Check Measurement System Analysis

Before we use our Resume data that collected from measurement activities, we must ensure that the data is representative shows the condition properly. The wrong data will be drive wrong decision and improvement direction that opportune to waste cost, waste time, etc.

How confident that our measurement is represent our condition? It question become our baseline to conduct the Measurement System Analysis.

Picture M.2 Condition I

Picture M.3 Condition II

The measurement system is influenced by many factors. Operator, gage/tool/instrument, environment, etc have an affect to the result that might not represent the actual condition of specimen.

Picture condition I shows that measurement result have bigger defect products (defect – out of spec) that actual is. If we trust with our data without verify it, may we decide to improve the condition because of the error occurred. We called this phenomenon is waste by company because the company will improve something that should be does not need improvement.

While picture condition II is contrary with the picture I. In this condition actual condition has bigger defect products that measurement result. If we trust our measurement without verify it, it will be no improvement on this process. This is very dangerous than condition I because customer will get low product quality than what company expectation. Whereas Company could be survive because of their customer. This gap caused by variation.

Observed process/product variation includes measurement variation as well as actual process/product variation. Our objective is how to reduce or eliminate the variation comes from measurement variation, so the variation that appear is comes from the product/process variation. Too difficult (or impossible) to eliminate measurement variation in our process, but we could be control it by determining good measurement system.

The Variation that comes from the gauge/tool/instrument is caused by Bias, Linearity and Stability. Bias is the differences between the observed average of measurements and the reference value. Linearity is the differences in the bias values trough the expected operating range of the gage. While stability is the total of variation in the measurements obtained with a measurement system on the same master or parts when measuring a single characteristic over an extended time period. The variation that occurs from the gage will affect to the accuracy of the measurement system. *) detail how to use and case study of this tools could be learned in Tool categories.

Variation that comes from the measurement actually can be caused due to operators and the gage/tool/instrument. Variation caused by operator could be comes from repeatability and reproducibility. Gauge Repeatability occurs if an operator makes different judgment when measuring same products/process that uses the same gauge (same products/process measured repeatedly use same gauge). While gauge reproducibility is a variation of measurement when operators (two or more) measure the same product/process use same gauge has different judgment. This variation will affect the precision of the measurement system.

Refer to the condition above, we must ensure whether our measurement system have appropriate represent the actual or not use the statistical study. We also ensure that the gauge repeatability and reproducibility not have significant effect in our measurement. To analyze this matter we can use statistical study. We call it the Gage Repeatability and Reproducibility Study (GR&R Study). This method will able to justify whether the measurement system appropriate to represent the actual or not. If the result is acceptable, so we can use the data resume that has been collected. While if there are not acceptable, we must re-design the measurement system in order to get the appropriate result. The way how to implement GR&R will be detailed on Tool categories.