• State TI Thinks Resolved
  • Locked Locked
  • Replies 2 replies
  • Answers 1 answer
  • Subscribers 55 subscribers
  • Views 177 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.

LAUNCHXL-F28379D: ADC cross talk issue

Kayala Naga Venkata Sandeep
Prodigy 90 points

Part Number: LAUNCHXL-F28379D

Hello all actually i am using LAUNCHXL- f28379D development board for my project i am using ADC B and ADC C channel AB2,AB3,AB4 in AC2,AC3,AC4 while i am sampling the third channel AB4 and AC4 iam getting exactly half of count in first channel .means when i am applying 3.3V to AB4 I am getting 2200 Count in AB2 channel ok how should resolve this cross talk I am ADC clock divided by 4 and sample window 500 and sample time 5000nonosec I am triggering these using ePWM 5kHz  how resove this issue ?

  • 0
    TI__Expert 7160 points

    Hello,

    The relevant expert is out of office by next week. Please expect reply when he comes back to office.

  • 0 in reply to Stevan Duraskovic
    TI__Expert 8450 points

    Hello,

    The behavior you describe is most likely not PWM-related crosstalk, but ADC sample-and-hold charge sharing on the F28379D, especially if the analog source driving AB2/AC2 is high impedance.

    When the ADC switches from one channel to another, the sample capacitor can retain charge from the previous channel. We added note that for even-to-even or odd-to-odd channel transitions, the new channel can start close to the voltage of the previously converted channel, while the acquisition window must be chosen based on the actual source impedance. In your case, B4 → B2 and C4 → C2 are both even-to-even transitions, so if AB4/AC4 are at 3.3 V and AB2/AC2 are weakly driven, AB2/AC2 can read artificially high.

    I would recommend checking the source impedance on AB2/AC2, increasing ACQPS further, trying a dummy conversion and using the second sample, and reordering the SOC sequence to avoid B4→B2 / C4→C2 transitions. If the input comes from a resistor divider or weak sensor output, buffering it with an op-amp is often the best fix.

    Best Regards,

    Masoud