Just doing what we always did

Jake surveyed the menagerie of dissimilar devices sharing his home network: two doorbells, a new microwave, stove and dishwasher, a Costco weather station, cable boxes, thermostat, his family’s cell phones, and a few laptops and PCs. Once he input the SSID (network name) and password, these devices all connected, and appeared to function happily despite their different protocols and purposes.

Perhaps you remember Jake from last year, who had a compelling vision of how digital integration of field devices would turbocharge his instrument reliability goals. Instrumentation was always “guilty until proven innocent” for any number of surprise upsets, going off spec or even causing a spurious trip. Instead, his team started seeing the diagnostics of every device, in one place, and hopefully foresaw impending issues—and fix them—before they impacted their production process.

There were some issues with retrieving and presenting diagnostics. Some devices had few diagnostics, and most were meaningless until the device was obviously broken. Fieldbus devices often lacked capabilities beyond the HART version, which wasn’t all that useful for preemptive fault detection. Jake’s legacy, HART-installed base wasn’t tuned to suppress redundant or useless alerts. Retrieving device information—much of it interesting—seemed like a throwback to the days of the 56K modem. His new system had visibility into practically any intelligent device, but often connected to his system through a tortuous path—passing through controller I/O to a database and configuration server, or through HART multiplexor/strippers, which communicated over RS-485 to a similar node. Profibus devices had yet another pathway, as did several hundred WirelessHART devices. Consequently, doing a comprehensive daily assessment of a few thousand devices was frustrated by bandwidth constraints.

Naturally, the bandwidth and speed offered by Ethernet-based solutions had great promise to lift Jake from his trough of despair. With the new Ethernet Advanced Physical Layer (APL), Jake could potentially communicate with field devices at a couple megabaud or more. If an intelligent device management appliance could sample a pressure signal at even 40 Hz, the system could assess for process noise attenuation and deviations from normal, alerting him to possible process connection issues like freeze-ups or plugging of its impulse lines. Ethernet supports parallel communication, so like Jake’s microwave, stove and dishwasher at home, the signals can coexist with other transactions on the same media, like his phones and computers.

Get your subscription to Control's tri-weekly newsletter.

It’s not hard to imagine network architectures, where asset management tools are peers with controllers and operator interface workstations, but this isn’t likely to be an easy transition from the installed control system—unless you employ Profibus/Profinet, extensively. It lends itself to APL quite nicely. However, Jake had a diverse installed base with relatively few Profibus devices. Most process distributed control systems (DCS) systems use a customized, proprietary version of Ethernet, and make special accommodations to ensure priority messaging for alarming, near-determinism for communications with PID controller instances, and near-real-time updating of process graphics. 

It’s a guess on my part, but I think a good one, that layering parallel, intervening device communications on such proprietary networks is unlikely to be embraced. Instead, the field Ethernet will be hosted on dedicated interface cards, as protocols like Modbus TCP, EtherNet/IP and Profinet are today. The DCS controller and its I/O subsystem can remain the conduit through which field device diagnostics must pass.

Jake will also face a difficult brownfield problem. The 1990s-era fieldbus pioneers and dreamers envisioned a gradual path to adopting fieldbus, which is why ordinary twisted-pair cable was an approved physical layer. Jake has a higher hurdle to invest in APL infrastructure, and after doing so, he will still have primarily conventional, HART or legacy fieldbus devices. APL field devices are still limited, so the pace of introductions, as well as the funding for replacements, could span the rest of Jake’s career.

The challenge of carving out funding to invest in an instrument reliability project may be the highest hurdle, even if there’s a visionary with six-figure approval authority, who’s jazzed about Industrie 4.0 or the Industrial Internet of Things (IIoT). The inertia of “just doing what we always did” is powerful. Not to mention, there are competing paths to providing preemptive fault detection, spanning more than just the measurement and control system.