• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 1 answer
  • Subscribers 33 subscribers
  • Views 978 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:

LMH1218: Device Register settings vs. IBIS-AMI model parameter settings

LMH1218: IBIS model parameter settings:

Jay Skender
Prodigy 170 points
Part Number: LMH1218

I am noticing apparent load differences on the LMH1218 RX inputs when I simulate with Hyperlynx v9.4.2  and I select between

Slow-weak, Typical, and Fast-Strong

The typical is a much smaller well behaved eye, but choosing either Fast-Strong or Slow-Weak results is x4 taller eye and not as symmetric transitions.

I have tried replacing the LMH1218 with a 120 ohm resistor and tied it directly to the transmitter : and I get the same size eye, but time shifted for each setting.

Any ideas what may cause this behavior?

  • 0
    TI__Guru* 93670 points
    Hi Jay,

    We use Agilent ADS to validate IBIS-AMI model. I am not sure exactly what is occurring here. Is it possible to check this with Mentor?

    Regards,,,nasser
  • 0 in reply to Nasser Mohammadi
    Prodigy 170 points

    Hi Nasser,

    I provided info to Mentor and asked them if they thought the issue is with the model or the simulator.

    Do you have access to someone who can run a simulation as described?  It appears the LMH1218 RX input termination/loading is changing when switching between Slow Typ Fast

    If they can reproduce the problem then that would be helpful.

    Thanks,

    -Jay S.

  • 0 in reply to Jay Skender
    TI__Guru* 93670 points
    Hi Jay,

    LMH1218 RX IBIS-AMI model is typical and it does not simulate extreme cases like slow-weak or fast-strong.

    We believe Mentor ideal transmitter behavior changes when selecting slow-weak or fast-strong. You can confirm this by putting a probe on LMH1218 package_rx1.s4p file and compare typical versus the other two cases.

    Also, please note we mainly used ADS to validate LMH1218 IBIS-AMI model versus using Mentor Hyperlynx.

    Regards,,nasser
  • 0 in reply to Nasser Mohammadi
    Prodigy 170 points
    I would recommend that the User Guide be documented to state only "typical" is supported in simulations (this may apply to many TI parts).
    If I missed this in the documentation please let me know.
    thanks for our support,
    -Jay S
  • 0
    Prodigy 170 points
    So feedback from HyperLynx is the RX model should be modified if the corner cases are not covered, then instead of using zero place an NA as shown below will cure the problem. ( It may be of value for TI to partner with Hyperlynx for recommendations on IBIS-AMI models)
    I ran with the mods below and the eyes look much much better: amplitude was comparable between typ slow fast.
    The NA would also serve as user flag that slow fast are not supported.

    Best Regards,
    -Jay S.

    [GND Clamp]
    |-6.6 -0.132 -0.0 -0.0
    | 0.0 0.0 0.0 0.0
    | 6.6 0.132 0.0 0.0

    | Just use TYP for all corner settings:

    -6.6 -0.132 NA NA
    0.0 0.0 NA NA
    6.6 0.132 NA NA


    [Power Clamp]
    |-6.6 0.132 0.0 0.0
    | 0.0 0.0 0.0 0.0
    | 6.6 -0.132 0.0 0.0

    | Just use TYP for all corner settings:

    -6.6 0.132 NA NA
    0.0 0.0 NA NA
    6.6 -0.132 NA NA