Post Go back to editing

AD598 - Error budget analysis

Category: Hardware
Product Number: AD598S

Hi,

i have an LVDT acquisition chain using LVDT sensor -> AD598SD -> ADA4099-1 configured as diff. amplifier with gain =1 and AD7980S as ADC.

I need to calculate the whole chain accuracy and resolution.

Based on AD AN-539 all errors must be referred to the same point (e.g: input of the AD598) and compared with the FS to calculate the ppm contributions.

I have basic knoledge about error budget analysis, so here are my questions:

- If i refer all errors to the input point (output of the LVDT sensor=input of the AD598), which is the gain of the AD598S to be considered?

- AN-539 separate error contribution for ABSOLUTE ACCURACY, DRIFT and RESOLUTION for the opamp and INA, so i think this is the same approch i had to follows. For the AD7980 which are the error contribution to be considered for the accuracy and which for the resolution?

- As above, for the AD598SD, which are the error contribution to be considered for the accuracy and which for the resolution? (the DS of the S version specifcy as error sources; Signal output non-linearity, Signal output gain error, Signal output gain drift, Signal output offset, Signal output offset drift, Signal output PSRR gain, Signal output PSRR offset, Signal output CMRR gain, Signal output CMRR offset.

Thanks for the help.

  • Hi  ,

    Good day.  I have moved this question about AD598 to the High Speed ADC community.  Someone here should be able to assist you.


    Regards,
    Gilbeys

  • Where did this thread go? Response says it was moved to High Speed ADC community but I don't see it when I search for that topic

  • JR7/BestUser

    Your Thread was moved to the Precision ADC community.  Which is where it should be addressed. With regards to your questions let me try to point you in the right direction.

    - If i refer all errors to the input point (output of the LVDT sensor=input of the AD598), which is the gain of the AD598S to be considered?

    In my opinion you've made a good first selection which is to refer everything back to the mechanical displacement sensed as this is really what you care about in your system design.

    As for your specific question about the gain for the AD598, that will depend ultimately on a few factors you select for your design, most importantly the sensitivity associated with the LVDT you are using and the value of R2 selected to set the full-scale voltage swing at the output of the AD598.   For the latter selection I would direct you to equation 1 on page 7 of the product datasheet to get started.  

    For the AD7980 which are the error contribution to be considered for the accuracy and which for the resolution?

    Based on the quoted tolerances and stability of the AD598 I would focus my efforts on understanding and minimizing those.  I would recommend selecting a voltage reference with a 10 ppm/C or less maximum temperature drift and as small of an initial error in absolute error as possible such that the AD598 errors dominate.  By doing so the most significant error from the ADC is the initial offset error which I would assume could be removed digitally in your processor during a calibration step and otherwise should remain fairly constant.  Other ADC errors such as non-linearity and initial gain error are likely going to be in the noise of the intrinsic errors of the AD598.  Power Supply Rejection shouldn't be an issue here if you use a linear drop out regulator and the recommended supply decoupling, but do let us know if you are using a switching regulator and we can think some more about potential ripple rejection issues.   Common Mode error should not be an issue for the ADC as the measurement is single-ended.

    As above, for the AD598SD, which are the error contribution to be considered for the accuracy and which for the resolution?

    The error terms you listed in your post all fall under accuracy error terms.   The absolute resolution of the system is going to be a function of the sensitivity of the selected LVDT and the absolute resolution of the data converter.

    Hope that helps.

    Sean