Automation World wanted to know how fieldbus technology is holding up in the age of Ethernet. As it turns out, the answer is “very well.” Many users are racing to adopt Ethernet-based networks and devices, but established fieldbus protocols won’t be going away anytime soon. Both groups of technologies seem to be finding a place in the automation landscape.
More than 450 people responded to the Automation World Fieldbus survey, which was conducted online in November. Seventy percent of respondents reported installing industrial fieldbus devices and systems in the past five years, while 86 percent reported installing industrial Ethernet-based devices and systems during the timeframe. Some respondents pointed out that industrial Ethernet networks are fieldbuses and of course function that way. But for the purposes of this survey, “fieldbus” encompasses legacy protocols like Profibus, Foundation Fieldbus and DeviceNet. Ethernet protocols include EtherNet/IP, Profinet, Modbus TCP/IP and others.
Responses made it clear that legacy fieldbus technology continues to be chosen, installed and optimized. Thirty-four percent of respondents say it is used either as their sole industrial networking technology or as the primary network in conjunction with some Ethernet for higher-level operations and enterprise systems. But respondents were evenly split in their opinion of whether or not Ethernet will replace most fieldbus technologies (49 percent say yes) or the two will continue to exist side by side for the foreseeable future (47 percent say yes).
Commented one respondent, “Fieldbus has the advantage of a ‘more’ hardwired approach, which is typically easier to troubleshoot locally. Ethernet is quickly overtaking [fieldbus with its] speed of information and flexibility, but doesn’t have ‘direct’ wired I/O capability. Making these work closely together will be a great challenge, but can only make their advantages even stronger.”
“I have been working for more than a decade with Profibus network,” said another respondent. “In recent years, we have been adopting Profinet for new projects. For me, it seems to be the natural choice or evolution.”
Commented a third respondent, “Fieldbus is too finicky. I think the plug-and-play of Ethernet is the way to go, as long as we can get the diagnostic information from the field devices and we don’t have distance limitations, which I guess leads us in the direction of using fiber to connect the field devices.”
One respondent went so far as to say, “Fieldbus technologies will eventually be replaced by Ethernet-based technologies as embedded technologies enable the transition. Physical media and simplicity of connecting directly to a port on each device will win the day.”
This same respondent added that the survey’s question on this topic should have included a choice that stated that the two networks will continue [to exist side-byside] until embedded technologies enable the wholesale replacement of the fieldbus technologies. “Remember,” said this respondent, “the camera folks did not see the added option of a ‘free’ smartphone camera killing their businesses either.”
Making a move to Ethernet is no small task for existing installations, so many users are making the move toward Ethernet incrementally. EtherCAT, a high-speed form of industrial Ethernet, “can be viewed as a flexible way to allow end users to continue using, yet reducing, their reliance on legacy fieldbuses, without having to make the ‘leap’ to a complete [Ethernet-based] system,” says Joey Stubbs, North American representative for the EtherCAT Technology Group (ETG). “Whether the networking strategy is ‘all-in’ or ‘partially-in,’ EtherCAT can address the intensity of system upgrades determined for each application. We find that many users of EtherCAT have identified that it can be part of a logical progression of moving away from a legacy fieldbus into a complete industrial Ethernet system.”
PI North America is an organization that exists to educate users about Profinet (an Ethernet-based network protocol) and Profibus (a group of legacy fieldbus protocols), and assist device makers in adding the protocols to their products. PI spokesman Carl Henning attests to the proliferation of interest in both industrial networking approaches.
“We see Profibus continuing to grow, but at a slower rate, while Profinet adoption is accelerating. We finished 2012 with 44 million Profibus nodes installed and 6 million Profinet nodes. This year promises to advance both numbers significantly,” says Henning. “Note that there are two physical layers for Profibus, one used for discrete (factory) automation and the other—Profibus PA— for process applications. Profibus PA will be around for along time since it is suitable for hazardous environments while no Industrial Ethernet is suitable.”
