Fieldbus Wars continue

Black adds instruments with an Ethernet interface can serve the same function as a fieldbus. “There isn't anything a fieldbus can do that Ethernet can't, so there isn't a good case why Ethernet can't be used,” he notes. “There are many advantages of Ethernet, such as openness, cost of installation, and hardware cost.”

Rockwell’s Appleby agrees: “The value of open standards like Ethernet is that they give people a common format to work from. The trick can be determining which standards are truly open.”

Meanwhile, Moin Shaikh of Siemens disagrees: “The biggest difference between fieldbus and Ethernet is that fieldbuses are deterministic and standard Ethernet is not,” he explains. “Certain Ethernet protocol changes must be made to achieve deterministic behavior.”

According to Lee Neitzell, an engineer with Emerson, while Foundation fieldbus is bus-powered, and can be used in intrinsic safety applications, Ethernet can’t claim these capabilities. “First, it’s not intrinsically safe. Second, even in non-intrinsically safe applications, Ethernet as used today is no longer a bus. That means there has to be a hierarchy of active switches and each device has to have a separate, Ethernet cable run to it from its switch. This results in significant installation and maintenance costs, as well as reliability costs. Third, Ethernet isn’t bus powered,” he says. “Therefore, each device also has to have power run to it, resulting in even higher installation and maintenance costs.”

Barbour of Pepperl+Fuchs says Ethernet with a fieldbus protocol may be the answer. “To date there’s no power on Ethernet,” Barbour says. “Developers trying to realize an Ethernet fieldbus solution will make Foundation fieldbus technology part of the TCP/IP Protocol Stack. Ethernet will be a welcome addition to Foundation fieldbus to increase bandwidth and network speed. There are 10-12 Ethernet-based instruments on the market today that were built on the FF protocol.”

Skip Hansen, I/O systems product manager at Beckhoff Automation, says it’s only a matter of time before the Ethernet juggernaut takes over. “Industrial Ethernet continues to rise to dominance, starting with the open Modbus TCP/IP protocol, and progressing to the latest technologies with the advent of next-generation Ethernet fieldbuses such as EtherCat,” he says. “These protocols reduce equipment costs and reduce installation time by using commercial, off-the-shelf (COTS) components and cabling. Ethernet solutions differ from each other in performance: Modbus TCP reaches the ‘tens of milliseconds’ control level, EtherNet/IP delivers millisecond level control, and EtherCat (See Figure 5 below) gets down to microsecond level control. Other future Ethernet-based solutions available or under development include Profinet, Ethernet Powerlink, SERCOS III and SynqNet. Devices with Ethernet interfaces can and most certainly do handle the same functions as devices with proprietary fieldbus interfaces.”

	FIGURE 5: SUPERFAST ETHERNET
	Ethernet is getting faster and faster. This PC (left) and Ethernet switch (right) provide microsecond response time. Source: Beckhoff

While the arguments against Ethernet as a fieldbus are powerful, Ethernet has proven to be a steamroller technology over the past 30 years, overcoming every obstacle, objection, and argument thrown up from competing LANs, WANs, and process control networks. All those proprietary networks are gone, and Ethernet remains. Today, the big dogs are lining up on both sides of the Ethernet fence: Automation IT with its array of big time companies versus the splintered association of squabbling fieldbus vendors. If the fieldbus people could agree and take a united front, they might have a chance against the steamroller.

What about Wireless?
Wireless is coming on the scene faster than a souped-up Corvette. As noted above, 40 vendors exhibited wireless automation products at Hannover Fair.

“Almost all industrial instrument and process control vendors have adopted some form of wireless technology,” says Saunders, of Moore. “Whether it’s 802.11 or some proprietary version boasting better security and throughput, the fact remains that wireless networks will continue to gain momentum throughout the process control sector. Ethernet and wireless Ethernet have become the de facto industrial wiring standard for both process control backbones and factory automation networks. Its ability to handle higher throughput, multiple protocols, low-cost of implementation and multiple vendor support will further allow Ethernet to displace many of the proprietary control networks that once dominated our industry. There’s no doubt that wireless Ethernet will continue its aggressive growth curve within our industry.”

But will it replace fieldbuses? Not in our lifetimes, says Emerson’s Lansing. “Wireless is a complementary technology to HART and fieldbus, and represents just the physical layer of the communications stack,” she explains. “It’s another ‘pipe’ to move the bits. Wireless Ethernet is actually also complementary, representing the physical and data-link layers. We have to remember that we still need a standard application layer on top of Ethernet, wired or wireless. That’s why almost all communications protocols offer an Ethernet version, such as Foundation fieldbus HSE, Modbus TCP/IP, Ethernet/IP, and Profinet.”

We have reservations, too. Wireless is a fine way to monitor data and transmit setpoint information, but it may not be suitable for real-time control when response is critical. Too many variables can upset the timing of wireless transmissions, including other wireless traffic, interference, electrical noise, and so on. We have no doubt that these problems will be solved some day, and wireless Ethernet will prove to be a dominant force in fieldbus.

It also seems that industrial vendors ought to look at consumer products as a guide to the future. After all, nearly every electronic consumer product from cell phones to microwave ovens has an embedded computer. For just a few dollars, it’s possible to embed a processor and a Windows CE operating system into every fieldbus device, router, and interface. Such a system would allow any process instrument to support Foundation fieldbus, Profibus, DeviceNet, HART, EtherNet/IP and wireless interfaces, all at the same time, if necessary. It could hold all the protocol stacks, device description (DD) files, EDDL files, FDT files and configuration files that anyone could possibly need.

With such a system, end users would not have to worry about vendors “playing nice” with each other, and they would not have to worry about which fieldbus to buy. The squabbles would be over, and we could all get back to doing process control.