TLV320AIC3254: AIC3254 loud hissing/noise from ADC with certain devices

Hello David,

That is a fair amount of noise coming from the Macbook and iPhone.
What is your ADC Fs and AOSR? which Decimation filter are you using on the AIC3254?
Can you provide a FFT of the MacBook output? Also for the idle noise you are seeing from the ADC when connected to the MacBook?

best regards,
-Steve Wilson

Its a 4x decimation filter. CODEC is working at 44100Hz sample rate with 5.6448Mhz input clk and PLL running at 90Mhz. and here are some applicable settings I'm using:

// #### setup clock
0,0x00, // page 0
60,0x00, //; Use miniDSP_A for signal processing
61,0x00, //; 8x Interpolation
17,0x08, //; 4x Decimation
23,0x04, // miniDSP decimation factor = 4
15,0x03, // miniDSP_D instruction control settings
16,0x88, // miniDSP_D instruction control settings
21,0x03, // miniDSP_A instruction control settings
22,0x88, // miniDSP_A instruction control settings
0,0x08, //; page 8
1,0x04, // adaptive mode for ADC
0,0x2C, //; adaptive mode for DAC
1,0x04, //
0,0x00, //; page 0
5,0x91, //; P=1, R=1, J=8
0x1A,0xA0, // Clock out = MCLK / 32 = 156kHz
0x37,0x06, // MISO is clock output
0x34,0x0C, // GPIO is output (low)
//#### need to set PLL J=16 because of half speed clock input! ######
6,0x10, //; J=16 (MSB)
7,0x00, //; J=16 (LSB)
8,0x00, //; PLL_clkin = MCLK, codec_clkin = PLL_CLK, PLL on
4,0x03, //; MDAC = 8, divider powered on
12,0x88, //; DOSR = 128 (MSB)
13,0x00, //; DOSR = 128 (LSB)
14,0x80, //; NADC = 2, divider powered off
18,0x02, //; MADC = 8, divider powered on
19,0x88, //; AOSR = 128
20,0x80, //; NDAC = 2, divider powered on
11,0x82, //
//##### custom code for outputting BCLK and WCLK
28,0x01, // 1 BCLK data offset
29,0x01, // BCLK is derived from DAC_MOD_CLK
30,0x81, // power up divider, set to 1
27,0x0C, // BCLK and WCLK is output from the device
//27,0x00, // BCLK and WCLK is input from BTM
//##### Enable GPIO
48,0x04, // int1 = miniDSP int
52,0x14, // enable GPIO = int1

//##### analogue intput config
0,0x01, // page 1
51,0x00, //; Mic Bias disabled
52,0x40, //; Route IN2L to LEFT_P with 20K input impedance; Route CM1L to LEFT_M with 10K input impedance; Route IN2R to RIGHT_P with 10K input impedance; Route IN1L to LEFT_P with 10K input impedance
54,0x40, //; Route CM1L to LEFT_M with 20K input impedance
55,0x40, //; Route IN1R to RIGHT_P with 20K input impedance
57,0x40, //; Route CM1R to RIGHT_M with 20K input impedance
58,0x00, // disable inputs IN1L and IN1R
//59,0x00, //; Enable MicPGA_L Gain Control, 0dB
//60,0x00, //; Enable MicPGA_R Gain Control, 0dB
0,0x00, // page 0
81,0xC0, //; Power up LADC/RADC
82,0x88, //; mute LADC/RADC

Below are some screenshots of the Macbook output and FFT. There is no noise problem with other portable speakers I've tested with. So it's either a lack of filtering (which should be covered and advised in your documentation) or a problem with the ADC settings. 

I hope you can offer a solution. The audible noise is VERY bad - David

I have done some more investigating. I tried the EVM and the audible noise wasn't there. So I started looking and what additional components there were on the ADC inputs in my circuit compared to the EVM. I started removing components until I found the cultrit.

It was a 10nF capacitor I place between the jack audio inputs and ground for ESD/EMC/EMI purposes.

I placed two 10nF capacitors between IN1L and ground and IN1R and ground on the EVM (after the 1uF caps this time) and exactly the same noise was heard. It seems to be creating an oscillation at 877kHz. Its so strange, it doesn't occur with any other phones or music sources and these caps have never caused a problem before. Obviously moving forward I will change/remove this cap. But is there anything I can do in software to fix this? 

Here is how the FFT and noise changes when the caps are added.