Twenty years ago, Federal Express introduced overnight shipping and changed everyone's definition of "fast delivery." Now the entire catalog industry delivers in days, not weeks. McDonalds not only set the standard for "fast food" but literally rewrote the rules of speed and convenience for the entire restaurant industry. These two companies have changed modern business.

Innovations redefine expectations. Changes in one business affect another. The Internet has made information available anytime, anywhere, on the user's terms. In the industrial equipment world, this has exposed an obvious incongruity between what should be possible, and what is.

In a highly connective world, it is no longer acceptable to purchase a drive or temperature controller, only to discover there is no easy way to communicate to the programmable controller, industrial computer or distributed control system. It is no longer acceptable to find out that a $2000 "black box" is needed to link two seemingly straightforward devices!

Vendors perpetuate fiefdoms to protect their market share, creating confusion and harming their customers. Linking industrial devices today is much too difficult, expensive, and time consuming. It impedes productivity instead of enabling it. This paper was written to clarify the urgency of breaking down the brick walls in automation.

Consider the Internet revolution itself. In only five years, the Internet has gone from a bleeding-edge curiosity to a mainstream necessity. The gains and benefits have been monstrous. When legions of people have the tools to easily connect and communicate, the result is a massive surge in productivity and the proliferation of entirely new opportunities.

The revolution is underway. The days of Fieldbus wars, single-vendor solutions, and proprietary protocols are numbered. The following details Synergetic's mission to revolutionize factory communications from the bottom up -- at the device level.

A Simple Vision: Easy Interoperability and Integration of All Devices

Despite the fact that the Internet and TCP/IP protocols are far more complex than any industrial protocol, ordinary people can buy a PC, bring it home and connect it to the Internet in thirty minutes. This simple fact creates an expectation of what should be, and great frustration over the hours, days, and even weeks of effort which are required to configure automation systems and networks.

Automation professionals must refocus their energy on the task at hand -- improving the process of manufacturing automobiles, food products, semiconductor wafers or chemicals -- not wrestling with compatibility problems and configuration nightmares! The Internet has redefined the PC from computing device to communication tool. Ten years ago, a typical PC had WordPerfect, Lotus123, and a few games. The only connection to the outside world was the keyboard and printer. Today, a PC without a modem or network card is almost considered useless.

Similarly, controllers which offer no connectivity to business LANs are of very limited value. It is no longer possible for communication to be an "afterthought" in new product designs. Industrial devices need to be easy to configure, just as small office LANs and PC dial-up connections are today.

How the Automation Industry Got To Be the Way It Is

In the automation business, PLCs began replacing relays in the 1970s, and at the time, there was no such thing as an open computing environment. Allen-Bradley, Modicon and Texas Instruments introduced PLCs; Honeywell, Bailey and Foxboro introduced DCS systems, and soon the need for distributed I/O and networking became apparent.

Proprietary networks were a natural way to 1) focus responsibility of functioning equipment on the equipment vendor, and 2) protect the vendor from competition, because once a customer chose a proprietary control and networking platform, it was impossible for the customer to give his business to a different vendor.

In some cases this was a good way for the vendor to be assured of customer loyalty in exchange for provision of on-site support and service. In other cases, it has created situations bordering on extortion. The customer finds himself in a headlock. He has made irreversible investments in proprietary technologies, then discovered that products are discontinued or do not live up to their promised performance. He reluctantly pays the vendor again to fix the problem which the vendor himself created!

Every time the customer purchases new equipment, he discovers that his choices are limited to the same single vendor, or to a small number of "approved" vendors who pay large licensing and membership fees to participate in an exclusive partnership program.

Fieldbuses Enter the Picture

In addition to the major proprietary networks like Allen-Bradley Remote I/O, DataHighway, Modicon Modbus Plus, and GE Genius IO, hundreds of other serial protocols have been defined by various vendors, for reasons known only to themselves.

Because of the diversity of the industrial marketplace, it's natural that there are dozens of ways to do just about everything. But it's clear to any observer that most of these protocols are totally unnecessary, and proprietary protocols are counterproductive for just about everyone except the vendors who control them. In the early 1990s, open Fieldbuses became the center of discussion. Network specifications maintained by trade organizations instead of vendors, and accepted by standards organizations like IEC, promised interoperability between hundreds of devices.

Over a dozen major Fieldbus standards -- Interbus, Profibus DP / FMS / PA, DeviceNet, LonWorks, Seriplex, WorldFip, Foundation Fieldbus H1 and H2, HART, CANopen, Modbus, SDS, Sercos, AS-Interface, ControlNet, and even Ethernet -- contended for market dominance. Users struggled with the decision of which one to use, and vendors struggled with the question of which one to support first.

One Network or Many?

In reality, none of these buses individually comes close to satisfying all, or even most, networking requirements. Tradeoffs between speed, distance, functionality, determinism, intrinsic safety, noise immunity, topology, implementation cost and complexity make this impossible. A close look at the capabilities shows that no one network will be able to handle all situations -- not even Industrial Ethernet.

However, 99% of all factory networking applications could be addressed by a combination of only three or four, and it's perfectly realistic for the industry to widely adopt and support a small plurality of standards. Synergetic's opinion is that for discrete manufacturing, DeviceNet, Profibus, and Ethernet together have nearly "won" the Fieldbus wars. Add Modbus on RS232 or 485 as the defacto standard for serial communication, and Foundation Fieldbus for intrinsically safe process applications, and the list of open networks is complete for the foreseeable future.

Other networks will continue to have their place, but will be relegated to specialized applications and industries -- i.e. Sercos for motion or AS-I for simple discrete devices.

