• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 30 subscribers
  • Views 288 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

TLV320DAC3100EVM-U: White noise audible with no audio input

James
Prodigy 10 points
Part Number: TLV320DAC3100EVM-U
Other Parts Discussed in Thread: TLV320DAC3100

Hi,

When using the TLV320DAC3100EVM-U, I have found that with an 8ohm speaker connected I can hear a constant white noise sound coming from the speaker. It is not loud but it is noticeable at arm's length in a quiet office environment. This is without any analogue audio being plumbed to the output and without any I2S digital audio input data being provided from an external source. 

Using CodecControl interface, I can see that a number of factors influence the volume of this white noise:

  1. muting the DVol (DAC volume control register) removes some noise
  2. opening the switch (0x30, Page 1, Reg 35, bit 6) that plumbs DAC audio into the analog attenuator or setting the analog attenuator to maximum attenuation has the same effect of further removing noise
  3. reducing gain of the Class-D speaker driver further reduces noise

CodecControl config

If steps 1, 2 and 3 are taken then the noise is reduced to such a level that I need to be within 10cm of the speaker to hear any 'hissing' in a quiet office environment; I'd consider this acceptable. In the figure above, none of those steps are taken and the white noise is 'maximised'.

If I observe the positive and negative sides of the class D output driver as it connects to the speaker in the time domain and trigger against one of those sides as a reference then I see the other side moving around relatively. It would seem that there is a degree of correlation between the quality of the 'edge synchronisation' of the drive waveform and the audibility of the noise. One edge can be often seen to be ~30ns misaligned with the other such as in the figure below.

spkr+ & sprk- with >1sec persistence time

Just wondering if anyone has had the same experience when using a TLV320DAC3100 or the TLV320DAC3100EVM-U and was there a way around it? 

  • 0
    TI__Guru* 95335 points

    Hi James,

    Do you have test equipment to measure the noise on the output, after an LC filter, and check if it is as expected based on the device spec?

    In addition to that, can you try checking if the noise is the same before and after playing audio? I mean to power and initialize the device, then check the noise, then play something and stop it and check the noise again.

    Regarding the Class-D is basically a BD modulation that's changing the duty cycle reciprocally, so I think your observations may be as expected.

    Best regards,
    -Ivan Salazar
    Applications Engineer

  • 0 in reply to Ivan Salazar
    Prodigy 10 points

    Thanks for getting back to me Ivan.


    The figure below shows the noise measured without any applied input signal. The yellow trace is with the class-D driver at 24dB gain, the red trace has it set to 6dB gain, attenuator is 0dB in both cases. I should note that this is taken after an LC differential output filter (10uH, 1uF). I suppose in the region up to ~10kHz then it would largely agree with the 87dB SNR spec for 6dB gain as shown in the datasheet. That said, at 6dB class-D driver gain the audible white noise was not really an issue - mainly at the higher couple of gain settings. Perhaps this is normal after all, I was quietly hoping there might be some tips for improving it though.

    6dB (red), 24dB (yellow) class-D gain

    Noise before and after audio is interesting. Before any audio has been played, changing the DVol setting doesn't seem to affect the white noise volume. After sound is played using the "beep" generator the noise volume seems to have increased slightly and now muting the DVol setting does lead to a slight reduction in noise volume. I'd like to better understand why this might happen.

  • +1 in reply to James
    TI__Guru* 95335 points

    Hi James,

    Thanks for sharing your test results, like you mentioned this looks like the expected results.

    Regarding the noise before and after playback, this is basically due to the DAC, once it's feed with some data it will keep a bit higher noise floor at the output.
    This is expected as well.

    Best regards,
    -Ivan Salazar
    Applications Engineer