iSixSigma

MTTF

Viewing 7 posts - 1 through 7 (of 7 total)
  • Author
    Posts
  • #54040

    Mike Russ
    Participant

    Explain to me how MTTF works in minitab. When I use MTTF =30 for 5 samples, the test time I got 29.11. When I use 6 samples, I got 27.882. Why the requirement for test time is reduced with increase samples. If I test all my 6 parts up to 27.882 time, how do I meet my target of 30. Also, it is average time when we use MTTF=30. That means I may have one part failing at time 25 , 2nd at time 32 and so on and my average will be time 30? I am not clear about this especially when Minitab showing to test up to time 27.882 with 6 samples. Explain to me please.

    0
    #193097

    gomezadams
    Guest

    Please let this die!
    Reference your other discussions.

    0
    #193101

    Mike Russ
    Participant

    Thanks for detail answer in other discussion but it was not relevant to my lst question.

    Minitab question
    I am using “Demonstration Test Plans”. I want to test my part to atleast meet min requiremnt of 30 degree bend. Explain to me how MTTF works. When I use MTTF for 5 samples, shape paremerer=4, failure allowed =0, weibull distribution, I got 29.11 degree bending (testing for each uinit). When I use 6 samples, I got 27.882 degree bending. Why the requirement for bending is reduced with increase samples. If I test all my 6 parts up to 27.882 degree bend, how do I assure my 30 degrees bend? Also, as I undestand MTTF=30. That means I may have one part failing at 25 degree, 2nd 32 degree and so on and my average will be 30? I am not clear about this especially when Minitab showing to test up to 27.882 degree with 6 samples. Explain to me please what does it means.

    0
    #193106

    gomezadams
    Guest

    Read the attached article on the link.

    0
    #193107

    gomezadams
    Guest
    #193111

    MBBinWI
    Participant

    @mikeruss – the short answer is that you are using a statistical approach with confidence intervals and distributions. You are being given the least number of no failure test samples that statistically demonstrate some minimum distribution curve. To understand this further, you will need a course in reliability statistics.

    0
    #193123

    MBBinWI
    Participant

    @mikeruss – since both Joel and I recommended a reliability course, we really owe you a recommended & reputable source. The best that I’ve participated in was from ReliaSoft (and no, I’m not directly affiliated with them, just have partaken of their training and have used their product).

    0
Viewing 7 posts - 1 through 7 (of 7 total)

You must be logged in to reply to this topic.