In Part 2 of this blog series, the author points out that too many shortfalls can occur from using a non-qualified measurement system, the two biggest being cost and time.
This is Part 2 of a three-part blog series on the science of measurement system analysis. To read Part 1, click here.
Measurement system development is a time-consuming task that process development teams often underestimate. The amount of information and data collected during development can be overwhelming. Measurement system analysis (MSA) helps to quiet the noise and allows the team to focus in on meaningful information, which gives everyone peace of mind. MSAs ensure data from the process, and are as accurate as can be with the chosen system/method adding limited variation.
For example, a developmental team tasked with building four class 102 pilot tools must define:
* material/part shrink information (critical to quality and to function steel safe dimensions)
* gate design/fill characteristics (melt flow)
* cooling profile
* overall production tool designs
A developed measurement system’s data accuracy plays a considerable roll in allowing a team to make correct decisions for each information point. Based on the critical need and accuracy of the information, team members should choose data from a challenged and qualified robust system. Too many shortfalls can occur from using a non-qualified measurement system, the two biggest being cost and time.
In order to choose, develop, and qualify the measurement system, the team needs a sound procedure guiding them. Part characterization plays a vital role in identifying the proper piece of equipment for gauge development. Measurement system variation is determined by conducting Gauge Repeatability and Reproducibility (GR&R). Characteristics such as human interaction, equipment and setup, gauge/fixture/tooling, and software all have variation. This variation can be isolated through GR&R in the following categories:
* Repeatability (variation of the measurement system)
* Reproducibility (variation of the individuals operating the system)
* Stability (processes free from special cause variation)
* Bias (influence factors that cause the data population sampled to appear different than it is)
* Linearity (measurements statistically different from one end to the other of the measurement space)
A few areas to focus on when developing the measurement system include types of data (variable or attribute), types of measurement (in-process, capability, fully automatic installation, etc.), measurement equipment selection/gauge development, measurement order, part orientation and whether the part would need to be cut to measure specific features.
Once a system is developed, it is time to challenge it. There are many methods available to perform the GR&R, so take care in choosing the right one. One of the most common methods used is ANOVA (analysis of variance). MSAs can include capability data (Cpk, Ppk), work instructions, qualification protocols, maintenance procedures, and measurement variation response plans. Typical (MSA) GR&R acceptability or confidence ranges are:
<10% robust system
10-20% acceptable system
20-30% may be acceptable based on importance of measurement; however, it should be followed up by a continuous improvement project
>30% system fail, not robust
MSAs also are useful during other phases of development and production. Once a robust system is qualified, it can be used for training personnel; to predict maintenance or end-of-life on mold components for tool refurbishment; re-execution annually to verify no degradation in the system or individual components; re-establishing data accuracy and annual training certifications for operations personnel.
Many process-improvement or cost-reduction projects can be derived from the in-process data collected using a qualified measurement system. One example would be in-process dimension-check reduction through a direct dimensional correlation study. For example, if dimension B shifts or goes out of tolerance, so do D and E. The data can be used potentially to propose removing dimensions D and E from the in-process checks, saving time and money.
Properly executed MSAs, equipment maintenance, and training will assist with eliminating measurement error or non-conformance investigations and reports; reduce unscheduled downtime, shorten validation cycles, and empower front-line employees. The end result is a robust measurement system that helps to direct sound, data-driven decisions.