Henning says users can bring Profibus PA into a Profinet backbone using a proxy. “A proxy is like a gateway except that it is defined in the Profinet specification. The proxy concept has been so successful for us that almost any legacy fieldbus can be proxied in to Profinet,” he says.
The almost 50-50 split between the belief that Ethernet will replace fieldbus and that both will coexist side by side reflects concerns that are hard to shake for many end users. Concerns include: network simplicity, suitability for the purpose, elimination or reduction in the varied number of fieldbus systems, performance, convergence with IT systems, future proofing, maintainability, scalability, and ability to leverage IT resources for support. But the concerns that rose to the top for most users revolved around three core issues: security (cited by 15 percent of respondents), reliability (11 percent) and speed (7 percent).
“We don’t consider the Ethernet network as secure a network as a fieldbus network,” said one respondent. “If a simple stack attack were to take place on layer 3 of an Ethernet network, this would disrupt all Ethernet network communications, resulting in no I/O communications. Fieldbus networks could not be disrupted or disabled in this manner.”
Another concern of experienced fieldbus users is the non-deterministic nature of commercial Ethernet. Said one survey respondent, “The fieldbus we use is deterministic: we can be assured of a precise interval for all communications on the network, even considering diagnostics and acyclic communications. Ethernet requires a lot of hardware and high-end switches to even approach that, and there are still too many ways you can swamp the network with defective equipment. As Ethernet hardware becomes hardened for the factory floor, and cheaper, then maybe.... But not now. When we need to start or switch a pump, we don’t want any delays.”
Proponents of industrial Ethernet protocols, however, say that for many applications the determinism problem has essentially been solved by the higher speeds supported by Ethernet. “The robot can’t react faster than a retry for an Ethernet packet,” says Mike Miclot, vice president of marketing for networking products supplier Belden. With EtherCAT and other high-speed variants of industrial Ethernet available, “motion over Ethernet gets solved and other new applications for Ethernet are possible,” he says.
Stubbs adds that numerous successful examples of EtherCAT implementations support a variety of application uses for Ethernet, from motion, generic I/O, high-speed data acquisition, safety and condition monitoring to an easy way to manage legacy fieldbus devices.
“There will always be a variety of legacy fieldbuses in the field for a variety of reasons,” Stubbs says, “such as ‘copy exact’ requirements from more conservative customers, lack of resources at the supplier or not enough demand to port a device over to a newer fieldbus. Perhaps the most critical aspect is for users to be able to continue using the devices they need, regardless of the backbone fieldbus.”
Devices are key
The support of specific industrial networking technologies by distributed control system (DCS) vendors and device manufacturers is a key influencer of where industrial networking is headed in the future. An important concern about Ethernet among Automation World’s survey respondents is finding the devices they need supported by an Ethernet protocol.
Process instrumentation vendor Endress+Hauser is steeped in legacy Profibus technology, having been involved with the Profibus International working groups since 1991. Currently E+H operates a PI Competence Center and a PI Training Center, and its people are also active in the working groups and many regional PI associations. But the company is also leading the way in bringing out devices that connect via the EtherNet/IP protocol.
In 2012, E+H introduced its Promass Coriolis flowmeter with industrial Ethernet connections. This year, E+H showed its new Liquiline CM44 multichannel transmitter with EtherNet/IP connectivity at Rockwell Automation’s Automation Fair.
“EtherNet/IP is not only a very capable fieldbus-type network, but also provides benefits beyond what typical-level fieldbuses deliver today and in the future,” says Michael Robinson, solutions business manager for Endress+Hauser.
For example, “devices on EtherNet/IP can also be polled by a condition monitoring system to determine if there are any diagnostic messages that need to be sent to maintenance personnel as an alert. An industrial PC equipped with asset management, maintenance, condition monitoring or HMI/SCADA software can access all the I/O and diagnostic information it needs directly from the devices via the Ethernet interface. With fieldbus, the same software has to access the information from the process historian or database in a DCS—at considerable extra cost,” says Robinson.
