DEM-PCM2912AEVM: Frequency response of Mic input on PCM2912A Eval Board.

Hello,

There should be some high-pass filtering associated with the AC-coupling capacitors, but this is not correct.  Can you confirm the jumper settings on the EVM?

Hello there,

Yes 0 dB on the gain setting and 100 mA for power. 

T

Also, Mic bias is floating in this scenario.  

Can you confirm how you're connecting the input signal, and what you're using for it?

We are coming out of the BNC connector of a prism sound device, with 0.5 Vrms single ended at 1 KHz, going to a mini TRS jack into the PCM2912A eval board, with signal routed to tip and sleeve is GND. The ring is left floating (Bias is not connected, could the bias left floating affect Mic input).

Thanks.

Quick correction: Swept sine 0.5Vrms, 20 Hz to 25KHz.

Hi Taoufiq,

Have you modified the EVM at all? This is definitely an abnormal response and not something we generally see from the chip. Out of curiosity what version of Windows are you running this on and have you had any issues with the device enumerating as a USB audio codec? You may want to start by verifying your Windows sound settings are appropriate and there isn't a a digital filter being applied somewhere in the measurement.

There shouldn't be an issue with floating the micbias. the input signal is AC coupled so as long as your generator provides an unbalanced floating signal I wouldn't expect any problems. One thing to note though is that the device has a mono input and it's intended that you would connect a mono 3.5mm jack to the mic input. A mono jack would short the tip and ring and this is how the bias is provided. If you are using a stereo jack from your equipment then one channel may be sinking current from the micbias, I'm not sure if this would impact your measurement equipment but you could try removing the 1kOhm resistor (R19 on EVM) to see if this helps.

Best,

Zak

Thanks a lot guys. The TS is definitely an issue, but we will retest with removal of R19 and will advise.

Thank you again.

After more testing yesterday, we're getting results that match the datasheet. However, I have a question about the user guide of the PCM2912A EVM where it say that the ADC high pass filter is programmable and it has a 0.078Fs as a cut off. Is the ADC really programmable? If so, is there a range, or is it just on and off.

Thank you.

Hello again,

I think the high pass spec was probably just a typo. But I have two last questions pertaining to this PCM2912A chip and would appreciate a reply.

1) The VCCA pin is regulated internally to 3.3V. Is it safe to drive it with an external LDO to say 3.85V (to increase performance) in a fashion similar to the VCCCI pin in the stereo version of this chip (PCM2906C - photo attached). I am assuming both PCM chips probably have similar internal regulators built on same process that can withstand same sink capability.  

2) In the datasheet it says that the input Full Scale on the PCM2912A is 0.43V VCCA, is that spec with PGA gain set to 0 dB (which would translate to the input to the ADC itself being Full Scale at 0.43VCCA). Just making sure we have a recommended gain setting for our customers to achieve best performance. 

Thank you so much.

attached). 

Hey Taoufiq,

Yes I don't think programmable is really the write word for the HPF, scalable is probably a more apt description since the cutoff will move with your sample rate. 

While in general I would agree that a lot of the architectural blocks for this family of devices are going to be similar, I don't have any information on the sink capability of the VCCA pin, only that the absolute maximum input voltage for this pin is 4V and this matches the spec for PCM2906. The PCM2906 actually provides THD+N data for an external regulator so it seems this was more of an intended use case for the PCM2906 where this is not mentioned at all in PCM2912 datasheet. If you notice though, the THD+N specs of the PCM2912 at 0,01% match the spec of the PCM2906 with an external regulator, suggesting that the PCM2912 actually has a better regulator and there would not be any significant gains in implementing an external regulator.  

Yes, the input range is assuming 0dB PGA gain and is the full scale input range of the ADC.

Best,

Zak