Lost in translation

Today, myriad proprietary design and inspection languages exist, and none can talk to another without a translator. It's a situation sources liken to the Tower of Babel.

Quality managers at certain industrial heavyweights have a dream: for all digital part data—including elements related to design, manufacturing, and inspection—to be compatible and, in essence, speak the same language. It's something they call interoperability.

To those outside manufacturing, that goal might not seem so lofty. Windows® Office dominates cubicles around the world, so it takes little effort to share an Excel®, Word, or PowerPoint® document. But for those dealing with multiple CAD formats as well as vendor-specific languages for everything from coordinate measuring machines to laser trackers and portable arms, the feat would be nothing short of conquering the Tower of Babel.

Babel is an apt description, said John Horst, an engineer with the National Institute of Standards and Technology in Gaithersburg, Md., "because everybody in the city speaks a different language, and no one understands the other without a translator. All information is embedded in files that are in different languages. But it's really all the same design information: features, tolerances, cylinders, planes, points, and so on."

For years NIST has supported the interoperability cause, collaborating with organizations that form an alphabet soup of acronyms. The Automotive Industry Action Group (AIAG), a global organization of OEMs and suppliers, has its Metrology Interoperability Project Team (MIPT). European automakers have gotten together to form an organization they call I++, which promotes the I++DME, a kind of foundational command code for CMMs and, potentially, other metrology devices. There's the International Association of Coordinate Measuring Machine Manufacturers (IA.CMM). And then there's the Dimensional Metrology Standards Consortium (DMSC), a group that promotes the Dimensional Measuring Interface Standard (DMIS), a language that puts measurement data in a standard format. DMSC leaders come from companies like Chrysler, Siemens, and Deere & Co.

The interoperability issue not only affects users of CMMs, but other digital inspection devices as well, including laser trackers and portable arms. Often used for large parts, these devices compare points based on a CAD model. Horst described an interoperability issue a precision sheet metal fabricator might experience.

"You always need design information, and you need a drawing of some sort to do your inspection plan," he said. "Typically, these devices do their alignment based on an internal model they have of the CAD. They collect data and perform best-fit algorithms to align the measured points with the actual model.

"There can be an interoperability problem right there. Let's say one of your customers works only with a particular version of CATIA®, and you want to buy a new laser tracker. But for some reason, it doesn't work well with that particular version of CATIA, so you have to use a translator to translate all your models. Not only will the translation itself be costly, but errors and omissions in the translation may create quality problems. You also may choose not to purchase the new tracker, which may cost you a competitive advantage."

Most CMM software uses what are called direct CAD interfaces, a software kernel that communicates with specific CAD packages. As Horst explained, CMM software vendors pay license fees to CAD vendors, and these fees are passed on to the shops that purchase the CMM software. Beyond this, small manufacturers serving multiple customers often find themselves in a costly situation; each customer requires the company to support a different CAD package. The result is that "each shop must pay for additional training along with multiple license fees," Horst explained, "and this can get really expensive."

Interoperability advocates claim that the current situation costs billions annually, revenue that, quite literally, gets lost in translation. Because the largest OEMs now have supply chains spanning the globe, and because they must have interoperable systems, manufacturers and suppliers expend a lot in translation software and labor to ensure digital creations carry over to different software platforms.

"The mother of all interoperability problems happened at Airbus in 2006," said Horst. The issue involved a wiring harness for a jet prototype of the A380. "Well into the building of the prototype, they discovered the wiring harness would not fit into the plane" because of a conflict between different CAD software—and not between different software packages, but different versions of the same software.

Design Information Standardization

Interoperability starts with standard part data. Outside the digital realm, the industry has ASME Y14.5, which defines standard definitions for product structures, including the geometric dimensioning and tolerancing (GD&T) marks common on paper blueprints across industry.

Horst used this to give a hypothetical example. "Let's say we had a computer-readable equivalent to the Y14.5 that every user required, and every CAD supplier used only this—sort of like everybody speaking English. The savings would be huge, because a host of unnecessary costs would be eliminated, like format translation, multiple software license fees, additional training costs, and so on."

