• State TI Thinks Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 1 answer
  • Subscribers 30 subscribers
  • Views 382 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.

Original question:

TAS2557: Feed Forward Protection

TAS2557: Feed Forward Protection

David Smith62
Intellectual 290 points
Part Number: TAS2557

I've read through post from December 9-11 about feed forward protection. 

Feed Forward Protection algorithm uses speaker characterization parameters.
Both modes, Feed Forward and Feedback Protection use speaker characterization parameter. The difference is that Feed Forward Protection does not use I/V sense feedback from the speaker in real time.

and 

The parameters entered in speaker model are used along with Learning Board to generate the characterization parameters.
The algorithm uses the latest ones, which are the parameters obtained from the speaker during characterization.

Question 1: Could you verify that using feed forward protection does still protect the speaker from overload?  It sounds like using feed forward protection, there's a mechanism where the TAS2557 keeps track of how much power and how fast that power is sent to the speaker and limits the output to the speaker based on that information (based on the output power vs. voice coil temp determined during speaker characterization).

Question 2: In our voice audio application,

* with feed forward turned OFF: when we talk loudly, the output starts loud, but drops way off and stays low for ~20 seconds before it comes back up. 

* with feed forward turned ON: when we talk loudly, the output starts at the same loudness, but the volume never falls off.

Is the speaker protected and "safe" if we use the feed forward protection turned ON?

Thanks.

  • 0
    TI__Guru* 84915 points

    Hi, David,

    My colleague will take a look at this thread. He will be answering as soon as possible.

    Best regards,
    Luis Fernando Rodríguez S.

  • 0
    TI__Guru** 100695 points

    Hi David,

    Let me add some comments about your questions:

    • Some speaker specs are provided during characterization like excursion and thermal limits, SPL and frequency response are captured as well. This data is used by the algorithm aiming for a flat frequency response, keeping the device under the specified limits.
      Feedforward protection use this data as well, however it cannot correct any difference between the ideal expected speaker performance and the actual real-time performance, some of the possible changes are due to aging, offset due to long playback times, speaker connection sensitivity and so on.
      It is recommended to use real-time protection is possible.
      In addition, the EQ can be used to fine tune the output audio.
    • The time you mention for the reduced output seems too long. What are your speaker details? Perhaps you can try to disable the Leak Detect, or reduce its attenuation to 0dB.

    Hope this helps to solve your questions.

    Best regards,
    -Ivan Salazar
    Applications Engineer - Low Power Audio & Actuators

  • 0 in reply to Ivan Salazar
    Intellectual 290 points

    Changing the Leak Detect didn't make a difference.

    These speakers are brand new.

    The other engineer found the performance of the EVM and our board to be different.  The only difference we can find is the EVM doesn't have the output ferrites / 1nF caps on the SPK_P and SPK_N outputs.  We have Ferrite beads and the 1nF caps on the output. Also, when I did the speaker characterization, it was on the EVM, which didn't have the ferrite beads on the outputs.

    He replace the ferrite beads on our board with 0 ohm resistors and this fixed the volume issue.  Could the ferrite beads make enough difference to change the performance that much?

    I had an additional related question:  if we remove the ferrite beads, which we put there for EMI reasons, how do you adjust the edge-rate control?  The data sheet mentions it on page 1 (Adjustable class-D switching edge-rate control) and again on page 97 in Section 10.2.2.3 - EMI Passive Devices.  The nearest description for a register setting is the RAMP_CTRL Register (book 0x00, page 0x02, addr 0x06).  It has 2 bits for RAMP_FREQ and 2 bits for RAMP_FREQ.  I couldn't find any registers that directly set the edge-rate or ramp rate.

    Thanks.

    Dave Smith

  • 0 in reply to David Smith62
    TI__Guru** 100695 points

    Hi Dave,

    Regarding the edge rate control, it is on Book 0, Page 0, Reg 0x06.

    What are your ferrites spec? Do you have the data sheet?
    Does the ferrite spec follow the recommendation from TAS2557 data sheet table 128?

    Best regards,
    -Ivan Salazar
    Applications Engineer - Low Power Audio & Actuators

  • 0 in reply to Ivan Salazar
    Intellectual 290 points

    Ivan,

    Our ferrite bead is a 180 ohm @ 100MHz part with 0.09ohm DC resistance.  The bigger difference we just found was the filter caps after the ferrite beads were 0.1uF instead of 1nF.  We replace the caps with 1nF parts and the volume stays up.  The extra capacitance really clobbers the volume.  At first thought, I would think it would just slow the edges down, but it must cause problems with the feedback/compensation network.

    We're planning on removing the ferrite beads and caps on the TAS2557 output.

    Thanks for your help.

  • 0 in reply to David Smith62
    TI__Guru** 100695 points

    David,

    Thanks for sharing your findings on the filter capacitor.

    You can remove the ferrite/cap filter and handle the edge rate for EMI mitigation or leave it on its default settings if EMI is not a concern.

    Best regards,
    -Ivan Salazar
    Applications Engineer - Low Power Audio & Actuators