Connect Your Islands of Automation

Nov. 8, 2012
Once Isolated Machines and Production Lines Now Link Up to Improve Performance
About the Author
Jim Montague is the executive editor for Control, Control Design and Industrial Networking. Email him at [email protected] or check out his Google+ profile.
Learn More

Get the white paper, Defining 'The Internet of Things' - How to Connect your Machine, in our Convergence Knowledge Center. 



"No man is an island," says old-school, English, metaphysical poet John Donne. Now, in many cases, no machine is an island either.

More capable software, improved network protocols, Ethernet, robots and other devices can form new and unexpected links between formerly standalone machines and production lines with modular components. And they're creating these ties, not just to improve product efficiency, but also to aid setup, inform operators, and update enterprise systems.

"We've been installing and automating robotic transfer units (RTUs) on overhead gantries or rails for about seven years, and it used to take a lot of engineering time to present 5–1,500 kg parts for aerospace, automotive and oil-rig applications to the six-axis robots and machining centers we work with," says David Flora, elecro-maintenance technician at system integrator Gosiger in Dayton, Ohio. "We often have one of Fanuc's RTUs serving eight to 12 machining centers, such as Okuma lathes, grinders and milling machines, but once we began using EtherNet/IP about four years ago to do the handshaking for loading and starting, it cut our implementation labor and time by 20–30% and made these lines much more efficient on average. EtherNet/IP and Fanuc's devices also saved us a lot of time on integrating sensing, vision components, cameras and infrared lighting."

Flora explains that each machine still uses its own CNC program, interface and DeviceNet network protocol to make products, but during a typical handshake it also notifies the RTU that it's done with a part and needs unloading and loading, which tells the robot to get ready to work in the software-defined space they both share. These communications go from I/O blocks in the back of each machine to the RTU, function blocks and PLC controlling the line (Figure 1).

"Conversely, the robot also can upload CNC programs to the machines, tell them to perform certain operations, and then select the right binary-output program from their own lists," Flora explains. "Every job and solution is unique, but everybody wants more automation. So, our job is to make the RTUs and machine controls more compatible, and EtherNet/IP and our other protocols help us do it more simply and easily."

Serving Stations
Figure 1: Nick Ullmer, sales engineer at system integrator Gosiger, uses a robotic transfer unit (RTU) playing the shell game to show how it can load and unload multiple machining stations, and use EtherNet/IP to coordinate implementation and operations more efficiently.Brian Sides, technology director at Okuma America in Charlotte, N.C., adds, "Many users are beginning to realize they're not employing their machines up to full potential. They might believe a machine runs at 80% of maximum utilization, but if it's only running for one shift, that's really 20–30% of maximum utilization. Likewise, others want to improve maintenance quotas, but find it's hard to connect Ethernet to their machines and difficult to download and set up software programs. So many are moving to PC-based controls like our Intelligent Numerical Control (THINC). They also want a black box they can connect to all their legacy controls, so they recently used small PLCs with I/O modules. We responded by making this type of module standard in all our machines, so they can link to their legacy devices without having to add other components."

Longer-Distance Links
Though many machine connections occur between a few closely clustered devices, other links tie together components over much more lengthy production lines. For example, system integrator PeakLogix in Midlothian, Va., builds conveying and sorting systems for some of the largest U.S. medical supply distribution centers.

During the past five or six years, PeakLogix helped build conveyors at 15 distribution centers for one client, and is presently building two more. The typical conveying and sorting line usually has about eight to 10 different spurs and about six picking zones, which is another 100 ft conveyor with a bidirectional diverter. These zones can be on one to three levels (Figure 2). Lately, these material handling structures are integrating hundreds of proximity sensors, solenoids, limit switches, photo eyes, stack lights and horns into centralized PC-based controls via Ethernet, but retrieving all those signals didn't look easy at first.

