By Kent Lennartsson, Research Manager, Kvaser AB
A Controller Area Network (CAN) is a message-based protocol that allows microcontrollers to communicate with devices without the need for a host computer, similar to today’s Internet of Things communication. The use of CAN began in the automotive industry and has since branched out into a multitude of sectors, including industrial automation. However, CAN in itself is not equipped to handle the control processes needed for this complex industry. Instead, a secondary, higher layer protocol was required to make adoption possible.
In the early 1990s, Allen-Bradley (now a part of Rockwell International), developed DeviceNet, a higher layer protocol for control systems specifically aimed at automation. While not the only higher layer protocol in use at the time, Allen-Bradley’s swift introduction of the specification through the Open DeviceNet Vendor Association (ODVA) made it more widely available. It was this availability that allowed CAN to become a viable option for the purposes of industrial automation (today there are several, customizable options available. For a comparison of the most popular, click here).
Classic CAN in Industrial Automation
Classic CAN is used in nearly every industrial sector, including manufacturing and assembly lines, heavy equipment, such as cranes and forklifts, and even simple sequence control demands like mail sorters. The protocol has remained the industry standard for a myriad of reasons, but the primary advantage to this system has been its proven history of reliability. Industrial automation of today requires a multitude of control systems moving in precision. In an industry that relies on schedules and accurate production, speed and efficiency are a requirement. Because of this, errors must be limited and should be quickly detectable. In a CAN system, there is constant communication between devices. Messages deemed a high priority are given precedence over others, removing any issues of message collision. Additionally, error messages are given high priority, allowing for swift error detection and correction. This ensures small mistakes do not become larger, more costly ones. Despite CAN’s prevalence in industrial automation, new communication technologies loom.
The Introduction of Industrial Ethernet
Today, Ethernet is the global standard for network communication. Developed in the 1970s, this technology is similar to CAN in that it allows devices to communicate with each other without the need for a central processor. It was not until relatively recently that this technology began to be fitted for settings of industrial automation.
Industrial Ethernet is simply Ethernet for industrial applications. It differs from the technology used in our homes and offices in that the messages communicated between devices use deterministic protocols, promoting the high likelihood of a desired outcome (e.g., the end product of an assembly line matches the initial design specifications). Some of the primary advantages of the system over traditional CAN are:
- Higher data transfer speeds (up to 100 MB/s) at higher network lengths
- Faster data transfers mean more data per second can be accessible from “the cloud”
- More individuals are familiar with Ethernet versus CAN
- More software support is available when used with Linux or Windows
Of note, while Classic CAN is limited to 1 MB/s with a cable length of 40 meters, it is possible to extend these figures using a high-performance CAN driver in combination with standard crystal oscillators. And, despite the strides Industrial Ethernet has made toward improved speed, a few big concerns remain.
Drawbacks of Ethernet
For industrial automation, the speeds gained through Ethernet have not transferred into improved efficiency when compared to Classic CAN. The minimum frame size for the communication protocol is 80 bytes, with 46 bytes of data. In a sector that relies on transferring 0–8 bytes, the excessive overhead can actually make data transfers less efficient, even with impressive bit rates. The security of this system has also been met with scrutiny.
Because Industrial Ethernet requires an IP address, it is automatically a less secure method of transferring data than Classic CAN. This means that automated systems may be vulnerable to cyber attacks. While connections can be secured with routers, bridges and firewalls, the vulnerability of this system is one that the industrial automation sector has not had to face under Classic CAN. And these safety assurances add to the costs associated with Industrial Ethernet.
On the surface, Industrial Ethernet and Classic CAN may appear relatively similar in price. However, there are other considerations to make when switching to Ethernet. For one, since CAN has been the standard for industrial automation, those looking to embrace the technology may face substantial costs in retrofitting existing systems. Secondly, there is the extra expense associated with ensuring the security of this new technology. And finally, at present, Industrial Ethernet has no dominant system. This means that, if one model becomes more popular than another, users could find themselves with obsolete technology sooner rather than later.
CAN FD to the Rescue
CAN FD (CAN with flexible data rate) was created to respond to the automotive industry’s need for increased bandwidth. The protocol now alternates between short and long bit times when transferring messages, rather than relying on a standard bit time for all communication. This new and improved extension to the Classic CAN protocol allows for data transfers of 8 MB/s, even with cable lengths beyond 40 meters, and CAN FD can transfer up to 64 bytes of data in a single message, all without the giant overhead inefficiencies associated with Industrial Ethernet. Additionally, the cost to upgrade existing CAN communication protocols to CAN FD is far simpler and cost effective than overhauling entire systems for the needs of Industrial Ethernet. While still in its infancy stage in the industrial automation, CAN FD has improvements that point to big benefits for the sector.
Industrial Ethernet and CAN Working Together
While the two systems of communication seem to be in competition, there are ways to capitalize on the advantages of both. A CAN-to-Ethernet gateway allows CAN systems to communicate through Ethernet methods. For many years, these gateways have proven cumbersome; however, simpler versions have recently been introduced to the market. For example, the Ethercan, a lightweight CAN-to-Ethernet gateway, has been used in conjunction with rugged CAN hardware to connect terminal operating systems with shipping port cranes. This use of CAN and Ethernet in harmony allows operators to access real-time data on crane systems information (e.g., tire pressure or load weight) using Ethernet portals. The reliability of CAN means errors are detected quickly, while the use of Ethernet allows CAN-collected data to be accessible from anywhere in the world.
The use of CAN in industrial automation is nothing new. While new challengers, like Industrial Ethernet, may pose threats to the dominance of the communications protocol, the many benefits of the system, including reliability, security and affordability, cannot yet be outweighed by this new technology. Even further, the potential for CAN FD to improve communication speeds mean that those committed to CAN may not need to go much further for Ethernet-like advantages. And for those looking to cross the two methods of data communication, CAN-to-Ethernet gateways are an option that have finally entered an era of improved ease of use, making them more adaptable for the industrial automation sector.