• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 2 answers
  • Subscribers 48 subscribers
  • Views 326 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.

ADC Recommendation for 1.5V to 4.5V Input with Low Error

Jessica Riley2
Prodigy 220 points

Other Parts Discussed in Thread: ADS1118

Hello - I'm looking for a recommendation for an ADC to measure a 1.5V to 4.5V input signal with no more than 1mV total measurement error.  The supply voltage is 5V (4.75V to 5.25V).  Does TI have any parts that will do this?

  • 0
    TI__Guru**** 182601 points

    Hi Jessica,

    How many channels are you looking for?  What sort of interface (SPI, I2C, etc) do you have at your disposal?  Can you also tell us what data rate you want to use?  Some basics about your application would help us identify a possible solution.

  • 0 in reply to Tom Hendrick
    Prodigy 220 points

    Ideally the ADC would have two single-ended channels, or one single-ended (1.5V to 4.5V) and one differential (~4.5V +/- 100mV).  SPI communication is preferred, but the update data rate is not critical.  I am planning to use this ADC as feedback for an adjustable LDO -- accuracy is important, but the voltage will not fluctuate while measuring.  

  • 0 in reply to Jessica Riley2
    TI__Guru**** 182601 points

    Hi Jessica,

    I was thinking about the ADS1118 here and apparently you were too!

    https://e2e.ti.com/support/data-converters/f/73/t/979505

    let us know if you want to explore other options.

  • 0 in reply to Tom Hendrick
    Prodigy 220 points

    Thanks for the suggestion and as you pointed out, yes I was already looking at the ADS1118.  However, when I calculate the total error I am getting more than 1mV total error.  Is the calculation below correct?

    Integral nonlinearity = 1 LSB max

    Offset error (differential) = ±0.1 LSB typical, ±2 LSB max

    Gain error = 0.01% typical, 0.15% max (includes PGA and voltage reference)

    Noise @ ≤ 128 SPS data rate = 1 LSB

    Total error = sqrt [ (1 LSB)^2 + (2 LSB)^2 + (0.15%*FSR)^2 + (1LSB)^2 ] = 670μV typical, 9mV max @ FSR = ±6.144V, LSB = 187.5μV.

    This ignores the effect of temperature drift, which is okay for my application, but is still higher than my target of ±1mV.  Is that correct?

  • 0 in reply to Jessica Riley2
    TI__Guru 56611 points

    Hi Jessica,

    The datasheet accuracy values are specified at the 2.048V range, so I would use this FSR to calculate the overall accuracy.

    At FSR = 2.048V, 1 LSB = 2*2.048V / (2^16) = 62.5uV.

    Therefore, you can consider the error of the noise, offset and INL negligible since these will all be at least an order of magnitude beneath the target (1mV).

    The main challenge here will be the gain error, which is 0.15% max. This corresponds to 0.15/100*2.048V = 3.07mV of error, which is above your target. But, keep in mind that these are max values and therefore statistically unlikely (but obviously possible). The typical gain error value (0.01%) is well below your target.

    -Bryan