• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 83 subscribers
  • Views 488 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

2407A ADC cross-talk

Simon Schoch
Prodigy 35 points

On the ADCIN01...ADCIN13 I have on my target several analog inputs. The sequencer samples this inputs in cascaded mode with the following sequence:
CHSELSEQ1=0x2121, CHSELSEQ2=0x8765, CHSELSEQ3=0xCBA9, CHSELSEQ4=0x000D

Now I can notice a crosstalk from ADCIN09 to ADCIN01. This crosstalk will disappear, if the ADCIN09 is not sampled or if the ADCIN09 is sampled at the end in the sequencer, like this: (CHSELSEQ1=0x2121, CHSELSEQ2=0x8765, CHSELSEQ3=0xCBA0, CHSELSEQ4=0x009D).

What is the reason for this behavior? I have excluded a crosstalk on the Board, because the crosstalk has disappeard if I changed the order in the sequencer.

Regards,

Simon

  • 0
    TI__Guru 60865 points

    Simon,

    What does the impedance look like for the driver of ADCIN01 pin?  Is it any different that what you are using to drive the other pins?

    The most likely issue is that the driver for the signal that is receiving the error (ADCIN01) has too high of an impedance and cannot charge the S+H capacitor to 1/2LSB in the allocated S+H window.  In this case, whatever the S+H was charged to last will show up on the subsequent conversion.  You can check if this is the case by either increasing the ACQ_PS value or double-sampling the second signal to see if the coupling goes away.  

  • 0 in reply to Devin Cottier
    Prodigy 35 points

    Devin,

    To load the S/H capacitor in the ADC we have on the target directly at the ADC input a 10nF capacitor. The S/H capacitor has 30 pF. There can be a crosstalk from one channel to the next by about 3 LSB. In this area, I can measure the crosstalk. But why is this happening by Result8 on Result0? Obviously the Mux before the S/H is set to this channel (defined with bits 0-3 of CHSELSEQ3), if the sequencer is finished. Then will charge the S/H capacitor on this signal and transmits on the first channel. This is although nowhere documented, but it must be so. If we know how this works, we can do something about it.

    Thanks for the suggestion.

  • 0 in reply to Simon Schoch
    TI__Guru 60865 points

    Hi Simon,

    Can you confirm that:

    You do see cross-talk when the channel sequence is ...,x,x,x,9,x,x,x,x,0,0,0

    And you do not see cross-talk when the channel sequence is ...,x,x,x,0,x,x,x,x,9,0,0

    I think when I read the first time, I was thinking the cross talk was occurring here:

    ...,x,x,x,0,x,x,x,x,9,0,0

    Also, can you elaborate a little bit on what you mean here:

    Obviously the Mux before the S/H is set to this channel (defined with bits 0-3 of CHSELSEQ3), if the sequencer is finished. Then will charge the S/H capacitor on this signal and transmits on the first channel. This is although nowhere documented, but it must be so. If we know how this works, we can do something about it.

    I didn't quite follow this.