• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 2 answers
  • Subscribers 82 subscribers
  • Views 664 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

Problem with watchdog

Patrick Schneider1
Intellectual 655 points
Other Parts Discussed in Thread: DRV8301, UNIFLASH, MOTORWARE, CONTROLSUITE

Hello everyone,

at first my setup:

Hardware: Own Hardware (similar to LAUNCHXL-F28027 with BOOSTXL-DRV8301) and a custom BLDC motor

IDE: Code Composer Studio 6.0.0.00190 and GUI Composer 1.0.0.201308252300 (rcp-main)

Software: LAB10A

So here is my problem:

We developed an own hardware board. It is designed like the schematics for launchpad and boostxl_drv8301, should really be a 1:1 copy of it. We can program the controller and test connections with code composer studio and uniflash are perfectly working. But the software is somehow not starting.

After a few days of investigation we found out, that the watchdog is repeatedly triggered. The XRS also shows that. The timing is perfect and matches the predefined watchdog value. So the software seems to try to start but then the watchdog resets the controller. It's just LAB10A, the exact same software running on a launchpad is working.

Is there anything else than the out file we have to program? I think I remember that I read something about preflashed math routines on the launchpad, is that correct and do we have to program them on our own?

I thank you in advance and hope for an answer soon!

BR,

Patrick

  • 0
    TI__Guru** 118000 points

    Patrick,

    This sounds more HW related to me, usually the power supply (or external crystal if you were using). Or it could be the boot mode pins. Everything for the LAUNCHPAD examples in MotorWare is set-up to program and boot to flash, but in default configuration you need to be connected to JTAG. On the LaunchPad there are switches that allow it to boot and run from flash if you don't have a JTAG connection.

     

  • 0
    TI__Mastermind 22025 points

    Hi Patrick,

    From my reading it sounds like you're able to run the program when CCS is debugging the board, but you're unable to run standalone, correct?

    If that's the case then I concur with Chris.  You boot pins are probably telling the device to boot to RAM which is zeroed at power up.  This is causing the device to go to an ITRAP and then the watchdog is resetting the device.  Look at the Boot ROM reference for the F28027 device to find the boot pins and the proper state to drive them to enable a boot to flash.

    BR,

  • 0 in reply to Trey German289
    Intellectual 655 points

    Hello,

    thank you both for your answers. I'm aware of the three boot pins and different boot modes. They are pulled up and down according to the pdf.

    Thats not quite right Trey, even if I try to debug the software is always somewhere in the first lines of code, always resetting. The clock is internally (complete HW copy of the launchpad) and the clock seems to be right, because the watchdog reset interval is exact the amount of time it should take.

    So unfortunately that doesn't solve the problem. Do you have any other ideas or have had a similar problem somewhere else?

  • 0 in reply to Patrick Schneider1
    TI__Mastermind 22025 points

    Patrick,

    My next guess is that you are forgetting to copy code from flash to RAM before trying to execute out of RAM.  Do you have a call to a memcopy function near the top of main?  If so are there any calls to any other functions before it?

    BR,

  • 0 in reply to Trey German289
    Intellectual 655 points

    Hey guys,

    just to close this one: we finally found the error.

    Something got mixed up while ordering all the parts for the PCB. At the end we got the TMS320F28027PT on our boards instead of the TMS32F28027FPT, which - as we learned - has the math algorithms programmed to flash. So you're examples and our final software couldn't work because these codes were missing (explains the watchdog though).

    Unfortunately all this happened because in the schematics from controlSUITE the controller-package is also named TMS320F28027PT, so we lost merely one to two weeks because of that little typo. You should definately fix that in the next version of controlSUITE.

    Nevertheless thanks for your help and suggestions!

    BR

  • 0 in reply to Patrick Schneider1
    TI__Guru** 118000 points

    you know, I almost mentioned that but figured there was no way you had the wrong silicon...

    glad you found it....sorry it took so long!