• State TI Thinks Resolved
  • Locked Locked
  • Replies 9 replies
  • Answers 1 answer
  • Subscribers 99 subscribers
  • Views 1356 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.

AM5726: DDR DQS output ringing

UNA
Mastermind 7980 points

Part Number: AM5726

Hi,

My customers have performed DDR3 compliance testing but the results are failing.
Looking at the waveform, there is a disturbance of the DQS waveform at the beginning and the end of the data communication respectively.
At the start of data communication: DQS rises steeply
End of data communication: ringing occurs in DQS after data communication
What can be considered about the cause of this problem?

Best Regards,
H.U

  • 0
    TI__Guru* 95721 points

    HU, you may not have properly tuned your drive strength and output impedance according to your board layout and simulations.  Please refer to the AM57xx EMIF tool app note () to properly configure these settings.

    Regards,

    James  

  • 0 in reply to JJD
    TI__Expert 7895 points

    Hi James,

    I sent this detail EVM wavelength and EMIF register information to you.

    Could you please check it ?

    Regards,

    Kz777

  • 0 in reply to Kz777
    TI__Guru* 95721 points
    Looking at the results of the EMIF tool, how did your customer come up with the values in 1C) and 1D) of the spreadsheet? Are these results of board simulations? Why are they completely disabling ODT on both ends?

    Regards,
    James
  • 0 in reply to JJD
    Mastermind 7980 points
    Hi James,

    They calculate values of 1 C) and 1 D) from waveform simulation based on IBIS.
    The simulation result is passed, we will confirm to them whether they can provide it.
    They want to disable ODT to reduce power consumption.


    I have additional questions:

    Do you have waveform measurement data with AM57xx EMIF tool default setting by using AM572xEVM and can you share it?

    What kind of modification should we make to solve this problem?


    Best Regards,
    H.U
  • 0 in reply to UNA
    TI__Guru* 95721 points
    Did the simulations also comprehend disabling ODT? They may want to try to experiment with these values along with output impedences to see if they can mitigate the ringing. The board sims don't seem to be correlating with what you are seeing.
    I'll try to see if we have info on the AM572xEVM.

    James
  • 0 in reply to JJD
    TI__Guru* 95721 points
    HU, we don't have any data on the EVM. Has your customer resolved the issues?

    Regards,
    James
  • 0 in reply to JJD
    TI__Expert 7895 points

    Hi James,

    Actually, when we enabled ODT, waveform result didn't fix correctly.

    Regarding waveform and this configuration table by private message.

    Regards,

    Kz777

  • 0 in reply to Kz777
    TI__Genius 9655 points
    H.U.

    0) Is the layout compatible with all the layout recommendations provided in the datasheet? Specifically - impedance, reference plane and spacing requirements? Please review the datasheet to make sure the layout complies to the datasheet requirements
    1) Where are you measuring the waveforms? Ideally, the waveforms should be measured at the far-end of the transmission line. If you are measuring closer to the AM57x during a WRITE operation (or) closer to the DDR memory during a READ operation, they will not look clean due to near end reflections
    2) Did you enable ODT on both the ends i.e. DDR and the AM57x device?
    3) Did you try different ODT settings to see if it makes a difference?
    4) Can you confirm if there are any memory errors or suspecting issues with the signal quality

    Regards, Siva
  • 0 in reply to sivak
    Mastermind 7980 points

    Hi sivak,

    sivak said:

    0) Is the layout compatible with all the layout recommendations provided in the datasheet? Specifically - impedance, reference plane and spacing requirements? Please review the datasheet to make sure the layout complies to the datasheet requirements


     - I confirmed this detail to customer lastweek. They copied datasheet requirement. And they also passed TI team circuit review.

    sivak said:

    1) Where are you measuring the waveforms? Ideally, the waveforms should be measured at the far-end of the transmission line. If you are measuring closer to the AM57x during a WRITE operation (or) closer to the DDR memory during a READ operation, they will not look clean due to near end reflections


    -I attached this waveform to this message via local TI FAE. Do you think this waveform is un-expected ?

    sivak said:

    2) Did you enable ODT on both the ends i.e. DDR and the AM57x device?
    3) Did you try different ODT settings to see if it makes a difference?
    4) Can you confirm if there are any memory errors or suspecting issues with the signal quality


    - I also submit DDR compliance test result. Is it possible to help above question?

    Best Regards,
    H.U