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.

TPA3255: Shutdown pop and voltage difference

Part Number: TPA3255

Hi,

I've designed an amp module with the TPA3255 and have built a prototype.

The TPA is in BTL mode.

During measurement the following issues were found:

  • When the reset pin is pulled down, there is an audible pop on the output. (sometimes but not always a startup pop is there as well)
  • When I measure on a speaker output, there is a small DC voltage (~1V measured between the pins of the speaker.)
    • This voltage is also present when the speaker is disconnected.
    • Both channels have this. (one is worse than the other at 1.8V for channel CD and 1V for channel AB)
  • The IC gets hot when idle and will go into thermal shutdown after some time with a very audible pop.(probably because of the DC voltage)
    • It is actually better to play music than to have it idle at this point

CSTART is 47nF

The strange thing is that when music is playing, everything works perfectly and when idle, the noisefloor is stable and low (as expected)

Questions:

What could cause this DC voltage on the output and how can I fix this?

What can I do to reduce the pop?

  • Update:

    I just measured with my oscilloscope and there is an oscillation (so not DC) on the speaker output at around 460kHz which is the PWM frequency of the chip. (The chip is in Nominal master mode at this time)

    Update 2:

    I've changed the Frequency adjust resistor to 10K (was 30k) and the module is now in Master AM2 mode. This has reduced the amplitude of the oscillation on the output and the idle power consumption.

    I've also changed the OC limit resistor to 22K (maximum current in cycle by cycle current control mode) This made it possible to reach higher output levels.

    One remaining issue is the shutdown pop.

  • Hi Sven,

    Could you post a waveform on the "oscillation"? 

    The 1V offset is abnormal. Is it happens when there is no load connected, and the device in switching status? 

    Thanks!

    Regards,

    Sam

  • Since I changed the frequency adjust resistor to 10K, the oscillation has increased in frequency to around 600kHz (which is consistent with the AM2 Master mode switching frequency). Since the output filter is a second order filter, going from 460kHz to 600kHz decreased the amplitude of the waveform by around 4dB.

    Scope is in AC mode to see the signal!

    on DC mode, the signal looks like this:

  • Hi Sven,

    The waveform you showed are normal switching ripples in class-d audio amp. The ripples are in switching frequency, which is well beyond audible frequency range.

    Thanks!

    Regards,

    Sam

  • Thanks Sam!

    That is what I thought about this measurement but wasn't sure that it was normal since this is the first class D that I'm making.

    The only issue I have now that remains is the shutdown pop.

    Best regards,

    Sven

  • Hello Sven, 

    Can you provide your schematic.

    Best regards,

    Luis

  • Hi, Sorry for the late reply, I was on holiday.

    The schematic is the following:

    I've been looking at the C-start capacitor but the value is what is in the reference schematic for BTL mode. I was thinking to play around with this capacitor to see if it fixes the issue. (though from my understanding, the C-start capacitor is only for start up behavior and not shut down)

    Best regards,

    Sven

  • Hi Sven,

    Yes, please slightly increase the C_start value(like 68nF, 100nF, etc...) to see if there is any improvement. It controls the slew rate for both start-up and shut-down.

    Regards,

    Sam

  • I found that the capacitors on the input, for which I chose electrolytics, were placed incorrectly.

    My input stage before the amplifier is powered with dual supply and the reference is ground for this stage, the input of the TPA3255 goes to 4V at startup, causing them to be reverse polarized.

    When switching these around, the circuit behaves the same as the evaluation module. I've found that there was no need to increase the value of Cstart.

    I guess the lesson here is:

    1. Check the voltages on the buffer stage and amplifier IC when using elco's

    2. If possible use ceramics or non-polar capacitors for this connection.