Welcome to lecture 33 on measurement system analysis. On completion of this lecture, we will understand about the concept of measurement system analysis such as gauge, r&r, measurement, correlation, bias, linearity, percent agreement, precision, tolerance, etc. As we had discussed in the previous lectures that one of the objectives of measure phase is to identify the critical outputs and inputs using process knowledge such as key process output variables, key process input variables etc. But what After identifying those parameters we need to measure the variation of these parameters for this purpose, define a data collection plan and collect data to measure the current performance of processes. However, before we start measuring, we need to validate the measurement system is capable of generating accurate data. We can validate it conducting an analysis of our measurement system MSA stands for measurement system analysis.

In fact, it is difficult to find out a perfect measurement system. All measurement systems are subject to variation. The main source of variation in any measurement system our operators, gauge our equipment and the part itself. Two major issues of measurement system are repeatability and reproducibility of measurements. Every time we take measurements by MSA, we study the repeater. ability and reproducibility of gauge known as gauge rNr gauge r&r stands for gauge repeatability and reproducibility.

Repeatability can be explained as the variation in measurements of same piece of product taken by one operator with the same gauge. When same operator produces different measurements while using the same gauge what will be the reason either the gauge is faulty or it may not be calibrated that means The operator could not repeat the results because of a faulty gauge. In the similar way, reproducibility can be explained as the variation in the average of measurements taken by different operators, while using the same gauge to measure same piece of product when different operators uses same gauge to measure same piece and the measurement vary, it means operator is the source of variation. Hence, the gage could not reproduce the result because of untrained operator. The common MSA terminology are bias which is a systematic difference between mean of the measurement result and through value linearity which is a difference and bias through the range of measurements.

Precision to tolerance ratio is a measure of the capacity of the measuring system. percent agreement refer to the percent of time in an attribute gauge r&r where the appraisers agree with themselves. Let us learn how to conduct a gauge r&r. A simple way to conduct gauge r&r study is select the gauge used in the process under study, select three operators and then different parts get two or three measurements each for these 10 parts calculate each variation and total variation in measurements. Let us see how MSA is conducted as we have just discussed You can select three operators who are regularly involved in the process which is under improvement. Now, select 10 different parts say nuts, bolts, shafts etc.

Let each operator take two or three measurements of each part with same gauge, say with Vernier caliper are a micrometer we get around 60 or 90 measurements. Can you expect that all those 60 or 90 measurements will be exactly the same? Never Am I right. Each measurements are different There may be one or two same measurements to Now, let us see what are the three types of variation in any measurement system. Let me illustrate all three types of variation First of all, each part measurement will be different. This is called part to part variation.

What do you think? Should there be this variation? Yes, of course, because each part is different in itself. Now, the second kind of variation is in the measurements taken by same operator Same part with same gauge. If radiation is more, we see the results are less repeatable. And as we discussed, gauge is the source of variation for less repeatability.

Finally, the variation between average measurements taken by different operators using the same gauge and the same part what was the source for this kind of variation, but discussed in the previous slide. Here, the source of variation is the operator himself. If this variation is more, we say measurement system risk salts are less reproducible as we have just discussed, the measurements will show three types of radiation part to part variation it must be there as each part is different repeatability variation in same part measurements by same operator reproducibility variation in same part measurements by different operators. Now, what is the total variation in any measurement system? It is the contribution of all the above three variations. Your MSE is set to be validated If the combined variation of repeatability and reproducibility is less than 14 percentage of total variation at times 14 to 28% is also acceptable depends upon the cost of repair of cages.

Fine, how about calculating the variation that is similar to the calculation of mean and range. However, it is very tedious and let us learn to use Minitab for this purpose download and double click the file one dot m s A it consists of two sets of measurement data 60 measurements by three operators each, for each gauge micrometer and volt gauge. We show you the MSA for micrometer, you need to do the MSA bubble gauge yourself, practice. Click Start Menu, quality tools, click gauge study and gauge r&r in bracket cross gauge r&r study dialog box will open. Double click on C one sample c two operator c three micro meter respectively Click OK to see the results. The results are shown in two windows, the session window results and graphical window results.

Let us analyze both in session window area, you can see the statistical results What are we interested in the total gage r&r, just nothing but the total variation by repeatability and reproducibility under the percentage s we call them check by the total gauge r&r is less than 28% Now check below of session window results the number of distinct categories it should be more than five. If both conditions are met, we can validate that our measurement system is good. Let us analyze the graphical result. The graphical results are seen as session window results. Only difference is that the results are shown graphically. As you can see from the graph, total variation is contributed by the high percentage of part to part variation as well as total gage r&r repeatability The reproducibility and separate column graph it is also seen from the graph that this measurement system r&r is within the permissible range.

That's all for this lecture. Now, it's time to proceed to the next lecture on process capability study.