Introduction

Tolerance Analysis has traditionally been thought of as an activity that happens toward the end of detailed design. The goal is to appropriately set the dimensional controls or tolerances on the individual components of a product in order to balance product quality and cost. This narrow view of quality control misses an important factor in the quality of the product, the robustness of the design itself to variation. How sensitive are the critical parameters of the product to the variation of the controlling elements of the design. This paper will introduce how Sensitivity Analysis combined with Tolerance Analysis can have a greater impact on overall product quality at a lower cost than traditional Tolerance Analysis.

Find a Robust Design Concept

Quality can be addressed in the concept phase of design by understanding the sensitivity of the critical product parameters to the location and orientation of the critical interfaces of the design. Even before dimension schemes are known and process selections are made important decisions that affect the eventual product quality are being made. For example, a combination of a 4-way and a 2-way locating pin is a common design approach to locate two parts relative to each other. The distance between these locating features will have an impact on how sensitive the relative orientation of these two parts is to the location and size variation of these same features. The closer these features are to each other the tighter the tolerance will need to be to achieve the same relative orientation control between these parts. Having a model of the design that allows for the analysis of these sensitivities provides an important first step in controlling quality. Various design concepts can be analyzed to find the concept that will be the most robust to the variation that will be introduced during manufacturing.

Model the Manufacturing Process

As a design concept progresses through detailed design and pre-production further decisions are being made that influence the product quality. Manufacturing datum schemes are established along with associated dimensional controls. Understanding the sensitivity using tolerance analysis of the critical product parameters to the identified datum and dimension schemes will allow for an intelligent analysis of the trade-offs of these various options. The same principle applies to the decisions surrounding assembly. Fixture locations and contact types will have different effects on the propagation of variation through the assembly. Being able to appropriately model and understand the sensitivity of various fixture configurations to the product quality is important to making objective decisions on various datum and fixture schemes.

Quantify the Product Quality

The sensitivity model described above has been used for making trade-off decisions but has not given an indication of expected product quality. Once we have a robust design and manufacturing concept, final process selection decisions can be made. Actual process data can be used with the sensitivity model to make an accurate prediction of the expected product variation. Objective decisions can be made to find the process that minimizes cost, reduces cycle time and still meets the quality objectives of the product. This is the area that is traditionally thought of as tolerance analysis, but now there is a robust sensitivity model that forms the foundation for this analysis. In addition, the steps of a traditional tolerance analysis are starting from a better starting place. There is less risk of needing major design changes to accommodate quality goals because the current design is robust to variation.

Understand the Root Cause of Defects

Investing in the creation sensitivity model not only gives a foundation for doing more accurate predictive tolerance analysis, but it also provides a foundation for doing a more analytical root cause analysis. If an unforeseen failure is manifest during production the sensitivity model is easily extended to add this new product parameter. The sensitivity model can then be analyzed to identify the design parameters that have the greatest effect on this new product parameter. Adjusting one of these high sensitivity design parameters will have the greatest impact on fixing the identified product failure. In addition, since the sensitivity model is still intact all previous sensitivity and tolerance analyses can be run to insure they will not be adversely affected by the new design change.

The Role of Geometric Dimensioning and Tolerancing

Geometric Dimensioning and Tolerancing, or GD&T represents a way to define the size, location, orientation, and form of a part feature. A special set of symbols are used to define relationships between a feature and a measurement reference. The GD&T control, along with a tolerance, any modifiers, and a datum reference frame are grouped together in a feature control frame as shown below.

The feature control frame, sometimes referred to as a GD&T callout, communicates not only the allowable feature variation and the inspection references, but also the design intent of the part. GD&T is created within the CAD system, such as Pro/E, CATIA or SolidWorks, and may also be created by GD&T Software such as GD&T Advisor by Sigmetrix.

Conclusion

Having a sensitivity model that relates the critical product parameters to the various parameters of the design provides a foundation for improving quality even more that traditional tolerance analysis. A design concept that is robust to variation can be identified. The proper dimension schemes and assembly fixtures that will help control variation can be put in place. Objective decision related to product quality and process selection can be made. Finally, root cause analysis of unforeseen defects can be performed more efficiently.