• State
  • Locked Locked
  • Replies 5 replies
  • Subscribers 82 subscribers
  • Views 768 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.

SDFlash programming fails

Patrick Poulin
Prodigy 230 points
Other Parts Discussed in Thread: TMS320F28027, UNIFLASH

Hi

Here is my situation.  I am using the SDFlash emulator to program some devices mounted with Piccolo F28027 on them.  Some of the cards work perfectly.  Some others, they are giving us trouble at programming stage.  When I check the Erase+Program+Verify checkboxes, the process does not work the first time then works once every 2 times.

  • fails programming with error -1041
  • pass
  • fails programming
  • pass
  • fails programming
  • and so on

A workaround I found to prevent this bug on these cards, is to do indepedently the Erase, program and verify steps.  When I do so, in addition to the fact that a delay is added between the steps, the SDFlash seems to disconnect and reconnect itself for each step.  This way, I get no error for each programming step (100% success).

After doing so, I even got some cases when the Erase+Program+Verify sequence is then a pass at each iteration, event with the combined operations.  If I reset the card, I get back to the previous behaviour, where the first E+P+V sequence fails at programming.

It's really weird and since we haev 50% of the devices having this behaviour, it is quite annoying.  Can anyone help me find a way to solve this?

Thank you.

Pat

Additionnal information

  • SDFlash version 1.66
  • CCSv5 working perfectly (erase/programming) always good
  • using TMS320F28027 rev A
  • Flash API version used for SDFlash28027_xxxxxx.out seams to be v1.00  (didn't find 2.00 versions to try)
  • Using XDS510LCUSB rev A
  • 0
    TI__Guru* 91235 points

    Patrick,

    You may want to try the CCS Uniflash tool.  It is based on the same back-end as code composer, but doesn't require a full code composer install and is kept up to date along with CCS.  It will work with the XDS510.

    http://processors.wiki.ti.com/index.php/Category:CCS_UniFlash

    Otherwise, I'm not sure what is going on with SDFlash.  I have seen it be flaky when programming using the serial option, but not JTAG.  With the release of Uniflash we have been moving away from SDFlash.

    -Lori

  • 0 in reply to Lori Heustess
    Prodigy 230 points

    Lori

    I will try Uniflash but this is far from being a great solution for me right now since we have programmed a lot of cards with F2809 and F28027 on it in the past and the production testbench uses the command prog28x.exe from the SDFlash utilities which is automatically launched from the testing software.

    Just to add some information, this bug wasn't occuring before we added a new manufacturer for our card, and since it is supposed to be exactly the same pcb and same electronic, we are quite worried about the programming process which is failing here and there.

    Uniflash may be a way to validate that the failure comes from SDFlash itself, but would not give us all the answers.

    *EDIT*     I just tried Uniflash...  since it is more than 250Mb, and consist in a large installation process (which failed by the way on my side because of either a corrupted file or anti-virus limitation), it will surely not be a suitable solution for our production and the testing department at our manufacturers site.

    Thanks

    Pat

  • 0 in reply to Patrick Poulin
    TI__Guru* 91235 points

    Patrick Poulin said:
    Just to add some information, this bug wasn't occuring before we added a new manufacturer for our card, and since it is supposed to be exactly the same pcb and same electronic, we are quite worried about the programming process which is failing here and there.

    That is very strange.  I have no idea why a board change would effect SDFlash but not CCS.  The under-the-hood algorithms are the same, just the interface is different.  Perhaps SDFlash is pointing to different drivers.  

    Patrick Poulin said:
    Uniflash may be a way to validate that the failure comes from SDFlash itself, but would not give us all the answers.

    If I understand, CCS is able to program the devices without issue, so this seems to point back to SDFlash.

    -Lori

  • 0 in reply to Lori Heustess
    Prodigy 230 points

    Yes Lori, it seems to point back to SDFlash but I wanted to seek for answers here to since SDFlash is working properly with the other cards.  I was looking for answers about some delay or disconnect/reconnect information between steps of programming.

    I have also opened a ticket at SD support but wanted to look for answers with the pros here...

    Pat

  • 0 in reply to Patrick Poulin
    TI__Guru* 91235 points

    I understand. I wish I had another suggestion at this time, but I will post back if I think of anything.