Technology Forecast '98

Here's a look at some of the key technology trends in computer hardware and software, electronics, fluid power, materials, and motion control that are definitely worth taking a serious look at this year.

COMPUTERS

Dual-processors swing onto desktops

Michael Puttre, Associate Editor

Wider might be better for snowshoes, surfboards, and automotive wheel-bases, but when it comes to computer processors: More is better! At least, that's the conclusion of a select but growing cadre of engineers.

Intergraph Corp. (Huntsville, AL) reports dual-processor configurations account for 35% of the sales of its Pentium II-based workstations (as opposed to servers, which routinely have multiple processors). And quad processor systems account for 5% of sales. The remaining 60% today are single-processor models, but Intergraph says it expects the number of multi-processor workstations it ships in the future to grow steadily.

Why two processors? At first glance, it might seem obvious that a floating-point-intensive program would run twice as fast with twice the power. However, in order to distribute operations between multiple processors, a software program must be written with this capability in mind. Not many engineering applications have this so-called "multithreaded" functionality. Nevertheless, Windows NT supports multithreaded operations, and industry observers say software vendors are just waiting for a critical mass of multi-processor workstations to reach the market.

A Catch-22 could be in the offing, with hardware and software vendors waiting on each other, except multiple processors confer other advantages as well that are driving their acceptance. Basically, dual- processor workstations can walk and chew gum at the same time. "It's like having two computers on your desk," notes Steve Pesto, manager of Intergraph's TDZ workstations. "It is possible to run, say, an FEA analysis using one processor while continuing work on a solid modeling CAD program unhindered on the other."

A raft of new multi-processor workstations suggests the idea is catching on. The new IBM IntelliStation M Pro workstations from IBM tackle complex calculations with dual Pentium II processors in configurations up to 300 MHz. On-board 512 Kbyte L2 cache and SDRAM memory promise improved application performance. Power workstation provider NetPower is offering its new Symetra3 line of workstations with dual Pentium II 300 processors for high-end video content creation and rendering applications.

The trend is not limited to the Intel-based desktop.

The Sun Ultra 2 is Sun Microsystems' highest-performing desktop system of the UltraSPARC product family. The 64-bit UltraSPARC microprocessors are housed in plug-in modules, enabling single- and dual-processor versions to share the same chassis. Silicon Graphics makes its Octane workstations, unveiled earlier this year, available with single or dual MIPS R10000 processors.

Hewlett-Packard has gone to great lengths to be able to pack a second specially designed J282 CPU module into the same desktop "footprint" as a single processor model.

ELECTRONICS

Flexible circuits roll on more apps

Rick DeMeis, Associate Editor

Electronic-device packages get smaller almost daily. So stuffing circuits into available nooks and crannies is more and more of a design challenge. Now the advantages of flexible-film mounted circuits have been known for a long time--resistance to shock and vibration and temperature stability due to the forgiving component carrier--but this market looks to grow geometrically by the year 2000 because of rising demand for consumer products that are small, lightweight, and portable. Flex circuits conform to a variety of geometric volumes, thus they can be crammed in just about anywhere. And users are demanding more performance from smaller devices, insuring that circuit capabilities will be growing as their size shrinks.

Starting in the early 1980s, efector (Exton, PA) began using flex circuits. Now its product line includes photoelectric sensors, flow meters, actuator/sensor interfaces, and electronic pressure switches. PC applications, beginning with touch pad switches, have grown. And cell phones and telecommunications devices demand flex use to allow circuits to fit in ever-shrinking device envelopes--and to survive the shocks of portable handling.

Bob Boyes, marketing manager for Poly-Flex Circuits (Cranston, RI), says his company relies on two basic flex circuit technologies. One is a copper-Kapton^TM flex circuit for high-temperature conditions. This process is subtractive--the circuits are etched out of the copper deposited on the Kapton substrate. The other technology features a polyester substrate over which printed-circuit lines are screen-printed using a thick-film silver-loaded polymer ink. This technique is about 20 times cheaper and faster than etching the copper, but electronic resistance of the ink is higher than for copper metal, and the temperature resistance is lower. Circuit pitch is as low as 0.02 inch.

