If you haven't checked out the ControlGlobal discussion group on LinkedIn, you should. A lot of smart and savvy people check in there, and some pretty interesting discussions get going. One of the most recent ones was started by Jean Carl, PR and social media consultant at Macro Sensors. She asked, "Do you prefer analog or digital output to monitor operating parameters?"
The question generated a lot of talk. Quinton Christensen, engineering/IT manager at Pilkington Metal Finishing, said, "I prefer digital where possible. If digital is not possible, I try to use a transmitter as near the signal as possible. Less induction and loss as the data traverses the plant."
On the other hand, Francis Malinosky-Rummell, PE, owner, Portland Engineering, industrial automation consultants, voted analog. "Networks can be very finicky and are much more difficult to troubleshoot if they're not working. You do have to make sure you're wiring correctly to avoid induction, but that goes for network wiring as well. If you're using 4-20 mA signals (vs. voltage), you should not have any signal loss"
Jorge Edgar Cano Carmona, global process technical consultant at Rockwell Automation, added that, "Please, we're in the 21st century. I prefer digital (FF or PA), of course."
Not surprisingly our own fieldbus guru and process control specialist at ISP Lima LLC, John Rezabek prefers digital. He replied to Malinosky-Rummell. "@Francis, 'networks can be very finicky' I read as 'digital networks are not robust.' I can't say that that's been borne out by my experience.
"We were 'surprised' last year when the long-obsolete 3Com switch interconnecting controllers and HMI, et. al. via Ethernet finally failed after nearly 13 years of non-stop 24-7 operation. We were surprised because it functioned quietly and flawlessly non-stop. This year we preemptively replaced its twin with a more modern switch, i.e. it was still functioning.
"At the same time our plant is full of two-wire digital buses, both fieldbus and RS-485 Modbus, which likewise have been amazingly robust for as many years.
"Sloppy, unreliable or otherwise deficient craftsmanship when installing the 'physical layer' is an issue for analog 4-20 mA as well as digital. You have more chances to screw it up with analog, since you're running a pair to every device."
Malinosky-Rummell replied, "I stand by my statement. Sure, there are plenty of working networks, and I've put in a bunch of them myself, including various flavors of Modbus, Profibus, DeviceNet, ControlNet, EtherNet/IP and Ethernet. I haven't worked with Foundation fieldbus or CANbus. They can be very functional, stable, and these days can provide more data out of a given device than you would normally get. That said, networks are often difficult to configure at the beginning, very difficult to troubleshoot when they go bad, and can be difficult to modify. If I have the choice of relying on a 4-20 or a Profibus node, give me the 4-20 any day."
Dan Schwab, automation manager, projects at BP, is another analog fan. He says, "Francis is right on. I prefer a 4-20 mA signal, leveraging HART—the best of both worlds. You can easily troubleshoot a 4-20 mA signal (loose wire for example), digital either works or it doesn't and is therefore difficult to pinpoint a problem."
Rezabek observes in another reply, "I can see where the simplicity and familiarity of analog makes your projects feel less risky, and that's what pays the bills. Maybe at the end of the day, the client who settles for 4-20 mA never notices any difference. But, there are end users who could and would benefit. Putting the kibosh on all things digital won't eliminate risk, and may discourage end users who would really benefit."
For more such discussions, go to www.linkedin.com/groups/CONTROL-Global-2230245.