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.

AM3352 Board Bring up issue

Other Parts Discussed in Thread: AM3352, TPS65910

Hi,

We need help in debugging our AM3352 custom board.
The board powers up properly but we are not able to debug the board.

We have modified the Starterkit GEL file to our DDR3 device.
Whenever we try to run the GEL script CCS gives below error.

CortxA8: Trouble Writing Memory Block at 0x4c0000e4 on Page 0 of Length 0x4:
(Error -1065 @ 0x4C0000E4) Unable to access device memory.

It seems the GEL script is struck at writing the register "EMIF_DDR_PHY_CTRL_1_REG".

Searching through several E2E posts we have tried some procedures like
Run->Reset->CPU Reset(HW) and then run the script or disabling the MMU, data cache etc.
But nothing seems to be working.

In the GEL script tried to Comment out the "EMIF_DDR_PHY_CTRL_1_REG" register write
and tried to access the 0x4C0000E4 memory, but it says Target failed to read

> > 4c0000e0:   Target failed to read 0x4C0000E0 [code=0x1]
> > 4c0000e4:   Target failed to read 0x4C0000E4 [code=0x1]

Could you please let us know what could be the reason that GEL file is
not able to execute EMIF_DDR_PHY_CTRL_1_REG write?

Best Regards
Kummi

  • Hi,

    Can you try with the original GEL file?
  • Hi Biser,

    Thank you.

    We have tried with the original Startekit GEL as well as
    AM335x EVM(rev 1.5) Gel file, but we get the same error.
    "Trouble Writing Memory Block at 0x4c0000e4"

    Best Regards
    Kummi
  • I have asked the DDR experts to comment what could cause this.
  • Likely there is no domain power.

  • Hi,

    Thank you for the info.

    As per your suggestion we have checked the Power domains.
    Including VDDS_DDR, all the power domains seems to be fine.

    VDD1          1.105V
    VDD2          1.086V
    VDIG2         1.8V
    VAUX33       3.3V
    VMMC         3.317V
    VIO_SMPS  1.503V (VDDS_DDR)
    VAUX33       3.243V
    VDAC         1.801V
    VPLL          1.820V
    VAUX1        1.811V
    VAUX2        3.317V
    VRTC         1.821V

    We are using TPS65910A3A1(VIO=1.5V) with DDR3L(MT41K256M16TW-107) device,
    the DDR3 datasheet mentions it is backward compitable, so I think it will work at 1.5V.
    Just in case please let us know if there is any special consideration required
    when usign TPS65910A3A1 with DDR3L.
          
    Without clearing this DDR3 issue, we are not able to proceed with
    this project..we appreciate your help to solve this issue.

    Best Regards.
    Kummi

  • There should be no issue regarding the PMIC or running DDR3L at 1.5V. Is this just one board you have tried or are you seeing the same issue on multiple boards?
  • Hi Biser,

    Thank you for the information.
    Currently we have tried 2 boards and both have this problem.
    Tomorrow will try the 3rd board and let you know the result.

    The AM335x is not able to access its own register memory.
    When tried to acess EMIF Register area "0x4C000000" through the
    CCS memory window it says "Target failed to read 0x4C000000".

    If possible please let us know what else could be checked to know the problem.

    Best Regards
    Kummi
  • Please post the JTAG part of your schematics. Have you tried to boot the board from SD card or whatever your boot source is? Do you see anything on the debug terminal?
  • Hi Biser,

    Thank you for the advice.
    We were able to solve the issue.
    It was the CAP_VDD_RTC connection. The board bringup tips Wiki page was very helpful.

    In this application, RTC feature was not used and the
    CAP_VDD_RTC was connected to GND(VSS).So after connecting 1.1V to CAP_VDD_RTC
    the AM335x worked properly.

    But this circuit was exactly same as our previous(2 years old) one in which
    CAP_VDD_RTC was connected to GND and RTC_KALDO_ENn connected to VDDS_RTC
    but it was working properly.



    I understand if RTC_KALDO_ENn is connected to VDDS_RTC
    then CAP_VDD_RTC should be connected to 1.1V, but if that is the case
    why was our previous board was working even though CAP_VDD_RTC was connected to GND.
    Is there any possiblity that AM335x would work in this connection also?

    Thank you
    Kummi,

  • You can see the correct connections here: processors.wiki.ti.com/.../AM335x_Schematic_Checklist I cannot comment on your previous board without seeing the full schematic, however I'm glad you solved your problem.
  • Please note that you cannot submit schematics for review on this forum. There is an official review request procedure, which must be done through your local TI representative/FAE. Schematic reviews are done by a different team.

    However from what I saw you have a signal named VRTC connected to RTC_KALDO_EN and the PMIC power rail is named PMIC_VRTC. As the schematic is not in text-searchable PDF format I cannot find if these two rails are the same. I think you should search in this direction.
  • Hi Biser,

    Thank you so much for the information.
    I am sorry for the schematics review. As the production time
    is nearby your suggestions are very helpful.

    I shall check on this connection.
    These two rails are same.

    Actually the RTC_KALDO_EN pin is connected to VPMIC_VRTC
    this VPMIC_VRTC comes from the PMIC(TPS65910) VRTC pin.
    And VPMIC_VRTC is also connected to VDDS_RTC.

    With regards to "CAP_VDD_RTC", is it OK to connect it
    directly to VDD_CORE(1.1V) when the RTC feature disabled?

    Best Regards
    Kummi
  • "With regards to "CAP_VDD_RTC", is it OK to connect it directly to VDD_CORE(1.1V) when the RTC feature disabled?"

    Yes, this is the correct connection.