• State Resolved
  • Locked Locked
  • Replies 1 reply
  • Subscribers 31 subscribers
  • Views 676 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.

PCM3168A performance issue

Todd F
Intellectual 590 points

Other Parts Discussed in Thread: PCM3168A, DIR9001, THS4521, THS4524

My customer has designed in the PCM3168A Audio Codec and is experiencing poor in circuit performance.

 I am on my second PCB spin, and I am getting very poor performance.  For this new board, we have only populated the clock, one CODEC, and input buffer, and an output buffer. Input has a low pass filter at around 15K, the output has a 7K low pass.  We are shooting for a passband of 150 Hz to 7KHz.

We are using a 32K effective sample rate, the 8MHz Codec sample clock has 2.5ps rms of jitter. I would like to see 90db or better SNR.

I have successfully interfaced a DIR9001 to the CPU and to the Audio Precision 2700 system. When I use the audio precision to inject an analog signal to the front end of the codec (the buffer amp is a THS4521 quad differential amp), the resulting SNR+noise curve results in only around 65 db signal to noise ratio. I was expecting at least 90 - 95 db from this arrangement. See the bitmap labeled "-002 codec A .bmp".

When we allow the audio precision to inject a digital test signal into the CODEC via the DSP, the output measurement of the codec, when buffered by an AD8397 is only around 60 db as well. The plot labeled "-002 codec A output amplifier.bmp" shows the output plot from the Audio Precision 2700. Again, I was expecting SNR in excess of 90 - 95 db.

Can you think of anything off the top of your head that can compromise the CODEC performance? I don't think the problem is with the buffer amps. I would not expect the input and output performance to be so similar if it was a buffer amp.  If I had not measured the jitter, I would have expected a noisy clock was the culpret, but I think 2.5 ps of jitter is quiet enough to deliver a 7K signal bandwith at over 100 db SNR.

Can you also forward this email to one of your audio experts, and have them send me a list of information they would like to see in order to get to the bottom of this.

(let me know if I need to get the schematics from the customer to forward on for you to review)

Here are the AP screen shots:

  • 0
    TI__Genius 16270 points

    Todd,

    On the input filters, capacitors such as C209 (0.047uF) are hooked up to both outputs of the THS4524 devices.  When connected this way, the capacitor's center is rotating around the common - mode, and this results in the capacitor's value effectively being double.  This would mean the cutoff frequency is much higher than 15kHz, which would create problems at the input.  I would recommend reducing the capacitors to a value of 22nF for example, to ensure the cutoff frequency of the low pass filter is correct.

    Please let me know if there are additional questions/concerns.

     

    Thanks and best regards,