• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Subscribers 45 subscribers
  • Views 553 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.

ADS1118-Q1: why the accuracy of ADS1118-Q1 is worse than ADS1118?

Howard Zou
Genius 12365 points
Part Number: ADS1118-Q1
Other Parts Discussed in Thread: ADS1118,

Hi,

In the datasheet,

the temperature accuracy of ADS1118 is ±0.5C(max) at 0-70C,

while ADS1118-Q1 is ±1C(max) at 0-70C.

My customer has been using ADS1118, and due to the shortage, they want to use ADS1118-Q1 directly, but this spec prevent them from using Q grade version.

  • 0
    TI__Mastermind 23275 points

    Hi Howard,

    the temperature sensor performance of ADS1118 and ADS1118-Q1 is actually identical. However for Q100 devices we use a larger CPK value for specifying the MIN/MAX values based on the statistics of the characterization data. This is where the discrepancy comes from.

    Regards,
    Joachim Wuerker

  • 0 in reply to Joachim Wuerker
    TI__Genius 12365 points

    Joachim,

    what does CPK value mean?

    So we can directly use ADS1118-Q1 to replace ADS1118?

  • 0 in reply to Howard Zou
    TI__Mastermind 23275 points

    Hi Howard,

    it will probably be best to google what CPK means. In short you could probably say it determines how many "sigmas" of the statistical distribution we use to derive the MIN/MAX values.

    In my opinion you can use ADS1118-Q1 in place of ADS1118 with no concern. You should not see any performance differences.

    Regards,
    Joachim Wuerker

  • 0 in reply to Joachim Wuerker
    TI__Genius 12365 points

    Joachim,

    so how many "sigmas" do we use for ADS1118-Q1 and ADS1118 respectively?

  • 0 in reply to Howard Zou
    TI__Mastermind 23275 points

    Hi Howard,

    I cannot share those details with you here. But for Q100 products we usually use a CPK >= 1.67.
    In the end we only guarantee the MIN/MAX values in the datasheet to the customer. All I can say here is that they should see no difference in actual performance between the two versions.

    Regards,
    Joachim Wuerker

  • 0 in reply to Joachim Wuerker
    TI__Genius 12365 points

    Joachim,

    Since I have to explain to the customer, I have to know the theory myself.

    CPK= Min[ (USL- Mu)/3σ, (Mu - LSL)/3σ]

    My understanding is: for either ADS1118 or ADS1118-Q1, σ is the same, Mu is the mean value which is also the same. 

    ADS1118 will have smaller (USL-Mu) and (Mu-LSL), thus CPK is smaller.

    USL-Mu is the temperature accuracy error shown in the datasheet, right?

  • +1 in reply to Howard Zou
    TI__Mastermind 23275 points

    Hi Howard,

    I think we can use that as a good explanation.
    LSL and USL would be the MIN and MAX values shown in the datasheet or the test limits used in production.
    In this specific example we should assume Mu = 0°C temperature measurement error.

    Regards,
    Joachim Wuerker

  • 0 in reply to Joachim Wuerker
    TI__Genius 12365 points

    Joachim,

    if that's the case.

    Then my understanding is: if we find an ADS1118 whose temperature measurement error is 0.8C, we will discard the device(since it's worse than spec in datasheet), but if we find an ADS1118-Q1 whose temperature measurement error is 0.8C, we will still keep the device and it will release to the market.

  • +1 in reply to Howard Zou
    TI__Mastermind 23275 points

    Hi Howard,

    there is a little more to this. This parameter is not tested across temperature at final test. That means we use some guardband based on the device characterization results to determine the test limits for final test which is performed at room temperature.
    If the same test limits are used for both ADS1118 and ADS1118-Q1 but we use a larger CPK value for ADS1118-Q1 that would then mean we have to widen the MIN and MAX specifications in the datasheet.

    Regards,
    Joachim Wuerker