Poly-Flex has perfected Poly-Solder®, a silver-loaded polymer adhesive to attach surface-mount devices (SMDs) to the substrate while limiting thermal stress by maintaining thermal-expansion compatibility between chip and board. Some components, such as resistors and thermistors, can be "printed" directly on the substrate. Multi-layer and double-sided flex boards are produced on a single substrate layer by alternating conductive and dielectric layers. The conductive circuits can function while bent around a 0.125-inch radius, while in the SMD areas, this radius is 4 inches. And flex circuits may be directly laminated to a housing, eliminating fasteners.

Boyes cites increased flex circuit uses where low-cost, lightweight solutions are needed for small package geometries. "Our circuits have found new uses in conforming to case contours of printers and other devices while incorporating LED indicators. And a new laptop keyboard use cuts the volume previously taken by three hard PCBs by up to a factor of five. Copper-Kapton flex circuits are being used for dispensing ink in a cartridge printer, while handling 8A of current in a hot environment." He adds other markets include appliances, medical sensors (DN 6/9/97), and automotive applications. Because of the booming market for flex circuits, the company more than doubled its plant size in 1995 from 3,000 to 7,000 ft².

In looking ahead, Boyes notes, "The flexible circuit market is growing dramatically. As consumer electronics get smaller and more ergonomic in design, the need for electronics adaptable to the profile of the goods increases. As a point of reference, just think of the first 'fat' cellular phones versus the palm-size Motorola StarTAC®--quite a difference in size and shape."

FLUID POWER

Hydraulics board the bus

Charles J. Murray, Senior Regional Editor

Taking a cue from the pneumatics industry, hydraulics component manufacturers expect to increase availability of fieldbus-based products in 1998.

Though they don't foresee a giant increase, manufacturers believe it should be significant for users. The reason: Demand for fieldbus-compatible hydraulic valves and controllers is on the upswing in industrial automation and mobile equipment design.

The move toward such products is no surprise; indeed, fieldbuses have been available in industrial automation circles since the mid-1980s. And most pneumatic component manufacturers have offered fieldbus products for at least five years.

But the hydraulics side has been slower to market with such products, mainly because of operating power concerns. "Pneumatic valves usually use two- or three-watt coils," notes Don Caputo, director of training for Parker Hannifin's Hydraulic Valve Division. "Hydraulic valves use 80-watt coils."

To solve that problem, hydraulics component makers have developed valves that can operate on the lower currents carried on the fieldbus. Vickers, in joint development with Festo Corp., has introduced multivalve assemblies for two fieldbus protocols, INTERBUS-S and PROFIBUS. Similarly, Parker Hannifin plans to roll out an industrial valve remote I/O package for DeviceNet in spring, 1998.

On the mobile equipment side, development of fieldbus-based valves and controllers is even further along. The reason: Mobile hydraulic valves require less power than industrial valves. As a result, suppliers have rolled out products that are compatible with the CAN-based buses used in mobile applications. Vickers, for example, markets its EMV-611e mobile valve bank for CAN-based systems. And Parker Hannifin last year introduced the MMC100 CAN-based motion controller for mobile applications.

OEM engineers are applying such products to the design of cranes, backhoes, and aerial lifts. There, fieldbus-based components play a key role: They provide "coordinated motion." By employing coordinated motion, mobile equipment designers can prevent the unbalanced loading conditions that can cause cranes and backhoes to tip over.

On the industrial side, the prime advantage of fieldbus technology is wiring reduction. By employing a fieldbus node, a machine's components can share a central four-wire cable, instead of being individually wired. Fieldbus marketers say that the process helps cut costs and save assembly time.

