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.

ADS1119: No listed SNR

Part Number: ADS1119

Hey team,

I'm asking this purely out of curiosity: why do some datasheets, like the ADS1119, not list SNR as a spec but some other ADCs do? Is it because other specs simply tell more than SNR like those given in the System Performance block of the datasheet? I'd love to hear your thoughts

Thanks in advance!

Marco

  • Hi Marco,

    The ADS1119 is a slow-speed Delta-Sigma ADC that is primarily used in DC or near DC applications.  If you take a look at the datasheet and the filter responses you will see that there is an FIR low-pass filter.  The Delta-Sigma uses over-sampling as a modulated bit-stream and then is followed by a low-pass digital filter.  The end result is the quantization noise is pushed to the higher frequencies and then the high-frequency noise is filtered out.  Very short and brief explanation but in the end you get a very precise (or repeatable) measurement.  You should take a look at the TI Precision Labs and ADC sections 2 and 3.

    So what it comes down to is DC measurements compared to AC measurements.  AC measurements can have a signal of interest that is relative to the noise floor and you can measure signal-to-noise.  The DC measurements are relative to the conversion noise of the ADC.  So if you have a DC (or near DC) signal, the noise of the ADC conversion process is the limiting factor.  In this case the noise is determined by shorting the inputs and the results are given in the ADS1119 datasheet in the noise tables.  So the tables show what the best possible resolution is for the measurement.  Additional factors, such as reference noise and EMI/RFI, will limit the total resolution further. 

    So whether the SNR is given in the datasheet is relative to the type of measurement primarily used for a particular ADC.  The digital filter limits the frequency response which is not linear and makes it difficult to specify an SNR unless the ADC uses a flat-pass band filter as opposed to a sinc response filter.

    Best regards,

    Bob B

  • Bob,

    Thanks for the response! I took some time to review the digital filter concepts and I have a couple follow up questions

    1. Do most DS ADCs use an FIR filter as mentioned above or are there other digital filters sometimes used such as IIR? If so, what kind of applications? If not, why are they not used at all?

    2. It appears that input-referred noise is the key factor here. One thing I was curious on when looking into this is that increased gain will actually reduce input-referred noise and I'm not sure I'm understanding conceptually why that is? I would think that gain would just gain up the input-referred noise to the output and not be a factor other than that.

    3. You mention "the digital filter limits the frequency response which is not linear and makes it difficult to specify SNR", however in my research I've found that FIR filters are usually credited for being linear. What assumptions am I making that isn't connecting the dots properly? 

    I really appreciate the level of detail you've provided, thanks for the help and I look forward to your response

    Marco

  • Hi Marco,

    1. Do most DS ADCs use an FIR filter as mentioned above or are there other digital filters sometimes used such as IIR? If so, what kind of applications? If not, why are they not used at all?

    Many types of filters can be used with D-S ADCs, but FIR is frequently used.  Sometimes the filters are even cascaded.  It really depends on the desired outcome.  Some filters add special characteristics such as 50/60 Hz notches.  You might take a look at the appnote Digital Filter Types in Delta-Sigma ADCs.

    2. It appears that input-referred noise is the key factor here. One thing I was curious on when looking into this is that increased gain will actually reduce input-referred noise and I'm not sure I'm understanding conceptually why that is? I would think that gain would just gain up the input-referred noise to the output and not be a factor other than that.

    Here I refer you to another document on the Fundamentals of Precision ADC Noise Analysis.  This information is also contained in a TI Precision Labs video on ADC noise.

    3. You mention "the digital filter limits the frequency response which is not linear and makes it difficult to specify SNR", however in my research I've found that FIR filters are usually credited for being linear. What assumptions am I making that isn't connecting the dots properly? 

    What I was referring to was the output response of the filter itself.  A sinc filter, for example, has a response that looks like a bouncing ball.  This means that the input signal compared to the ADC output is not flat across frequencies. The ADC conversion result will not be flat (and is non-linear) in the response as compared to the input.  There are other filters that are wide-bandwidth and have a flat-passband response followed by a brick wall filter.  These D-S ADCs are meant for AC type measurements and will have the SNR specs.  Again I refer you to the appnote I mentioned earlier in this post.

    Best regards,

    Bob B