• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 1 answer
  • Subscribers 82 subscribers
  • Views 254 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.

TMS320F28377D: Incorrect acquisition

Clément Letonnelier
Expert 2986 points

Part Number: TMS320F28377D

Hello,

In our production batch of 30 boards, we have one board providing incorrect results.

Our design is summarized as follows:

  • All acquisitions are single-ended
  • An external 3V reference is used
  • VDDA is supplied by a 3V3 regulator
  • The acquired voltage is stabilized and is issued from a voltage divider with an OP-amp as buffer

On the board, I have measured 0.940V at the via that is interconnecting to the ball of the component.

However the acquired value is equal to 935 LSB which corresponds to roughly 0.685V so off the chart.

We could suspect an issue with components soldering maybe, however all our other interfaces of the microcontroller are working perfectly.

An idea/suggestions wrt. what could be wrong ?

Clément

  • 0
    TI__Mastermind 41690 points

    Hi Clément,

    What is the SYSCLK, ADC speed and ACQPS selected?  Also what channel was used during conversion and did you try increasing the ACQPS to see if the conversion improved?

    Thanks and regards,

    Joseph

  • 0 in reply to Joseph Casuga
    Expert 2986 points

    Hi Joseph,

    The answer are as follows:

    • SYSCLK: 200MHz
    • ADCCLK: SYSCLK / 8.5
    • ACQPS: 200 SYSCLK periods

    I did not try to increase ACQPS as it works perfectly fine on the 29 other boards and the 200 SYSCLK periods are, from a computation standpoint, considered sufficient to have a correct result.

    I have the issue on all the channels that I use on the ADC for that board, ie. most of the TMS320F28377D channels.

    Clément

  • 0 in reply to Clément Letonnelier
    TI__Mastermind 41690 points

    Hi Clement,

    Thanks for the information.  From what you provided, a sampling time of 1uS should be more than sufficient since signal is also buffered by an op-amp, and I'm assuming you have also measured the 3V reference between VREFHI and VREFLO pads.  Only thing that comes to mind is probably contact/resistance issue between the ball and the pad, but would be hard to prove without removing and re-soldering the part.  I do not know what this process will also do to the rest of the balls on the component.

    Regards,

    Joseph

  • 0 in reply to Joseph Casuga
    Expert 2986 points

    Hi Joseph,

    This is what I was suspecting.

    I had measured the VREFHI reference indeed, not the VREFLO but this is a ground plane so only if we had an incorrect PCB I wouldn't expect the reference to be incorrect.

    So yeah I tend to think like you it is most certainly a soldering issue, we are going to investigate this way and try to perform a repair operation to remove and resolder the part. We already did it on another board.

    Clément

  • 0 in reply to Clément Letonnelier
    TI__Mastermind 41690 points

    Hi Clement,

    Closing this thread for now and in case resoldering the part did not solve the issue, please post more details in the forum (like channel# and ADC module or snapshot of the schematics - if at all possible) so we can have a more detailed review on the issue.

    Thanks and regards,

    Joseph