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.

TLV320AIC3104: White Noise problem

Part Number: TLV320AIC3104

Hi, TI support team

Please review below issues.

We are currently reviewing the TLV320AIC3104, and the mode works well without any problems.

One problem was identified during operation verification: white noise appears at high volume levels.

According to what has been confirmed so far, it is confirmed to be the DAC's own noise.

• Even if the DAC front end is muted, white noise occurs.

• If you connect the rear end of the DAC to another input and set it to the same level as the DAC vol, there will be no white noise.

Considering this situation, it is judged to be the DAC's own noise.

The confirmed details are as follows.

1. As a special feature, if you increase the quiescent current (current default) of R109 to 50% or 100%, the white noise becomes larger as it increases.

Currently, R93 is being used as a single speaker phone in the configuration below, and DAC_L1 and PGA_L are enabled in the Right-LOP/M mixer.

2. The volume of DAC-related R89 is set to 0dB, and when the volume is lowered, the white noise level also decreases.

If the gain of R89 is 0 to -3dB, white noise occurs. It is believed that R43 has no effect.

3. When a call is connected while reviewing the function, gain/path is set, but if the same 0dB is written to R89 again while noise is occurring, it disappears. As a result, when setting call path and gain/path, there is no effect

even if 0dB is written last in R89 or used interchangeably with path connection. If you write after the call path is set and noise occurs, it will disappear.

Q. What causes white noise?

Q. Please tell me how to improve noise.

[additional information]

DAC settings (R37) are set to 0xA0.

Only Left DAC is powered on and Right DAC is not used.

0x00@ R41.

If you look at the overall block diagram on page 21 of the TLV320AIC3104 datasheet,

As shown below, “*DACs must be powered down in order to use ADC processing” is written.

Q. What is the scope of ADC processing?

Q. Please check what settings and when to power down the DAC during operation. (Please tell me how to set the power control of the DAC according to the ADC processing operation.)

Thanks.

Regards,

MJ

  • Hi,

    -. It's hard to say what causes that, are you seeing this on several units or just one setup? The SNR spec is with 0dB setting on all control in the path. Maybe use AP and measure the SNR for this unit.

    -. Noise could come from anywhere, but here is an apps. note that talks about noise in audio. 

    https://www.ti.com/lit/an/slaa749/slaa749.pdf?ts=1694723618295&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FTLV320AIC3104

    Have you tried in your case, reduce volume in R43 and gain up in R89?

    -. This is referring to record path shown below, in this case DAC must be powered down. ADC processing is referred to the ADC path/block - the decimation etc....

    In order to use this recording path, playback will not be enabled.

    Regards

  • read_reg.zip

    Dear TI Support Team.

    For this issue(White noise from DAC_L1(R89), I'm attaching the reg value for our product when it happens.

    So far, according to our testing, the white noise source is DAC_L1(R89) .

    If I do mute R43 in the previous register, I can hear the white noise.

    In our app, DAC_L1 is used for a melody ring play from CPU.

    This DAC_L1 is mixing for Right LOP/M with PGA_L(R88). 

    The PGA_L(R88) is connected to R15 PGA / MICx by bypassing ADC and 3D processing block. This is used for voice path.

    So the melody ring or voice can be played on Right LO .

    Under this condition, If we disconnect or mute DAC_L1(R89), white noise is disappear. 

    Seemingly, the issue seems to be DAC_L1(R89) itself. (Currently 0dB setting)

    The white noise is not a specific narrow band sound. It's pure white noise.

    In a case, when I tried to do below, the noise is disappeared.

    1. powered down - DAC_L1(R37.7) 

    2. Set  volumes, gains in that audio path.

    3. Route DAC_L1(R89.7)

    4. Set Volume - DAC_L1(R89.6:0)

    5.powered up DAC_L1(R37.7).

    --> I can't hear the white noise in this case. However, this way is always not effective in every voice path case.

    And the noise level is depending on R109 (0--> 50% -->100% is more loud).

    Currently we are using DAC_L1(R89, 0dB) , Right LO(R93, +1dB). for the melody ring path

    Decreasing R89 and increasing R93 are another option to improve the white noise. But I'd like to improve DAC_L1 noise first.

    Please advise me about this issue

     This is our audio block.

    Best regards

    Byunghee Sohn

  • 6012.read_reg.zip

    Dear TI Support Team.

    For this issue(White noise from DAC_L1(R89), I'm attaching the reg value(read_reg.zip) for our product when it happens.

    So far, according to our testing, the white noise source is DAC_L1(R89) .

    If I do mute R43 in the previous register, I can hear the white noise.

    In our app, DAC_L1 is used for a melody ring play from CPU.

    This DAC_L1 is mixing for Right LOP/M with PGA_L(R88). 

    The PGA_L(R88) is connected to R15 PGA / MICx by bypassing ADC and 3D processing block. This is used for voice path.

    So the melody ring or voice can be played on Right LO .

    Under this condition, If we disconnect or mute DAC_L1(R89), white noise is disappear. 

    Seemingly, the issue seems to be DAC_L1(R89) itself. (Currently 0dB setting)

    The white noise is not a specific narrow band sound. It's pure white noise.

    In a case, when I tried to do below, the noise is disappeared.

    1. powered down - DAC_L1(R37.7) 

    2. Set  volumes, gains in that audio path.

    3. Route DAC_L1(R89.7)

    4. Set Volume - DAC_L1(R89.6:0)

    5.powered up DAC_L1(R37.7).

    --> I can't hear the white noise in this case. However, this way is always not effective in every voice path case.

    And the noise level is depending on R109 (0--> 50% -->100% is more loud).

    Currently we are using DAC_L1(R89, 0dB) , Right LO(R93, +1dB). for the melody ring path

    Decreasing R89 and increasing R93 are another option to improve the white noise. But I'd like to improve DAC_L1 noise first.

    Please advise me about this issue

     This is our audio block.

    Best regards

    Byunghee Sohn

  • Dear TI Support Team.

    I have tested this one for another path below.

    I tested DAC_R1 instead of DAC_L1. But it's the same result

    And after disconnecting DAC_L1, DAC_R1, I routed DAC_R3 with PGA_L.

    However, white noise became louder than DAC_L1/ DAC_R1.

    Can you advise me about this?

    Best regards

    Byunghee Sohn

  • Hi,

    I'd expect the same from either DAC. Here are a couple of suggestion you can do:

    • Change your word length or bit depth from 16 to higher even to 32
    • Minimize your digital volume side and use your analog gain to compensate

    Regards.

  • Hi 

    Thank you so much for your reply.

    I'd like to know if the word length(16bit) can have an noise issue on DAC.

    Is there an issue on 16 bit word length ?

    In case of our application, data coming from Host I2S is fixed to 16bit.

    I also think that adjusting the digital gain and analog gain become an solution.

    However, I want to find out what the issue of the noise is.

    Regards

    B.H. Sohn

  • Hi,

    It's not that 16 bit causes noise issue on DAC, but rather the quantization noise is better with larger bit depth.

    You can search the web on this topic.

    As proposed above, you can use whichever method that works better for you.

    Regards.

  • Hi 

    Thanks for the exploration.

    We now used Low-pass filer in 2.4 Out-of-Band Noise of "SLAA749" that you recommended.

    And we can see the noise is almost removed under loud noise condition (high vol, 100% @ R109)

    You can close this ticket.

    Many thanks

    Byunghee Sohn

  • Good to know.

    You're welcome.