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.

AM2612: AM2612 OSPI sampling and register issues

Part Number: AM2612
Other Parts Discussed in Thread: SYSCONFIG

Hi experts,

 

I am asking for customer.

For AM2612, 

  1. Is it possible to change the current sampling timing for OSPI?
  2. Is it possible to change the slew rate of these OSPI ports?

Currently, I cannot find a place in SysConfig to modify the OSPI polarity and phase mode. My customer would like to change the sampling rising/falling edge and sampling delay ticks without modifying the OSPI mode.

 

These register descriptions are unclear. Can you explain them? What do the values 0/1 for the specified bits mean respectively? Also, what is the difference between the sampling edge described in mode0/mode3 and the sampling edge described in this register? Which configuration does the actual OSPI data sampling truly depend on? 7a360e41a0eedd9ef25078e1895117d4.jpeg

Why does the SDK configure the OSPI mode as mode0 (rising edge sampling), but the capture register is set to 1 (which should be falling edge)?

 

Best Regards

Kita

  • Hi Kita,
    Checking it, will get back to you

  • Hi Lakshya,

    Could you help to reply ASAP?

    Best Regards

    Kita

  • Lakshya,

    Please check the updated questions:

    1. Both OSPI_FLASH_CFG_RD_DATA_CAPTURE_REG and the OSPI mode can configure the OSPI sampling edge (rising/falling). Which one takes precedence? (In the SDK, the OSPI default configuration sets mode to 0 for rising-edge sampling, while OSPI_FLASH_CFG_RD_DATA_CAPTURE_REG's SAMPLE_EDGE is configured to 1 — if 1 means falling edge, then there is an inherent conflict.)
    2. The register description for OSPI_FLASH_CFG_RD_DATA_CAPTURE_REG is unclear — what do values 0 and 1 respectively represent for the sampling edge register? Is the sampling delay register based on the OSPI reference clock (166 MHz)? Can sampling only be delayed, not advanced?
    3. Both AM2612 TRM and the external Flash documentation state that only modes 0 and 3 are supported — so why does configuring mode 1 still allow normal communication?

    Please reply ASAP, thanks!

    Best Regards

  • Hi Kita,
    Please find my answers below

    Is it possible to change the slew rate of these OSPI ports?

    Yes, in sysconfig there are options for high and low .




    Currently, I cannot find a place in SysConfig to modify the OSPI polarity and phase mode.

    OSPI SysConfig module → check "Show Advanced Config" → a "Frame Format"







    These register descriptions are unclear. Can you explain them? What do the values 0/1 for the specified bits mean respectively? 



    Will correct this part in Register Addendum to make it unambiguous for customers.

  • Hi Lakshya,

    Thanks for you update here.

    In Friday I posted some new questions about OSPI from customer, please also answer these questions:

    1. Both OSPI_FLASH_CFG_RD_DATA_CAPTURE_REG and the OSPI mode can configure the OSPI sampling edge (rising/falling). Which one takes precedence? (In the SDK, the OSPI default configuration sets mode to 0 for rising-edge sampling, while OSPI_FLASH_CFG_RD_DATA_CAPTURE_REG's SAMPLE_EDGE is configured to 1 — if 1 means falling edge, then there is an inherent conflict.)
    2.  Is the sampling delay register based on the OSPI reference clock (166 MHz)? Can sampling only be delayed, not be advanced?
    3. Both AM2612 TRM and the external Flash documentation state that only modes 0 and 3 are supported — customer tried to configure as mode 1 and it works normally. So why does configuring mode 1 still allow normal communication?

    Best Regards

    Kita

  • Hi Kita,

    Both OSPI_FLASH_CFG_RD_DATA_CAPTURE_REG and the OSPI mode can configure the OSPI sampling edge (rising/falling). Which one takes precedence? (In the SDK, the OSPI default configuration sets mode to 0 for rising-edge sampling, while OSPI_FLASH_CFG_RD_DATA_CAPTURE_REG's SAMPLE_EDGE is configured to 1 — if 1 means falling edge, then there is an inherent conflict.)

    There is no precedence conflict, the two settings don't compete because they control two different clocks in two different places:

    - OSPI Mode (CPOL/CPHA) controls the sampling/valid-data edge on the external SPI clock pin (sclk). The agreement between the controller and the flash chip.
    - SAMPLE_EDGE_SEL_FLD controls the sampling edge of the internal reference clock (ref_clk). Used only for the controller's own internal data-capture logic, after data has already arrived back from the flash chip.

    The SDK default (Mode 0 + SAMPLE_EDGE_SEL_FLD = 1) is not a conflict, both happen to mean "rising edge," just of two unrelated clocks. The premise that "1 means falling edge" is incorrect; it means rising edge.

    Is the sampling delay register based on the OSPI reference clock (166 MHz)? Can sampling only be delayed, not be advanced?

    Yes. The delay field in RD_DATA_CAPTURE_REG (DELAY_FLD) counts its delay in units of ref_clk cycles.
    It can be only delayed (need to confirm)

    Both AM2612 TRM and the external Flash documentation state that only modes 0 and 3 are supported — customer tried to configure as mode 1 and it works normally. So why does configuring mode 1 still allow normal communication?
    Customer ran some examples using this configuration?
    I will get back to you on this

  • Hi Lakshya,

    For the third question

    Customer ran some examples using this configuration?

    Customer is configuring mode 1 in their own project, no our sdk demo. The external flash is GD25WQ32E.

    Best Regards

    Kita

  • Hi Kita,
    I need to check this with the hw team.

  •  
    Can help in Question 3

  • Hi Rijohn,

    Is there any update on this question?

    Both AM2612 TRM and the external Flash documentation state that only modes 0 and 3 are supported — customer tried to configure as mode 1 and it works normally. So why does configuring mode 1 still allow normal communication?

    Best Regards

    Kita

  • Hi Kita,

    Apologies for the delay!

    • While the internal OSPI controller hardware can technically handle other modes, the AM26x device design is only officially validated and timing-closed for Mode 0 and Mode 3.

    • Because the paths for other modes are not timing-closed, we cannot guarantee reliable hardware performance across varying temperatures ,voltage and higher operating frequencies. Even if Mode 1 appears to work correctly in a specific test environment, it risks becoming unstable under different production conditions.
    Therefore, to ensure complete system reliability, device support remains strictly limited to Mode 0 and Mode 3 as outlined in the TRM. We highly recommend configuring your design to use one of these supported modes





    Thanks & Regards,
    Rijohn