Problems in correlating porous coating pull-off strength test results through the use of test coupons led to the use of Taguchi methods to isolate the variables that contribute to test result differences between the customer and supplier of these coatings.
In the past, the supplier would test the coupons representative of the parts and ship the parts believing the pull-off strength requirements were met. The customer then tested additional coupons and usually found a lower strength (sometimes below specification) which resulted in rejection of the material. This, of course, led to problems. If both testing sources had correlated at a particular strength level, whether it be high or low, there would be no question as to the disposition of the material. Therefore, the major concern was that the numbers from both testing sources agree.
Since there are scores of variables that could affect the pull-off test, general categories were selected to determine major sources of variation that could be further analyzed at a later date if necessary. The four sources of variation (factors) that were selected included: (1) who performed the gluing operation on the coupons (the supplier or the customer); (2) at which facility the pull-off test was performed (the supplier's or the customer's); (3) what testing fixture configuration was used (two different types); and (4) at what speed was the tensile machine run (two different speeds).
Two analysis methods were used which concluded that the difference in fixturing was probably the most important source of varination While the other factors were significant, it was discovered that the highest pull-off strength values were obtained by having the customer glue the sample and the supplier run the test. This would mean that neither the customer nor the supplier had an overall better technique, further verifying the main source of vanation was in the fixturing.