TAS6424M-Q1: Maximum output power vs Input voltage (PVDD)

Hi Jesus

Jesus Rojas said:

The maximum output power I am getting is approximately 12W

12W is only around 10V peak output. May I ask how do you calculate the output power? By using AP testing, or using oscilloscope to check the output voltage?

There's maybe several possibility:

1. It's possible your input digital audio signal is not full scale. Could use the Gain settings to verify. If use larger Gain like 21V/FS could increase the output, it means your input is too small.

2. Your power supply has some limitation, can't provide enough current, will make the PVDD voltage drop at the peak of sinewave. You could use oscilloscope to check both PVDD and output to verify.

Hi Shadow He,

To get the maximum output power I am using the oscilloscope as you said, measuring the diff voltage and the output current.

Regarding the possibilities:

1. I am using a 1KHz sinusoidal wave, but not sure if it is a full scale. I will validate your comment increasing the gain to 21V/FS.
2. The external power supply seems is not limiting the output current to the system, I've already checked the PVDD voltage on the power pins and is not dropping.

Do you have a tone/file (1KHz sinusoidal waveform) to test the device?

Hi Jesus

Below is a full scale 1KHz. But you'll also need to take care your testing platform, when transferring USB into I2S, would it create any zoom into the audio signal?

Also please provide your tested output voltage and current waveform, I could check if anything unusual.  

Hi,

Thanks for sharing the .wav file. it was useful to test the device.,

Attached the picture as reference:
VBAT: 14.2V, Volume ch: CF,  0dbs, Gain 15V/FS, and output channel as BTL.

VBAT: 14.2V, Volume ch: C1,  0dbs, Gain 15V/FS, and output channel as BTL.



It was not possible to get the 20WRMS using a speaker (voice coil impedance 4ohms nominal) I guess because of the frequency response of the speaker... To get the output power according to the spec, it was necessary to use a 4ohms resistor load.. Could you confirm what kind of load is recommended to calibrate/test the device? I was checking the information in the datasheet but it only defines 4ohms but not the load type.

Hi Jesus

Jesus Rojas said:

VBAT: 14.2V, Volume ch: C1,  0dbs, Gain 15V/FS, and output channel as BTL.

This results indeed have some problem. As you could see, the output only reach to 8Vpp, far from the Max output, which should be around 14Vpp. I would still think your testing system, which transfer audio .wav file  into I2S format signal, will give some depression on the amplitude.

Jesus Rojas said:

VBAT: 14.2V, Volume ch: CF,  0dbs, Gain 15V/FS, and output channel as BTL.

This one also used Gain 15V/FS? The difference is ch:CF. Could you please help me understand what is the major difference with another one? This waveform looks good.

Hi,

Both are using the same Gain 15V/FS, register 0x01 = 0x031.

The differences between both plots is the volume configuration, the first plot you shared is writing in the register 0x07 = 0xC1, it means -7dbs and the second plot is 0x07 = 0xCF 0dbs (0x07 because I was using the CH3).

Hi Shadow He,

I have some questions;

Could you confirm what type load (speaker/resistors) is recommended to calibrate/test the device? 
is there a way to confirm if the module is transferring the .wav file to I2S is attenuating or depressing the amplitude?
Do you have some recommendation for algorithm to control the volume assuming CF is 0dbs (100%) and avoiding suddenly changes in the input voltage to have better fidelity sound? 

Hi Jesus

Jesus Rojas said:

Could you confirm what type load (speaker/resistors) is recommended to calibrate/test the device? 

For general performance test, use a 4ohm resistor, could guarantee the results are stable.

Jesus Rojas said:

is there a way to confirm if the module is transferring the .wav file to I2S is attenuating or depressing the amplitude?

Provide a larger PVDD, like 18V. Set internal Volume 0dB, and Analog Gain 15V/Fs. Then the full scale input, should have a 15V peak sinewave output. Check if you could have it.

Jesus Rojas said:

Do you have some recommendation for algorithm to control the volume assuming CF is 0dbs (100%) and avoiding suddenly changes in the input voltage to have better fidelity sound? 

Try add smooth in MCU for the audio input suddenly change. Like for each of FSYNC cycle, limit the amplitude change to 0.1dB, and ramp to the target value.

Hello Shadow He,

Could you please elaborate more in the comment:"Try add smooth in MCU for the audio input suddenly change. Like for each of FSYNC cycle, limit the amplitude change to 0.1dB, and ramp to the target value".?

Hi Jesus

Sorry, I might misunderstand of your need. You are talking about the input voltage change, which is PVDD suddenly change, right?

Shadow He said:

avoiding suddenly changes in the input voltage

Could consider using the PVDD foldback control, when detecting PVDD change, use AGL (Automatic Gain Limiter) to change the audio signal level. In below link, chapter 16.9 have some introduction of AGL.

https://www.ti.com/lit/an/slaa786a/slaa786a.pdf?ts=1740360538775&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FTAS5825M

Hi Shadow He,

Regarding to your question, yes, I was talking about suddenly changes on PVDD. I think is clear now. Thanks.

I have another question,

Assuming 20Wrms at 13.5V on PVDD (fixed input voltage) and the speaker register has been defined 0xCC as default gain (-1.5db).

The application is looking the following:

volume -> POUTrms

100% -> 20Wrms

75% ->15Wrms.

50% -> 10Wrms

25% ->5Wrms

I performed some measurements using the 4ohms resistors changing the registers using steps of 0.5db according to the datasheet, but I think the change from 20Wrms to 10Wrms is to short and it makes sense because is not linear... To get 20Wrms I used the 0xCB and 0xC5 to get 10WRms (Approximately).

Do you know if already exist a algorithm for this module?

Do you know how we can control the gain of the speaker's registers to have the configuration desired?