The network is seeing more use as a backbone, linking automation equipment together while providing compatibility with front office systems.

Ethernet works well to connect products that use fairly powerful microcontrollers. Proponents say that as 32-bit chip prices drop, they’re making it easier for equipment designers to implement Ethernet’s protocol stacks. That makes the network viable in more areas.

“Around 85 percent to 90 percent of controllers have Ethernet connectivity. It’s being extended down to remote input/output (I/O), motor control, variable frequency drives and other products that used to use RS232, DeviceNet, Profibus and others,” says Bill Black, Controller Products Manager at GE Fanuc Intelligent Platforms, Charlottesville, Va.

But for components such as sensors and a range of I/O devices where even far smaller 8-bit controllers must be near the low end of both pricing and physical size limits, alternatives are often far more cost effective. Many networking specialists feel that there’s still room for a hierarchy of networking schemes.

“Ethernet provides an enormous amount of bandwidth for data that may not be fully utilized on the device level. A controller area network (CAN)-based network is more efficient for the device level. By using a gateway to bridge the CAN-based device level information to the Ethernet control level, the user can experience all the benefits of Ethernet with maximum efficiency,” says Chris Vitale, senior product manager with vendor Turck Inc.’s network and interface division, in Plymouth, Minn.
 
While CAN fits in some applications, it’s not viable in others. When fieldbuses will link directly to Ethernet, integrators need to make sure data flows seamlessly between the networks. That can be accomplished with bridges that are widely available.

“There are methods available to take the devices on the fieldbus level to Ethernet without having to physically change out lower-level devices to make them Ethernet compatible,” Vitale says. “This method bridges existing fieldbus systems completely onto Ethernet, making the devices appear as if they’re on an Ethernet network.

“These gateways are self-configurable and the process data is presented on the upper-level network in its native protocol.  This seamless transition is also scalable, allowing for future expansion,” Vitale says.

There are also issues on the Ethernet side. One is to make sure that input from the fieldbuses doesn’t overload the central Ethernet network. Avoiding collisions due to bus overloading is a well-known factor, but it’s perilous to overlook it.

Properly managed switches can prove their worth by managing this characteristic of the network. “Ethernet switches can help to segregate information data from I/O data, prioritizing messages from the slaves/clients, and limit broadcast storms,” says Wilson Lee, product specialist at Turck.

Related Feature - Builder Beware
To read the feature article relating to this story, go to www.automationworld.com/feature-4874