TPA3223: THD+N issue

＊Test Conditions：

（1） 通過直流電源向 DUT（Net_PVDD 和 Net_5VP）供電。

    (2) AMP Output(CH_1 & CH_2) connect to Loads(8 ohm).

    (3) Audio Precision Output(GND) connect to DUT_GND.

    (4) Use Audio Precision(AP-525) to measure.( Include AUX-0025)

    I attach my circuit diagram & Equipment wiring diagram & gerber file. Pls. see below：

1. Circuit diagram-1 (Equipment wiring diagram).pdf
2. Circuit diagram-2 (Equipment wiring diagram)_No EMI Parts.pdf
3. TPA3223 AMP PCB - Bottom Side.pdf
4. TPA3223 AMP PCB - Top Side.pdf

Circuit diagram-1 (Equipment wiring diagram).pdfTPA3223 AMP PCB - Bottom Side.pdfTPA3223 AMP PCB - Top Side.pdfCircuit diagram-2 (Equipment wiring diagram)_No EMI Parts.pdf

Current measurements show that THD+N is obviously on the high side overall. Various situations have changed, but the cause and solution have not yet been found. The conditions for change are as follows:

• BST capacitance (C47、C48、C49、C50)： Change from 33nF (EVM Board used value) to 220nF (Datasheet used value).
• Change PWM frame：Change R65 from 10k (PWM=480k Hz) to 50k ohm(PWM=600k Hz).
• Remove EMI Parts of AMP Output： see Circuit diagram-2 (Equipment wiring diagram)_No EMI Parts.pdf

       P.S. Noise of DUT (ClassD AMP)：Measured value is about 130uV (AMP Gain 20dB). IC specs is 100uV. There is not much difference between the two.

The measurement curve of the above changes is similar to the following：

Measurement curve：POWER OUTPUT vs THD+N

POWER OUTPUT vs THD+N (42V).pdf

2. In addition, the voltage of PVCC has been adjusted from 42V to 12V (24V / 39V / 45V).

     (A) PVCC = 12V & 24V & 42V.  Measurement curve：POWER OUTPUT vs THD+N

      POWER OUTPUT vs THD+N (12.24.42V).pdf

   (B) PVCC = 39V & 42V & 45V.  Measurement curve：POWER OUTPUT vs THD+N

     POWER OUTPUT vs THD+N (39.42.45V).pdf

But the distortion is also bed.

Can you give me some suggestions to improve this problem? Thank very much.