Local I/O Convey Data
Figure 2: PeakLogix uses bus couplers to localize I/O points from the sensors and photo eyes on its long, multi-level conveying systems, which means far fewer long cable runs are needed back to the application’s main controls. This sorter uses a spiral conveyor to reach picking zones on three levels."We had lots of small devices and sensors over a wide area, and we just couldn't bring them all back to where we wanted," says Matt Cummings, technical services director at PeakLogix. "However, we found that our PC-based Steeplechase VLC control software could use Phoenix Contact's bus couplers with IP addresses to localize I/O points in each zone on the conveyor, and then send back one Ethernet cable with multiple TCP/IP signals, instead of sending back wires from 20 sensors. We also learned that we could do the same with motor controls, contactors and protectors in our distributed panels. Besides saving thousands of feet of cable, wiring installations that used to take at least a month now take one or two weeks. I also found that our Value Line PC's Windows environment allowed me to run some Visual Basic code over an OPC connection back out to descriptive reference tags in our distributed controls, and email alerts and activate stack lights. And Windows also means we can connect and do remote diagnostics from anywhere using pcAnywhere or remote desktop software."

PeakLogix's users also wanted statistics on sorting performance by the conveyors in individual spurs, Cummings adds, but this level of detail wasn't possible before. "So we use the connections on our new network to show the daily load on the overall conveyor and on each spur," he says. "If one spur gets too busy, it will show up on their panel, and they can shift some of the load up or down the overall sorter for more efficient performance."

Monitoring Moves to Standards
Of course, machines connecting and transmitting data beyond their envelopes and cells is nothing new. Point-to-point hardwiring and dial-in phone lines made basic remote monitoring, diagnostics and troubleshooting possible for many years. These quickly evolved to include more and faster Internet, Ethernet and wireless links, while protocol overlaps and clashes increased and pushed builders to standardize how their equipment communicates.

"We historically used NetOp protocol via a dedicated phone line and modem to dial up remote diagnostics from our machines," says Lem Linder, CMfgE, national applications engineer at Mazak in Florence, Ky. "You could see screenshots, transfer files, and adjust parameters — though you also had to reboot the machine. We used NetOp for about 10 years, and still apply it where users want to be especially careful and secure. It works more like pcAnywhere now."

More Inclusive Monitoring
Figure 3: Darren Daniel, applications engineer at Mazak, checks specifications and operating parameters on its Mega 8800 horizontal machining center, which is monitored by MTConnect open-source, extensible, factory-floor communication protocol based on HTTP and XML programming.While many builders use basic TCP/IP and other Ethernet varieties such as Profinet, EtherNet/IP and EtherCAT to enable machine connections, interoperability problems persist. As a result, some builders helped develop the 4-year-old MTConnect open, factory-floor communication protocol, which was used for machine monitoring, status reporting and other details, but is growing to include alerts and alarms, temperature and speed.

"MTConnect lets us have intelligent nodes on our machine networks," says Neil Desrosiers, Mazak's software developer. "OEMs develop adapters and push them to MTConnect, so their middleware gets signals that can upload to MTConnect. This allows it to serve as a black box on our overall Windows software."  

Mazak also uses MTConnect to transfer sophisticated machine health and other information, such as data from System Insights' vimana software or IGear's performance monitoring software. For example, MTConnect linked and monitored more than a dozen of Mazak's machining centers at the recent IMTS 2012 tradeshow in Chicago, including the company's Mega 8800 horizontal machining center (Figure 3).

In addition, Curtiss-Wright Controls Flight Systems in Shelby, N.C., uses MTConnect to collect data from its machine tools for making flight controls and utility actuation products, and employed vimana's classification engine and real-time dashboards to uncover subtle production-loss periods. Danny Cooper, Curtiss-Wright's senior manufacturing manager, reports that his team implemented vimana on a production cell with four CNC-based horizontal-milling machines, compared performance across different tools, and improved overall equipment utilization by more than 20% in about 10 weeks.

Simpler, More Integrated Builds
Beyond improved machine health and performance, one of the big benefits of closer, more integrated machine connections is building simpler machines that still have increased capabilities.

Ted Rozier, electrical systems engineer at Doosan Infracore Machine Tools in Pine Brook, N.J., reports that he and his colleagues recently combined separate HMI platforms for the robot and machine tool on their small DooCell packaging machine into one unified interface panel. The machine uses IO-Link to talk to its Fanuc M10 robot about its own tasks, but then uses EtherNet/IP to coordinate joint activities by the machine tool and robot.

