• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 45 subscribers
  • Views 359 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.

Analog input range for ADS1243

David Pfaltzgraff
Expert 2445 points
Other Parts Discussed in Thread: ADS1243

This is question on the input 'limits' for the ADS1243. The Vdd is at 3.3V and Vref is 1.5V. (Vref- is connected to ground and Vref+is at 1.5V.) If one input (Ain8) is tied to ground, how close can another input (Ain0) come to ground and have the specified accuracy and linearity?

Another way to ask the question might be: Are the inputs truly rail-to-rail (or in this ground to Vref)?

A secondary question: For this scenario, if Ain0 is tied to Vref+, would the reading (between Ain8 and Ain0) always produce the full scale reading of 0x7FFFF? (Of course allowing for noise variances.)

Thanks.

  • 0
    TI__Guru** 113765 points

    David,

    I'm going to make the assumption that the Buffer is off in your case, Range is 0 and the PGA is equal to 1.  If you need to calculate for other gains, see Table VI in the datasheet.  I am also assuming that you mean AIN8 is really AIN7.  So, if the Mux Control Register has AIN0 as your positive input, and AIN7 as your negative input, then you will get a result in the positive range of values where 1 LSB is VREF/[PGA*(2^24 - 1)] or in your case 1.5/(16777215) or 89nV.  You will get a negative reading if the AIN7 is set as your positive input and AIN0 is set as your negative input.  In other words you cannot go below ground.  If the reference input is 1.5V, then you will be able to measure from 0V to 1.5V.  If you tie AIN0 to VREF, with the given situation, you should read full-scale (0x7FFFFF.)

    Now, barring input offset errors, drift and noise you should be able to read all codes within the range.  Offset error can be self calibrated and minimized to within about 15ppm of full-scale (about 22.5uV), but noise is also going to be the key factor.  See the typical characteristic curves on page 7 of the datasheet.  The default data rate at 15sps has an ENOB of 20, and about 20.8 bits for the slowest data rate.  So a best case scenario would be 1.24uV of noise.  There are some possible ways to reduce the noise by digital filtering, but we'll assume none in this case.

    So if you do a true system calibration to set the actual zero point, the absolute best you can do with no other system noise (input, reference, supply, etc.) is read a value down to 1.24uV above ground.  That is the theoretical, but I doubt that you will come close to that due to other noise factors, including ground noise.

    Best regards,

    Bob B

  • 0 in reply to Bob Benjamin
    Expert 2445 points

    Hi Bob,

    Thanks for the input. It more or less verifies what I had thought, but I wanted to be sure that I was interpreting the datasheet correctly.

    Since I put in the query, I went back over the circuit and was able to see that measuring low voltages was not that important. The key thing for me is that Vref results in a full scale reading and I can use that as a reference point.

    Thanks,

    Dave