CAN Versus Ethernet: Advantages, Disadvantages and Alternatives for Industrial Automation

Kent Lennartsson, Research Manager
Kvaser AB

In industrial automation settings, like manufacturing plants and assembly lines, computing systems are used to replace the need for human involvement in the decision-making process. For example, rather than having operators at a cola plant use a machine to fill soda bottles with the CO2 levels needed for correct carbonation, the machine itself can be programmed to complete this task. While some lament the limited human involvement, there are several advantages to this method of manufacturing, including:

  • Increased productivity: machines can perform tasks at a much higher rate of speed than humans
  • Improved product quality: well-programmed machines remove human error
  • Higher profits: a result of faster and higher-quality production
  • Easier flexibility: changes in production require zero training
  • Improved safety: individuals are less involved in the functionality of dangerous, heavy equipment
  • Less downtime: interruptions or performance errors can be detected before they degrade system performance (i.e., a proactive service model)

Keep in mind, these advantages rely upon properly functioning, deterministic protocols (those that only permit outlined functionalities and outcomes based on incoming data). Classic CAN (Controller Area Network) has long been relied upon for this use. However, the adaption of Ethernet to industrial settings, providing higher data throughput and standard communication software, has begun to challenge this status.

KVRA Brief Introduction to CAN
CAN is a communication hardware that allows communication between parts of a system without the intermediary of a central computer. This concept is similar to today’s Internet of Things, in which devices like our smartphone can turn on our house lights or unlock our front door without a PC working as the middleman. In the automotive industry (a widely used application for CAN), electronic control units (ECUs), such as anti-lock brakes, engines or fuel injection systems, can communicate easily with one another, maximizing the performance of a vehicle. In industrial automation, this communication process can be used to transfer information between parts of a robotic arm to perform manufacturing duties or to adjust machinery performance due to changing line speeds, along with a wide variety of other, much more complex, automated manufacturing and production processes. Errors in a system can also be detected and diagnosed through this method of internal communication. And, while multiple messages may be transmitted at any given time, the CAN protocol determines the priority of each, allowing for an easy flow of communication between parts.

The CAN Protocol: Reliability at Its Best
CAN is considered one of the most reliable methods for transmitting real-time data. This is hailed as one of the protocol’s greatest advantages and, arguably, is one of the most important considerations when determining the system best fit for industrial automation. In manufacturing, unreliable data communication can have a direct effect on production, making it difficult to adhere to time-sensitive schedules and to ensure optimum product quality. A breakdown in communication can also mean that errors go undetected, turning smaller maintenance issues (like a loose arm) into larger problems down the line (like damage to products or machinery). The reliability of Classic CAN is achieved through a multitude of factors, including the impossibility of message collision and a very short error recovery time, ensuring large delays in production or damage to equipment is highly unlikely. However, this protocol is not without imperfections.

One of the primary drawbacks to CAN is the protocol’s limited bit rate (the speed at which information is communicated) in relation to the length of the network (the distance that information can be communicated). Classic CAN communication can reach up to 1 megabit per second at a cable length of 40 meters. Notably, it is possible to extend the cable length or the bit rate by using a high-performance CAN driver in combination with standard crystal oscillators—users simply need to be aware of what delays their application can afford to accept. While this “limitation” has not made a notable interruption to processes in the past, they are surpassed by the impressive bit rate and network lengths of Industrial Ethernet.

The Rise of Industrial Ethernet
Developed in the 1970s, Ethernet is now the global standard for network communication. This system allows our devices to “talk” to one another. Industrial Ethernet is exactly as it sounds: Ethernet for industrial applications. This method of data communication utilizes Ethernet systems to send information between automated machine parts. This form of Ethernet differs from its traditional counterpart, in that specialized, deterministic protocols are used to ensure a high probability of a desired outcome (e.g., quickly and accurately filling soda bottles with CO2).

Compared to Classic CAN, Industrial Ethernet offers high speeds of data transfer (up to 100 megabits per second) at higher network lengths, allowing for the collection of more data per second that can be easily bridged to the “Cloud.” And, while less reliable than CAN, Ethernet expertise is more widespread, with lots of software support when used in combination with Linux or Windows. However, these advantages do not come without issues.

