• State TI Thinks Resolved
  • Locked Locked
  • Replies 10 replies
  • Answers 1 answer
  • Subscribers 31 subscribers
  • Views 766 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.

TAS5825M: Unwanted noise at low frequency input signals

Stephane Mevel
Prodigy 30 points
Part Number: TAS5825M
Other Parts Discussed in Thread: , PUREPATH-CMBEVM

Dear TI team

We are integrating the amplifier IC TAS5825M into our new products in both BTL and PBTL modes.

We have encountered a peculiar phenomenon when we play sounds integrating low frequency signals with modulated amplitude. It happens when the general volume reaches a certain level.

At first, we were wondering if it was caused by our own electronics but we managed to recreate the same phenomenon on the EVM TAS5825MEVM + PUREPATH-CMBEVM in housekeeping configuration (USB input from PC, 24V DC power supply, 4 ohms load, same as our electronic system). Attached the EVM configuration file, filename “tas5825 config.ppc3”.

It seems that when the sound volume is at fade-in and fade-out phases, and at a certain general level, some faint noise can be heard on top of the amplified original signal. It is not easy to describe this noise, it is like a cracking noise, like if some quantification bits were added or missing. So, it generates some unwanted higher frequency noises, not acceptable especially for our high-end application.

We have generated a dedicated sound file to analyse this issue. Please find it attached “20HZ FS_Fade in_1Sec_Fade Out_1Sec.wav”.

Also attached for TI team to investigate, a recording of the output of the TAS5825M, filename “REC_20HZ FS_Fade In_1sec_Fade out_1Sec.wav”.
To ease the understanding of that issue, we have made a 50Hz high pass filtering of this file, attached as “REC_20HZ FS_Fade In_1sec_Fade out_1Sec_HP filter 50Hz 48db.wav”. One can hear clearly that there is unwanted noise at fade-in and fade-out phases, but not during constant volume phase. And one will notice that this effect is not random, it follows the input sound wave with the same pattern all the time.

We would be very grateful if you could help us understanding the root cause of this issue.

Many thanks for your support.

Best regards,

Stéphane.tas5825 config.rar

  • 0
    TI__Guru 52460 points

    Hi Stéphane

      About the PPC3 file, as I checked it seems the default settings of Housekeeping(2.0) process flow. Do we changed anything like enabled PVDD AGL/OTFB or anything else? 

      This process flow seems specially for 192KHz sample rate, may I ask that do we need to use such high frequency? And the audio source we used for testing, is also 192KHz sample rate? This frequency is not commonly used, how about we consider using like 48KHz or 96KHz process flow and have a try?

      You mentioned tested BTL and PBTL. The PBTL seems not having this 192KHz process flow, what we used for PBTL? And PBTL also has this noise problem?

  • 0 in reply to Shadow He
    Prodigy 30 points

    Hi Shadow

    Many thanks for your reply.

    Sorry for this late reply, I was out of the office.
    Here our answer below.

    “About the PPC3 file, as I checked it seems the default settings of Housekeeping(2.0) process flow. Do we changed anything like enabled PVDD AGL/OTFB or anything else? “

    PVDD AGL/OTFB is disable.
    But we noticed that enabling it does not affect the unwanted noise.
    We use the housekeeping mode to get as little TAS5825 processing as possible (low latency).

    We have tried different modes (process flows) and configuration on the EVM but still we get the same noise.

    “ This process flow seems specially for 192KHz sample rate, may I ask that do we need to use such high frequency? And the audio source we used for testing, is also 192KHz sample rate? This frequency is not commonly used, how about we consider using like 48KHz or 96KHz process flow and have a try?”

    We are developing a high-resolution audio system hence the need for 192kHz frequency.

    We have tried 48kHz and still the same noise during volume variation (fade in, fade out).

    “ You mentioned tested BTL and PBTL. The PBTL seems not having this 192KHz process flow, what we used for PBTL? And PBTL also has this noise problem?”

    Please correct us if we are wrong, but it looks like the housekeeping mode allows for 192kHz in PBTL mode in the PPC.

     

    Have you managed to replicate the issue on your side? And if so, could you guide us on how to solve this issue?

    Many thanks for your support.

    With kind regards,

    Stéphane

  • 0 in reply to Stephane Mevel
    TI__Guru 52460 points

    Hi Stéphane

      Seems not having much similar complain for this devices, I'm still trying to find out the possibilities. The noise issue also possibly related with speakers, not all the speaker type could hear clearly.

      When we adjusting Volume, are we adjust the input audio source, or do we adjust device's Volume settings? If we are adjusting device, we could have a try with the Volume Alpha settings. Set with longer Alpha time should able to help.

  • 0 in reply to Shadow He
    Prodigy 30 points

    Hi Shadow

    Many thanks for your support.

    Just a comment on your reply: we don't use or change the volume of the source or the TAS5825 or PPC during the test. The fade in  and out is part of the sound file.


    What we discovered is when we play music with low frequencies that are volume modulated (like wobbling if you see what I mean) this cracking noise can be heard clearly. It is not related to speaker type. It happens only during low frequency wobbles, only during fade-in and fade-out. When the volume is constant, there is no noise or distortion.
    This is why we generated this test file that can replicate what we hear with some music tracks.

    The recorded files that we provided you (see first post) have been recorded directly from the output of the TAS5825 EVM using a high quality sound card. So those recordings are not speaker related.

    Hoping this helps.

    Regards,

    Stéphane

  • 0 in reply to Stephane Mevel
    TI__Guru 52460 points

    Hi Stéphane

       Thanks for these info, I understand now. And I realized it quite possible related with 1 internal block. Could you please have a try with below I2C script? Put these script at the end of your initialization would be fine (below script use 0x98 as the device address, change it according to your board). I would explain more if it helps. Thank you.

    w 98 00 00

    w 98 7f 00

    w 98 3a f9

    w 98 3f 0f

  • 0 in reply to Shadow He
    Prodigy 30 points

    Hi Shadow

    Many thanks for your help.

    We have tried your suggestion and I am afraid the noise is still there.
    We are sending the full cold-boot + the 4 lines above at the end.

    Is there anything else we can try?

    Regards,

    Stéphane

  • 0 in reply to Stephane Mevel
    TI__Genius 15200 points

    Hello Stephane

    Can you run the following before you enter the device into play mode

    w 98 00 00

    w 98 7f 00

    w 98 3a f8

    best regards,
    Luis

  • 0 in reply to Luis Fernandes de Pina
    Prodigy 30 points

    Dear Luis

    Many thanks for your support.
    Good news, this has fixed the issue.

    Could you explain what this fix does? Many thanks.

    Best regards,

    Stéphane

  • 0 in reply to Stephane Mevel
    TI__Genius 15200 points

    Hello Stephane,

    You can send me an email at luisvf@ti.com if you want additional information. At a basic level its an internal register related to dynamic range on this device. Please check to see if your output noise is still within tolerance of your system design with this register change.

    best regards,
    Luis

  • 0 in reply to Luis Fernandes de Pina
    Prodigy 30 points

    Dear Luis

    I have sent a private email to you last Friday. I am just checking that you have received it.

    We would be very grateful if you could explain the effect of that solution.

    Kind regards,
    Stéphane