onsemi advances industrial Ethernet with 10BASE-T1S controller

Higher data rates have been the main highlight in Ethernet progression, but not every networking application requires high speed. The IEEE 802.3cg standard was created to address several additional important factors. Deterministic real-time performance, cable reach, installation complexity, and network maintenance costs are also critical for industrial use cases.

10BASE-T1S is defined in the IEEE 802.3cg standard. It aims to further enable Industry 4.0 by taking Ethernet all the way to edge node devices, enabling a wider set of industrial applications to use Ethernet, lowering the complexity and cost of industrial network installations, and replacing legacy multi-drop communications.

A Short Profile of 10BASE-T1S
10BASE-T1S is the short-reach, multi-drop variant of the 10 Mb/s single-pair Ethernet PHY defined in the IEEE 802.3 standard. 10BASE-T1S PHYs use unshielded single-pair cables, that are cheaper and easier to deploy compared to multi-pair and/or shielded cables. According to the specification, any 10BASE-T1S implementation supports at least 25 meters reach with at least 8 nodes.

Cables are often the highest cost factor in a network installation. Single-pair cables have just two wires compared to traditional Ethernet cables which use up to 8 wires (4 twisted pairs). 10BASE-T1S can substantially reduce the cost, size, and weight of cables needed for network installations.

The multi-drop capability further reduces the number of cables by enabling a single pair to connect to multiple nodes. Additionally, multi-drop significantly reduces the number of PHYs and switch ports needed.

The NCN26010 exceeds the IEEE 802.3cg specs by enabling up to 50 meters reach with 8 nodes and up to 40 nodes on a 25-meter segment.

The 10BASE-T1S technology yields even more cost reduction opportunities by simplifying network maintenance and reducing the number of standards that must be supported. In fact, 10BASE-T1S can reduce network installation costs by removing up to 70% of the cables and up to 80% of deployment costs. Moreover, it eliminates the need for large switches, gateways, protocol translators, and the additional wiring and power they require.

Figure 1: In a new 10BASE-T1S installation, an unshielded, single-pair cable can replace all the yellow wires in the cabinet.

Another important advance brought by T1S is determinism, thanks to the invention of PLCA (Physical Layer Collision Avoidance). A deterministic system is one in which no randomness is involved in the development of future states of the system. In other words, for a given initial state, a deterministic system will always return the same output or reach the same future state. In the context of Industrial Ethernet, deterministic communication is the ability of the network to guarantee that an event will occur (or a message will be transmitted) within a specific amount of time. This is sometimes referred to as a “bounded response.” An application is considered deterministic if its response time can be guaranteed to fall within a certain margin of error. Determinism provides a measure of reliability that the communication or output will not only be correct but will happen in a specified time. For a system to be considered real-time, it must specify a maximum time in which it responds to an event or transmits a message. A non-real-time system, on the other hand, is one that runs at a consistent speed, with no deadline. It’s important to note that determinism is a defining quality of a real-time system.¹

Networks using CSMA/CD (Carrier Sense Multiple Access/Collision Detection) are not deterministic. Instead, they rely on statistics to get access to the shared medium. The fundamental principle of CSMA/CD is that the PHY detects whether another station is transmitting (carrier sense), letting the MAC defer any transmission until the line is free. However, if no carrier is detected, multiple MACs may initiate a transmission concurrently, creating a physical collision. This happens especially at the end of a transmission, where multiple deferring MACs may detect the end of the carrier at the same time, or more precisely, within the same time window, called the collision window. The collision window is determined by the network propagation delay, including the cable length and the PHY implementation characteristics. When a collision occurs, the PHY reports it to the MAC, which in turn aborts and reschedules the current transmission. Before the MAC makes a new attempt at transmitting the frame, it pauses for a random time, called backoff, where upper limit increases exponentially with the number of (consecutive) failed transmission attempts. This progressively lowers, but never eliminates, the chance of having subsequent collisions. The random backoff leads to huge variability in the access times, which could even result in the packet not being transmitted (dropped) after ten failed transmission attempts. As a result, CSMA/CD shows very poor real-time performance. CSMA/CD networks are known to operate reliably in non-real-time environments if the average load does not exceed 30% of the available network bandwidth. Another side-effect of CSMA/CD is the so-called “capture” effect, where one node may repeatedly get access to the network for a very long time. Potentially, for as long as the MAC has frames to be transmitted. Therefore, fairness of access is also a big concern for CSMA/CD when considering real-time performance. Finally, the collision detection mechanism in the PHY can be quite complex and may yield very poor performance in the presence of high noise, which is often found in industrial environments. Consequently, CSMA/CD networks are not suitable for many industrial applications that require deterministic and reliable real time performance.

