# MTBF

Six Sigma – iSixSigma › Forums › Old Forums › General › MTBF

- This topic has 6 replies, 4 voices, and was last updated 18 years, 5 months ago by Ovidiu Contras.

- AuthorPosts
- March 6, 2002 at 7:19 pm #28941

Ovidiu ContrasParticipant@Ovidiu-Contras**Include @Ovidiu-Contras in your post and this person will**

be notified via email.Please help me with this one :

Data : 15 machines tested

: machines had the same design

: machines not tested in the same time ,but over a longer period of time

: each machine ran 500h without failure

Question : does this mean that we proved an MTBF of 2503h ,with 95% confidence?

Your help is much appreciated ,as usual.0March 8, 2002 at 12:58 pm #72982Remeber that statistically you only have a 38% chance of meeting calculated MTBF. Therefore if you have a goal of 2500 hours you must demonstrate 6579 hoours.

0March 8, 2002 at 1:02 pm #72983

Ovidiu ContrasParticipant@Ovidiu-Contras**Include @Ovidiu-Contras in your post and this person will**

be notified via email.Thanks for the reply ,and please forgive my ignorance (I am not a reliability expert) :

from where the 38% comes from ?0March 8, 2002 at 1:16 pm #72984I think the recognised way of doing this is involves the chi-squared function.

I do not have details to hand, but any textbook with ‘Reliability’ in the title will have it.0March 8, 2002 at 2:18 pm #72991March 8, 2002 at 3:01 pm #72995

A FriendParticipant@A-Friend**Include @A-Friend in your post and this person will**

be notified via email.You are correct IF you are willing to assume the failures follow an exponential distribution.

If so, the lower 95% confidence interval for the Mean Time To Failure (which for exponential is the same thing as MTBF) is 2503.56 hours, according to MINITAB’s Parametric Distribution Analysis – Right Censoring. This means that there is a 95% probability (confidence) that the actual, but unknown, MTTF is above 2503.56 hours.

This result from MINITAB matches the recommendations from Meeker & Escobar’s “Statistical Methods for Reliability Data” from Wiley.

From an exponential perspective, you have already tested 15×500=7500 hours without a failure.0March 8, 2002 at 3:14 pm #72997

Ovidiu ContrasParticipant@Ovidiu-Contras**Include @Ovidiu-Contras in your post and this person will**

be notified via email.Thank you very much for your responses .It is very much appreciated .

0 - AuthorPosts

The forum ‘General’ is closed to new topics and replies.