Because of the Thanksgiving holiday in the U.S., TI E2E™ design support forum responses may be delayed from November 25 through December 2. Thank you for your patience.

  • State
  • Locked Locked
  • Replies 2 replies
  • Subscribers 30 subscribers
  • Views 433 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

PCM4202: Is there any info on the Vcom reference output noise spectrum? And/or impedance looking back into that pin?

Michael Steffes
Guru 48395 points
Part Number: PCM4202
Other Parts Discussed in Thread: THS4131, OPA1632, THS4561, TLV333

I was wrapping up the next article on using FDA's to drive these Audio ADC's noting the possible risk of higher low frequency (<10Hz) output differential noise due to unnecessarily high common mode control noise. The FDA's themselves have very high output CM noise if the Vocm pin is simply floated. I was going to suggest if the ADC Vcom output is to be used, at minimum at a low F RC filter from that pin to the FDA input control pin, what input noise spectrum hits the Vocm control pin gets to differential output noise through feedback ratio mismatch - normally not an issue, but if the <10Hz noise is very high, could be. 

Looking at the PCM4202 datasheet there is very little info on this reference output. The buffered version is recommended with an external cap. So, personally, I would suggest maybe a series 1kohm (4nV R noise) into that 1uF cap to the FDA Vocm control pin, but again, there is no way to know what CM noise spectrum will appear with that interface. Just adding the cap alone seems odd as no mention of the buffered output stability or source R is made? Also, there are some mistakes here, this mentions the OPA1632 FDA having a low Vocm input impedance - that is not easy to find, but the THS4131 is the same die and it shows 15kohm input R equivalent in Section 7.2 - probably do not need the buffer amplifier and in Section 8.2 of the THS4131 it actually shows what the text of the PCM4202 here is describing - with the 1kohm series R. Which is I think better than simple cap on the Vcom pin in the PCM4202 datasheet. 

So, to show what concern I want to bring up here, check this THS4561 CM noise plot out - just floating it would be a mistake it seems, easy to fix with external cap or buffer amp, but easy to miss too 

this CM noise floated is remarkably high at 10Hz (measurement mic apps)

  • 0
    TI__Guru 62895 points

    Hi Michael,

    Nice to hear from you and cool articles.  The Vcom output impedance in the PCM4202 is pretty high and it mentions on page 13 that it should be buffered if connected to nodes with impedances < 1MOhm.  Therefore, 15kOhms is too low and needs to be buffered.


    Can you draw out the connections you're asking about?  I'm pretty sure I know what you're asking but want to verify.  The internal circuitry recommends the single 0.1uF cap to GND so it's likely a simple reference type structure and the cap will produce an output noise spec similar to what you'd likely see in a REF50xx type product.  I agree additional filtering can be accomplished by adding a series RC after the cap to GND before the Vocm buffer.

    From datasheet:

    Although the VCOML and VCOMR outputs are internally buffered, the output current is limited to a few hundred µA. It is recommended to connect these pins to external nodes with greater than 1MΩ impedance, or to buffer the outputs with a voltage follower circuit when driving multiple external or low impedance nodes.

  • 0 in reply to Collin Wells
    Guru 48395 points

    Well it does say the outputs are internally buffered? is that not an amplifier with it own noise and 1/f info. 

    So the REF parts have their own noise plots, but if there is an added (likely low power) buffer stage, will that not also get into the noise spectrum appearing at those VCOML and R outputs? Essentially, how would I arrive at the minimum required effort to handle the low F noise spectrum from the buffer output with or without and RC. 

    here is the full up circuit I ended up suggesting if an added op amp is ok - the TLV333 is not low noise, but no 1/f. The noise needs to get real high to show up through feedback R mismatching in the FDA,