• State
  • Locked Locked
  • Replies 4 replies
  • Subscribers 30 subscribers
  • Views 255 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.

TAS2553: output still active after sound file is played

Jens Tyslik
Intellectual 1610 points
Part Number: TAS2553


Hi Ivan, 

I received the below issue, could you please take a look at it? 

We realized by measurements that after a sound file is played, the output of TAS2553 still shows some activity. In the following picture you see the sound file first (signal with higher level) which is used as feedback to the user after a button is pressed. Directly after that we could observe some activity for approx. 3sec, with lower level. The signal was measured directly at one of the two speaker outputs. Used is a 8 Ohm speaker, there is no audible noise, since the output volume is already rather low. 
Is that a feature of the device, or did we configure it in a wrong way? 

Thanks, 
Jens, FAE

  • 0
    TI__Guru** 110125 points

    Hi Jens,

    Could you please capture the input clocks during this test? I wonder if the input clocks stop after this 3sec delay and that's what disables the output of the amplifier.

    Best regards,
    -Ivan Salazar
    Applications Engineer

  • 0 in reply to Ivan Salazar
    TI__Intellectual 1610 points

    Hi Ivan, 

    currently the PSR-13N08SM-YQ is used as speaker with 8 Ohms. No audible noise can heard, however this may be because the gain is quite low and the sound itself is already pretty quiet. 
    In addition to the output signal (OUT-) also the DIN of the device (pin D3) and the I2C bus were recorded:
    - CH1: OUT- (connection to the speaker)
    - CH2: DIN (Audio Interface from CPU)
    - CH3: SDA (I2C Data)
    - CH4: SCL (I2C Clock)



    The output of the codec already shows activity, although on the DIN (Audio Interface) of the CPU nothing happens yet. This causes some confusion. Could the I2C access activate anything in the codec before the activity on the output starts? 

    Below also what happens on the I2C bus: 

    Add.   Register

         Data

    1.

    0x80   0x06

          0x00

    2.

    0x80   0x02

          0x69

    3.

    0x80   0x02

          0xE9

    4.

    0x80   0x0D

          0xA9

    5.

    0x80   0x0E

          0x28

    6.

    0x80   0x02

          0xE8

    7.

    0x80   0x01

          0x04

    8.

    0x80   0x01

          0x00

    activity 1. to 3. seem to be configuration (blue), the actions 4. to 8. then show the completion of the configuration (yellow): 

    What we do not really understand is why in the configuration two times in a row different values are written into the same register: 

    2.

    0x80   0x02

          0x69

    3.

    0x80   0x02

          0xE9

    And why in the yellow block (completion of the config) again two times different values are written to the same register: 

    7.

    0x80   0x01

          0x04

    8.

    0x80   0x01

          0x00

    Could you please help to shed a bit more light on this? 

    Thanks, 
    Jens 

  • 0 in reply to Jens Tyslik
    TI__Guru** 110125 points

    Hi Jens,

    I'll check at your latest inputs and provide further comments by the end of Friday.

    Best regards,
    -Ivan Salazar
    Applications Engineer

  • 0 in reply to Ivan Salazar
    TI__Guru** 110125 points

    Hi Jens,

    Please noticed this is a Class-D amplifier, the output will start switching idle as soon as it is enabled, regardless of data input.
    The device has an internal boost, seems like the output starts with boost enabled right at the beginning and then when input data stops it disables the boost. This behavior should change based on the configuration for auto-pass through register 0x14.

    These commands you shared come from EVM?

    • 0x02 = 0x69 disables Class-D
    • 0x02 = 0xe9 enables Class-D this may be just to reset the Class-D?
    • 0x01 = 0x04 set device in Mute and disables SWS
    • 0x01 = 0x00 takes device out of Mute, while keeping SWS disabled

    The EVM GUI suggests this as init script:

    # 32-bit 64-fs
# Slave Reg Value
w 80 01 12
w 80 08 10
w 80 02 EA
w 80 03 5D
w 80 04 00
w 80 05 10
w 80 06 00
w 80 07 C8
w 80 09 00
w 80 0A 00
w 80 12 15
w 80 14 0F
w 80 01 10

    I'm not sure on what is the issue you observe, can you clarify?

    Thanks and Best regards,
    -Ivan Salazar
    Applications Engineer