HOW PRODUCTION VOLUME CAN AFFECT A PRODUCT’S DESIGN —> Part 2: Improving Reliability

BY MIKE JONES

die cast transmission

INTRODUCTION FOR IMPROVING RELIABILITY

Improving reliability starts with design. Product failures due to poor reliability can create customer dissatisfaction. This in turn may drive customers to alternative choices and reduce sales. When warranties are in effect, replacement or services costs can erode profitability.

Reliability modeling and testing can be helpful to estimate failure rates. In combination with production volumes, the cost of unreliability can also be estimated. With high-volume products the financial motivation to improve reliability can be very significant. Improving reliability begins with the design. Testing can generate a better understanding of the design and create strategic opportunities to make design changes and improve reliability. Here are some methods one may use to improve reliability.

Simplify the design
  • Frequently there are multiple concepts that can achieve a product’s performance goals. An important step in achieving high reliability is creating that “elegant” concept that does the job most simplistically. Compared to alternatives, this concept may use a different and more robust technology. Or, it may have fewer parts. For example, achieving a necessary motion by using a pivot rather than a linkage. Part count can also be reduced by integrating more features into fewer parts. This is often seen in sheet metal, injection molded, and cast parts.
  • Reducing the number of product features will reduce the functional requirements and supporting component count. To do this requires a critical review of the product requirements and finding the right balance between features, performance and potential sales.
die cast transmission
Example: A die cast transmission housing with integrated fluid passages. Image courtesy of NADCA.
Example: On the left a three piece sheet metal assembly with two fasteners. On the right a single piece die cast part. Image Courtesy of NADCA.
Making the design insensitive to noise
  • We are using the term “noise” to categorize factors that are not controlled and may affect the product’s function. Often experiments are used to identify noise factors and their effect. This may be combined with system modeling to develop a better understanding of the effects and to generate potential solutions. Here are some examples:
    • Experiments with shielding to reduce EMI or RF emissions
    • Modeling a robotic arm to develop a controller that is robust to varying payloads
    • Experiments on a media feeding system to understand the effect of humidity, temperature, thickness, etc.
Example: A Simulink model of a drive system used to predict behavior with various disturbance loads.
Design of key components
  • The application of engineering methods to the design of the components can be critical to achieving system reliability. On mechanical systems this is often determining the loads on a component and sizing the component to achieve a stress and or strain level set by a safety, functional, or fatigue life goal. Here are just a few of the application areas where we can apply this. Sizing of bearings, threaded fasteners, power screws, belt drives, weld joints, rivets, adhesive joints, springs, gear trains, and shafts.
Example: determining bearing life based on loading

FMEA

A common tool to improve quality is Failure Mode and Effect Analysis. Initially, this technique was developed to improve aircraft safety. Subsequently, it has been adopted by many industries to improve product quality. FMEA is a process that can be applied during product design. Most often it begins after concept selection. The process has several steps in which a team identifies potential failure modes and rates them for severity, rate of occurrence, and detectability. Each failure mode gets a Risk Priority Number (RPN). The result is a prioritized list of potential product failure modes. The goal is to address high priority failure modes through design modifications that eliminate or reduce occurrences, decrease severity, or increase detectability. As the design changes the RPN’s are updated. This is a proactive approach to product design and reducing failures and warranty costs.

Testing

There are many forms of testing used to prove and improve product performance and reliability. Results fed back to engineering create opportunities for design changes and improvements. Testing can be costly and time consuming. These factors have influenced the creation of accelerated testing.

Why do accelerated testing?

  • Reduces test time
  • Reduces quantity of test samples (prototypes)
  • Improves reliability and reduces warranty exposure
  • Increases yields
  • Increased product durability/robustness

Below, are two common types of accelerated testing used during development:

  • HALT – Highly Accelerated Life Testing: combines cyclical stressing with aggressive changes in factors such as increasing or decreasing temperature, humidity and vibration. This is to reduce test time and increase the probability of failure, often used to identify simple fixes such as connectors coming loose or bracket fatigue, may identify effects due to thermal expansion mismatches, delamination, galvanic reactions, changes to viscosity, etc.
  • ALT – Accelerated Life Testing: applying higher stresses to force failures and identify weaknesses, uses fewer prototypes than conventional life testing done at moderate stress levels, identifies design defects more so than wear out modes, used to determine the “weakest links”. Although it is not desirable to add the expense of testing to production, this form of accelerated screening may be useful
Example: A belt drive ALT test applying higher stresses with acceleration and load torque.

Although it is not desirable to add the expense of testing to production, this form of accelerated screening may be useful

  • HASS – Highly Accelerated Stress Screening: similar to HALT but implemented during production, uses aggressive temperature change rates and vibration that are near or beyond the operation limits. This may force quick failures without impacting the product life. Customers only receive the robust units that pass the test. Thus shipped product reliability is improved.

CONCLUSION

We hope Part 2 of this 3-part series gives you a better understanding of Improving Reliability with consideration of production volumes.

Stay tuned for Part 3 and 3 of this series, coming soon. We’ll discuss more about reliability and stability of the product supply chain, minimizing changes, the safety and health of those involved, and the environmental impact. Feel free to contact us if you have questions.

Image of Mike Jones

About the Author: Mike is a Systems Engineer at SIGMADESIGN with a background in controls design and system integration. He is a certified “Black Belt” in “Design for Lean Six Sigma”. He enjoys the creative phase of the design process and working cross-functionally with the teams to achieve customer goals.