The newest requests are for controls that offer open, modular architectures to gain life-cycle economies. Presently, users look at their controls and see that software is still unique to each controller. Upgrades and peripheral equipment are limited to components designed for a particular unit, or the customer must call in systems integrators to customize the control. Making even minor changes in a control’s capabilities requires an engineer and can take hours of reprogramming. It is also costly in terms of retraining, installation, and in new parts.
The world of the office personal computer is different. This device has hundreds of software programs ready to plug in and use, with more coming every month. Users can go to their local computer store and buy peripherals, such as additional memory, I/O interfaces, additional drives, CD-ROM, and programs in the form of plug-and-play cards or disks. Thus, devices can quickly offer the latest technological advances with economical life-cycle costs.
Several large manufacturers, especially General Motors, have noted this difference between the office PC and manufacturing controls and the engineers are asking why. Two years ago, engineers at GM reviewed their manufacturing operations and came up with strategies to cut their development and manufacturing time. But to implement these strategies, the engineers needed a control that they could upgrade with minimal additional investment.
“Such a control will have to be based on an open architecture that ultimately will facilitate the same rapid revolution for the shop floor that has occurred in office computing applications,” says James Albus of the Intelligent Systems Div., National Institute of Standards and Technology (NIST). The goal of this innovation is to enable suppliers to develop plug-in hardware and modular software for manufacturing.
Success breeds change
While GM is one of the primary drivers for an open architecture, other drivers include relatively small firms:
• Searching for ways to meet customer needs of mass production while simultaneously manufacturing a highly custom product.
• Searching for the best-of-breed among computer-control components. Users want to choose from the best I/O modules, the best peripheral interfaces and the best software programs — not just those components that work with a specific control.
• Demanding solutions that seamlessly integrate control and information throughout a company.
• Distributing microprocessor control and I/O to other devices.
• Developing innovative network strategies such as device-level networks, Figure 1.
• Seeking to leverage the technology shifts found in electronic component, software, and network technologies.
Because there is such a huge installed control base, the opportunity for open, modular systems revolves around application interfaces (APIs). These consist of software or hardware that will let logic engine, I/O, and human-interface components send data back and forth, Figure 2. There are no standard APIs yet. GM has taken the stand that defacto standards are acceptable.
In the works
NIST is working on a plug and play control it calls the Enhanced Machine Controller (EMC). It is currently under evaluation by the GM Powertrain Group in Pontiac, Mich. Some of its features include:
• Operator display with a look and feel that is common with other controls. Yet, operators can tailor the display to their preferences. The goal is to eliminate costly retraining.
• Common set of core features and interfaces to enable load-and-run software.
• Windows software.
• Software that links Windows-based programs to the control’s real-time operating system. An electronic message system shuttles communications among the modules and addon software and hardware.
Other approaches
The Allen-Bradley division of Rockwell Automation is moving from its present control strategy of distributed processing to a new strategy of distributed control, Figure 3. Their vision of the future control has no central controller, instead there’s a logic engine in every device, along with diagnostics. Says Bryan Parker, Business Manager, Open Control, “the company sees the need to maintain the independence of alternatives in control solutions.”
Cutler Hammer is working on a line of industrial PCs that it labels iPCs. Part of the new line includes a PC compatible software-hardware system, NetSolver, for configuring, viewing, and troubleshooting all control devices running on a DeviceNet network. “Combined with our other control components,” says Dave Williams, marketing manager, “this software can reduce life cycle control system costs by 40%. Cost reductions result from easier system design, less installation, and less downtime when compared to traditional control systems.”
Honeywell developers have been working on their open, modular control architecture for several years. The first step was the SDS device-level network. Now, they offer an industrial PC that will be an interim step between a bus network and their vision of true distributed control.
Their solution consists of open modules that can be packaged together or sold as individual pieces. These modules consist of SDS communications, I/O drivers, a real-time control engine, human-machine interface, and interfaces to other PCs, networks, and information systems.
Into the future
It’s a law in the world of computers that chip speed and capability double about every 18 months. It’s taken a little longer (about 20 years) for factory personnel to accept the computer, as a computer, on the plant floor. We are now at a point of dramatic (for manufacturing) change in the world of factory controls. The computer is fully accepted as a factory control device. All other computer-based controls, including PLCs, are changing to incorporate the more user-friendly functions and capabilities. Many of these traditional controls’ features still remain, as they are optimized for specific factory applications. But soon, all factory controls will be considered as, and labeled as, industrial controls, not PLC or PC.