PmodI2S2 bill of material (BOM).

[MagicianT]

Hi,

Is there any info I can get on the type of components for PmodI2S2 codec module.

I observed unusually high THD level at the DAC outputs, that make me think about C4 & C5 (3.3uF).

[jpeyron]

Hi @MagicianT

Welcome to the Digilent forums!

We do not publicly share our BOM's.  Looking at our internal BOM it uses generic capacitors and resistors and would fall into this statement I got from one of our design engineers.

"For discrete parts without part number see following requirements:

All ceramic capacitors are temperature coefficient X5R, X7R or C0G, unless specified otherwise.

Capacitors without maximum voltage rating are 6.3V or greater.

Capacitors without tolerance specification are 10% (20% if 10% is unavailable) or better.

Resistors without power rating are 1/16W or greater.

Resistors without tolerance specification are 5% or better."

If there are specific components your are want more information about please let us know.

best regards,

Jon

 

 

[MagicianT]

Quote

All ceramic capacitors are temperature coefficient X5R, X7R or C0G, unless specified otherwise.

I 'd like to refer to TI, page 5, 6, 7 https://training.ti.com/system/files/docs/1324 - Low Distortion Design 4 - slides.pdf

or Maxim-IC https://pdfserv.maximintegrated.com/en/an/AN4333.pdf

Those docs explain in details why X5R & X7R are NO-GO into audio design, especially with high quality codecs  (-100dB THD level). 

My test setup shows almost -40 dB for 2-nd harmonics at lower frequency end, and -80 dB in the middle band. It's not even close to what I was expecting reading cs4344 data sheet.

 

[jpeyron]

Hi @MagicianT,

I have reached out to a co-worker to try to get more specific information about the capacitors and resistors on the Pmod I2S2. In the meantime, I did reach out to one of our design engineers about this thread. They responded that the input is meant to be line level high impedance.  

Are you trying to use a microphone?  Which typically has low impedance.

best regards,

Jon

[MagicianT]

1 hour ago, jpeyron said:

Are you trying to use a microphone?  Which typically has low impedance.

 In my first message, I mention 

Quote

that make me think about C4 & C5 (3.3uF).

than if you open schematic from this link: https://reference.digilentinc.com/_media/reference/pmod/pmodi2s2/pmodi2s2_sch.pdf

you see that C4 & C5 at the OUTPUT of the DAC CS4344. Outputs left & right channels were loaded with standard 47k impedance of the USB sound card.

I have another codec from the same manufacturer (Cirrus Logic) CS4354 that doesn't require DC blocking capacitors, and I'm pretty sure my test setup is perfect, since I get -94dB THD running two codecs in parallel simultaneously. Same data bus, same clock etc. Since DAC are always slaves on I2S, nothing prevents line up 2 or more units for comparative testing. 

I would expect same issue for the INPUT to ADC CS5343 ( C14 & C15), but I didn't test this part yet.

[jpeyron]

Hi @MagicianT,

I was able to get more information about the capacitors 

C4- C5 3.3uF, ceramic, 10%, 16V, X5R, 0805

C14-15, 1uF, ceramic, 10% 10V X5R, 0805

best regards,

Jon

[MagicianT]

It's a pity, that X5R type of capacitors was used in the audio path.  Cirrus Logic missed to put a note into CS4344 data sheet on this occasion, as they did for CS4354 

https://statics.cirrus.com/pubs/proDatasheet/CS4354_F3.pdf

page 12, figure 5:

Quote

Note 1: Capacitors must be C0G or equivalent.

I can add, that even common aluminium electrolytic capacitors would be better choice to keep THD at low level. 

[kwilber]

For readers not familiar with why C0G is better than X5R in this application, you can find a simple explanation here. (The links @MagicianT posted above go into much more detail.)

In particular it explains (C0G have class 1 dielectrics. X5R have class 2 dielectrics)...

"If you design audio devices, or if you simply prefer quiet PCBs, you have another reason to choose C0G over X7R or X5R: Class 2 caps exhibit piezoelectric behavior that can cause them to function as both microphones (that will convert sound into electrical noise) and buzzers (that will convert AC signals into audible noise). Class 1 capacitors don’t have this problem."