[Sound click]

pmdown_time=-1

[Sound click]

pmdown_time=500

 

[Sound click]

I want to clarify the essence of the problem.
If the audio subsystem is turned off, the output potential is zero potential.
In working condition, a voltage of about 2 volts is set at the output of the audio subsystem, which corresponds to a zero level of the output audio signal.
If you install /sys/devices/platform/soc/1c22c00.codec/cdc/pmdown_time
a value of -1, then after the playback of the file is finished, the output voltage is stored corresponding to the zero level of the output audio signal.
But when playing a new file, the audio subsystem is completely reset.
First, the zero potential is set at the output.
And then the voltage of 2 volts is restored, corresponding to the zero level of the output audio signal.
The duration of such an impulse is short, but a click is clearly audible on the speakers.

[Sound click]

29 minutes ago, guidol said:

where did you put pmdown_time in?

 

I did only found this information:

Put the following line at the end of /boot/cmdline.txt

snd-soc-core.pmdown_time=-1

at https://support.hifiberry.com/hc/en-us/community/posts/201847631-Disabling-auto-shutdown-output-has-no-effect-anymore

I try this way

echo -1> /sys/devices/platform/soc/1c22c00.codec/cdc/pmdown_time 

It works.
But when the playback is initialized, there is a reset and restoration of voltage at the processor's audio output. We saw this on an oscilloscope.

 

[Sound click]

I mean internal DAC

[Sound click]

Good day,

 

We use NanoPi Neo Core in voice information systems.

We found a problem with the implementation of Dynamic Audio Power Management for Portable Devices in H3.

Our program plays voice files using the standard call of the operating system - "aplay".

Before starting the reproduction of each file the voltage is discharged and restored at the output of the DAC.

This leads to a loud click on the output of the audio system.

Setting pmdown_time = -1 does not save the situation.

Can you help in solving this problem?

 P.S. This problem appeared in version 4 of the kernel.