"We were trying to give some machine operators in Mexico alerts and alarm displays that were pictures, instead of the usual text or verbal message, so more of them could respond to operating conditions," Rozier explains. "We found that the robot and machine tool's platforms had enough in common, including Ethernet and IO-Link networking, that we could integrate them into one control panel, which we're calling the Doosan Touchscreen Control Cell (DTSCC). Now users can pick or develop the status or alarm graphics they want, but they only have to do it on one panel and not two, and they only have to interrogate one screen during operations."

Shake Hands, Pass
Figure 4: Two long-armed Fanuc robots coordinated by a R-30iB controllers use software-defined envelopes and its Intelligent Interference Avoidance strategy to safely and accurately pass parts between its Robodrill D21MiA5 machining centers.

Robots Run on Rails — Safely
No doubt the most dramatic example of machine connectivity is watching two long-armed robots roll back and forth on rails or an elevated gantry, pulling finished parts from machines, passing them between each other, and putting bar stock in to start the cycle again. Not surprisingly, these graceful, ballet-like moves are coordinated very tightly by integrated and intimate machine connections.

For instance, Fanuc Robotics America in Rochester Hills, Mich., reports the two robots passing parts in its overhead Robodrill loading system use its new Intelligent Interference Avoidance strategy to work in such close proximity (Figure 4). The robots' software-defined envelopes, zones and motions, as well as the Robodrill D21MiA5 machining centers they service, are coordinated by a Fanuc controller, which has a high-speed V12 microprocessor, four Ethernet ports and a teach pendant interface. All of Fanuc's robots, machines and related devices are networked via Ethernet and IO-Link.

"When we link the CNC to the robot by IO-Link, the two CPUs perform an automated handshake, in which the machine tells the robots that it's clear to go in and load or remove the part," explains Richard Johnson, Fanuc's senior account manager for national distribution. "To make sure a robot never makes an unsafe contact, Intelligent Interference Avoidance draws software-defined geometric cylinders around each arm, while customers can add more shapes for their tools. These allow the robots' encoders to give their exact position, monitor how close they're getting, and decide which one has jurisdiction in each location. The shapes also expand as the robots run faster to allow for longer stopping distances and avoid collisions."

Johnson adds that redundant safe I/O functions that once were performed by many relays now are accomplished in software. This allowed Fanuc to develop its Dual Check Safety (DCS) method, which establishes software-defined walls that the robots can't ever penetrate because their programming won't allow it. "We know these robots and our Intelligent Interference Avoidance capabilities are catching on because we installed 180 Robodrills at Chrysler just two years ago, and they're serviced by robots running on three miles of track," Johnson adds.

Instrumental App
Figure 5: iBlue is a Bluetooth transmitter that connects via USB to gauges, calipers, micrometers or other devices, and then wirelessly acquires, records and shares their data by linking with apps on smartphones, tablets or other PCs.

Smartphone Apps for Machines
Inevitably, once machines touched Ethernet-based networks and the Internet, it was only a matter of time before they showed up in digital widgets and icons on smartphones and tablet PCs. One of the latest apps for machine monitoring and optimization was just released by one of Okuma's end users, Sides reports.

Indiana Technology and Manufacturing (ITAMCO) in Plymouth, Ind., is a job shop and contract manufacturer of large machine components such as final drives, transmissions and mining parts. Joel Neidig, ITAMCO's information systems engineer, says it uses Okuma's MC50 and MC60 machines and other equipment, which it monitors with ShopViz software that's based on MTConnect. He adds these new tools inspired ITAMCO to make its own innovations.

"We also developed iBlue, an industrial Bluetooth transmitter that can connect via USB to gauges, calipers, micrometers or other devices, and then wirelessly acquire, record and share data by linking with smartphones, tablets or other PCs," explains Neidig, who adds that about 60 iBlue transmitters are now in beta testing (Figure 5). "The mobile application let users perform mobile material hardness testing, check thermocouples and temperatures, and collect more data via the interface. For instance, an operator can use iBlue and an iPad to check the quality of a part before even taking it out of the machine."

ITAMCO will soon upgrade iBlue to work with control inspection sheets in ShopViz, and then export them via MTConnect for immediate streaming and access by anyone. "These sheets are usually in XML files, where data is entered manually, but now they can be filled out automatically," Neidig says.

About the Author

Jim Montague | Executive Editor, Control

Jim Montague is executive editor of Control. He can be contacted at [email protected].