TLV320AIC3204EVM-K: Decoding Issue

Hi Arya,

The code you are using assumes a 48kHz sample rate with a 12.288MHz MCLK and does not use the PLL. Since your configuration is different, you may need to adjust some of the dividers. How did you select the clock rates you are using? Unless these values are rounded, they do not seem to be valid multiples of one another, which is required if you are not using the PLL.

Another thing to note is that all of the inputs have AC coupling capacitors, and unless you are bypassing these, a DC input signal will be filtered out. You could start with a 0.1Vrms sine wave to see if this is captured correctly.

Best,

Zak

Hi Zak,

My input MCLK is 3.125 MHz. So using the same value of dividers as in the code, I should get the same sampling frequency by selecting MCLK as the source. More precisely, my sampling frequency is 12.20703125 kHz.

Regarding the AC coupling capacitors, I do not realize it until I read the schematic. Thanks a lot! That's why I feel strange as my data sampling never works with low frequency rectangular wave. So now I change the input now to 3 Vpp sine wave with the frequency of 2 kHz. Now the bit sequence is more explainable. Here is the result of decoded bit sequence (left channel):

Just to confirm, is this 2's complement representation of 16 bit integer? I can't find any details about the bit format in the datasheet but If I use this representation, I can convert it to the correct analog input voltage.

Best,

Arya

Hi Zak,

I have another question regarding the code that I am using currently, particularly these registers on Page 1:

Is this procedure must be executed regardless the input signal? I will sample a vibration sensor later, should I make any changes?

Best,

Arya

Hi Arya,

I think you will need to do some testing to decide what settings work best with your sensor. The script provided is for stereo single-ended inputs, if you are using a differential sensor then you would want to change some of the routings to support a differential input. You will probably also want to apply some gain with the PGA to amplify the sensor signal and possibly adjust the input impedance of the ADC depending on the impedance of your sensor. There's also the DAC settings if you intend to use this functionality of the codec. If not, then you may be better off using one of our ADC products that does not include the DAC.

Best,

Zak

Hi Zak,

Thanks, that is very helpful! My lab will design a pcb using the same ADC chip on this evaluation board, so I need to gather a lot of information. Regarding the impedance, how should I adjust the impedance? The impedance of my signal generator is 50 Ohm, is there a formula/guideline to pick the correct value?

And this is my final question for my current test. I set the input to 3 Vpp sine wave @ 10 Hz. Here is the signal captured from oscilloscope:

I do the sampling with 12,2 kHz frequency. Then it is turned out that the reconstructed signal is saturated like this:

When I use the raw data and plot them, I have the same phenomena. Is this possibly because quantization error due to resolution limitation or something else?

Best,

Arya