It’s Only Common Sense: Customers Need Confidence More Than Low Prices
Dan’s Biz Bookshelf: Revisiting 'The Popcorn Report'
- Copper anode supplier
- Copper film supplier
- What shift the product is run
- Relative humidity
- Day of the week
The reason these variables were chosen to be part of the noise variables is the company wants to develop a process robust enough to be run in a normal production environment. The reality of most any production environment is there are more noise variables than identified in this example.
So the high and low level settings could be:
Once the noise level variable is determined, the experimental design is similar to a fractional factorial design. Some statisticians have trouble with the “ambiguity” of the noise conditions. The goal is not to understand the noise conditions, but to create two distinct noise levels and see if results are affected.
The Taguchi experiment not only helps in determining what variable settings will yield optimal results, but also what parameter settings are the least affected by the noise. If noise parameters are correctly identified, and process settings are selected with zero noise interaction effects, then a robust process window can result.
Process Robustness vs. Optimal Results
It is possible a set of parameters could yield optimal results, and another set of parameters yields the most robust results. Achieving the optimal output, might require extreme or impractical controls on input variables. Alternatively, robust settings may achieve results that do not meet business goals. Management has to determine if the expense required to achieve optimal parameters justifies the added value provided to the business.
Fractional factorial experiments are very complex and require a good statistical analysis software package. The design and analysis can involve sophisticated statistical calculations. More important is the data interpretation. Management resource allocation is needed to support these types of process improvement efforts. Training employees on proper design experimentation is a key investment and can often result in significant improvement of the factory capability.
Dave Becker is vice president of sales and marketing at All Flex Flexible Circuits LLC.
Page 2 of 2
