• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 46 subscribers
  • Views 996 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.

ADS8684: Offset error temperature drift and internal reference accuracy

John Griffith
Mastermind 21811 points
Part Number: ADS8684

Hello,

I have a customer that has 2 questions about the ADS8684 device. 

  1. On page 8 the "Offset error temperature drift" parameter. Is ppm/°C the same as µV/°C? So if you wanted to calculate the maximum offset error at a certain temperature would you multiple the offset error temperature drift by the number of degrees from 25°C you want. Take this and add and subtract it to the max offset error at 25°C (EO)? The customer wants to make sure that this is not with reference to AVDD, DVDD or the reference voltage. 
  2. From the customer: What is the accuracy of the actual reference used by the adc? Not the level appearing on the REFIO pin. At the top of page 10 it just states Vref = 4.096 (internal) without any tolerance, which is incredible UNLESS the tolerance is perhaps already taken into account in the various offset and gain tolerances. In other words, perhaps we take “4.096” as a 0-tolerance number, knowing that whatever tolerance it actually has will be accounted for when we do our calculations of gain and offset.

Thanks you in advance!

John

  • 0
    TI__Mastermind 25020 points

    Hello John,

     

    1. The offset error temperature drift (Off_td) given in ppm/C is not the same as uV/C, note the latter has an added unit of Volts, while the given specification does not include a voltage term.

    To calculate the offset error at a specific temperature, T1, then you would need to the find the delta from the ambient temperature (Ta = 25C) and T1 and multiply this by the offset drift (Off_td), as follows: Off_td *(Ta-T1)

    This will result in ppm/C, you can use the conversion factor of 10,000ppm = 1%  to convert to percentage.

     

    I found the following that touch on this, perhaps they can prove to be helpful. 

    ADC accuracy: Effect of temperature drift on ADC signal chain (Part 3)

    WHAT DESIGNERS SHOULD KNOW ABOUT DATA CONVERTER DRIFT

     

    2.The internal reference is measured at the REFIO, thus the tolerance and specs given at this pin is for the internal reference. At Page 10, the internal reference is specified at 4.095 to 4.097, this device is factory trimmed with an initial accuracy of +/- 1mV. Within the datasheet, section 8.3.8, Reference, starting on page 26, there is more detail on the reference performance, including histogram of internal accuracy, variations across supply and temperature, and drift histogram

     

    Regards

    Cynthia

  • 0 in reply to Cynthia
    TI__Mastermind 21811 points
    Hi Cynthia,

    Thanks for the help!

    For 1) is this then a percentage of the measurement or a percentage of the full scale range? The Burr-Brown note assumes it is PPM of FSR – which I take it would be the Vref, fundamentally, but that this would then be multiplied by the gain selected for the channel, with the result being that we could just take it as PPM referenced to the full dynamic of the signal chosen?

    For 2) Should the customer treat Figure 60 as just an example of the drift or is it fairly deterministic in which case they could measure the temperature and then compensate for the drift?

    Thanks again!

    John
  • 0 in reply to John Griffith
    TI__Mastermind 25020 points

    No  prolem John,

    1. This would be in terms of the full scale range being used. As you stated, if you take it in terms of the reference and then multiply it by the factor for the full range, and you still end up at the same result as if you would have used the full scale range initially.

    For example, if the offset drift is 3ppm/C, and the full scale range is +/-10.24V, then, Off_td * FSC = 3*10^-6*(2*10.24) = 61.44uV/C.

    If you use the reference instead and gain it up based on the range desired then the equation would be:

    Off_td *(4.096*2.5)*2=61.44uV/C

    2.  This figure shows typical performance. It would be best to use the specs in the electrical characteristics to compensate for the drift based on the measured temperature.


     

     

    Regards

    Cynthia