If we only measure Useful Life

Does that mean the early life failures and wear out failures don’t count?

Designing to keep the useful life failure rates low is good design practice. This generally means a design that is robust, operates smoothly, incurs little temperature rise, and is as simple as it needs to be to function.

Setting reliability goals for the useful life period (the duration over which customers expect the item to provide the desired function) is common practice. Over this period we expect no dominate failure mechanisms and the mix of possible failures mechanisms provide a seemingly low and random failure rate.

It takes a lot of knowledge about the product and possible failure mechanisms to make this assumption and in my experience is rarely truly low and random. Making the assumption is easy. It supports the notion the design is and is going to be good.

Besides Useful Life

There are two other groups of failure mechanisms generally discounted when focused on the useful reliability goal: Early life and wear out failures.

All three types of failures, early life, useful life, and wear out, only provide a convenient way to group common failure mechanisms within three durations of interest to the business and customer. The consequence of a failure, independent of the underlying mechanism is still a failure and denies the customer the value of the product. Technically, a product can fail due to a wear out mechanisms in the first week of operation or fail after a long and useful life of operation due to damage incurred during initial shipment.

Products fail. They fail due to the mechanisms that wins the race to cause failure. All of the possible ways a product can fail exist in the product and it generally only takes one mechanism to cause failure. If that mechanism didn’t exist, the product may still fail the next hour of operation due to the next most likely mechanisms to cause failure.

We use the terms early life, useful life and wear out to roughly group failure mechanisms and link the durations to customer satisfaction and business objectives.

Early Life Failures (ELFs)

Imagine just opening and starting your new item (motor, cell phone, etc.) and later that day it fails.

Bummer, eh?

As a customer I still remember the purchase, the expectation of value the purchase promised and now it doesn’t work.

The failure could range from the promised function is not what you expected to a faulty solder joint. As the customer, we just know it doesn’t work for what we made the purchase to accomplish. Our thoughts turn to product return, troubleshooting, or the purchase of a different solution.

The common causes of product failure in the first days or weeks of product use generally (and I mean only very generally) occur due to:

• Incorrect or poor assembly
• Shipment or installation damage
• Faulty components
• Lack of margin in design for specific environment
• Unanticipated use or customer expectation

The duration of the early life period is arbitrary and commonly set by the customer’s expectations. When does the product lose the new car smell. When does the product enter common use and is beyond the point when the customer is particularly upset with a failure of a new item.

Some organizations define the early life failures as quality failures and focus on the supply chain, assembly and shipment/installation activities to resolve failures. It is common that process variation does lead to both products that fail early and those that work very well. While not concerned or aware of the useful variation, the customers let us know about early life failures. Process control and stability is often the means to reduce adverse variation that leads to early life failures.

Common durations include

• Setup and commissioning sign off
• First use
• First month of use
• Three months

The duration is arbitrary and should be set to help you and your team monitor and resolve failures. The failures and reliability goals for early life periods should encourage design and manufacturing team attention to the minimization of failure causes that commonly lead to early failures.

Thus a good reliability goal includes the probability of successful operation over the early life period.

Useful Life Failures

The purchase of a product generally comes with a common understanding of how long the device should operate. According to my dentist my toothbrush has a useful life of about 6 months. A cell phone should last the period of the service contract (currently 2 years). A car should operate at least as long as the duration of the loan payments.

The manufacture also has useful life expectations. For inkjet printers and razor blade handles the product should last long enough that the customer buys the consumable ink or blades for many years. The useful life should be well after the warranty period and not so long as the customer has no need to purchase another product in the future. The product should last long enough that the customer believes the product is reliable and provided sufficient value given the purchase price and cost of ownership.

Failures do occur during the useful life period and may have many causes. Failure may occur due to:

• Early onset of wear out failure mechanisms
• Excessive use leading to early wear out
• Latent defects caused by assembly or shipping damage finally causing a failure
• Excessive stress (lighten strike, dropped into pool, etc.)

Ideally the failure rate is relatively low and stable over the useful life period. Accidents will occur that destroy a product. With adequate design the assembly and shipping stresses will not weaken a product. And with proper design the onset of wear out mechanisms will not increase the failure rate significantly.

The duration of the useful life period is also commonly set by customer expectations and business objectives. Warranty may or many not define the useful life period, that is also a market and business decision. Common durations include:

• 30 years for solar panels
• 2 years for cell phones
• 1 hour for disposable coffee cups

The duration is arbitrary. The period should reflect your customer’s expectations. The duration provides the design team a goal to monitor when considering the onset of wear out failure mechanisms. Combine the useful life period and use models to consider number of operations for elements of a product.

For example consider the hinge on a flip phone. The hinge should function as expected over two years. If we assume 5 calls per day, and each call requires an open/close cycle on the hinge, the hinge may experience 5 cycles x 365 days x 2 years = 3,650 open/close cycles. We could test the hinge design and determine the design’s ability to remain in specification (I.e. Not wear out) when operated over 3,650 cycles.

The variation of stress induced failures (dropped off the side of building, run over by a train, or lightening strike, etc.) may occur. Depending on the product and how it is used the frequency of excessive stress may or may not significantly cause problems. If the use conditions include frequent dropping, then the design may need to include features to reduce drop damage.

Understanding the use conditions, use rates, expected range of stress and nature and types of over stress situation, along with the expected wear out mechanisms, define the design for reliability activities.

Wear out Failures

Product’s like people get old. As products age they may experience an increase in the probability of failure due to:

• Corrosion
• Material degradation
• Abrasion
• Creep or crack propagation

And, the product may still failure after the useful life period due to any of the failure causes listed above.

When defined correctly the useful life period defines the point when customers with a failure agree the product was reliable. A failure after the useful life period confirms the product has lasted long enough.

When considering product reliability in design, it is common to consider the range of factors that lead to wear out type failures.

Summary

What ever the failure mechanism, all products will fail. When a product fails matters. The customer response and the companies obligation change over the life of a product. Early life failures adversely impact brand perception and customer satisfaction. Useful life failures diminish the value the product produces for the customer. Wear out failures should occur as expected and may improve brand loyalty.

Failures occur and provide a means to learn about the appropriate product improvements.