Despite the obvious advantages of such systems, however, the trend toward fieldbus technology in hydraulics is likely to be gradual, particularly on the industrial side. The main reason for the slow increase is the presence of too many bus protocols. By some estimates, there are now more than 60 different protocols. "We can't develop products that plug into every possible protocol," notes Fred Phillips, director of advanced technology for Vickers. "Industry needs to converge on a limited set of standards that we can realistically support."

Still, experts say that the need for fieldbus technology will continue to grow in 1998. "Most of our major automotive customers are either experimenting or dedicating themselves to fieldbus technologies," Phillips says. "And all of them want to connect hydraulic equipment to those buses."

MATERIALS

Plastics and metals battle it out

Gary Chamberlain, Senior Editor

There's an all-plastic car in your future! Yeah, you've heard that story before. But that time may be coming sooner than you think.

Late last year, Chrysler unveiled the Composite Concept Vehicle (CCV) with a body made completely from plastic. Each inner and outside panel of the body was produced in three minutes by Cascade Engineering (Grand Rapids, MI) using a 9,000-ton injection-molding machine. The project also features a new grade of 15% glass-filled Impet PET developed by Ticona Corp. (formerly Hoechst Technical Polymers; Summit, NJ). It is impact modified, but has a low enough melt viscosity to fill out the large panels.

Chrysler built the five-seat prototype for such developing countries as China, India, and Mexico. The four large panels are bonded with adhesive and attached to a steel frame. The left and right inner and outer panels provide structural support as well as decorative finish, according to Ken Mack, executive engineer and CCV program manager. The plastic doors and hood are molded separately. If the developing-country project takes off, North American models may be just around the corner.

Other polymer advancements to keep an eye on this year include: A new class of lower-price thermoplastic elastomers (TPEs) said to provide exceptional oil resistance and very rubber-like properties--a combination previously available only in TPEs costing 25 to 50% more per pound. The product: Flexalloy® OR from Teknor Apex (Pawtucket, RI). Expected applications include automotive, wire and cable, and other products exposed to hydrocarbon-based oils.

Meanwhile, Applied Sciences Inc. (Cedarville, OH) has debuted Pyrograf 3TM, a patented carbon fiber made by a combination of catalytic and chemical deposition processes that could significantly reduce the cost of certain composite materials.

Also on the composites front, Dow-United Technologies has just introduced a process that enables the manufacture of complex aerospace components with "shaped unidirectional fiber preforms." The process substantially increases the quality of the parts at the point where two or more sections are molded together.

But don't count metals out. Last fall, the American Iron and Steel Institute (AISI) and the Steel Manufacturers Association (SMA) published the Steel Industry Technology Roadmap. The report will serve as an agenda to guide public and private research investments toward the projects of greatest value to the North American steel industry.

Moreover, copper is on the attack against aluminum in the auto world. "Many in the automotive industry assume that aluminum is the most cost-effective choice as the material for radiators," exclaims Johan Scheel, vice president of the International Copper Association (ICA). "We are out to prove them wrong."

Universal Auto Radiator Manufacturing Co. (UAR) will begin using a new manufacturing process, CuproBraze®, developed by ICA to make the first commercially available brazed copper-brass radiators for cars and trucks. Scheel says the process "offers significant advantages over the Nocolok process used for aluminum radiators, including less energy usage, lower scrap rates, wider temperature margin in the brazing furnace, and substantially greater throughput rates."

At last year's 5th International Conference on Advanced Materials & Processes, Co-chairman F.H. Froes from the University of Idaho reported on new RST aluminum materials that offer high strength, corrosion resistance, high modulus,low density, and elevated temperature strength. Expanding on powder-metal advancements, he also revealed exciting new developments in mechanical alloying, aluminum foam as a replacement for wood or sandwiched materials, and a nanostructured Ti bone implant, which combines strength with benign behavior.

Plastics lead a ‘quiet’ revolution

In 1998, plastics will continue to move into more automotive applications once dominated by metals. For example, thermoplastic compounds offer design engineers the opportunity to reduce automotive noise, vibration and harshness (NVH) problems at a lower cost than metals, says Emile Homsi, senior design engineer for AlliedSignal Plastics, Morristown, NJ.

