Remember the world-wide-wait? You may recall a time “before Google,” when a frequency-shift-keying (FSK) device called a “modem” connected to a broader network, be it CompuServe or AOL. Modems let users employ their twisted-pair phone lines as a “segment” to communicate with other computers. Speeds of 300 baud (bits per second) were possible with FSK, and later technologies stepped up to 1,200 and 2,400 baud and beyond.
At the same time we heard our modems sing a tune over the phone lines, the HART protocol was created to interact with smart transmitters. HART also employs FSK, using twisted-pair cable employed for 4-20 mA analog signals, and superimposing a low-amplitude, 1,200-baud signal consistent with contemporary technologies. Not coincidently, HART was derived from a Bell protocol (for after-Google folks, that’s the phone company). Aside from compatibility with 4-20 mA signals, HART’s other important property was hazardous-area capability.
The idea of using existing infrastructure was not lost on the ISA SP50 committee, tasked with creating a specification for fieldbus to become the digital equivalent of 4-20 mA. Speeds of 31.25 kHz preserved some of the same priorities as HART’s FSK, enabling it to use existing cables and limit voltages for hazardous-area classifications. Speeds that were 25 times faster than 1,200 baud FSK seemed adequate, especially when the nerdier among us were using a 56K dial-up modem. Today’s Profibus PA and Foundation fieldbus physical layer communicates over twisted-pair cable, using a pure digital protocol and trapezoidal, nominally one-volt peak-to-peak “square” wave that increases its robustness and noise immunity.
After first being deployed on “fat” and “thin” coaxial cables, Ethernet began its rise to ubiquity when it moved to cheaper, more readily available, twisted-pair copper cables. Named for the “luminiferous ether” that 18th century scientists postulated to explain transmission of light waves, Ethernet was created as a network for office computers. It didn’t take control manufacturers long to develop Ethernet networks for industry.
Beginning with the operator interface (HMI) and later extended to controllers and I/O subsystems, Ethernet is employed in almost every modern control system. Nearly all have their own “customizations,” and limit the choice of topologies to ensure time-critical, deterministic communications.
If you haven’t noticed yet, there’s been a little arms race going on to offer Ethernet-connected field devices. The latest incarnation of leading Coriolis flowmeters are all touted for their easy interconnection to control systems that support “industrial Ethernet” protocols like Profinet and EtherNet/IP. These products have found a market among end users and system integrators in food and pharmaceuticals, and some in upstream oil and gas. They’re finding it easy to integrate Coriolis flowmeters using a familiar Ethernet connection.
But for many, deciding to deploy field devices with Ethernet connectivity means a potentially substantial investment in infrastructure and power supplies. But what if analog, twisted-pair cabling could be repurposed for industrial Ethernet? Consumer products that use phone or power lines have been available for years, and the most recent generations are achieving reasonable transmission rates.
The obstacle to using this technology in process plants has been hazardous area capability—keeping energy levels low enough to prevent it from becoming an ignition source. Today, innovations to allow 4-20 mA or fieldbus infrastructure for high-speed Ethernet in hazardous areas are being prototyped and demonstrated.
Pioneering end users aiming to employ device intelligence have been reliving the “world-wide-wait” when gathering and viewing device information. The network isn’t always to blame, but any improvement in bandwidth for accessing field devices would be welcome.