• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 30 subscribers
  • Views 425 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.

PGA460: DSP Frequency Accuracy

Jerry%20Leung
Intellectual 905 points
Part Number: PGA460

Hi Team,

I have a customer planning on using the PGA460 in an ultrasonic distance sensor for robotic applications and have a question they would like to clarify:

The PGA460 data sheet indicates that the DSP frequency accuracy is +/-4%, and this is used as the basis for setting the drive waveform frequency delivered to the ultrasonic sensor.  However, piezo ultrasonic transducers generally specify +/-1% frequency accuracy.  The datasheet is somewhat ambiguous, but I’m wondering if the PGA460 tunes the actual drive frequency to the resonance peak of the transducer.  If not, how does one ensure that the transducer is driven at resonance?

Thanks for the help and insights.

Best regards,

Jerry

  • 0
    TI__Mastermind 27950 points
    Hi Jerry,

    The PGA460 automatically sets the coefficient values for the DSP's bandpass filter and lowpass filter based on the user selected values of the frequency, bandpass bandwidth, and lowpass cutoff. Therefore, the user does not need to worry about manually setting these values.

    The only scenario where the DSP is not automatically updated with the correct coefficient values is when the high-frequency (180-480kHz) mode is enabled. Do you know which frequency value your customer is using? For robot applications, I assume the low-frequency range of 30-80kHz, where the DSP is automatically set, so there should be no concern.
  • 0 in reply to Akeem Whitehead
    TI__Intellectual 905 points
    Hi Akeem,

    Thanks for the help, want to follow up with some clarifying questions from our customer:

    • Most piezoelectric transducers have a high Q, and they specify operation at +/-1% of the specified frequency for optimal performance. Performance drops off as you move off resonance.
    • The PGA460’s drive frequency and the FIR receive filter in the DSP are always matched because they are derived from the same master clock frequency. However, the overall accuracy of the master clock is +/-4%.
    • If the master frequency is 4% off, the transducer won’t be driven at resonance. Equally important, the reflected signal will not be at the transducer’s resonant frequency.

    The PGA460 data sheet contains the following sentence:

    "Transducer frequency measurement During the decay stage of the record interval a frequency measurement on the transducer node is performed to validate the performance and proper tuning of the transformer and transducer matching."

    Even if the transducer is driven with a frequency off-resonance, the ring-down frequency will be at the transducer’s resonant frequency. Measuring that frequency should indicate if the drive frequency is off-resonance (a frequency will be measured that is different from the programmed drive frequency). It appears that the PGA460 does not automatically adjust the drive (and FIR) frequency; however, by reading the measured frequency it should be possible to adjust the drive frequency until it matches the measured ring-down frequency.

    Would the above description be correct?

    Thanks for the help and best regards,
    Jerry
  • 0 in reply to Jerry%20Leung
    TI__Mastermind 27950 points

    Hi Jerry,

    The customer is correct about the majority of their points:

    1) Yes, some transducers have a high-Q, and performance will diminish if the transducer is excited away from the resonance. However, the center-frequency of this high-Q can drift across multiple units.

    2) Yes, the PGA460 drive frequency and DSP filters are always matched as they are from the same master. The master does have an accuracy tolerance of +/-4% at 16MHz. This translates to approximately +/-1kHz of error in the common low-frequency transducer range (30-80kHz).

    3) If the master frequency is offset to its margin extremes, the transducer's performance may vary in both generated sound pressure level, and maximum return amplitude of the reflected echo.

    4) The transducer frequency measurement is a system diagnostic tool intended to be used as an in-system diagnostic tool to check the status (temperature drift, long term usage drift, sensor impediment), and as a calibration tool as you've described. Be aware that transducers typically have a a mechanical and electrical resonance which may not always be matched. During the rind-down, the natural frequency should be observed, but you may find that a slightly offset driving frequency will yield improved echo amplitude.

    All of you concerns are valid with regard to PGA460 clock tolerance and the transducer's optimal resonance. That is why it is crucial to implement testing and calibration at production to check the performance of each module. This production check should include the ability to optimize the drive frequency based on the SNR of a known object for instance, or, alternatively, the SPL recorded by an external free-field microphone. The PGA460's system diagnostics of the transducer's frequency, decay, and voltage can also be used as checks at production.