Design News: How will this come about?

Homsi: Engineering plastics materials allow this to be achieved at lower cost because they can be made at net shape, rather than requiring machining-to-shape as metals do.

Design News: We have already noted reduced NVH and increased performance provided by thermoplastics in such underhood applications and air-induction, climate-control, and engine -cooling systems. What’s next?

Homsi: Automakers are now considering applications in muffler and exhaust systems as well.

Design News: From an engineering standpoint, what makes this conversion possible?

Homsi: Noise and vibrations can be reduced through careful systems analysis and design planning that takes into account all noise-source factors of the automotive component. AlliedSignal’s design engineering team provides strategic and complete technical development support, including systems analysis and servicing. We accomplish this by placing our engineers on customer work teams, then allowing them to fully participate in all phases of the design program.

Design News: Where else might we see this NVH analysis being applied?

Homsi: This same technology is being used to reduce NVH problems in a wide range of other applications: lawn and garden, power, and consumer products. Typical applications include chainsaws, blower fans, and air-conditioning fans. In the case of fans, our engineers adapted a strategy of increasing air delivery performance by optimizing flow distribution across the fan blades.

MOTION CONTROL

Motion devices make the connection

John Lewis, Northeast Technical Editor

Faster, more precise actuators and increased focus on software development are on tap for motion control in 1998. Together, they'll help engineers improve performance, and cut development time. Moreover, the trend toward device-level connectivity continues to gain momentum. Primarily because the peer-to-peer communication it provides between motion controllers and machine controllers translates into big savings when integrating different types of technology in complex applications, as is often the case.

Although only a few device network applications in motion control have actually sprouted, the seeds are spread for the coming year. "Virtually every supplier already has or will introduce something with device-level capability," says Sal Spada, senior analyst at Dedham, MA-based Automation Research Corp. Vendors of intelligent drives, PLCs, PC-based motion cards, and even companies that purchase board-level product to integrate into proprietary systems, will include more device-network content with their products.

Nobody knows for sure which network protocol will win. Each is designed to address specific needs, such as communication with PLCs, PCs, or EMI immunity. That's why vendors anticipate selecting one or two initially, and then maintaining multiple network support. In fact, there is increased interest in using TCP/IP or EtherNet to transfer data up from the controller to the host, and DeviceNet down from the controller to the machine.

Industry guru Tom Bullock of Fond du Lac, WI-based Industrial Controls Consulting (ICC) claims that SERCOS is here to stay in the motion industry. "They have been waiting for something better to come along, and it just hasn't happened." Bullock's assertion is backed by one of the fastest-growing companies in the business, Santa Barbara, CA-based Motion Engineering Inc. (MEI). MEI's Product Manager Bates Marshall explains that their new MCD/SERCOS-4100 MultiDrive not only incorporates a SERCOS interface with DSP technology, but higher data-rates of up to 10 Mbits are now available.

No matter what track the industry takes, Ken Crater, president of Hopkinton, MA-based Control Technology Corp., sees the underlying trend as tighter integration of motion control and communication technology. "Motion controllers don't exist on an island anymore, they must function not only in the trivial sense of making the physical connection to the information infrastructure, but must also function as a full participant in the network."

Successfully extracting information from the machine and distributing it to all the different stakeholders in the plant hinges on tighter integration and reduced complexity. Crater envisions integrating motion control and communications the same way that Control Tech. has integrated motion control with PLC capability-- by using the technology that provides ease-of-use, integration, and unlimited access to information for the masses: the Web! In fact, Control Technology Corp. will offer an embedded web server in their existing controller line later this year.

Motion control vendors will continue to ride the coattails of the computer revolution, electronics advances, and software development to help engineers design systems that are easy to use, and tightly integrated.

