ADS124S08: Input resistance, capacitance, accuracy

Hi Ray,

Unfortunately we do not not have specific input capacitance for the ADC itself, but we do have the package IBIS model which can be used as this most likely will dominate.  The IBIS model can be downloaded from the ADS124S08 product page.

As far as achieving accuracy of 0.1 deg C, this depends on many factors and the particular sensor being used.  What generally dominates here is noise and drift.  Some additional considerations for thermocouples is the cold-junction measurement accuracy.  For RTD measurements the best method is to use a ratiometric measurement where the reference is excited by the same excitation source as the sensor.  I would suggest having the customer review the material in A Basic Guide to Thermocouple Measurements and A Basic Guide to RTD Measurements.

You may also want to share the code library for temperature measurements.

Best regards,

Bob B

Hi Ray Vincent,

I'm not sure what types of ICs the customer is looking at that shows this information, but it may be op amps as opposed to ADCs.  Sometimes there will be an impedance value given for some of our older ADCs, but more often than not these days there is typical input current given as opposed to R and C values.  The reason being the input is much more dynamic and there are multiple contributions.  The input current can be from the input sampling, input chopping and leakage.  The input currents are given in the Electrical Characteristics Table on page 9.  

Giving R and C values doesn't give the full story.  As already stated the operation of the input is dynamic and will change relative to the applied voltage at the input.  To give a better representation you can observe the behavior shown on the graphs starting on page 17 of the ADS124S08 datasheet.  As devices such as the ADS124S08 are highly integrated, it is difficult to create a simulation of the behavior as the response differs with gain and data rate.  Some simulations are available for devices not as highly integrated such as the SAR or Delta-Sigma devices that do not have a PGA at the front end (strictly switched cap input sampling).  To analyze error it is best to use a root-sum-squares (RSS) of the error.

In the end how the ADC will affect the measurement result will depend on the sensor.  For a thermocouple the measurement is a voltage measurement and will depend on the accuracy and drift of the reference.  For the ADC there will be offset and gain error as well as the drift.  Another potential error is INL.  There can also be error in the measurement related to noise.  The Delta-Sigma ADC is an oversampling device with a low-pass digital filter.  Quantization noise is pushed to the higher frequencies and the low-pass digital filter removes this noise from the result.  As bandwidth is increased (higher data output rates) the noise begins to become a part of the measurement.  So configuration of the ADC becomes important when determining the error due to noise (such as gain applied, muxing of channels and data output rate).

For RTD measurements we recommend that the measurement is ratiometric where the same excitation for the RTD is also used to establish the reference using the built in IDAC current sources.  The reference is established by using a resistor with a voltage created by the IDAC current through the resistor.  The benefit of using this method is the noise and drift error of the excitation cancels in the measurement.  However the accuracy and drift of the reference resistor will add error its own error.  There will also be error differences with respect to whether a 2-wire, 3-wire or 4-wire RTD is used.  In the case of the 3-wire RTD measurement, two IDACs are used for lead cancelation.  Unless IDAC chopping is used, any error related to mismatch and drift of the IDAC sources will add error.

To look at the total error, I find it much easier to use a spreadsheet and RSS analysis of the error sources.  The error sources must use the same units (examples are ppm, % of full-scale, uV, etc).  Some of the errors can be calibrated out such as offset and gain.

Even though the ADC is capable of 0.1 deg C, doesn't guarantee this performance as there are many other external error contributions (including the sensor itself).  This error can include EMI/RFI, poor PCB routing, noisy power supplies and power line-cycle pickup.  There can also be input signal loss due to ESD/EOS protection devices.

As far as the EVM is concerned, the development prioritized types of connections (voltage, current, bridge, thermocouple, RTDs, etc.) to demonstrate possible configuration as opposed to performance.  In other words, there was no particular input configuration that was optimized.  At this time we do not have specific data regarding the EVM performance and temperature resolution.

Best regards,

Bob B