Good Monday All,


One of our customers has requested Mean Time Between Failure data. If you have knowledge of this, please explain what this is and how it is calculated.

Thanks,

Is this for a new design or do you have failure data?

This is new design. We have not produced this part yet. They also want HALT data (we should be able to provide this as we have had testing performed on similar parts).

This is new design. We have not produced this part yet. They also want HALT data (we should be able to provide this as we have had testing performed on similar parts).
If you have produced similar parts used under similar conditions, you can base your prediction from that data.

MTBF(hours) = Total Part Operating Hours/Total # of failures

I am not sure if we will be able to calculate MTBF because do not have enough data to give an accurate calculation. This type of product is new (we have only been producing it for about 3 months).

Mshell,

Please let me know what your product is.

Mean Time Between Failures is as Al stated. HALT is Highly Accelerated Life Testing.

You can estimate potential MTBF using a reliability block diagram (or a simulation) if you know some information about the components going into your product. I'll explain more when you provide info about your product.

We manufacture dome enclosures for the CCTV industry. The product in question has been manufactured and coated with a hard coat solution. These parts are produced using acrylic sheet and/or plastic resin the coated with a solution to help increase the life of the product.

As an aside, with your product they probably want MTTF (Mean Time To Failure) not MTBF, the subtle difference meaning once your product fails, it is not repaired, but replaced.

Anyway, If similar dome enclosures have been produced by you or other companies in the past, you may be able to get historical failure data that you can use to calculate MTTF (NOTE: very often the collection of historical failure data is woefully inadequate. When this is the case, interviews with end users, customers, and such may yield you anecdotal data [educated guesses]). If you have made improvements over past products you can estimate the degree of improvement. For example, if typical dome enclosures last 2.5 years (MTTF) and a new coating process is predicted to extend the life 1 year, you now have a 3.5 year MTTF estimate.

When reliability is extremely high (i.e. years before failure on average), then some reliability practitioners suggest accelerated life testing (i.e. tests at higher than normal temperatures, speeds, stresses, etc.) to estimate what the average failure times might be. These tests, in my opinion, are often suspect at best.

I hope some of this helps.

Thanks for the help. We are meeting later this afternoon to discuss this and I present this information to the group.

:thanks:

Mshell,
Did you finally find the failure data for the mentioned components? OR decided to perform your Accelerated Life Testing (ALT)?

If this would help,Reliability handbook from Dodson (1995) reference various failure data sources:
GIDEP
FARADA, IEEE,LER,NERC,NPRD,NPRDS,PVP, etc.
We use RELEX-PRISM software in our organization. I understand that RELEX also has links to various failure database.

In my personal opinion, If the component/module is identified critical to the function of an overall subsytem, I would rather prefer conducting our own Accelerated Life Testing (ALT).
Let us know if need any help from this forum or this thread is considered closed.
Govind.