Mean Time Between Failures (MTBF) Basics - Thoughts on Product Life Cycles

[Armtzcor]

Hello Everybady,

I need your help. Our company (within the field of manufacturing of electronic components for pc-boards, power supplies) have to implement Mean Time Between Failures (MTBF) but don't know anything about about how to start.

Do somebody there outside knows where can we find information about this or some formula, standard procedure (MIL, ANSI, etc...), method, etc...?

Please help.

Regards,
Arnulfo

[Bill Ryan]

Arnulfo,

"Reliability Statistics" by Robert Dovich (considered one of the "gurus" in the reliability field) is available from Quality Press (for $20-25 - US). This book contains just about every formula you may ever use with some explanation for each application (I recommend this book in any company's library involved in reliability testing!)

Another good reference (IMO) is "Reliability, Maintainability, Availability Assessment" by Michael Locks (from $61 to $75 - US).

Bill

[Randy Stewart]

MTBF can be determined by dividing the number of failures occuring over a given period into the operating time during that period. It is the average time that the equipment will operate without failure. Similarly, maintainability is expressed using Mean Time to Repair (MTTR). Using MTBF and MTTR and estimates of downtime costs, the Cost of Corrective Maintenance can be determined over the life of the equipment. Lets say Plant ABC estimates downtime costs of $300.00 per hour in machining operations and as high as $3,000 per hour in assembly. These costs account for the cost of maintenance manpower, the cost of lost productivity and increased cost due to additional overtime required. These $ figures can then be used in your MTTR calculations.

By using these figures along with the ones listed below you will be able to determine the "Life Cycle Cost" of any equipment. This can be very useful in justifying more cost for a new piece of equipment.


Life Cycle Cost Relationships

Life Cycle Cost (LCC) = Acquisition Cost + Ownership Cost

Acquisition Cost = F & T Cost

Ownership Cost = Cost of Operations + Cost of Maintenance

Cost of Operations = Cost of Manpower + Cost of Expendables

Cost of Manpower = Operator Cost x Number of Operators

Cost of Expendables = Cost of Electrical Power + Cost of Tooling + Cost of Lubricants, etc.

Cost of Maintenance = Cost of Preventive Maintenance + Cost of Corrective Maintenance

Preventive Maintenance Cost = Maintenance Labor Cost + Cost of Materials + Cost of Logistics Support System + Cost of Lost Productivity

Corrective Maintenance Cost = Maintenance Labor Cost + Cost of Failed Item + Cost of Lost Productivity + Cost of Logistics Support System

Maintenance Labor Cost = Cost of Maintenance x Number of Maintenance Actions

Hope this helps.

[Armtzcor]

Hey Guys,

Thanks a lot for your help, it was very helpful all of your comments.

Best Regards,
Arnulfo

[Manoj Mathur]

I was just thinking to start a new discussion and just through search I found this article on MTBF.
My Question to you is "IS MTBF ALWAYS CHANGING PHENONMENON?" In the year 2001 I had Three Brake Downs . I have calculated MTBF for that In the Year 2002 this M/C was so bad and troubled 8 Times. My MTBF was reduced. While In this Year It has never gone Down?
Can you please tell me What Is MTBF of this M/C?

Manoj Mathur

[Marc]

MTBF is an estimate or prediction. A component rated at 10,000 hours MTBF can fail 1 minute after it is first turned on. OR it could last for 30,000 hours.

The attached is from a paper from 1988 (scanned from my copy - can't find the disk file off hand) and presented to US Army TACOM in Warren, MI. that year. It's a typical life cycle 'bathhtub' curve. When you do a MTBF calculation, it gives you a location on the curve.

The original for this was derived from data from various aerospace (jet and turboprop) LRUs and found to be similar no matter what the LRU. It also correlated with manpack radios we made for field use and the early field computers.

The important part is to understand this is an estimate or prediction.

Just wondering - anyone have any current thoughts?

[Al Rosen]

Originally Posted by Marc

MTBF is an estimate or prediction. A component rated at 10,000 hours MTBF can fail 1 minute after it is first turned on. OR it could last for 30,000 hours.

The attached is from a paper from 1988 (scanned from my copy - can't find the disk file off hand) and presented to US Army TACOM in Warren, MI. that year. It's a typical life cycle 'bathhtub' curve. When you do a MTBF calculation, it gives you a location on the curve.

The original for this was derived from data from various aerospace (jet and turboprop) LRUs and found to be similar no matter what the LRU. It also correlated with manpack radios we made for field use and the early field computers.

The important part is to understand this is an estimate or prediction.

Just wondering - anyone have any current thoughts?

Marc:

I wanted to add that a reliability prediction is a way to compare designs for reliability and as you pointed out only a statistic.

I have MIL-HDBK-217F, RELIABILITY PREDICTION OF ELECTRONIC EQUIPMENT (14MB) and MIL-HDBK-472, MAINTAINABILITY PREDICTION (5.6MB).

If anybody wants a copy of either, send me a PM with your e-mail address and I will be glad to send them to you. They are too large to attach.