Ethernet as a leading machine automation protocol
Although there are still dozens of industrial fieldbus protocols used in machine automation, Ethernet is starting to become the norm with EtherNet/IP and Modbus TCP becoming leading protocols in North America.
Fieldbus technology was a welcome advance from point-to-point wiring when it emerged during the last few decades of the 20th century, and it’s had a nice run in industry since then. Many fieldbus protocols have come and gone, but all have connected sensors, input/output (I/O) devices, and other field devices to automation systems.
For today’s industrial networks, Ethernet can be a more attractive option than competing protocols as performance can match and exceed fieldbus technologies. Setting up an Ethernet network is also typically less expensive and easier to configure than with other protocols.
A brief history of fieldbus
In the early years, there were the basic open-communication serial standards such as RS232 and RS422/485. These, among others, were the basis for better-defined standards such as Modbus, which used serial communication standards as the foundation for what became the leading industrial protocol.
Ethernet was not yet mature, and fieldbus protocols offered sufficient performance and reliability in many applications. However, fieldbus technology was often expensive and difficult to setup, and different protocols were incompatible on both the hardware and software levels.
Many initial fieldbus implementations were designed for connecting remote I/O to programmable logic controllers (PLCs). Many of these communication standards were based on the RS422/485 standard, but most were proprietary initially before being transferred to independent foundations and made open.
For example, DeviceNet, Modbus, and Profibus DP each started out as proprietary protocols with Rockwell Automation, Modicon, and Siemens respectively. Each eventually became an open standard administered by an independent foundation.
Today, there are several fieldbus options for linking PLCs to remote I/O and other simple devices such as motor starters, sensor manifolds, and pneumatic valve manifolds.
Current fieldbus protocol options
While Industrial Ethernet’s growth is exceeding fieldbus growth for accessing devices and is expected to become the more dominate technology over the next 15 years according to IMS Research, other fieldbus networks have a very large installed base because they were the only option before Ethernet technology matured. Many applications still benefit from the highly deterministic architecture that fieldbus networks such as DeviceNet and Profibus DP offer. With these and other similar protocols, it can be very convenient to add more devices using field-mounted I/O blocks to an existing network that is still satisfying the needs of the application.
Another popular fieldbus protocol is IO-Link, which is a point-to-point (P2P) network used for tying field devices to controllers, often through a converter (see Figure 2). Some newer devices can provide process data, configuration data, identification, operating parameters, and diagnostics. The ability to transmit diagnostic data can reduce machine downtime by intelligently diagnosing the exact item of failure or, in some cases, preventing downtime by providing information that failure is imminent.
Although IO-Link and low-level fieldbus protocols work well for linking simple devices to controllers, more complex connections can benefit from the power, speed, and flexibility that Ethernet offers.
The case for Ethernet
An industrial Ethernet protocol may be considered instead of fieldbus communications in many new machine automation applications that require a high degree of information exchange, such as linking a vision system to a PLC. As hardware costs drop, it’s becoming more cost-effective for even simple applications such as remote I/O and for fieldbus device connections.
Many people still consider Industrial Ethernet as something different than fieldbus, but if one considers what fieldbus technology has traditionally accomplished in the past with what Ethernet can do today, they are really one in the same. Fieldbus technologies have traditionally boasted highly deterministic data delivery, and Ethernet can do the same with its increased speed and low level time synchronization methods. Power can be delivered to devices over many fieldbus cables, and the same is true with Power over Ethernet. There are very few industrial networking applications where a properly specified Industrial Ethernet protocol won’t work.
With early Ethernet networks, determinism was poor and jitter was significant, which resulted in slower processing speeds. Less intelligent network devices such as simple Ethernet hubs were the norm. As a result, data collisions and retries were frequent. With the advent of cost-effective, industrial Ethernet, unmanaged switches and then eventually managed switches, collisions have become a nonissue. Processing power has increased, and it has reduced data transmission days to an insignificant level in most applications.
Even with standard, off-the-shelf Ethernet chips, jitter is low enough for most applications as these can utilize scheduling mechanisms such as Class 1 I/O Messaging in EtherNet/IP. For applications that are even more time critical, protocols such as EtherCAT use precision time protocol synchronization (IEEE 1588).
While most Ethernet networks utilize a star type of topology where the Ethernet switch or switches forms the center and the devices branch out from there, many devices include switch ports with multiple RJ-45 interfaces to make daisy chaining more cost-effective.
