Measurement Research Resolving CMM Uncertainties

A new company, called MetroSage, has been formed in California by some of the nation's leading metrology and software researchers to address this tricky, technical challenge. The company is a result of a decade-long research commitment made by the Consortium for Advancement Manufacturing--International (CAM-I). Work conducted by the National Institute of Standards and Technology also contributed to the technology driving the new company.

"Three years ago, CAM-I and NIST recognized that each was conducting complimentary metrology research work," says Woody Noxon, president of CAM-I. "The real measure of this successful collaboration between CAM-I and NIST is shown by the ability of a start-up company to take the technology and develop a new product that will have a significant impact on the manufacturing industry."

CAM-I was focusing on optimum measurement techniques assuming a perfect measuring machine and an imperfect part. NIST was studying uncertainties assuming a perfect part and an imperfect measuring machine. Working together, the two organizations have produced a successful result for industry.

MetroSage is developing a software program that will provide manufacturers with a far better understanding of the quality of their measurement processes. The company's product, called PUNDIT, an acronym for Predicts UNcertainty in Dimensional Inspection Techniques, is based upon complex mathematical formulas that calculate the effectiveness of a company's measuring systems.

Requirements for understanding and disclosing the accuracy of measurements on parts continue to become more severe, as manufacturers must produce parts of ever-increasing precision. "As part tolerances become tighter and tighter, issues that we used to just sweep under the rug and not worry about are becoming more important," says Steven Phillips of NIST's Manufacturing Engineering Laboratory, and one of the world's experts on manufacturing metrology.

For example, there is a well-established rule that a company's CMMs should be 10 times more accurate at measuring a part than the machine tool that produces it. This ten-to-one ratio is becoming more difficult-- and far more expensive-- to fulfill as tolerances become even more severe. It can become virtually impossible to know if a measuring system can determine the conformance of a part to engineering specifications. "Some experts have made enough measurements to have a good feeling for it, but for average users there are situations where the world conspires against you, and your measurement accuracy can be much worse than you expect," says Phillips.

CAM-I's work over the past decade was instrumental in creating the Dimensional Inspection Techniques Specifications (DITS). Now MetroSage hopes to take the research to the next step and introduce a commercial product.

The company's four founders, who also did the research work for CAM-I, have nearly a century of combined experience in software programming and metrology. "We're trying to provide insights into metrology that aren't presently available," says Kim Summerhays, a principal in MetroSage LLC, and a professor of computer science and chemistry at the University of San Francisco. "When you measure a particular part dimension, you come up with a number, but there is an uncertainty in your measurement. If the uncertainty is large, you can't say with great assurance whether or not the part is within tolerance."

Metrology issues are growing more important by the day. "Virtual" manufacturing enterprises are contracting out manufacturing capability on an ad-hoc basis, with one entity designing parts, another producing them, another performing additional processes on them, and yet another doing the assembly. "The question of whether the final product is good or not is leading to a lot of finger pointing," Summerhays points out. "People are saying, 'I measured it on my CMM and it looked good and you measured it on your CMM when you got the parts and now you say it's not good. So who's at fault?'"

The PUNDIT software simulates the part measurement given all available information. In many cases, it will tell manufacturers they are over-measuring their parts, and don't need to buy such an expensive CMM. This in turn will lead to lower costs and faster throughput. "It's probably true that for some folks, PUNDIT will be the bearer of some bad news; they don't know as much about their products as they thought they did," says Summerhays. "But there is a price to be paid for having one's head firmly planted in the sand. Improvement can only come when the truth is known."

The International Standards Organization is pushing the issue. In its current year 2000 re-write of the ISO 9000 standard it states that: "Measuring and monitoring devices shall be used in a manner that ensures that measurement uncertainty is known and is consistent with the required measurement capability." The United States is concerned about this language "because they realize how difficult it will be today to meet that standard," says Summerhays. "But it's coming; it has to come."

For more information about the CAM-I Technical Programs, contact Bailey Squier at 817/860-1654 Ext. 114 or visit the CAM-I Web Site: www.cam-i.org. For more information about MetroSage and the PUNDIT Project, contact Kim Summerhays at 415/422-6142 or visit the MetroSage Web Site: www.metrosage.com.