• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 3 answers
  • Subscribers 46 subscribers
  • Views 863 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.

RE: ADS124S08: How to convert the high-side reference to low-side on ADS124S08EVM

user6471363
Prodigy 140 points

Other Parts Discussed in Thread: ADS124S08, ADS124S08EVM

Dear Bob,

Could you please share how to configure the ADS124S08 EVM to design a three-wire measurement circuit with low-side reference and two IDAC current sources? Since we have met the same problem these days.

Thanks,

Tianze Li

  • 0
    TI__Guru** 113765 points

    Hi Tianze,

    Welcome to the E2E forum! I split the thread from the previous thread as the topic changed.  To use the ADS124S08EVM for a 3-wire low-side measurement, you would use the same current path but instead use the bias resistor (R70) for the reference.

    To make R70 the reference you would use the REF0 input pair for the refence input.  To avoid any voltage contention, remove jumper JP8 and make sure that JP11 is connected to the AGND position. You would then need to add in addition to the 3-wire RTD a jumper wire from J7 pin 4 to J8 pin 1.  Make sure that you change the reference configuration to REF0 for the conversion measurement.

    Best regards,

    Bob B

  • 0 in reply to Bob Benjamin
    Prodigy 140 points

    Hi Bob,

    Thanks a lot!!

    Best regards,

    Tianze

  • 0 in reply to user6471363
    Prodigy 140 points

    Hi Bob,

    I have tested the EVM according to the configuration you taught. And the first picture is the result of 3-wire high-side and the second picture is 3-wire low-side. As for the total error, I think low side is higher because of the IDAN mismatch at the Rref. But for noise, low-side is lower than the high-side. I don't know what leads to the lower noise. Could you please kindly help me with the question?

    Total Error

    Noise

    3-wire high-side

    120uV

    2*1 /(2* (2^24-1))*74=4.411uV

    3-wire low-side

    245uV

    2*2 /(4* (2^24-1))*52=3.278uV

    Best regards,

    Tianze

  • 0 in reply to user6471363
    TI__Guru** 113765 points

    Hi Tianze,

    I'm not sure what register settings you are using, but what I think you are seeing is the effect of using increased gain for the PGA in the low-side reference case.  If you look at the noise tables in the ADS124S08 datasheet, you will see that the input-referred noise lowers with increased gain.

    There is an advantage to using the integrated PGA with Delta-Sigma devices in that the input-referred noise lowers with gain.  Of course with increased gain the value of the LSB is smaller which appears to be worse with respect to effective resolution (ENOB in the statistics).

    In general with these comparisons it is better to look at the number of codes as opposed to the actual voltage.  A ratiometric measurement compares the code result to the total number of available codes times the resistance. There is no voltage computed as the ratio is codes to total codes for the reference resistor value used to determine the RTD resistance value.

    As to the total error it is not clear how you are making the computations. Here again you should be looking at the conversion code as opposed to the voltage as this adds computation error.  If you are using a fixed resistor value, then you can calculate the resistance from the conversion code and determine the error as a difference in resistance (which can then be compared to a RTD resistance lookup table to determine the accuracy relative to temperature).

    Best regards,

    Bob B

  • 0 in reply to Bob Benjamin
    Prodigy 140 points

    Hi Bob,

    You are right! !

    Yesterday, I tested the low-side reference case with Rref=1kohm, PGA=4 and IDAC=1mA. And today, I changed the PGA=2 and IDAC=0.5mA, and the PGA=2 is equal to that of my high-side case test. And the output Pk-to-Pk noise code does increase to 71, which is similar to the noise of high-side reference case. 

    But I don't understand why. I think the reason why the input-referred noise lowers with increased gain should be that input-referred noise=noise/Gain, so it decreases as the gain increases. But I think the test result of noise is the output noise. And the output noise should be the same. Why the output noise decreases as the gain increases?

    Thank you for all your help!

    Best regards,

    Tianze

  • 0 in reply to user6471363
    TI__Guru** 113765 points

    Hi Tianze,

    One of my colleagues has put together an excellent article on ADC noise analysis.  In particular you may be interested in Chapter 3, but I would suggest reviewing the complete document.

    https://www.ti.com/lit/wp/slyy192/slyy192.pdf

    Best regards,

    Bob B

  • 0 in reply to Bob Benjamin
    Prodigy 140 points

    Hi Bob,

    Thank you!

    And three questions need your help to check.

    1. Is the output Pk-to-Pk noise the input-referred noise?

    2. Is the ENOB calculated by log2(2^24/7.6)=21.1?

    3.If I change the EVM from high-ref to low-ref, will the R68 have impact on the circuit?

    Thanks a lot for your help.

    Tianze

  • 0 in reply to user6471363
    TI__Guru** 113765 points

    Hi Tianze,

    Here are my responses to your questions:

    1. Is the output Pk-to-Pk noise the input-referred noise? [Bob] The noise Pk-to-Pk is the total noise, which includes any external noise (input-referred), as well as quantization noise.  Reference noise would also be included, but with ratiometric measurements the noise will be greatly reduced in ratiometric case.

    2. Is the ENOB calculated by log2(2^24/7.6)=21.1? [Bob] That would be the effective resolution for DC or near DC measurements and in the past has sometimes been called ENOB, but ENOB is really an AC characteristic and is calculated differently.

    3.If I change the EVM from high-ref to low-ref, will the R68 have impact on the circuit? [Bob] R68 will only effect the compliance voltage and should not largely impact the result unless using a large value of current placing the the total voltage drop of the circuit path close to or above the compliance voltage.

    Best regards,

    Bob B