The role of CAI in competitive manufacturing

Inspect: from the Latin inspicere – to look deeply, carefully, within

by Peter Marks

Editor’s note: This is the first of a two-part article. In the first part, the author looks at the evolution of inspection and CAD within the context of manufacturing trends. The second part, to be published in the September issue, will examine tools that are beginning to revolutionize the middle ground between design intent and real-world production.

Recent advances in computer-aided inspection (CAI) aren’t likely to be at the top of most manufacturing CEO’s things-to-survive-and-thrive list. But, that’s soon to change for a wide variety of products.

It’s obvious that basic product manufacturing is moving toward countries with lower labor costs and overheads. The trend started decades ago with such processes as forging and steelmaking. More recently IBM sold its personal computer business to Lenovo in China. Korean companies such as Samsung and LG are becoming dominant in products ranging from appliances to consumer electronics. And, the world’s largest company (WalMart) has perfected a supply chain that runs from low-cost providers to low-cost buyers.

All of which leaves manufacturing executives with two choices. One is to become the low-cost provider of commodity products. The other is to design and build products that are worth more to customers.

The low-cost path heads toward newly industrializing countries. Here we find low labor costs, increasingly skilled workers, and favorable regulatory environments. While low labor costs often imply greater reliance upon manual processes, newly industrialized countries sometimes have more modern production equipment than their competitors in North America and Europe.

The worth-more path leads to intelligent decisions about what customers need and how best to deliver it. For some safety-critical components such as circuit breakers or aircraft engine parts, it may mean near-perfect reliability. When buyers see the product as an extension of their personalities, such as with Nike in shoes, BMW in cars, Harley-Davidson in motorcycles, or an iPOD among MP3 players, it may mean a differentiated design.

When the product is literally an extension of the customer, as in the case of medical prostheses or custom-fit sports equipment, it may mean mass customization. Even in a situation where components of an assembled product are sourced to low-cost providers, the assurance that everything fits and integrates can be the key to industry leadership. Dell Computer is an example, along with the entire auto industry.

It turns out that new CAI technologies are a key to successful manufacturing in all these areas.

CAD’s reverse twin

In a sense, computer-aided inspection is the reverse twin to computer-aided design. CAD systems help construct a product point by point, line by line, surface by surface, and solid by solid. CAI deconstructs a product from solids and surfaces right down to critical dimensions of lines and points. This ability to inspect is, of course, the key to creating interchangeable parts, producing derivative products, monitoring manufacturing processes, and ultimately, assuring quality and reliability. In a never-ending chain of “as is” and “to be,” inspection and design trade places as a product and its manufacturing process evolve.

The great advances in CAD came when we went from automating plotting points in 2D to 3D wireframe design to 3D surface modeling and 3D solids. Now, even mid-range CAD systems are capable of fairly rapidly modeling entire products, including fairly sophisticated surfaces and shapes.

Inspection processes are undergoing a similar evolution, though somewhat behind the pace of CAD. For centuries, right through to the mid 1900’s, the dominant means of inspection was a visual look combined with point-to-point measurement of selected dimensions using rulers, micrometers, height gages, and similar measuring instruments. Comparators and vision systems allowed entire 2D profiles to be inspected. Coordinate measuring machines automated point-by-point measurements in 3D space – which is about the equivalent of 3D wireframe modeling on the CAD side.

When complex surfaces or whole parts needed to be inspected, industry defaulted to laboriously measuring a small sample of points around the part. “Productivity” meant automating this point-by-point measurement process and working hard to assure that a limited number of points was a good enough representation.

Millions of points in minutes

What’s new about CAI is that millions of points can now be inspected in the time previously required to measure a handful of points using traditional coordinate measuring methods.

The technology includes a variety of white light, laser and other primarily non-contact scanning methods combined with new inspection software. Scanners acquire thousands to millions of points – a so-called “point cloud.” The software uses those points (acquired from any contact or non-contact method) to reverse-engineer part geometries. The software “sees” the part in the “cloud” and then inspects critical dimensions and various features, compares entire parts and tools to their original CAD geometry, and provides a starting point for derivative designs or improved tooling.

