• State Resolved
  • Replies 4 replies
  • Answers 1 answer
  • Subscribers 34 subscribers
  • Views 90 views
  • Users 0 members are here
Support feedback
Options
Options
Related

DS250DF230: ambiguous CTLE boost value from auto adapt

Liang Qiao
Prodigy 70 points

Part Number: DS250DF230

Hi,

When adapt mode is set to 2, and register 0x2F[1] is at the default value of 0, is there a way to tell if the CTLE gain used is 0.6dB or 9.2dB from some other register?

From the diagram below, it looks like that these two CTLE boost values have the same read back register value:

image.png

The way I'm currently handling this ambiguity is:

  1. manually record DFE tap values
  2. set DFE to recorded values and set CTLE to 0.6dB
  3. get FOM from eye (if recovered)
  4. set DFE to recorded values and set CTLE to 9.2dB
  5. get FOM from eye (if recovered)
  6. compare the two FOM from the two CTLE settings

Which ever FOM is better, then the equivalent CTLE setting can be determined.
This method is quite tedious and time consuming if you have multiple channels.
Are there easier ways to do this?

 

Regards,

Liang

  • 0
    TI__Mastermind 43158 points

    Hi Liang,

    If I understand correctly, the ambiguity only occurs if you read back 0x00 from register 0x03.  Is this correct?

    Could you check if 0x2E[5] acts as readback of CTLE bypass bit?

    Also, is it possible to share more about your use case?  I would expect that for most channels, you can identify if 0.6 dB or 9.2 dB of boost are being applied based on channel insertion loss.

    Thanks,

    Drew

  • 0 in reply to Drew Miller1
    Prodigy 70 points

    Hi Drew,

    Your understanding is correct.
    Reading back 0x00 from register 0x03 is giving me ambiguity without a figure of merit calculation based on HEO and VEO.
    The bit 0x2E[5] does not act as a readback of CTLE bypass.
    0x2E[5] stays 0 when met with the 0.6dB/9.2dB ambiguity.
    This was tested by forcing 0.6dB CTLE, reading back 0x2E[5], forcing 9.2dB CTLE, reading back 0x2E[5].
    I've also done an XOR on the register 0x37 (this register was labeled as CTLE_STATUS, reserved for future use) to see if forcing 0.6dB and 9.2dB changes anything.
    XOR came back as 0.

    The context for this is testing passive and active cables.
    If the cable is actively re-driven, then the CTLE required would be small since channel loss wouldn't be so large.
    If the cable is passive, then the CTLE required may be 9.2dB.

    Please let me know if there are any other methods to differentiate between these two boost values other than calculating a FoM?

    Thanks,

    Liang

  • +1 in reply to Liang Qiao
    TI__Mastermind 43158 points

    Hi Liang,

    Unfortunately, I'm not aware of an alternative register we can look at.  Perhaps in cases where retimer CTLE reads back 0x00, I'm wondering if it might be faster to try again with adapt mode 1 to see how the retimer adapts without DFE.  In this case, you might be able to skip manually setting DFE if you need to manually compare FOM.

    I'm still interested in better understanding your use case.  The retimer automatically handles adaptation in adapt mode 2.  Do you have some use case where you need to manually set CTLE?

    Thanks,

    Drew

  • 0 in reply to Drew Miller1
    Prodigy 70 points

    Hi Drew,

    Thank you for your assistance, I will keep using the FoM method from HEO and VEO.
    I have sent you a DM of a more detailed summary of my use case.

    Regards,

    Liang