The PLCA method, on the other hand, allows you to calculate the worst-case media-access latency as a function of the number of nodes and the MTU (maximum frame size) of your network, which is configurable. 10BASE-T1S PHYs (such as the one embedded in the NCN26010) typically employ PLCA, which is a key ingredient for real-time applications that require deterministic performance such as automotive, industrial, and building automation. PLCA is designed to provide collision-free operation on half-duplex, multi-drop networks. With PLCA in place, the transmission cycle begins with a beacon (a 2.4 µs physical layer signal) sent by the coordinator node (Node 0) that the network nodes use to synchronize. After the beacon is sent, Node 0 gets a transmit opportunity. If Node 0 has no data to send, it yields its opportunity to Node 1 after a very short time (by default, the time needed to transmit 32 bits, that is 3.2 µs). Otherwise, if Node 0 has data to be transmitted, it can send a packet of up to the configurable maximum allowed frame size (MTU) of the network which by default is 1500 bytes. This process continues until each node has been offered one transmit opportunity. A new cycle is then initiated by the master node, which sends another beacon. To prevent a node from blocking the bus, a jabber function interrupts a node’s transmission if it exceeds the maximum allowed frame size, allowing the next node to transmit. This solves the ‘babbling idiot’ problem which affects several multi-drop technologies. The result is that there is no impact on data throughput and no collisions on the bus. The PLCA architecture enables 10BASE T1S to be used in noisy industrial applications that require deterministic real-time performance.²

Another important requirement for many industrial networks with real-time and deterministic requirements is the ability to work reliably in harsh EMC environments. While many other Ethernet standards were not designed to support industrial EMC environments, 10BASE-T1S was designed with such requirements in mind. The result is that 10BASE T1S with unshielded single pair cables shows outstanding EMC performance compared to existing technologies. Using 10BASE-T1S it is possible to design systems that meet Class 3 IEC61000-4-6 EMI requirements (10 V_rms common-mode noise injection) with unshielded single pair cables. PLCA also plays a key role in electromagnetic immunity: knowing that the bus is collision-free, the PHY’s receiver can employ advanced techniques to recover the signal in the presence of high levels of noise.

Another interesting characteristic of PLCA is that although it changes the way the nodes access the media, it does not modify the MAC. 10BASE-T1S PHY implementations can be connected to a standard CSMA/CD MAC using a legacy MII port. This is possible because the collision avoidance mechanism is fully implemented in the physical layer (hence, the name PLCA), while the MAC reacts to carrier-sense and collision-detect signaling as normal. In other words, the PHY provides a sort of “augmented reality” to the standard MAC through the CRS and COL MII signals.

10BASE-T1S Ethernet Controller
The NCN26010 10BASE-T1S Ethernet controller is a vital building block in installing and maintaining cost-effective industrial networks that allow companies to achieve the vision of Industry 4.0.

