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.

ISO7821: ISO7821FDWWR, would like to confirm FIT Values on TI-Web

Part Number: ISO7821

Hi, TI-engineers,

I'm J.Komura of TI Distrbutor TED and we received question from our customer for TI indicated FIT's values on Web.
Customer are examining several deveices' FIT values for their new applications.Then below two FIT are different than their calucalation result.

(their others are as same as TI values.)

                                TI-Web / Cust-Cal
ISO7821FDWWR: 0.4FIT / 0.16FIT
AMC1311BDWVR: 0.8FIT / 0.17FIT
*Used parameters are as same as TI-Web data.
By my calucration, those are same as customer's.

So would you please give advice how caluclate those.

Thanks and best regards,
J.Komura@TED(Tokyo Electron Device Inc.)

  • Hi Komura-san,

    Thank you for reaching out. Please allow me to comment on ISO7821.

    I am giving below a link that describes in simplified version how we calculate various quality and reliability parameters including FIT.

    https://www.ti.com/support-quality/reliability/reliability-terminology.html

    Please note that FIT for a given device is calculated collectively across that device family which for ISO7821 includes all devices from the ISO78xx family. Hence, the 0.4 FIT listed in the website is different from what customer can calculate using the parameters available on the website. The actual calculations use a much larger sample set and FIT is dynamically calculated by internal tools and then pushed to the website. Hence, we recommend customers to consider TI provided FIT.

    I hope that answers your question, thanks.


    Regards,
    Koteshwar Rao 

  • Electron Device)

    Rao-san

    Thank you for your response and kindly advice for our customer question.

    It's little bit difficult your comment, so please check my understanding as below.


    " Customer calculation with TI-Website information is correct.
    However some TI device's sample size of FIT parameters are included all family products.
    And Website system can calculate each device FIT correctly. However used sample size for each
    device for FIT parameter are not reflect to Website."

    About 0.4FIT at TI-Website vs 0.16FIT(calculated with TI-Web size),,
    0.4FIT are worse than 0.16FIT,, this reason is used by smaller size from Website info.
    So TI-Website FIT info are correct values."

    If not enough, please give advice.

    Thanks and best regards,
    J.Komura@TED(Tokyo Electron)

  • Hi Komura-san,

    Please see my inputs below.

    " Customer calculation with TI-Website information is correct.
    However some TI device's sample size of FIT parameters are included all family products.
    And Website system can calculate each device FIT correctly. However used sample size for each
    device for FIT parameter are not reflect to Website."

    Your understanding is mostly correct here. FIT is calculated for the whole device family and all those devices share the same FIT value. Hence, individual device part number FIT cannot be calculated by using the sample size provided on TI website.

    About 0.4FIT at TI-Website vs 0.16FIT(calculated with TI-Web size),,
    0.4FIT are worse than 0.16FIT,, this reason is used by smaller size from Website info.
    So TI-Website FIT info are correct values."

    If we use the sample size, test duration and number of failed units for calculating FIT, we will get 0 FIT as there are 0 failed units. I am not sure how customer got a FIT of 0.16.
    TI uses a much larger sample size across product families that share the same process to find FIT for overall device family and that for ISO78xx is 0.4.


    Regards,
    Koteshwar Rao

  • Rao-san,

    Thank you for your advice again.

    I think it is possible to answer the difference in FIT value to the customer in the first half of the sentence.

    However, since the FIT value is the reciprocal of MTBF / MTTF (Mean Time Between Failures),

    the closer the FIT is to 0, the better the performance.

    So we'd like to know why the FIT numbers with bad values are indicated.

    And below URL can help to get FIT values easiler and it shows 0.16FIT of customer result with TI papameters.

    https://www.maximintegrated.com/en/design/design-tools/calculators/general-engineering/qafits.html

    And TI FIT data as below,

    www.ti.com/.../report

    Thanks and best regards,

    J.Komura@TED

  • Hi Komura-san,

    Thanks for sharing additional inputs.

    I understand that when you use a calculator from Maxim website, FIT is coming out to be 0.16 while on TI website it is 0.4. I didn't see Maxim providing any equation on how they calculate FIT and hence, I cannot comment how they have calculated 0.16. 

    Unfortunately, we will not be able to share specific details or equation related to calculating FIT. Hence, I recommend suggesting customer to consider TI estimated FIT rate for the devices. Thanks.


    Regards,
    Koteshwar Rao

  • Rao-san,

    Thank you very much for understanding customer issue.

    Today I explained your mention of FIT rate calualation. But it's not enough for them.

    They must estimate (technology)FIT values under their Ta condition for meeting customer requirements.

    And then, when 0.4FIT is calculated back, by about 3K sample sizes it is close such FIT.

    So can we say that they can refer to this numbers?
    Or if there is another way to get the right parameters, please let me know.

    Thanks and best regards,

    J.Komura@TED

  • Hi Komura-san,

    Thanks for the update.

    They must estimate (technology)FIT values under their Ta condition for meeting customer requirements.

    For calculating FIT for customer Ta, please use the below calculator. I am copying below a snapshot of an example calculated value for Ta of 75C and the FIT value is 1.7 with same activation energy of 0.7 eV. Thanks.

    https://www.ti.com/support-quality/reliability/temperature-change-FIT.html


    Regards,
    Koteshwar Rao

  • Rao-san,

    Thank you very much for your nice advcie.

    We could clear customer questions with this info.

    Thanks again and best regards,

    J.Komura@TED