SOFTWARE

CAD advances in functionality

Deana Colucci, Associate Editor

As 1998 rolls in, look for the features and functionality in mid-range CAD products to continue to advance, increasingly narrowing the gap in functionality between them and their big-ticket counterparts. Late '97 brought major enhancements to Solid Edge in version 4, while major updates to several other packages, including SolidWorks, are due out early this year.

The affirmation by mid-range developers of the importance of a Windows-native product pushes on. New entries into the field are sure to step up with native packages, as was the case with Computervision's recent DesignWave introduction. Even in the high-end, companies such as SDRC, EDS Unigraphics, Parametric Technology Corp., CoCreate, and FEA companies ANSYS and The MacNeal Schwendler Corp. have addressed the issue by offering Windows versions of their products (See DN 11/3/97 for more on the growing interest in Windows products).

Watch for the continued explosion of intranets in manufacturing organizations, as well. "People have discovered that this technology supports them particularly well in times of manufacturing and engineering crisis, such as increasing globalization and increased outsourcing," says Lisa Tomlinson of Cambridge, MA-based consulting firm Daratech Inc. "Intranet and Web technologies have become the lifeline of new distributed design teams." (See DN 5/19/97 for more on the Internet and engineering).

And vendors are trying hard not to be left out in the cold. "Both CAD and PDM companies continue the rush to web-enable their offerings as users find out that, for comparatively modest investments of time and money, they can have highly efficient and highly productive corporate and engineering intranets making data more widely available to more users than ever before," notes Daratech's Bruce Jenkins. Indeed, late '97 saw a number of announcements of enhanced web capabilities from major vendors, including Smart Solutions' new SmartWeb^TM PDM product. 1998 may bring major strides in the area of Web-enabled CAD with the promised delivery of Bentley Systems' Microstation/J--the first mechanical CAD application to feature Java as its primary programming language.

All of this will lead to increased investment on the part of mechanical users in solutions that they believe will increase their competitiveness and timeliness to market. If 1997 is any indicator--worldwide CAD/CAM/CAE user spending topped 18% according to Daratech--1998 will be a record year.

The multi-processor housing shortage

"Heat and power are the killers with dual-processor machines," says Guy Wagner, a senior engineer at Hewlett-Packard. Much of this challenge stems from the requirement that a dual-processor computer share the same chassis as its uni-processor forerunner.

 "We had a box that originally housed a 35W processor," Wagner relates. "This jumped to over 100W for the multi-processor version. Frankly, we weren't sure we could do it."

The solution proved to be a combination of clever internal packaging with an eye toward air-flow and innovative component design. HP engineers performed extensive thermal and electromagnetic interference analyses to discover just how tightly things could be packed. A "fan-sink" combining the cooling properties of a fan and a heat-sink was developed that could be mounted directly on the new J282 CPU module.

Silicon, sister,...got what it takes

The next step in flexible-film circuits looks to be the advent of devices using silicon layers thin enough (15 microns) to flex. Irvine Sensors (Irvine, CA) is developing these attached to Kapton layers, via a "flexible adhesive," for medical sensors that can be stuck on like a band aid. The circuits will also contain an RF or IR transmitter so no patient cords will be needed--ideal for operating rooms, and emergency and battlefield situations.

The company sees availability of the first device within 18 months--a nerve pulse sensor for use in rehabilitating patients with muscle injuries. It would track their progress and monitor pain. Processors in the sensor compare the muscle signals in sequence to determine if any muscle twitching occurs.

But first fruit of the flexible-silicon effort actually looks to hit the consumer-photography market in 1998. Bruce Totty, Irvine Sensor vice president, corporate marketing and sales, says the silicon-film technology "allows making a film canister full of electronics." Which is what the company has done with its E-Film--a 1,280 × 1,024-pixel, 24-bit color--imaging CMOS array and electronics "film can" that drops into a conventional film camera--instantly transforming it into a digital device. A higher-end professional model will have about 4,000 × 3,000 pixels and 32-bit imaging. Cost for the consumer version is initially targeted at well under $1,000. Software on the photographer's PC will correct any color saturation shortcomings compared to film and allow other image manipulation.

