PERFORMING THE INTENDED FUNCTION SATISFACTORILY | Six Sigma and Beyond: Design for Six Sigma, Volume VI

A product fails when it ceases to function in a way that is satisfactory to the customer. Products rarely fail suddenly in the way that a light bulb does. Rather, they deteriorate over time. This eventually leads to unsatisfactory performance from the customer's standpoint. Unsatisfactory performance can result from:

Excess vibrations
Excess noise
Intermittent operation
Drift
Catastrophic failure
And many other possibilities

Unsatisfactory performance must be clearly spelled out. The customer's perspective must be recognized in this process. There will usually be various levels of failure based on the customer's perceived level of severity. The levels of severity are frequently grouped into two categories such as:

Major
Minor

The severity of the failure to the customer must be documented and recognized in a Failure Definition and Scoring Criterion that precisely delineates how each incident on a system or equipment will be handled in regards to reliability and maintainability calculations. Such documents should be developed early in a design and development program so that all concerned are aware of the consequences of incidents that occur during product testing and in field use.

The design team must be able to use the failure definition and scoring criterion to address product trade-offs. If the severity of a failure to the customer can be lowered by design changes, the failure definition and scoring criterion should promote such trade-offs.

SPECIFIED TIME PERIOD

Products deteriorate with use and even with age when dormant . Longer lengths of usage imply lower reliability. For design purposes, target usage periods must be identified. Typical usage periods are:

Warranty period(s): A warranty is a contract supplied with the product providing the user with a certain amount of protection against product failure.
Expected customer life: Customers have a reasonably consistent belief as to how long a product should last. This belief can be determined through a market survey.
Durability life: This is a measure of useful life, defining the number of operating hours (or cycles) until overhaul is required.

SPECIFIED CONDITIONS

Different environments promote different failure modes and different failure rates for a product. The environmental factors that the product will encounter must be clearly defined. The levels (and rate of change) at which we want to address these environmental factors must also be defined.

ENVIRONMENTAL CONDITIONS PROFILE

The environmental profile must include the level and rate of change for each environmental factor considered . Environmental factors include but are not limited to:

Temperature
Humidity
Vibration
Shock
Corrosive materials
Immersion
Pressures, vacuum
Salt spray
Dust
Cement floors/basements
Ice/snow
Lubricants
Perfumes
Magnetic fields
Nuclear radiation
Weather
Contamination
Antifreeze
Gasoline fumes
Rust inhibitors /under coatings
Rain
Soda pop/hot coffee
Sunlight
Electrical discharges
And so on

Not all of these environmental conditions would be appropriate for a particular product. Each product must be considered in its individual operating environment and scenario. Environment must consider the environment induced from operating the product, the environment induced from external factors, and the environment induced by delivering the product to the customer.

RELIABILITY NUMBERS

The reliability number attached to a product changes with:

Usage and environmental conditions
Customer's perception of satisfactory performance

At any product age (t) for a population of N products, the reliability at time t denoted by R(t) is

R(t)	=	Number of survivors/N, which is equal to
R(t)	=	1 - (Number of failures/N) = 1 - Unreliability

This is the reliability of this population of products at time t. The real world estimation of reliability is usually much more difficult due to products being sold over time with each having a different usage profile. Calendar time is known but product life on each product is not, while warranty systems monitor and record only failure.

INDICATORS USED TO QUANTIFY PRODUCT RELIABILITY

Several metrics are in common use to indicate product reliability. Some of these actually quantify unreliability. Some of the metrics follow:

MTBF ” The mean time between failures, also MTTF, MMBF, MCTF. MTBF = 120 hours means that on the average a failure will occur with every 120 hours of operation.
Failure rate ” The rate of failures per unit of operating time. » = 0.05/ hour means that one failure will occur with every 20 hours of operation, on the average.
R/100 (or R/1000) ” The number of warranty claims per 100 (or 1000) products sold. R/100 = 7 means that there are seven warranty claims for every 100 products sold.
Reliability number ” The reliability of the product at some specific time. R = 90% means that 9 out of 10 products work successfully for the specified time.