A Comparison of Applications for SPE or Two-Wire Ethernet
Is single-pair Ethernet (SPE) on your radar? Originally receiving a lot of attention in the automotive market, SPE is gaining even more traction for factory automation applications.
In this article, I will review single-pair Ethernet technology (abbreviated to SPE) and show use cases of product integration. Industrial Ethernet is used to exchange data between a programmable logic controller and field devices on the factory floor (as shown in Figure 1), such as servo drives, sensors and actuators.
Most industrial Ethernet networks are based on 100BASE-TX Ethernet technology, which exchanges data at a speed of 100 Mbps and is based on two-pair (four-wire) CAT 5e Ethernet cable. Newer industrial Ethernet networks are capable of 1,000-Mbps speeds based on 1000BASE-T Ethernet technology, using all four pairs (eight wires) in a CAT 5e Ethernet cable. 100BASE-TX and 1000BASE-T Ethernet technologies support up to 100-m cable length between two Ethernet segments (or PHYs).
SPE uses one pair (two wires) for transmitting Ethernet frames over a cable. SPE is fully compatible with existing 100- and 1,000-Mbps industrial Ethernet technologies above the physical (PHY) layer, and therefore supplements existing 100- and 1,000-Mbps technology. Depending on the Ethernet speed, SPE can be used in Ethernet segments of up to 1-km in cable length. Figure 2 illustrates the data rate and required cable pairs for each Ethernet technology.
SPE brings-in advantages in terms of using Ethernet technology:
- Helps existing two-wire communication technologies (Controller Area Network, Process Field Bus [PROFIBUS], 4 to 20 mA, Foundation Fieldbus) increase data bandwidth while keeping the same cabling infrastructure.
- Makes communication robust in noisy environments by protecting the message with a frame check sequence.
- Secures communication of message payloads through encryption.
- Reduces cabling, resulting in lower costs for copper and reduced weight in new installations.
- Uses Ethernet technology in form-factor-constrained applications like robot arms.
- Let’s look at an industrial Ethernet node that consists of an application processor, a media access controller (MAC) and an Ethernet PHY transceiver, as shown in Figure 3.
Industrial Ethernet requires a dedicated MAC implementation to guarantee process data delivery within a given data exchange cycle time. Industrial Ethernet processors like the Texas Instruments Sitara™ processor family can operate multiprotocol industrial Ethernet standards in the MAC. The physical Ethernet technology is solely defined in layer 1 of the Open Systems Interconnection (OSI) model, which is the PHY device. SPE is implemented by using a dedicated Ethernet PHY in layer 1.
The Ethernet frames transfer over the Media-Independent Interface (MII)/Reduced MII/Reduced Gigabit MII between the MAC and the PHY. The application processor used in the media data input/output interface acts as a sideband interface for controlling the register configuration of the PHY device.
There are different SPE standards for addressing transmission rates and ranges:
Institute of Electrical and Electronics Engineers (IEEE) 802.3cg – 10BASE-T1: 10-Mbps transmission rate, 1,000-m range: Used for remote sensors and actuators with a low amount of data. Previously handled by analog devices such as 4- to 20-mA interfaces, as used in processing and automation in factories. Makes the reuse of existing two-wire cable infrastructures possible.
IEEE 802.3bw – 100BASE-T1: 100-Mbps transmission rate, 15-m range. Used for sensors, actuators and servo drives with a typical data exchange cycle time from 31.25 µs to 4 ms. Factory automation applications for shorter communication distances, robotics; shorter range than 100-BASE-TX Ethernet. Reduces cabling costs in the arm of a robotic system, for example.
IEEE 802.3bp – 1000BASE-T1: 1,000-Mbps transmission rate, 40-m range. Used for vision sensors with a high amount of data; an Internet Protocol (IP) network camera, for example. Factory automation applications with a shorter range than 1000-BASE-T Ethernet. Reduces cabling cost; requires one pair instead of four pairs. Figure 4 compares the range and speed of the different SPE standards.
All of the IEEE 802.3 specifications in the list above support point-to-point communication. IEEE is also working on a standard that supports multidrop, where multiple Ethernet PHYs connect to a SPE.
The PHYs developed for the 802.3 standards typically support one of the T1 protocols and are not backward-compatible to previous versions of the 802.3 standards. You will need to select the PHY device with respect to the required use case. PHY vendors sometimes offer pin-to-pin compatible versions of the PHY, which eases product development.
SPE also enables the coexistence of power and data on a single pair, which is referred to as power over data line (PoDL) and advanced physical layer (APL). With the PoDL/APL approach, it is possible to source a remote sensor or actuator at distances as far as 1,000 m with the same cable to transport data, similar to the previously mentioned 4- to 20-mA interface. SPE can be used in a star topology, a line topology or any combination that includes hubs and switches.
Example Implementations of SPE
Let’s look at a few example implementations of SPE. Example No. 1: A sensor or actuator device as shown in Figure 5 with low data rates in process automation applications using 10BASE-T1 as the interface, with up to a 1,000-m range.
Data rates and processing speeds are low in a sensor or actuator device; thus, it is possible to build such systems with a microcontroller that has an integrated MAC. The sensor or actuator uses an external 10BASE-T1 Ethernet PHY. Such devices are common in point-to-point applications where the control unit and sensor are hundreds of meters apart, and they are not typically daisy-chained into a line topology. It is possible to use PoDL or APL for such applications. This example can replace devices with existing 4- to 20-mA interfaces while reusing the existing cable infrastructure.
Example No. 2: A servo drive, sensor or actuator (as shown in Figure 6) with typical data rates in factory automation applications using the 100BASE-T1 interface, with up to 15 m in range.
The use of a servo drive, sensor or actuator is a typical scenario in which industrial Ethernet protocols (Process Field Net [PROFINET], Ethernet/IP) communicate in factory automation networks.
The device typically has a three-port switch with two physical Ethernet ports and one host port to the device application. Such a device requires more processing power and a dedicated MAC interface, and therefore uses a microprocessor.
Example No. 3: A media converter application, as shown in Figure 7, is a converter bridge between two- or four-pair Ethernet to SPE. A media converter is built without a MAC. Instead, the T1 Ethernet PHY connects over the MII back-to-back to a two- or four-pair Ethernet PHY. The media converter acts like a protocol converter.
Conclusion: SPE differentiates in layer 1 of the OSI layer model, which is the PHY layer. The layer 2 (MAC) and above are not affected by the different SPE standards. Based on the use case, the developer has to select PHYs according the different data rates and ranges. With SPE the network operator is able to re-use existing cabling or save cabling cost for new installations. PoDL offers the possibility to provide power to the remote devices; hence SPE can replace the analog interfaces with a packet-based digital interface.