•  Takes Ethernet all the way to edge
•  Allows for the design of deterministic real-time systems that many industrial applications require
•  Works with unshielded single-pair cables
•  Reduces the complexity and cost of network installations
•  Eliminates the need for large switches, gateways, protocol translators, and the additional wiring and power they require
•  Lowers software maintenance costs because you don’t need to maintain multiple technologies
•  Reduces network maintenance complexity by reducing the number of network standards that must be maintained. This can be done by replacing legacy point-to-point and multi-point standards (including RS-485, CAN, FlexRay, RS-232, and HART) with 10BASE-T1S Ethernet
•  Enables greater data throughput over existing cables, eliminating the need to run new cables, which is often the highest cost factor in a networking installation
•  Designed with EMC requirements for both industrial and automotive in mind
•  PHY compliant with engineered Power over Data Lines (PoDL)

Below are a few examples of how 10BASE-T1S can be used:

•  Elevators: Reduce wiring inside the car as well as on the per floor control units (displays and call button)
•  Industrial Cabinets: Reduce installation complexity and cost
•  Networked Sensors: Ethernet that can go all the way to sensor edge node devices with up to 40 nodes on a single 25-meter SPE segment, 5x the requirement of IEEE 802.3cg standard
•  Street lighting: add EV charging stations to street lighting
•  Intra-System Communications: 10BASE-T1S Ethernet can replace most of I2C, SPI, and other proprietary on-board busses and backplanes (for example, server motherboards and switches). Reduces software and maintenance efforts significantly. Simplifies the layout as well as power distribution without impacting data rates or latencies
•  Building automation: Simplify building access control systems
•  Train / Tram / Bus: Reduce the wiring at door modules and surveillance and emergency call systems. A unified Ethernet architecture makes maintenance and system complexity more manageable
•  Automotive: Reduce the weight, complexity, and cost of wiring. Additionally, allows for service oriented architectures up to the edge of the IVN

A typical 10BASE-T1S Ethernet PHY controller provides the physical layer functions needed to transmit and receive data over an unshielded single pair cable and supports communication with a MAC via a standard MII interface. However, onsemi’s NCN26010 Ethernet controller integrates the PHY and MAC on a single MACPHY device, bringing Ethernet to sensors and other industrial devices using mid/low-end MCUs with no integrated MAC. That, in turn, significantly reduces the system complexity and provides flexibility to repurpose nodes in forever-changing system configurations after the initial installations.

Figure 2: A MACPHY device can connect to industrial controllers, sensors, and other devices that may not include a MAC.

MACPHY Ethernet Transceiver
The NCN26010 from onsemi is the first 10BASE-T1S Ethernet MACPHY on the market. It is an integrated Ethernet MACPHY transceiver that can connect sensors, actuators, and other industrial devices without needing an external MAC device.

The NCN26010 offers two major differentiators. The first is enhanced noise immunity mode with superior bit error rate (BER) performance which enables up to 50 meters reach with 8 nodes, 2x the requirement of the IEEE 802.3cg standard. NCN26010 meets the IEC6100-4-6 conducted-immunity test at 10 V_rms, thus ensuring robust signal detection in noisy factory environments. The second is lower capacitance on the line pins which enables more nodes to be connected on a single segment. It allows up to 40 nodes on a single 25-meter single-pair cable, which is five times greater than the requirement set in the IEEE 802.3cg standard.

Figure 3: NCN26010 features low line capacitance to offer up to 40 nodes on a single 25 meter unshielded single-pair cable.

The NCN26010 also lowers software maintenance costs by following Ethernet’s layered approach. As a result, changing Ethernet PHYs doesn’t require modification in the upper software layers.

Industrial Ethernet to the Edge
The NCN26010 can help companies realize the benefits of Industry 4.0 by providing an Ethernet solution that can be used in places where it was not possible or practical in the past. 10BASE-T1S offers the potential to expand Ethernet connectivity to the very edges of industrial networks while simplifying architectures and lowering network installation and maintenance costs.

References
¹ Deterministic, real-time control: What does it really mean in motion control applications?
² Ethernet 10BASE-T1S and 10BASE-T1L are Reinventing Vehicle and Industrial Connectivity – Microwave Product Digest