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.

quickStartOMAPL1x_rCSL McaspEcho problem

Other Parts Discussed in Thread: OMAP-L138

Hi,

I'm trying to develop an audio application on LOGICs OMAP-L138 EVM board with TI's C6748 dsp.

Now I found the quickStartOMAPL1x_rCSL samples that would be a good starting point.

But there is a problem when running the MCASPECHO sample on my baseboard. The output is covered with a lot of in-band noise.

I've tried two versions of the LOGIC board, everytime the same problem. So the boards shouldn't have an error.

I also checked some clocking and DAC settings in the codec.c and mcasp.c but found nothing to improve  this issue.

I don't understand what the problem is because this is a TI sample that should work.

I added a Jpeg of to illustrate that (showing input (blue) and output (pink) of a sine).

Hoping for some ideas...

Thank you and regards

  • Hi Martin,

    I think that what you are seeing is the correct operation of the sigma-delta DAC in the AIc3106 codec and that the noise is in fact out-of-band.

    Regards,

    Donald

  • Hi Donald,

    Sorry for the delay in responding and thank you for your answer.

    I don't know what I've done. Last time I did a digital filtering of the output (fg= 20kHz) and there was a lot of noise left in-band.

    Now I tried again and, as you said,  the noise is gone.

    There must had been a mistake of mine last time (misconfiguration of my osci e.g.) , so thank you for your help.

    Regards,

    Martin

  • Hi Martin,

    Glad you've sorted your problem. It was bugging me that I couldn't find the attached images sooner, but I've just found them and thought I'd show you even though it's probably too late to matter.

    These images are of aic3106 codec output waveforms generated using samples of a 1 kHz sinusoid. The 'raw' output appears noisy on a 'scope but an FFT (performed using MATLAB) on data captured by the 'scope shows that noise to be out-of-band (blue trace). Filtering the codec output using a first order analog CR low pass filter cleans up the appearance of the sinusoid on the 'scope and the corresponding reduced noise spectrum is the red trace.

    Regards,

    Donald

    outofbandnoise.docx
  • Hi Donald,

    thank you for the attached pictures and the support.

    I did not see that you answered to my post again, so my answer is delayed, again.

    Unfortunatly my problem is not completely solved yet.

    I don't have an analog filter here and so I still use the digital one from my osci. As described above this reduces the noise in the range from 2 kHz to 20 kHz.

    But underneath the 2kHz boundary there is still noise within my signal. I don't know if that has something to do with the digital filter.

    IF I have the time to build one I'll test an anlog version.

    Regards,

    Martin

  • Hi Donald,

    I finally tested an anlog filter and the test results are are good, no audio noise left.

    I don't know how it had to do with the digital filter of my osci but there must have been the problem in any way.

    So, thank you again for your support.

    Regards,

    Martin