Other Parts Discussed in Thread: PCM1690
Tool/software:
Dear,
Customer uses 2pcs PCM4204 cascaded to achieve 8-channel ADC audio acquisition. During recording, there is a probability of sustained noise of around -45db, which can be restored after resetting the IC.
They conducted the following investigation (the problem is that white noise of -45dB is added to each channel, as shown in the recording file below, which appears probabilistically)
1. The CPU provides SCKI, LRCK, BCK, and connects SDOUT1 of PCM4204 to DIN1 of PCM1690, indicating the presence of noise probability.
2. Modify SCKI from 24.576MHz to 12.288MHz (LRCK is 48kHz, BLK is 12.288MHz, i.e. BCK=256fs), and noise probability exists.
3. The CPU resets PCM4204 during operation, with a reset time of 10ms, and there is a probability of noise.
4. Reducing the period_2 and buffer_2 of the CPU increases the probability of occurrence.
5. Connecting to the CPU through telnet may also result in noise.
Overall, we confirm that this noise is generated by PCM4204, but since SCKI, LRCK, and BCK are provided by the CPU, we suspect that the timing of SCKI, LRCK, and BCK from the CPU may have affected PCM4204.
SCH
Noise recording file
IIS waveform:
Channel 1: MCLK
Channel 2: BCK
Channel 3: LRCK
Channel 4: SDOUT1