When it comes to communications at the device level, a 2012 market study released by the ARC Advisory Group said that 47 percent of all field devices installed worldwide at year-end 2011—34 million devices—communicated using fieldbus—specifically, the HART Communication Protocol. “This is the largest market share of any field communications protocol—larger than all other fieldbus devices (Foundation Fieldbus, Profibus and other all-digital), proprietary protocols and 4-20 mA (analog/non-smart) combined,” says Ted Masters, president and CEO of the HART Communication Foundation.
Smart devices hardwired from the field level to the control system using HART is a way to combine the benefits of smart devices with the need to lower operating and maintenance costs, improve safety and improve process availability. Masters says suppliers today are “manufacturing HART products in record numbers,” and the total number of installed HART-enabled devices is currently estimated to be over 38 million.
Of course, process control users today “don’t have much of a choice, because their instruments by and large do not use Ethernet,” says Jonas Berge, a smart instrumentation expert with Emerson. “You just can’t get most sensors and actuators with Ethernet today.”
Berge notes that there are two distinct classes of “fieldbus.” He says H2 fieldbus protocols are used at level 1-1/2 of the Purdue reference model to connect remote I/O, PLCs, MCCs, variable-speed drives, motor starters and wireless gateways to the DCS. These protocols include DeviceNet, Modbus/RTU and Profibus-DP. H1 fieldbus protocols, on the other hand, are used at the lower level 1 of the Purdue model to connect the underlying sensors/ transmitters and positioners/actuators/valves to the DCS. They include Foundation Fieldbus H1, Profibus-PA, DeviceNet, ASI, IO-link and CompoNet.
“Keep in mind that, to date, only about 5 percent of process plants have adopted some kind of H1 fieldbus in a big way. The other 95 percent are only familiar with their H2 fieldbus,” says Berge. He agrees that use of Ethernet-based networks—EtherNet/IP, Modbus TCP, and Profinet—is growing, but only as a replacement for H2 fieldbus applications.
“I personally believe hardwired signals like 4-20 mA and on/off have to go away to move the industry forward, being replaced by digital, just like everything else around us is going digital,” says Berge. But there are several good reasons why there are so few Ethernet-enabled field instruments to replace them, and “H1 fieldbus does not have these problems.”
Wider adoption of H1 fieldbus would generate an order of magnitude more data from the field level, and an intermediate H2 fieldbus could not cope with this, Berge says. This is where Ethernet can come in with greater bandwidth.
Based on Berge’s comments, it’s logical to assume the accelerated adoption of fieldbus for sensors and actuators might, at the same time, accelerate the adoption of Ethernet at the next higher level of communications. “It is just like Ethernet and USB: They have mutually exclusive advantages that complement each other, rather than compete,” he says.
Endress+Hauser’s Robinson says, “Process devices will get more intelligent—the past and present demonstrate this. A process device will have a lot of information to share, and will need ever more network capacity and capabilities.” Users can meet these needs by leveraging Ethernet advances, and taking advantage of Ethernet’s huge economies of scale.
“This economy of scale and the tremendous technological advancements that go along with it is what makes [Ethernet-based networks] more capable than a fieldbus network, now and especially in the future,” Robinson adds.
While the Automation World survey data shows that users are not yet abandoning traditional fieldbus technology anytime soon (nearly 80 percent of respondents still see a place for legacy fieldbus networks), it’s clear they’re racing to install Ethernet at a faster rate.
“In the 1990s, fieldbus [installations] shot up like a rocket and enjoyed a steady climb,” says Belden’s Miclot. They’re still growing, but nowhere near the same rate. Ethernet has been pumping right along, but it’s a slower climb and it still has a lot of runway. It’s the B1 bomber, just taking off and gaining altitude.”