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.

DRA74P: moving from single die to dual die 8GB DDR issue

Part Number: DRA74P

Hi everybody  ,

on 8Gb memory   moving from 

From p.n.: MT41K512M16HA-125 AIT:A (DDR3 8Gb)   (

To p.n. MT41K512M16VRP-107 AIT:P based on a dual-die package.

 What consequences there are on the configuration of DDR controller moving from 2 p.n.

On  p.n. MT41K512M16VRP-107 AIT:P  I see  page size per DIE =1KB

On  p.n. MT41K512M16HA-125 AIT:A I see page size =2K

I need to configure inside Emif controller  Bank/columns/Rows with different value?

 

what about multi die  handling by EMIF ? any other concern / issue to take care ? 

Best regards

Carlo

  • Hi Carlo,

    As stated in the documentation, DRA74P does not support twin die.

    EDIT: The above statement is in relation to loading on the signals. DRA74P does not support dual loading on the data bus, or more than 4 loads on the CA bus. Please always refer to the latest datasheet for loading restrictions.

    Thanks,
    Kevin

  • Kevin,

    It's not twin die in the sense that you're thinking.  Here's a snippet from the data sheet:

    Uses two x8, 4Gb Micron die to make one x16 package

    From an electrical load perspective, this is equivalent to using two 8-bit DDR3 devices, which is a supported configuration.

    Best regards,
    Brad

  • Colombo Carlo said:

     What consequences there are on the configuration of DDR controller moving from 2 p.n.

    On  p.n. MT41K512M16VRP-107 AIT:P  I see  page size per DIE =1KB

    On  p.n. MT41K512M16HA-125 AIT:A I see page size =2K

    I need to configure inside Emif controller  Bank/columns/Rows with different value?

    You don't need to change the values from the EMIF perspective as it looks the same.

    Colombo Carlo said:
    what about multi die  handling by EMIF ? any other concern / issue to take care ? 

    This will change the loading on the address/control lines.  Is the customer using VTT termination on the address/control (hopefully yes)?  Ideally the customer should perform IBIS simulations for these two different devices to determine the optimum configuration of the drive strength, especially on the address/control lines where the loading will double.  Data lines may also change slightly due to the fact that it's a different device, but it will still be a single load.

  • Carlo,

    Does the customer have JTAG available on their board, plus a TI probe and CCS?  If so, I have a script that will scrape their DDR configuration and do some analysis.  The customer can download the script here:

    http://git.ti.com/sitara-dss-files/am57xx-dss-files/blobs/raw/main/am57xx-ddr.dss

    They should attempt to boot the board as usual (i.e. the DDR should be in its configured state), and then they should run the script by following the directions here:

    http://git.ti.com/sitara-dss-files/am57xx-dss-files/blobs/main/README

    The script will output a *.txt file to their desktop.  Please post it here.

    Best regards,
    Brad

  • Hi Carlo,

    Brad and I have discussed internally and agree that we don't see any specific issues with that device / topology. I apologize for the confusion.

    Thanks,
    Kevin

  • Closing this thread. We can re-open if needed.