Frustration for Users, Integrators, Device Manufacturers

Even with the consolidation that is taking place, the attitude many people have today is confusion and resentment. Too many networks. Too much mystery and difficulty. Not enough expertise on the part of plant floor personnel and integrators. So much fear that a particular network will become obsolete, in an environment where any given industrial product may still be in use 20 years from now. Too many promises of "openness" and "interoperability," only to find that there's a "catch."

People are tired of communication standards battles which seem to go on indefinitely. Users are confused because they are often not sure exactly what communication their existing equipment supports, or what a new device will support.

Small vendors experience perhaps the greatest frustration. There are a myriad of companies who make specialized devices and controllers -- vision systems, motion controllers, data acquisition instruments, chart recorders, moisture and gas detectors, and other specialized sensors -- who at the present time cannot possibly support a half dozen or more Fieldbuses.

Often they spend months bidding on projects and negotiating with customers, only to find after spending valuable resources and time, that a communication barrier makes participation impossible.

Network Based Systems Are Strategic

Industrial network installations tend to be large, involve many nodes, many dollars, and be "high visibility." Vendors report that they represent only 5-10% of their customer base but 20-25% of their business! The larger an installation, the more likely a Fieldbus will be the backbone of the system.Vendors cannot afford to ignore Fieldbuses, and the ones who offer full support of the technology will find themselves in an ugly, competitive situation. Today, the ability to provide multiple Fieldbuses is a differentiator in the marketplace.

Network Based Systems Represent a Shift in Control System Philosophy...

Old Paradigm:

• The controller, PLC or DCS is the "backbone" of the system, and the network is an accessory.


• The PLC or DCS vendor is the "expert."


• Strategic design efforts are centered on devices and the manufacturing cell.


• Devices are chosen based on their compatibility with the network vendor.


• Device vendors gain a competitive advantage by joining a proprietary network club.

New Paradigm:

• The network is the "backbone" of the system, and the controller, PLC or DCS is "just another node."


• The network / systems integrator is the "expert."


• Strategic design efforts are centered on the enterprise and Information Technology.


• The network is chosen based on its compatibility with best-of-class devices and multiple vendors.


• Device vendors gain a competitive advantage by supporting multiple open networks.

This New Paradigm has some implications and requirements:

• It forces a level playing field, where competitive advantages are based on core functionality, not external compatibility.


• Networking is a standard feature, not an option.


• No "secrets," surprises, or undocumented data formats should be allowed.


• Openness extends to database formats, configuration utilities, and programming languages.


• The user must clearly and deliberately decide "who is responsible for making this system work."


• The value of additional information supplied to the enterprise (ERP, MRP, MIS) far exceeds the cost of the network itself.


• Devices must have powerful networking capabilities "under the hood."


• Devices must support multiple networks.


• The cost of networking must be driven down substantially below the cost of the device itself!

A Completely New, Low Cost, Multi-Network Embedded Solution

In order to fully realize the vision of the "new paradigm," a multiple Fieldbus solution is needed, a solution which is an order of magnitude less expensive than what is presently available.The solution must:

• Have a BOM cost well below $50


• Support multiple Fieldbuses simultaneously


• Need no additional communication ASICs (no external CAN, Ethernet or Profibus chips)


• Present a common API to the host for all buses


• Pass all conformance tests


• Be simple and turnkey enough to be implemented without extensive "Fieldbus expertise"


• Use widely available, popular tools for those who wish to develop custom applications.

An embedded "System On a Chip" (SOC) was the solution that could meet these demands. An extremely ambitious R & D effort was mounted, the result of which was a communications "Swiss Army Knife":

• Turbo 80186 processor with 12.5 MIPS (192 MHz equivalent speed)


• 2 CAN channels


• Profibus FMS & DP Master or Slave with DPV1 and DPV2 extensions


• Ethernet 10/100 Mbps controller


• 2 Serial Ports & SPI Port


• 8K Dual Port Memory


• 256K RAM


• Watchdog timer, 186 peripherals (DMA, timer, chip select, interrupts)

Master and slave software protocol stacks were also developed for DeviceNet with Rockwell-compatible configuration tools; Profibus with Siemens-compatible configuration tools; Ethernet Modbus/TCP and Ethernet/IP protocols and configuration tools; web server and TCP/IP, UDP and FTP.Furthermore, the SOC supports up to 16Mbytes of addressable external memory and enough processor power to function as the exclusive host processor for the embedded device, with no other processor needed.

Fieldbus and Internet Communication: Option or Standard Feature?

One of the most powerful aspects of the SOC concept is drastic parts reduction. Depending on the application, this SOC replaces 5 to 10 board level components, including the host processor. In many cases, there is no need for RAM, glue logic, or communications peripherals. All that is needed is optical isolation and transceivers for the network.

When cost savings are factored in, the result is zero net cost for many designs -- the designer achieves compatibility with CAN, Profibus and Ethernet, together with several major application protocols, with little to no impact on cost. The design cost is nearly the same as a design with no network capability at all!

What this means is that every product can be shipped with powerful communications capability built in, under a single part number. This drastically simplifies product planning and manufacturing, and allows every device vendor to participate in the networking revolution. It fulfills the vision of the network as the backbone of the 21st century smart control system, and blasts a gaping hole in the brick wall of industrial communications.

"System on a Chip" combines the capabilities of 4-5 separate daughter boards on to one piece of silicon, with substantial processing power -- effectively solving the multi-network quandary of embedded communication.

Abbreviated Cost Comparison: Traditional Hardware Design vs. SOC for Master Implementation, 10,000 pc qty

Click here to view chart.