Here, the Standard for the Exchange of Product Model Data, or STEP, steps in. Supported by the International Organization for Standardization (ISO), STEP carries standardi- zation through all elements of designing, making, and inspecting a product. Within the overall standard lie various subsets that focus on different stages of design, manufacturing, and inspection. On the design end, for instance, is STEP's Application Protocol 203 (AP 203), an internationally standardized data file that captures the geometric and part attribute data from CAD needed for manufacturing.

Curtis Brown has high hopes for edition 2 of STEP AP 203. He's the current president of DMSC and principal engineer for National Nuclear Security Administration's Kansas City Plant, managed and operated by Honeywell Federal Manufacturing & Technology. The most recent permutation of STEP's AP 203 includes not only standard part dimension data, but also as Brown put it, "non-shape stuff," such as information regarding surface finish, the pressure at which a part will operate, and a representation of geometric dimensions and tolerances. In other words, he said, it exchanges more design-intent information to CAD; that data is clearly defined from the get-go, and then is flowed seamlessly through manufacturing and inspection based on a standardized, machine-neutral language.

At most job shops and small contract manufacturers, personnel mark blueprints to describe datums that need to be checked. But if complete inspection data becomes embedded into the CAD file, the story changes. The thinking goes that today's inspection devices, from CMMs to laser tracking devices, could tap into this design and inspection data from CAD to facilitate quality automation.

This happens already, but the process usually involves proprietary software platforms communicating with one another, like a certain laser portable arm communicating with CATIA. Interoperability advocates would like to use one standard inspection language that could draw from an intelligent, standardized design file, like STEP. The result: An engineer could design once, manufacture it, and inspect anywhere, regardless of the make and model of the inspection device, with no translating required.

Still, sources concede that the standardization push at the design level has had its roadblocks. STEP, for instance, has been a tough sell. "It lags the proprietary technology that's out there," Horst said. "Still, the issue starts with design, and it's probably the thorniest, most difficult problem for interoperability."

A Common Inspection Language

Interoperability supporters have pushed a set of standard languages into the metrology arena, and each represents a different level of machine control: a high-level measurement process plan, a low-level measurement plan, and a common operating command code underneath.

For that underlying command code, the I++ organization has promoted its I++DME command language for CMMs (and potentially other devices), basically giving machines the same digital foundation. For the high-level process plan—including basic part feature and tolerance information for automated inspection environments— metrologists on both sides of the Atlantic are developing several standards.

Between the high-level standard and underlying command code, a low-level standard spells out a programming language for dimensional metrology. It is this low-level standard that may matter most to a manufacturer's QA department, and here is where the Dimensional Measuring Interface Standard, or DMIS, enters in.

Ray Admire, the current DMSC secretary, is a staff quality engineer and lead CMM programmer for one of the world's best-known manufacturers, and DMIS, he said, would make his life easier.

Admire works for Lockheed Martin.

"One of the things I keep driving for," he said, "is to get everyone to require software and hardware providers to follow the standard. This doesn't mean all software has to execute directly in DMIS, but regardless of what offline programming system we use, we need to be able to export and transfer a file to DMIS.

"DMIS is like using the English language at the control tower at the airport," he added. Anywhere a plane lands, be it Chicago or Shanghai, pilots communicate with air traffic controllers in English. "If everyone is not communicating, you're not going to get good results."

DMIS would benefit suppliers too, Admire said. "Let's say we develop something in-house, and one of our suppliers takes over the product line. We would generate a standard file format measurement plan and provide that to our supplier, who would bring it into their system. Using DMIS allows me to do that."

Admire added that the standard is making significant headway in some areas, including CMMs, but it hasn't gained acceptance in other areas, such as portable arms and vision systems for measuring flat sheet metal parts. Nevertheless, he said, DMIS does incorporate programming elements integral to sheet metal part inspection. For instance, the latest version offers relative measurement, in which features are measured relative to each other.

An Ongoing Effort

Sources concede that the interoperability effort is an uphill battle. For it to truly take root, the world's largest OEMs need to step up and demand standardization of file formats. But some hesitate to do so and, according to Horst, for understandable reasons—particularly in the design phase.

"The design information is so important and so costly if it's messed up," Horst said, which is why most require their suppliers to work with certain CAD platforms.

"But what if you lost a particular CAD vendor? Think of the cost of translating all those files," he said, adding that if the industry standardized, translation wouldn't be necessary.

"To me, having a standard is a no-brainer."