• State TI Thinks Resolved
  • Locked Locked
  • Replies 2 replies
  • Answers 1 answer
  • Subscribers 33 subscribers
  • Views 269 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: DS100BR111 SI data/model questions

Kevin Shafer
Prodigy 60 points
Part Number: DS100BR111

Tool/software:

We plan on using the DS100BR111 configured as a 10G-KR repeater.  We are getting into SI analysis of the design and have the following questions:

JITTER

  1. Typical deterministic jitter values (DJxx, see Nov2014 datasheet section 7.4 page 8) are rather high to insert the repeater reliably into a channel with pre-existing jitter, hence it is imperative we understand the repeater’s inherent jitter. Can TI provide additional jitter values (max values, other EQ settings, etc.)? Is there any other application info and/or measured data on repeater jitter? Assumption: no additional data is available beyond the datasheet and model.
  2. The part’s associated IBIS-AMI models (called from model DS100BR111.ibs rev 6.0 Sep2014) do not specify jitter parameters in either the Tx or Rx .ami files, nor does the model’s User Guide address jitter. As such, it seems we should add jitter, per datasheet values, using EDA tool constructs?  Assumption: yes.
  3. The model’s User Guide quantifies “jitter” when comparing the AMI model’s performance to Cadence transistor simulation on pages 10 to 13. Can we assume these are simply values measured in the eye diagrams with no additional (datasheet-driven) jitter imposed in either simulation environment? Assumption: yes, the values in the Guide capture time variation that subtracts from eye width and calls this “jitter”.
  4. Simulating the above models into testloads, it appears that increasing EQ increases the device’s inherent jitter. Does TI agree with this observation? Assumption: yes.
  5. If our existing channel already has 0.30 UI jitter (channel only, no Tx or Rx), does TI agree that inserting the repeater will increase the end-to-end channel (repeater included) jitter linearly by the amount imposed by the repeater? Assumption: yes.

 

EQ

  1. The model and datasheet/User_Guide do not agree on either Tx or Rx EQ dB levels. Compare Rx min EQ and Tx max EQ levels. Which is correct? Assumption: datasheet.
  2. User Guide section 4’s VOD_Level Table (right column) says “De-emphasis value” which we believe to be peak-to-peak amplitude? Assumption: yes

 

TI can provide additional information on these questions, or simply confirm our assumptions are correct.

Thanks much,

Kevin

  • 0
    TI__Mastermind 27980 points

    Hi Kevin,

    I'm looking into this and will get back to you with feedback tomorrow.

    Best,

    Lucas

  • 0 in reply to Lucas Wolter
    TI__Mastermind 27980 points

    Hi Kevin,

    1. Your assumption is correct, no additional data is available beyond the datasheet and model. In case it was unclear to you, the deterministic jitter on the output is 0.09 UI when there is no media. DJ increases when a cable or trace is connected to the receiver because jitter comes from the longer input channel.

    Typically we recommend using a retimer to clean jitter because of the inclusion of a clock and data recovery (CDR) functional block. This enables a retimer to effectively reset the jitter budget, while a redriver can only compensate for jitter associated with channel insertion loss. But our retimers do not support link training and auto-negotiation protocols, so a redriver is necessary for 10G-KR applications unless you can disable LT and AN.

    2. I recommend to keep the jitter values as 0. I would only add jitter values to the simulation from the system Tx driving the DS100BR111. This will yield more realistic results as the model was verified in this way.

    3. Yes, jitter measurements provided in the Model Verification section of the IBIS-AMI user guide are extracted from the eye openings.

    4. I don't agree with this assumption. I believe jitter should be minimized when the optimal EQ setting is selected for a particular channel. If EQ is increased beyond the optimal setting, then I believe jitter will be higher.

    5. I don't agree with this assumption. A redriver should be capable of compensating for jitter associated with the channel insertion loss when the optimal EQ setting is selected.

    6. I do see the values provided for DE_Level in the IBIS-AMI user guide slightly differ from programmable de-emphasis levels provided in register 0x11[2:0] and 0x18[2:0] in the register map in the datasheet. Device de-emphasis levels match the datasheet.

    I don't see any differences between the values provided for EQ_Level in the IBIS-AMI user guide and the EQ values provided in Table 3 of the datasheet.

    7. Yes that is a typo, VOD_Level corresponds with peak-to-peak amplitude.

    Best,
    Lucas