Reliability Handbook

Hi, Im a reliability engineer, im reading the Reliability Handbook written by Analog. I have a question about the first example on the "Failure Rate Calculation" on page 14-15. 

Im plugging in the numbers on the equation to see if I get the same, and Im not. I do get the same results for the acceleration factor using the Arhenius equation, but im not sure how the "No. of Device Hours at Test Temperature" and the "Equivalent Device Hours at 55°C" are calculated. 

This is the formula for the first one:

Ndt = Nd × Nh × AT

Ndt=77*168*77 using the first component AD7357. 

The results are different on Table 4, and im not sure for which component are those calculations.

Is it possible to get this explained?

Thank you,

Daniela

Parents
  • 0
    •  Analog Employees 
    on Jun 8, 2017 8:07 PM

    Hi ,

    I think I see where the confusion lies.  Table 4 calculates equivalent device hours at 55C by the formula Ndt=Nd*Nh*At.  From the table, Nd*Nh=924000 (device-hours) and At=77.  Therefore equivalent device hours is 924000*77=71148000.  

    It sounds like you are OK with how the acceleration factor At of 77 is derived from the Arrhenius equation, so we won't go into that. 

    To understand where device hours of 924000 is coming from, we look at Table 3.  Table 3 is a sample of data collected on parts all on one process, so we are looking at the summation of all parts at each test temperature.   The table shows us that AD7357, ADA4830 and AD8229 were tested at 125C, so we will consider those in aggregate.   The zeroes in the table indicate when along the duration of the test that the devices were evaluated for rejects.  So we can infer that AD7357 and ADA4830 were tested for 1000 hours and AD8229 was tested for 2000 hours.   Each part had a total sample size of 77*3=231.    So are total device hours are 231*1000+231*1000+231*2000= 924000. 

Reply
  • 0
    •  Analog Employees 
    on Jun 8, 2017 8:07 PM

    Hi ,

    I think I see where the confusion lies.  Table 4 calculates equivalent device hours at 55C by the formula Ndt=Nd*Nh*At.  From the table, Nd*Nh=924000 (device-hours) and At=77.  Therefore equivalent device hours is 924000*77=71148000.  

    It sounds like you are OK with how the acceleration factor At of 77 is derived from the Arrhenius equation, so we won't go into that. 

    To understand where device hours of 924000 is coming from, we look at Table 3.  Table 3 is a sample of data collected on parts all on one process, so we are looking at the summation of all parts at each test temperature.   The table shows us that AD7357, ADA4830 and AD8229 were tested at 125C, so we will consider those in aggregate.   The zeroes in the table indicate when along the duration of the test that the devices were evaluated for rejects.  So we can infer that AD7357 and ADA4830 were tested for 1000 hours and AD8229 was tested for 2000 hours.   Each part had a total sample size of 77*3=231.    So are total device hours are 231*1000+231*1000+231*2000= 924000. 

Children
No Data