AGMA Tech: Bill Bradley

The idea that inspection is non-value added is correct. However, to conclude that all measurements should be eliminated is drastically wrong. Read on to learn more.


The age-old question when trying to improve corporate profits: "is a function or process value added?" Quality managers have always been told by business-school corporate management that inspection of the product is a non-value added function, a waste of money. So it could be asked:

• Is there economic or technical importance to knowing your "measurement uncertainty"?

• Is it important to have traceable measurement calibration to a prime metrological institute?

To some, the answer to these questions may be a very clear yes, or an uninformed no. To others, just trying to understand measurement uncertainty is very difficult. Trying to cover the importance of "measurement uncertainty" and traceable gear measurement "calibration" in a column may not be possible. However, it is impossible to discuss one subject without the other.

The Value of Knowing Measurement Uncertainty
Inspection of a product in order to determine conformance to a design or specification usually involves measurement. It is performed for the purpose of deciding whether a product meets specific criteria and is satisfactory for a customer's application. Everyone knows that the inspection function has been treated as "non-value added." For this reason, companies continue to try eliminating as much inspection as possible.

The idea that inspection is non-value added is correct. However, to conclude that, since inspection is non-value added, all measurements should be eliminated is drastically wrong. Correct measurement practices can supply very important data. With good measurement data, a company can improve designs and processes to produce products economically with the quality required for the application. Understanding and applying measurement uncertainty through a process of traceable calibration is the only method that can ensure that the measurement data is of sufficient accuracy so that decisions are made correctly.

Companies that are struggling to improve their current technologies–investing in new measurement technology, trying to limit "non-value added" inspection, and that do not confuse inspection with other forms of measurement–are those which seem to be headed in the right direction. These companies understand that in order to eliminate costly inspection operations, high-quality data from reliable measurements is required.

Accurate measurements for other than inspection can be very important to a corporate economy. These measurements can be used to control manufacturing processes, analyze product test results, develop new products, develop functional tolerances, improve existing manufacturing processes, and promote corporate profits. The measurement's value is not determined by the cost of the measuring instrument, nor is it determined by the measurement accuracy. The measurement's value is determined by the importance of the decisions that can be made with the resulting measurement data.

Application of Measurement Uncertainty
ISO 18653:2003 Gears – Evaluation of instruments for the measurement of individual gears and ISO/TR 10064-5 Cylindrical gears – Code of inspection practice – Part 5: Recommendations relative to evaluation of gear measuring instruments (soon to be published) have extensive information and specifications on the calibration of gear measuring instruments with the application of traceable artifacts of known measurement uncertainty. The primary purpose of these two documents is to assist the gear manufacturer and purchaser to understand measurement uncertainty as applied to gear products. These two documents were developed from four AGMA documents: AGMA 931-A02, ANSI/AGMA 2110-A94, ANSI/AGMA 2113-A97, and ANSI/AGMA 2114-A98.

Application of standards can be valuable, but ISO 14253-1 contains requirements that a tolerance is reduced by the amount of measuring uncertainty to prove conformance to a specification, and that the tolerance be expanded by the measuring uncertainty when attempting to prove non-conformance to a specification. These rules provide a formal rigor to the process of proving conformance or non-conformance by requiring a manufacturer to produce within a conformance zone, and a user to prove that the measurement is in a non-conformance zone. These two zones are separated by the "uncertainty" range (see figure). But the problem is that a tolerance needs to be established as a "functional tolerance," with the concept of the involved measurement uncertainty considered. In addition, one needs to know when their measurements are adequate for the purpose, and whether the uncertainty of the calibration is a major contributor to their result. Very few tolerances for industrial gears have been established with these concepts. In many cases the tolerance is the one that the last designer of that product used.

The Need for Traceable Calibration
Unfortunately, although the application of measurement uncertainty may be well accepted in many fields of metrology, it is just coming into practice in the "art" of measuring gear teeth. Understanding and applying measurement uncertainty through a process of traceable calibration is the only method that can ensure that the measurement data is sufficiently accurate to allow associated decisions to be made correctly.

To improve an existing gear application requires measurement and test knowledge. A prime example of this need is found in the field of high power density gearing. The need to put more power through a smaller, more-economical gear set is required in many industries: automotive, aerospace, marine, and industrial applications such as wind turbines.

A persistent problem in surface hardened high power density gearing has been micropitting wear. To solve this problem there have been many theories and tribological tests performed. To correctly analyze the results of a theory's tests requires very accurate measurements of gear tooth flank contour and surface finish. It does not take much measurement uncertainty to have a substantial influence on the analysis of the test results. In a set of tribology tests, a 5 mm difference in crown height across a test tooth flank increased the surface contact stress 12 percent and resulted in a noticeable difference in performance. Without a known uncertainty and traceable calibration, this cause and affect of a different contour measurement may not have been discovered.

Is it economically or technically important to know your measurement uncertainty? Is it important to have a prime metrological institute and traceable measurements to guarantee quality, or to improve a product or process? You can bet your future–and that of our gear industry–that it is. As gear power density continues to increase, the necessity for lower "measurement uncertainty" also increases. Those who do advanced research and development must work through properly accredited calibration laboratories to establish reliable uncertainties and traceability. In some cases they will have to go directly to a prime metrology institute for the lowest calibration uncertainty.