• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 1 answer
  • Subscribers 82 subscribers
  • Views 644 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.

F2812 Flash process stalls after erase

Francis Kwok
Prodigy 150 points


Hello,

I was hopefull somebody can help me out.  I have a set of boards in which some of then exibite this behavior:

I'm trying to use the On-Chip Flash Programmer (CCS 3.3)  to flash our bootloader onto the chip.  However, after the erase operation, the process stalls.

I scope the XRS pin and it seems like the watchdog is reseting the chip.  The XRS pin is solid high from the time I connect to the chip to right after the erase process.

I'm thinking maybe it's having troubles getting back into JTAG mode? 

I have the following on the JTAG port:   4.6k Pull-up on EMU0 and EMU1 and a 2.2k pull down on TRST.

Is there any other reasons anybody can think of that will make the chip come out of JTAG mode after the erase process??

Thanks in Advance,

Francis

  • 0
    Intellectual 305 points

    Hi Francis,

    I suggest you to refer to these documents and then observe if you are still observing the issues:

    TMS320F281x Boot ROM Serial Flash Programming

    Running an Application from Internal Flash Memory on the TMS320F281x DSP

    Thanks,

    Shilpa

  • 0
    TI__Guru* 92800 points

    Francis Kwok said:
    I'm trying to use the On-Chip Flash Programmer (CCS 3.3)  to flash our bootloader onto the chip.  However, after the erase operation, the process stalls.

    Hi Francis,

    When you program with the on-chip flash programmer it actually downloads an algorithm to the device and the CPU performs the programming.  When you say it stalls after erase - does the erase itself complete?  If so, when does it stall?  Or does the erase itself not complete?

    If the erase is not completing, I would double check the flash plug-in settings.  Make sure the clock is setup correctly.  I also recall the version of the API is reflected in the name of the programming algorithm (I can't check at the moment).  If so the latest version is 2.10.

    Francis Kwok said:
    I'm thinking maybe it's having troubles getting back into JTAG mode?

    The device doesn't switch out of JTAG mode. 

    Are you able to do anything with the device?  Reset the part, check memory contents etc?

    -Lori

  • 0 in reply to Lori Heustess
    Prodigy 150 points

    Thanks for the replies,

    I've tried a few new things.  I tried the Erase Only option and it works okay.  However, when I do the "Program, Verify" option it craps out right away.  The indication it crapped out is by monitoring the XRS pin and instead of solid high, I can now see the watchdog pulse the line.  On some of the board what we're having no problems with, the XRS remainds high. 

    The chip after the failed program is still in a running state and when I exit the programming dialog I can still preform operation like Reset the part, halt the chip, etc...

    I've confirmed that I'm using the latest Flash API (2.10).  The chips i'm trying to flash are "E" Rev. but from the Flash API documentation, the API should have no problems with it.

    Thanks in Advance!

    Francis

  • 0 in reply to Francis Kwok
    TI__Guru* 92800 points

    Francis Kwok said:
    I've confirmed that I'm using the latest Flash API (2.10).  The chips i'm trying to flash are "E" Rev. but from the Flash API documentation, the API should have no problems with it.

    Wow - that's a pretty old device there.  API V2.10 is backward compatible to Rev C silicon so that's not the issue.

    Were the devices obtained from a TI distributor or another source like a broker?  We have seen issues with devices that are not from a normal TI channel.   A note here from TI

    http://focus.ti.com/general/docs/gencontent.tsp?contentId=66266

    Francis Kwok said:
    The indication it crapped out is by monitoring the XRS pin and instead of solid high, I can now see the watchdog pulse the line.  On some of the board what we're having no problems with, the XRS remainds high. 

    It sounds like the devices may be fetching an illegal opcode or experiencing a power surge.  I suggest the next thing to check is that all connections are good - power/current is good - no glitches when you start programming - clocks are good.

    -Lori

  • 0 in reply to Lori Heustess
    Prodigy 150 points

    Hi Lori,

    Yeah, we dug up some old stock from a previous project so I think they’re not counterfeit.

    We isolated the problem to the 3.3V rail power supply.  We’re trying the power supply design from the 2812 ezDSP reference design.  The power supply (TPS767318) is sending a reset after the 3.3 rail drops from the flash operation. 

    Some boards show that it's regulating well (the boards that work) and the boards that don't work show a very slow response to voltage dip.  So I guess we'll investigate why some boards are can’t source enough power for the flash operation.

    Thanks for your help!

    Francis