enough residual equalization to compensate for the PCB losses in this case.  I think it is a good idea to reduce the Tx output voltage as shown in the TEST BIOS as well.

    Regards,

    Lee