Questions of Efficiency
Despite the fast transfer of data, Industrial Ethernet has not perfected the efficiencies of those transfers.

The minimum Ethernet frame size is 80 bytes with 46 bytes of data. When transferring 0 to 8 bytes, which is the data length used today in the Classical CAN communication frames, a large amount of overhead remains. This extra can make data transfers less efficient, despite the technology’s impressive bit rates.

Security Concerns
Thanks to the need for an IP address, Industrial Ethernet is by far a less secure method of transferring data. While the type of data communicated is usually less sensitive than the information we personally transmit through our own Ethernet interfaces, system controls are more prone to cyber-attacks (imagine the havoc manufacturing plants would face if entire operations were shut down thanks to a virus). Yes, firewalls, routers and bridges, along with other protective methods of securing Ethernet connections, exist, but they shouldn’t be considered foolproof.

Hidden Costs
While an Ethernet controller integrated in a microcontroller (MCU) is relatively similar in price to Classic CAN, these figures do not account for retrofitting. Due to the extensive use of CAN in industrial automation, configuring systems for Ethernet requires significant changes. Additionally, the cost of securing this technology from outside threats must be factored in. Industrial Ethernet is also relatively new, which means, at present, no single system stands out as dominant. These may be silly examples, but think about VHS versus Betamax or MiniDisc versus CD. Those aligned with the wrong technology end up paying more to catch up. For industrial automation, systems could face a costly upgrade sooner rather than later.

Implications for CAN FD
CAN FD (CAN with flexible data rate) is the latest installment of CAN. Improvements through CAN FD solve the limited bit rate issue by providing about 8 megabits per second even with cable lengths above 40 meters. CAN FD also ensures the transmission of larger data blocks (up to 64 bytes) in a single message, making real-time data closer to instantaneous, all while retaining system security. CAN FD improved these bit rate capabilities by alternating between short and long bit times during message transmission rather than a standard bit time for communication. Even more compelling, particularly to those weighing the price of Industrial Ethernet upgrades, CAN FD controllers can be used in existing CAN systems, making industrial automation upgrades simple and cost effective. While not yet widely used within this industry, the improvements of CAN FD make it an attractive option for industrial automation.

CAN and Ethernet in Harmony
While CAN and Industrial Ethernet are typically thought of as incongruous, the two systems can be adapted to work together. CAN-to-Ethernet gateways allow the two protocols to communicate. These gateways keep CAN at the heart of communications, but they allow for the inclusion of Ethernet where beneficial. While CAN-to-Ethernet gateways have previously been considered cumbersome and difficult to integrate, new improved versions of this technology are beginning to hit the market. For example, shipping terminal cranes used to locate, remove and replace large containers in a container port utilize Ethercan, a lightweight CAN-to-Ethernet gateway, to connect to terminal operating systems. This allows operators to access real-time, CAN-collected crane data, such as tire pressure, fuel levels and load weight, using high-speed Ethernet portals. Supervisors can also receive this data instantly at a centralized location to ensure all shipping terminal cranes are reaching top-level functionality. If issues are detected, such as low tire pressure, resolutions can be made quickly before small problems grow into larger, more time-consuming ones. Even better, this CAN data can be accessed in real time remotely so that information for the entire container port can be reviewed by interested parties anywhere in the world.

The Results Are In
Both Classic CAN and Industrial Ethernet offer an array of advantages to industrial automation. However, drawbacks to Ethernet, like a less secure connection and the lack of an industry standard, make this technology difficult to endorse. For now, CAN’s proven track record of reliability makes it the optimal choice for industrial applications. Additionally, the improved performance of CAN FD and the ease at which it can be integrated into existing systems make it a formidable challenger to widespread replacement of CAN systems by Industrial Ethernet. And with the rise of simpler gateways offering CAN-to-Ethernet compatibility, a farther-reaching solution leveraging the advantages of both may soon be on the horizon.

For more information, visit www.kvaser.com.

 

Comments are closed.