My wife and I hate the new washing machine. Our 26-year-old Sears Kenmore had failed in a way I wasn't motivated to diagnose or repair. The new machine promised "low water usage" and other high-tech features. What I didn't know was its safety interlocks lock the lid and prevent almost all interaction with the load of laundry as it buzzed, clicked and whirred mysteriously, never making the reassuring sounds of "clothes getting clean" like the old Kenmore. The routine we'd practiced for 25-plus years no longer applied.
It seems we aren't predisposed to embrace change. So it also has been with digital integration of field devices. I've heard numerous stories of people and service providers trying to practice the "old ways" with fieldbus devices and meeting with frustration. It's caused some pain, and has created some push back to the more widespread adoption of fieldbus in some cultures. Some have started reverting to technology of the same vintage as the old washing machine.
Then, last quarter, IMS Research released a study forecasting the end of fieldbus.
Can't say that I agree. As Fieldbus Foundation's Larry O'Brien noted in his blog post of April 3 (What's in a Node? Putting Market Research in Perspective), FF's technology and services have been sustaining double-digit growth. Sure, there has been some pushback, and some prominent players in the supplier community have been peddling proprietary solutions in lieu of fieldbus. But especially in the developing economies of India, China, Brazil and their peers, enthusiasm for fieldbus applications remains steady.
You needn't dig too far into the IMS Research paper to surmise that its relevance to the process industry fieldbus market may be tangential at best. It lumps the vast and diverse collection of technologies—everything from AS-i to Bitbus to DeviceNet to Profibus—into its definition of fieldbus. The competing communications technology that presumably will replace all these buses, including process fieldbuses, is Ethernet. The study doesn't seem to be aware of the fact that technologies like FF and Profibus also run on Ethernet. As O'Brien observes, even the predictions that foreshadow fieldbus' demise only see its market share diminishing from 75% (present day) to 69% in 2016—hardly staggering to the retirement home.
Years ago, it was fashionable to think that Ethernet had become so pervasive, so ubiquitous and consequently so cheap, that even your toaster would have an RJ-45 jack on it. My friend at a major instrument supplier would ask me routinely, "How about a pressure transmitter with Ethernet!?" Aside from the well-known obstacles of hazardous-area capability, two-wire power, topology and distance, there's the issue of how boring a well-specified, correctly installed and maintained transmitter really is. You don't run a 20-inch pipeline to the road for your gas line, even if it's cheaper than 2-in. galvanized. Delivering 31.25 kHz over a (single) twisted pair of copper provides ample bandwidth, power, flexible topology and up to 1900 meters of copper network with no powered repeaters, switches or media converters. Even when I deploy advanced diagnostics like statistical process monitoring solved at high speed in the device, I do so with negligible impact on network performance.
Ethernet is ubiquitous, and industrial protocols like EtherCAT and EtherNet/IP are indeed supplanting proprietary buses in a host of discrete parts manufacturing applications. The process industries have differing requirements and priorities, and the specifications on which process fieldbuses are built were created with considerable input from process industry end users.
Not like my washing machine, which, incidentally, still lacks an RJ45 jack. Maybe if it had one, we'd know what the heck it was doing!