The drop-in E-Film module permits a photographer to retain the cameras and precision optics acquired over the years and couple them to modern digital processing and image manipulation. The device typifies what Totty sees occurring more frequently in the future, "Now any open space in a device can have an electronic system built in without redesigning the package."

Fieldbus: more costly now, for some

By now, most design engineers know the reasons for employing fieldbus technology: wiring reductions, diagnostics, simplicity, and cost cutting. Those advantages are said to be even greater on huge automation systems with high I/O counts.

But what about smaller systems? Bruce Flesher, founder of Device Tech, a custom automation integrator in Hampshire, IL, says that in those systems, fieldbus technology may not always live up to its promises.

To determine the relative value of the technology, Flesher compared two automated systems: a connector assembly machine that employed a traditional PLC-based system and a circuit breaker assembly machine that used fieldbus technology. Both had I/O counts of approximately 250.

The contrast between the two was remarkable, he says. The fieldbus-based circuit breaker assembly machine, which used DeviceNet protocol, cost about twice as much to build and required 150% more labor. "By the time we finished writing the code, we'd invested more than 1,000 hours of engineering," Flesher explains. "Wiring time took another 250 hours." In contrast, the standard PLC-based system needed less than 500 hours of labor, including engineering, assembly, and wiring.

Much of the additional labor stemmed from software troubleshooting, Flesher says. "If you don't have a resident expert who knows C++ or Visual BASIC, you're in trouble," he adds.

Flesher has compared other systems, but none of those matched up as well in terms of size and I/O count. On average, he estimates that PC-based fieldbus technology is at least 40% costlier than conventional PLC technology.

Flesher believes that the advantages of fieldbus technology might be more evident on a very large factory automation system. But for the machines he builds, he's convinced that it's too costly today. Nevertheless, he see a definite future for fieldbus technology. "It just needs to be more defined," says Flesher, "and industry needs to find easier and faster ways of working with it."

Putting autos on a steel diet

Recent developments by the steel industry could put a damper on the rumor of steel's death in the automobile industry. In fact, steelmakers have committed to an ambitious technology development program to keep their product competitive for the long term. Some of the projects on tap include:

"Our (steel makers) goal is to prove to our customers the untapped potential of steel in a car," says Douglas Tyger, manager of applications engineering for AK Steel Holding Corp. (Middletown, OH).

Cyber contacts:

Embracing intranet technology

Like many companies in the OEM, Advanced Energy Industries Inc. (Fort Collins, CO) is making the move on-line and creating an intranet. Kevin Pennock, CAD administrator, offers a little insight into the company's changing situation.

"Right now we print drawings and store them in notebooks for people to use as reference. We wanted to eliminate that system and replace it with an electronic solution that would enable us to view drawings and components, but we weren't focusing specifically on creating an intranet. Our use of MicroStation (Bentley Systems, Exton, PA) just seemed to logically lead us in that direction.

"With ModelServer Publisher (also from Bentley Systems), our first goal is to use the intranet to view all of our mechanical design files, and to allow people who don't have Microstation to view them as well.

"We also plan to create component libraries of reusable parts such as brackets. Right now we have no way to see what a part looks like, other than going to document control to view the actual drawing, which is not always helpful. We want to categorize our parts in the system and be able to do searches. That reusability of components will probably be the biggest benefit from our intranet.

"In the future, we'd like to have our vendors dial in through the Internet, access all the parts they are working on, and be able to create and download IGES files. Right now, our solution for exchanging data with our vendors is time consuming and high-maintenance for administration.

"We do have a few obstacles to overcome to set up the system, though. For example, we need to get Internet access so we can have our vendors access our intranet. And we do have some concerns about the security issues involved with allowing external people to access our files. Plus, there is always some fear involved when implementing something new, and when dealing with a product, such as ModelServer Publisher, that is relatively new. Overall, though, we think we'll benefit a lot from our new intranet."