Popular Ethernet protocols
Unlike other communication technologies, Ethernet permits multiple protocols to run over the same network. Popular industrial Ethernet protocols include:
|Table: Industrial Ethernet Protocol Examples and Governing Organizations|
|Modbus TCP (www.modbus.org)|
The automation supplier often drives selection of the proper Ethernet protocol for the application as different suppliers favor different protocols.
EtherNet/IP, maintained by ODVA, was born out of the DeviceNet and ControlNet protocols developed by Rockwell Automation. EtherNet/IP has expanded to encompass many diverse uses in industrial applications (see Figure 3).
Modbus TCP was one of the first industrial Ethernet protocols to emerge on the scene. The Modbus TCP standard is maintained and updated by the Modbus organization. It is used frequently because of its simple but effective design.
Profinet is the Ethernet progression of Profibus DP and PA. It’s maintained by the Profibus organization, and it’s the Ethernet protocol of choice for Siemens devices. It has multiple available hardware architectures including wireless, making it suitable for a variety of applications.
EtherCAT is an Ethernet standard designed by Beckhoff Automation and rooted in IEC 61158. It uses a distributed clock mechanism and is designed for high-speed and synchronized motion applications.
SERCOS was one of the earliest Ethernet standards intended for motion control. The earliest versions used fiber optics as the communications medium. The latest versions, however, have adopted it to run on standard Ethernet.
Powerlink is an Ethernet protocol developed by B&R Automation, and like EtherCAT, it also performs with very low jitter by using low-level time synchronization methods.
Although each of the protocols listed above have their proponents, EtherNet/IP and Modbus TCP are emerging as leaders in North America.
EtherNet/IP and Modbus TCP lead the way
EtherNet/IP and Modbus TCP are the most prevalent Ethernet protocols in the industry with many PLCs and other controllers supporting one or both protocols. Both can run on standard, off-the-shelf Ethernet hardware, and the choice of which one to use depends upon the hardware being used and the type of application.
Most suppliers favor a specific Ethernet protocol, but they often aren’t just limited to one. It’s often advantageous to use the supplier’s favored protocol because this will yield the best support and widest range of compatible products.
If an application requires a mix of suppliers and both protocols are available in the controllers and field devices, timing and speed requirements are often the deciding factors. EtherNet/IP uses a scheduling mechanism called requested packet interval (RPI) that can make it more desirable for high-speed applications. If there are many different suppliers involved and low cost is a prime factor, Modbus TCP may be a better choice as it has been implemented in many devices due to its relatively small memory footprint and simplicity.
Pick the right protocol
The application will help define which fieldbus technologies to use. In many applications, it may make sense to use several protocols, with each applied at the most appropriate level. Mixing protocols and communication architectures to get field devices connected to controllers, controllers connected to human-machine interfaces (HMIs), and HMIs connected to higher level computing systems is often the right choice. To do this, it’s necessary to pick a controller that supports multiple protocols.
IO-Link or a similar protocol is often used to link multiple field devices to a controller because it’s a much more cost-effective option than hard wiring each field device to a controller input point. With IO-Link, it’s common for an IO-Link to Ethernet/IP gateway to connect directly to field devices and then connect back to the controller via EtherNet/IP.
The gateway acts as a multiplexer by connecting to multiple IO-Link field devices directly and making information from them available to the controller via a single EtherNet/IP cable. Although IO-Link supports a relatively short distance between the sensor and the IO-Link master device, it can connect to a variety of gateways and related protocols to extend its reach to the controller.
For example, in large manufacturing and warehouse conveyor applications, longer distances between the controller and devices can be achieved by using gateways that sit on an EtherNet/IP network. The use of easily available managed and unmanaged switches also helps distribute Ethernet connections.
For connections among controllers, or for controllers to HMIs, some variant of Ethernet is the preferred option, with selection of the particular protocol dependent on which is supported by the hardware. Although Ethernet dominates these higher-level connections, lower level connections from controllers to remote I/O and field devices are still made via a variety of communication technologies such as IO-Link, DeviceNet, Profibus DP, or some variant of Ethernet.
As Ethernet hardware declines in cost and as capabilities continue to expand, particularly with respect to higher speeds and more determinism, engineers should expect it to become ever more widely used in industrial communications.
Chris Harris is team lead-technology assistance group at AutomationDirect. Edited by Chris Vavra, production editor, Control Engineering, CFE Media, firstname.lastname@example.org.
- Ethernet’s performance can often match and exceed levels of fieldbus technologies today.
- EtherNet/IP and Modbus TCP are emerging as leaders in North America for manufacturers.
- More users will adopt Ethernet hardware as the cost declines and its capabilities expand.
What other Ethernet technologies and protocols might emerge in the future to accompany EtherNet/IP and Modbus TCP?
See related articles linked below.