In an article titled: “The Top 10 Myths About Industrial Ethernet Networks”, the following statement is made: “industrial Ethernet networks are more capable and just as reliable as the fieldbus networks they are increasingly replacing.”
Responding to that statement, an Automation World reader—Jonas Berge (a director with Emerson Process Management in Singapore)—pointed out the following: “One should be careful when saying Ethernet is replacing fieldbus because there are two distinct and very different levels of fieldbus: H1 fieldbus and H2 fieldbus. H2 fieldbus is used between the control system and underlying package unit PLC, remote-I/O, drives, motor starters, MCC, and wireless gateways at level 1-1/2 of the Purdue reference model (and ISA95) and includes protocols like Modbus/RTU, Profibus-DP, and DeviceNet, etc. Without a doubt their Ethernet counterparts—Modbus/TCP, Profinet, and EtherNet/IP—are gaining in popularity. On the other hand, H1 fieldbus is very different. H1 fieldbus is used for sensors and actuators at level 1 of the Purdue model, where the dominant connection technology is hardwired analog 4-20 mA and discrete on/off signals. We need digital communication down to the sensors and actuators too. I see a little bit of Ethernet used in magnetic and Coriolis flow meters as well as electric actuator/motor operated valve. However, Ethernet is struggling to take the place of 4-20 mA and on/off signals. For digital communication with intelligent pressure, temperature, level, and interface level transmitters, vortex flow meters, as well as pH, conductivity , and amperometric analyzers, control valve positioners, and intelligent on/off valves at level 1 of the Purdue model, I do not see any Ethernet. Instead, at this lower level, plants use one of the “H1” level fieldbus protocols such as Foundation fieldbus H1, Profibus-PA, AS-I, IO-link, or CompoNet, etc. to provide digital communication all the way including the “first meter”. I don’t see Ethernet and fieldbus competing at this level. Indeed H1 fieldbus and Ethernet complement each other to eliminate hardwired 4-20 mA and on/off signals.”
Judging by Berge’s comments and his affiliation with Emerson Process Management, it’s clear that he is very familiar with Foundation Fieldbus. But are his assertions essentially correct? Especially with regard to statements such as: “Ethernet is struggling to take the place of 4-20 mA and on/off signals.”
To help clarify his statement and get further input on this topic, I asked representatives from several fieldbus and Ethernet protocol organizations to comment. Here’s what they had to say.
“I would generally agree with the comments from the reader about Ethernet protocols not being competitive at the Level 1 sensor level if the term ‘Internet-based protocol’ is substituted for ‘Ethernet protocol,’” says Joey Stubbs of EtherCAT. “There is a huge difference between the typical Internet-based protocol examples that the reader references, and real Level 1-capable Ethernet fieldbus protocols that are based on the Ethernet physical layer, but don’t have the overhead of Internet protocols, such as EtherCAT.
“The reader seems to acknowledge the need for digital fieldbus connectivity at the device level, which, although not mentioned directly in his/her comments, can include features such as noise immunity, diagnostic information, accessing additional parameters, synchronization with other devices, parameter configuration, and firmware download.
“Additionally, utilization of a high speed fieldbus can permit the ‘dumbing down’ of the sensors and a move away from high cost ‘intelligent sensors.’ These costlier and more complex sensors take raw data and perform signal conditioning with an on-board processor, then provide a surrogate signal or value back to the control system in lieu of the original raw data. With a high-speed digital fieldbus such as EtherCAT, the raw data can be taken directly back to the controller, where the processing power of the modern CPU can be utilized for the same algorithms, if needed. Taking over what the former on-board processor of the Level 1 device would do, EtherCAT technology allows field devices to become much simpler, less expensive, and higher performing.”
Also agreeing that Berge’s statements are essentially correct is Carl Henning of PI North America (Profibus/Profinet). “I agree with him for the most part,” Henning says, but “factory fieldbuses are another story. Although Profibus DP and Profinet are complementary, Profinet can do whatever Profibus DP does … and better.”
To help explain his point, Henning provided the following example: If you start with a Profinet backbone you can “connect PACs, PLCs or DCSs, then I/O, vision systems, drives, etc. to the Profinet backbone. If you need a process fieldbus because you need power over the bus or intrinsic safety, you can use Profibus PA or Foundation Fieldbus and proxy it onto Profinet. WirelessHART can proxy onto Profinet as well. If you need to connect to smart sensors and actuators, there are IO-Link masters that connect to Profinet. And, if it’s not a greenfield plant, and there are legacy networks like DeviceNet, Profibus DP, HART, or others, you can proxy them onto Profinet, too.”
Eddie Lee of Moxa also agrees that Berge’s assertion is essentially correct, adding that “there is a difference between H1 and H2 fieldbus. When you hear people say that Ethernet is replacing fieldbus, most are referring to H2 fieldbus,” such as switching DeviceNet to EtherNet/IP, Profibus DP to Profinet, or Modbus RTU to Modbus TCP.
Finally, and straight from the Fieldbus Foundation, I spoke with Larry O’Brien, the global marketing manager for Fieldbus Foundation. Like the other respondents, O’Brien agreed with Berge’s core points, though he did add that “we do not use the term ‘H2’ any more. Our high speed network is called HSE, and is based on standard Ethernet and standard IP.”
O’Brien also pointed out that this very discussion has been heating up quite a bit lately following a report from IMS that made a similar assertion about Ethernet replacing fieldbus across industry. “The terms ‘Ethernet’ and ‘fieldbus’ are not mutually exclusive, O’Brien states in his Fieldbus Foundation blog post comment on the IMS study. “I do not think it was the intention of IMS Research to foretell the doom of fieldbus, and if you look at the actual numbers presented in the report, it does show continued growth for fieldbus over the next several years. The problem, in my opinion, is that people tend to latch onto the term ‘Ethernet’ and view it as a panacea for everyone's network requirements. Yes, many plant networks today, even Foundation fieldbus HSE, are based on Ethernet. So, the discussion is not one of ‘fieldbus versus Ethernet’. It is rather a discussion of Ethernet and standard IP-based networks in automation versus other networks.”