Adapting New Hampshire's motto, "Live Free or Die," controls designers must change to a distributed architecture to reduce cost and increase future flexibility or see their company lose its competitive advantage—"Live Distributed or Die." Distributing I/O modules closer to the action connected to the controller with one network quick-connect cable reduces total wiring costs, increases flexibility for future changes, enables OEMs to standardize, and reduces chance of wiring errors.

Charlie Cook, Wago (Germantown, Wis.) electronics product manager, puts it this way, "If an OEM is not proactively asking how to use networked I/O, the company will lose the marketing war. An engineer today must be open-minded toward the changes taking place and work with experienced vendors to shorten the learning curve."

Anyone who has visited a manufacturing plant is familiar with the current state of controls. Most noticeable when entering a manufacturing area is a wall of enclosures. Inside the enclosures will be power supply equipment, a PLC rack with controller and I/O modules, more remote I/O racks, banks of relays and motor starters, and lots of wiring duct and wire. The wire terminates in rows of terminal blocks, one wire per input and output. Wires attached to the other side of the terminal blocks meander along the machine or production line to various devices.

In a distributed system, there is still need for a power supply and a controller. The controller may still look like, or even be, a PLC. It may also be a computer in one of many disguises. The only other thing required at this point is a communications card.

Blocks of I/O are now independent of any PLC or custom rack. Most are DIN-rail mountable. Initially designed in blocks of analog, digital, or mixed signals, distributed I/O modules are increasingly granular. This gives the designer a way to place just as many I/O modules at the site as needed and add just what is necessary in future changes. Each distributed I/O block has a communications adaptor. The designer chooses the network optimum for the application, then matches communications adaptors for the system for that protocol.

Wiring now consists of one cable from the controller with drops (depending on the topology of the network) to each I/O location. Electricians now only run wires a short distance from the I/O module to the sensor, actuator, etc.

Derald Herinckx, GE Fanuc Automation (Charlottesville, Va.) remote I/O product manager, says, "When we conducted market research during development of the Versamax line, end-users told us they wanted to standardize on one communications layer so that it was easier to hook machines together. Meanwhile OEMs wanted to standardize on hardware as much as possible to cut costs. The answer is to develop an I/O system that can hook into any standard communications network."

Adds Laurie Davis, I/O product manager at GE Fanuc, "What both OEMs and end-users are trying to do is optimize control structure to each application." She suggests that engineers beginning this process analyze the physical layout of the equipment to determine the best component location, calculate required system response times, look at specific I/O needs like high analog or high-speed counters, then determine what network and I/O products support this.

Should anyone be surprised by the discussion of networks while talking about I/O devices?

As Siemens Energy & Automation (Alpharetta, Ga.) networks and I/O product specialist John Burns puts it, "You can't talk about I/O products anymore without talking about buses. Where once you talked about technology of the device, now you also look at the benefit of the wire. Since you can put the I/O device where it best serves its purpose, you don't have to build the machine around the control. You build the best machine for the application, then put the control where it works best."

This looks like more work for the engineer, and it is. The gain is freedom to design more functions in and more cost out of the equipment. Is this just a marketing dream of those seeking greater market share? No. It looks like many customer engineers are seriously weighing their options at design and specification time. Siemens' Mr. Burns continues, "We are seeing mo re head-to-head performance testing before deciding on the specification. We emphasize to engineers to look at real-world performance and to look beyond mere 'specmanship.'"

Graham Harris, I/O business manager at Rockwell Automation (Mayfield Heights, O.), sees increasing demand for openness and network compatibility from customers. He points to four trends:

• Collapsing architectures —I/O functions will be integrated with another object like a controller or device;
• Intelligent I/O devices create better information integration and commercially available technology;
• Distributed architectures permit shorter wiring runs with increased use of networks; and
• Reduced space —control panels will continue to shrink as more control equipment moves closer to the machine, and increased use of hardened I/O modules that don't require an enclosure.

Mike Rothwell, Advantech Automation (Cincinnati, O.) product manager, ventures a description of today's ideal I/O system. It is completely distributed with electronics and network interface integrated with the field device. This added intelligence would be cost-effective, not commanding a large premium price over its predecessor. Its high speed will enable control over the network. Intelligent devices handle local control of events without remote intervention. Finally, the system would be self-diagnosing, performing duties like reporting failures and requesting maintenance.

Danny Lease, Weidmuller (Richmond, Va.) technical support specialist for I/O products, acknowledges the strides diagnostics has taken with fieldbus connected I/O devices. "Testing out all the connections in those huge cabinets of control and I/O modules took a long time for a technician. Now the problem is located faster and more easily. Such common problems as short or open circuit are pinpointed saving much maintenance time."

Maria Piazza, senior marketing manager and former product manager, notes that Weidmuller looks at its business as connectivity. Open architectures have enabled it to move beyond terminal blocks and connectors. I/O modules are just another way of connecting field devices to the controller. The real benefit of this consolidation of functions is cost reduction—not just material, but also the labor cost of additional wiring and checking.

Bryan Moore, Entrelec (Irving, Tex.) product support engineer, lists savings obtained from new I/O strategies as cost and time savings. An OEM can standardize on one I/O device vendor, see inventories decrease, and have the ability to negotiate better prices with increased volume from the single source. A simple bus coupler change from one system to the next yields compatibility with customers' processes. Application software need not be changed from system to system, and standard configuration software for the low number of standard networks save programming time.

An OEM using a VME or CompactPCI controller lands a new customer using a different network. No problem, according to Mark Knebusch, industry and systems group manager for Phoenix Contact (Harrisburg, Pa.). Not only can OEMs realize benefits of machine optimization, but off-the-shelf network adaptors are available for these and other controller form factors.

Says Omron Electronics I/O specialist, Bill Arnold, "The most dramatic way customer acceptance of open I/O architecture has changed machine design is in final testing and shipping. Previously, an OEM would manufacture, assemble, test, disconnect, break down, and ship the machine, only to reassemble, test, and debug on the customer's floor. With distributed I/O architecture, each cell can be tested independently. Cells are connected with network cabling on the customer's floor and final system debugging is all that is left. This can save weeks, or even months."

Joe Benedetto, I/O business development operations manager for Schneider Electric's Automation Business (North Andover, Mass.), notes another benefit, both for OEMs and end-users—easy expansion of the system in the future. It is inevitable that changes will be made after the machine or process has been in production. Distributed I/O racks are now designed to allow easy insertion of additional modules. Most networks easily recognize the new point making it less painful to change.

As Greg Matthews, Wieland Inc. (Burgaw, N.C.) product manager says, "As far as I/O modules themselves are concerned, the major activity of companies is filling out a complete line of types. The challenge is to stay up with customer demand for the type of network desired. One other thing to watch is whether fiber-optic cabling catches on."

A note of caution comes from John Browett, product marketing manager of Mitsubishi Electric Automation (Vernon Hills, Ill.). "In our experience, most companies do not wish to fully exploit this flexibility due to the difficulties involved with working and negotiating with multiple vendors. Many companies want an open solution but wind up with a 'semi-open' one, because most purchases came from one vendor."

There will always be uses for PLCs with either plug-in I/O modules or built-in I/O points. Distributed I/O would not normally be useful in machines with a small footprint. Connect that machine to many others on a production line, however, and distributed I/O devices networked to each other and a controller makes great technical and economic sense.