While there are still many issues, we’re approaching the ability to “look” at a part and simultaneously measure millions of points. This is as important to inspection as the move from ambiguous 3D wireframe modeling to unambiguous 3D solids was to CAD.

CAI from a user’s perspective

We recently convened a distinguished group of CAI users and industry experts to discuss the implications of this technology for manufacturing. Most of the group worked in companies whose products must meet the highest standards for quality and reliability, such as circuit breakers and aircraft engine parts. There was additional experience in such areas as consumer product design, medical devices, quality assurance of outsourced molded parts, and the like.

A common theme throughout the discussion was the notion of intelligent process design. It’s one thing to copy a generic product design and aim for low-cost manufacturing through some combination of low-cost labor and off-the-shelf automation. It’s something else to deeply understand the relationships between product performance and manufacturing processes and evolve near-perfect processes for near-perfect products. That’s where the competitive edge, the worth-more for their customers, came for many in our group.

The transfer of responsibility from engineering to manufacturing is often described as throwing the design “over the wall.” The emphasis is usually on the difficulties in clearly communicating design intent. It turns out that this over-the-wall description understates the challenge, especially for products that must be manufactured to high tolerances. These include everything from snap fits in plastic housings to automotive, power generation, and aerospace equipment.

It appears that there is less of a wall and more of a moat between the world of nominal design and the world of quantity production. This chasm is filled with questions about tolerance design, GD&T (geometric dimensioning and tolerancing), process design, tool design, first tool inspection, first-article inspection, inspection planning, design changes – all followed by manufacturing and quality assurance over an entire product life cycle. Companies assume easy passage through this middle ground between design and mass production at their peril.

The ways in which different companies answer these pre-production questions are typically as much a matter of experience as science. A good example is seen when companies transfer or consolidate manufacturing at a different site. Even when moving manufacturing from one advanced facility to another, product performance and quality problems are common.

One example cited by a forum participant is moving production from a high-quality European supplier to an equally advanced U.S. supplier. Getting the process right involves many variables – it’s not just a matter of having documented designs and a modern manufacturing infrastructure. It’s even worse when manufacturing moves from a long-established site, with a tradition of quality manufacturing, to a low-cost labor region – nearly every company that has moved manufacturing has buried skeletons (and asked that we not mention the details).

CAD/CAM in the real world

The success of computer-aided design (CAD) and computer-aided manufacturing (CAM) is possibly part of the problem. Vendors promote the notion that passing a full 3D digital representation of the product (3D CAD) to manufacturing is an end to all those over-the-wall problems. If it’s that simple, why invest?

That might be true in an ideal world, where every designer foresaw the interaction of components in an assembly and where every part was precisely manufactured to its nominal dimension. In the real world, designers can’t foresee all possible interactions. Subtle differences in manufacturing methods and process capability lead to not-so-subtle product failures. As one participant noted, even seemingly simple cases – such as having two plastic cell phone housing parts snap together – can result in failures.

Business managers sometimes treat engineering design and production as two adjacent functions, perhaps separated by a wall in the old days and, today, connected by a broadband connection and “data exchange” or “collaboration” software. The notion of two distinct functions only works for the simplest and least mission-critical products. For everything else, there are dozens of critical processes, currently embedded in design, engineering, manufacturing engineering, tooling, inspection, quality control, and other functions that bridge the “moat” between nominal design and successful manufacturing.

It is now time, especially now with the rationalization of manufacturing worldwide, to give this middle ground the same attention, discipline and investment that has already been made with computer-aided design, manufacturing, and supply chain management.

Next issue: Exploring the middle ground.

Peter Marks is president of Design Insight, a consultancy for new product development. The firm’s methods of customer research, process improvement, and technology selection have helped clients add several billion dollars of incremental product and service revenue.