Because of the Thanksgiving holiday in the U.S., TI E2E™ design support forum responses may be delayed from November 25 through December 2. Thank you for your patience.

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.

DS100DF410EVK: How to reduce jitter?

Part Number: DS100DF410EVK
Other Parts Discussed in Thread: DS100DF410

Hello, This is my first post.

We are considering a 10GbE re-timer using the DS100DF410EVK.
What settings can be used to reduce jitter?

The above is input eye wave.

The above is output eye wave through DS100DF410EVK.

The current SigCon status is below.

Thanks.

  • Greetings,

    The above waveform shows jitter at the output of the DF410EVK. Please consider the followings:

    1). Please make sure device is locked - CDR lock indication is active.

    2). This incoming jitter could be within the CDR loop bandwidth. Please provide a block diagram showing the setup for this measurement. 

    3). Please provide register dump of DF410. Using GUI, you can go to low level and click on Save All.

    Regards,Nasser

  • Hi Nasser,

    Thanks for your advice.

    I will send block diagram and register dump data.


    Best regards. df410evk_1028.cfg

  • Hi Sim,

    Have you tried adjusting the transmit de-emphasis?  If so, any improvement?  Also, what is the best case jitter measurement that your scope is capable of making (i.e. how much jitter do you observe with a golden transmitter)?

    Transmit de-emphasis and CDR loop bandwidth are the primary settings which can be adjusted to change the jitter you observe.

    Thanks,

    Drew

  • Hi Drew,

    I haven't tried adjusting the transmit de-emphasis yet.
    Because it is a transmission system that does not have I2C control.
    I'll try it with a jig, so please wait.

    Below is the waveform of the FPGA output.

    There is jitter at the FPGA output, but this cannot be corrected due to the circuit fix.

    That's why I want to reduce jitter with a re-timer IC.

    Best regards.

  • Hi Sim,

    Please try adjusting transmit de-emphasis to see if this improves your measurements.  With that said, I am wondering if some of the jitter we are observing is a limitation of the scope.  Typically we expect a very clean output from the DS100DF410 when there is not much insertion loss between the EVM and measurement equipment.  See scope shot below from datasheet.

    Thanks,
    Drew

  • Hi Drew,

    Sorry for the late confirmation.

    Since the optical module used does not have a variable emphasis function, I changed the system to the one shown in the figure below and changed the emphasis of the FPGA.

    FPGA default waveform

    FPGA pre-emphasis setting waveform

    EVK output waveform at FPGA pre-emphasis

    It will not be a scope shot like the datasheet.

    Anything else to try?

    Best regard.

  • Hi Sim,

    The primary retimer settings that can be used to reduce jitter are the transmit de-emphasis and CDR bandwidth.  At this point, adjusting the CDR bandwidth seems like the most appropriate action since it doesn't appear that de-emphasis is required.

    In order to understand if adjusting the CDR bandwidth can help, it would be valuable if you can use a BERT to generate a signal and enable various amounts of jitter in the signal.  By doing this, you can observe if the retimer is responding to jitter as expected.

    At this point, do you have concerns about the eye opening you are observing from the retimer TX?

    Thanks,
    Drew

  • Hi Drew,

    Thank you for your advice.
    Using a BERT, we found that when the jitter bandwidth is about 5 MHz, the EYE opens and the jitter decreases significantly, but when the jitter bandwidth is about 1 MHz, the jitter remains the same and the EYE does not open.

    So question, is there a register to adjust the CDR bandwidth of the DF410?

    Best regards.

  • Hi Sim,

    I think your measurements correlate well with the retimer CDR bandwidth.  By default, it is around 5 MHz, so it makes sense that jitter > 5MHz would not be tracked by the CDR.

    The CDR bandwidth is adjustable (see below), although it cannot be reduced to 1 MHz.  The screenshot below is an excerpt from our 10G retimer programming guide, also linked below.

    https://www.ti.com/lit/pdf/snla323

    Thanks,

    Drew

  • Hi Drew,

    You taught me that I can adjust the CDR bandwidth.
    The original goal was achieved.

    Thanks,
    Sim