• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 2 answers
  • Subscribers 46 subscribers
  • Views 477 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.

ADS131A04: Missing differential input impedance Zin min/max specification

Stefan Olejniczak57
Prodigy 140 points
Part Number: ADS131A04
Other Parts Discussed in Thread: ADS131B04-Q1

Dear TI team,

i am missing a min/max specification of the differential input impedance Zin. Only the typical value (130 kOhm @ fMOD = 4.096 MHz) is specified. Can You provide such a min/max specification? I would even accept a large deviation from the typical value, as long as the limits are known.

Background:

I am using this ADC with passive RC filtering in front of the analog inputs AINxP and AINxN in a ratiometric configuration. The dominating gain error is therefore the attenuation of the voltage divider consisting of series filter resistor and ADC input impedance Zin. Since i can calculate this voltage divider, i can compensate the typical gain error. What i cannot compensate is the gain error deviation from the typical value, due to the unknown tolerance of the ADC input impedance Zin. Therefore i would like to calculate the maximum expected deviation from typical, to check, if my system still meets my accurracy requirements.

Thanks and best regards,

Stefan

  • 0
    TI__Guru 53416 points

    Hi Stefan Olejniczak57,

    See below from the datasheet. Note that the impedance is calculated using the average value of the current drawn by the inputs, and is also frequency-dependent. Therefore this is not a fixed value, but will change depending on the impedance of your source, the input voltage, the clock frequency, and the sampling capacitor (Cs) tolerance which is on the order of +/-15%. Or, to answer your question directly, there are no min/max values available for the input impedance of this ADC

    Therefore, I would not recommend connecting a voltage divider directly to the ADS131A04 inputs. Instead, install a buffer between your sensor and the ADC to ensure that the ADC inputs see a low source impedance.

    You can also check out the ADS131B04-Q1, which is effectively a buffered version of the ADS131A04.

    -Bryan

  • 0 in reply to Bryan Lizon86
    Prodigy 140 points

    Hi Bryan,

    thanks for your reply, which partially answers my question. Let me explain my application a bit more.

    I use a unity gain opamp buffer before the ADC input. The opamp output drives an anti aliasing 1st order RC lowpass filter of 470 Ohm and 4.7 nF (Rfilt and Cfilt as depicted in fig. 95 of the datasheet). There is no voltage divider, other than the inherent voltage divider formed by Rfilt and Zin.

    This means that overall gain is directly related to calculating this (inherent) voltage divider, hence my question.

    From my understanding, given that fmod is fixed, Zin tolerance is directly related to sampling capacitor (Cs) tolerance, which you state is in the order of +/- 15%.

    Furthermore, from my understanding, sampling capacitor (Cs) dependency over input signal amplitude should be far less than Cs tolerance, because otherwise there would be a severe degradation in THD performance (which i do not see in my application). Can you confirm this? Can you provide a rough estimation of Cs dependency over input signal amplitude?

    Thanks and best regards,

    Stefan

  • 0 in reply to Stefan Olejniczak57
    TI__Guru 53416 points

    Hi Stefan Olejniczak57,

    Thanks for providing more information about your design

    Note that the sampling capacitance tolerance is a rough estimate, and could change with respect to temperature. I do not have more explicit values / tolerances I can share with you however.

    With buffered inputs, the biggest source of error will be the gain error associated with the resistor divider, as you mentioned. Assuming the sampling capacitor increases by 20%, you can expect a gain error of 0.4% when fMOD = 4.096 MHz and Rfilt = 470 ohm. You might consider reducing fMOD or reducing the size of Rfilt if this is an unacceptable level of gain error

    You will also see some phase shift due to the filter if you are measuring AC signals

    -Bryan

  • 0 in reply to Bryan Lizon86
    Prodigy 140 points

    Hi Bryan,

    thanks for Your reply. These countermeasures (reducing fmod and/or reducing Rfilt) are exactly what i'm considering. Alternatively i'm considering calibrating part-to-part variation during our production process (but this obviously would not help with the temperature dependency).

    Thanks again & best regards,

    Stefan

  • 0 in reply to Stefan Olejniczak57
    TI__Guru 53416 points

    Hi Stefan Olejniczak57,

    I am glad we could help you resolve your issue

    -Bryan