• State
  • Locked Locked
  • Replies 1 reply
  • Subscribers 98 subscribers
  • Views 263 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.

omap3530 IBIS model

inanc
Prodigy 30 points
Other Parts Discussed in Thread: OMAP3530

Hello,

I am doing Hyperlynx simulations with the OMAP3530/25 CUS IBIS Model (Rev. B) which can be downloaded with the following link: http://focus.ti.com/dsp/docs/litabsmultiplefilelist.tsp?sectionId=3&tabId=409&literatureNumber=sprm324b&docCategoryId=4&familyId=1526

My main interests are the LPDDR and GPMC buses. According to my simulations, OMAP doesn't drive smooth, the resultatnt signal shape small spikes all over and I don't mean the overshoot/undershoot and the ringins. A TI employee recommended me to use series resistors to smooth out the signal. However the omap3530.pdf datasheet states series resistors are not recommended nor needed.

Secondly, OMAP's SLOW and FAST drive modes do not have obvious differences. In fact SLOW mode seems to be little faster than the FAST mode.

Has anyone tried out the IBIS file I am referring to and know what I mean? How should I interpret these results?

Thanks in advance.

  • 0
    TI__Intellectual 1280 points

    Inanc,

    I have seen results similar to what you are describing when connecting SDRC_D0 to a specific Micron LPDDR ibis model such as MT46H64M32LFJG.  Results were improved when I switched to another Micron model, MT46H32M32LFJG so keep in mind that other models/components included in your simulation could be impacting your sim results.  You may consider doing some simulations using 3530 IBIS model ->simple 50 ohm transmission line -> typical 1.5pF load for comparision.  In regards to series terminations, they will definately help reduce overshoots and help with EMI if needed but you are correct that our Data Manual does currently state that they are not required.  Note: If you are simulating with Micron LPDDR ibis models for example, make sure that you are probing at memory die at not at pin in your Hyperlynx simulations.

    In regards to fast vs slow driver modes, they should correlate with DS (drive strength) setting of IO bufer.  Slow mode has DS0_0 and Fast mode has DS0_1 which has been confirmed by IO modelling team.  I previously did some simple simulations of slow vs fast models for sdrc_clk (model1 (slow) vs model2 (fast)) and sdrc_nclk (model29 (slow) vs model 30(fast)) for OMAP3530 CUS IBIS model.  I did this using Hyperlynx 8.0 using IBIS model as an output driving a simple 50 ohm transmission line and 1.5pF capacitive load to compare edge response.  Model 2 output drive was faster than model 1 and model 30 output drive was faster than model 29 based on my simulations which agrees with fast vs slow designation on these models.  My simulations were done with existing single ended IBIS model using typical process corner selection for comparison in simulation tool. 

     Please reference IBIS Model Report below for more details regarding adaptive feedback IOs which may exibit non-monotonic characteristics dependent on load and simulator used...

    https://sps02.itg.ti.com/sites/PAI/prelimdocs/OMAP37x/Device%20Models/IBISModelReport.pdf