Transition paths in different industries
While the idea to enable IP based networks down to the sensor level is obviously beneficial, the question is, how SPE is able to deploy in the focused devices and installations. When looking at the current installed base, there is a large variety of fieldbusses (see Table below) and sensor networks in the field.
|Fieldbus||Transmission speed||Max. distance|
|Interbus||500kBd .. 2Mbit||up to 400m|
|Profibus DP||9,6kBd .. 12 MBit||100m .. 1200m|
|CANopen||62,5kBd .. 1 Mbit||30m .. 1000m|
|Devicenet||125kBd .. 500kBd||100 .. 500m|
|CompoNet||up to 4Mbit||1500m (@93kBd)|
|CC-Link||up to 10Mbd||100m|
|HART||1,2kBd||1500 .. 3000m |
(dep. on cable)
The table shows a mixture of fieldbuses in operation in the process and factory automation. Hence, there are different speed and distance requirements that are deployed today. These requirements define the specification that any following technology needs to support. On top there is the Ex area in process automation with specific requirements for intrinsic safety. The connection to the actual upper layer networks in brownfield installations bares a number of disadvantages as follows:
- demand for gateways between the “old” installation and new Ethernet networks
- lack of diagnosis and parameterization capabilities in some cases
- transmission speed and cycle time of some current networks limits performance
- maintenance and support effort increase for keeping “old” know-how alive
- installation effort to support multi-vendor configuration and tooling
- warehouse cost increase and availability challenged by potential end of life scenarios
In fact not all of the listed fieldbus standards are deeply affected as in the above list and some like for example IO-Link are just in ramp-up stage. But it is clear that the call for action and market pressure is different in the industries depending on the extent of use of old technology in the installations.
Hilscher expects an early adoption and deployment of SPE in the process automation sector starting around 2021. The reason is that the installations are based to a large extend on HART, PROFIBUS PA and similar fieldbuses that are not in the necessary extend supporting digital business models. The Namur Organisation and specially the FieldComm Group together with the PROFIBUS International (PI) and ODVA strongly drive the move to APL for process automation industries. The applications in these areas have usually lower performance and cycle time requirements and the deployment of diagnosis and parameterization through the network has not yet happened as much as in factory automation applications. On the other side, the deployment of SPE in factory automation might take a bit longer. Organization such as the ODVA, PI including the IO-Link group have started activities to evaluate the integration of SPE into their respective standards looking at positioning and the benefit in their applications. In parallel, there are two very active groups on their way to propose different plugs, connectors and cabling for the installation. Looking at all these very productive initiatives and the numerous open questions they address, a field deployment in factory automation would rather start around 2024.
As a conclusion we believe that those areas in the manufacturing industry, where equipment is still non-transparent to the upper layers in terms of status, diagnosis and parameterization, the overall system performance lacks and the advantages of digitization such as machine up time increase, availability, predictive maintenance etc. will not materialize. Therefore, the pressure in the process industry is expected to be much higher to act and move forward as quickly as possible.
Single-pair Ethernet at a glance
The question might come up why people have not moved much earlier towards a single twisted pair as the basic idea appears comparably simple and potential benefits appear also quite obvious. However, the change versus an existing Ethernet network is not so quite simple as the exchange of a cable implies. In addition, a number of requirements need to be added to enable the expected benefits for the different industries.
Different physical layer
The current Industrial Ethernet 10Base-T/100Base-TX that is the most deployed and adapted standard in industry, uses two twisted pair cables for unidirectional transmit and receive data. Therefore, in change to this, a single-pair Ethernet transmits and receives via in the same twisted pair and therefore requires a different physical layer as well as different coupling and transducers.
Long distance transmission
Especially the targeted sensors, actuators and other peripheral field devices in industrial automation require on top a much larger cable length between them. Hence, a strong demand came to enhance the cable distance between stations up to 1000m, versus the specified 100m that are available in 100Base-TX today.
A third aspect came from the process automation field. Next to the long distance requirement, an intrinsically safe transmission is needed, to support the Ex and hazardous areas.
In many of the actual sensor communication fieldbuses, a power transmission over the communication cable is possible. Hence, the single twisted-pair cable needs to also carry the necessary power to enable the powering of remote sensors and actuators.
Application specific bandwidth demand
Next to field level devices and sensors that are well covered from their bandwidth demand with 10Mbit transmission speed, the idea was to also roll-out SPE into higher bandwidth applications. Therefore, the IEEE also defined standards suitable for vision, motion or HMI including the physical layers.
IEEE Standardization and related applications
All these requirements and inputs led to several SPE IEEE standards in the different transmission speeds as shown in the following Table:
|IEEE Standard||PHY standard||Transmission Speed||Cable Bandwidth||Cable length||Applications|
|IEEE802.3 cg||10Base-T1L||10Mbit||20MHz||1000m |
|Sensors, actuators and peripherals, machine controls, train and bus networks, building automation|
|10Base-T1S||10Mbit||20MHz||15m (UTP) |
|Cabinet installations (no PoDL) half duplex|
|APL||10Mbit|| 1000m |
|Intrinsically safe and Ex equipment|
|IEEE802.3 bw||(BroadR Reach)||100Mbit||166MHz||15m (UTP) |
|IEEE802.3 bp||1000Mbit||600MHz||15m (UTP) |
|HMI, IPC, Camera, Motion & robotics|
|IEEE802.3 ch||2.5/5/10Gbit||4-5 GHz||15m (STP)||Vision sensing, IPC, HMI, Analytics, medical systems|
|IEEE802.3 bu||Power over Dataline (PoDL for SPE, max. 60W power transmission|
The Table "Overview of different SPE related Standards" shows a split into three definitions for the 10Mbit single-pair Ethernet Standard IEEE 802.3cg to reflect the different needs and demand from the different sensor, actuator and peripheral applications. The 10Base-T1L is most suitable for the requirements in sensors as it allows up to 1000m cable length in a point-2-point connection and fits very well into actual installations.
In terms of the physical layer definition, the APL is exactly same as T1L, but adds in cases the components for intrinsically safe transmissions in the Ex area. The 10Base-T1S allows in opposite to T1L a multi-drop set-up with much shorter cable length and a different PHY layer called PLCA (physical layer collision avoidance). Multi-drop has a good fit for example in cabinet installations or other short-range applications. Both system require a different physical layer as shown in the picture and following table:
|half-duplex multi-drop||full duplex|
Although the physical layer differs to some details, that connection to the upper layers is the same. The IEEE has taken effort to make sure that any actual system with a MAC and an MII connection can interface to the actual new PHY so that the major change remains in only one OSI layer. The following picture shows the set-up:
On top of this, the 802.3bu standard has defined a standardized transmission of power over the data line and it is capable to transmit up to 50W to the single endpoint. This feature allows backward compatibility to several existing sensor network standards that also power the connected sensor from a central power controller. The set-up in SPE is as follows:
The system requires a Power Sourcing Equipment (PSE) to deliver energy over the cable. Three different voltages are defined that are connected to specified power. On the receiver side, the Powered Device (PD) in the above case a maximum of 50W at 48V can be delivered in a point to point T1L connection. At 24V it is still a max. of 10W with a regulated PSE. The system is largely compatible to the trunk and spur topology in process automation networks.
The standardization of single-pair Ethernet, as it stands, is well suited to support the demand of the industrial automation requirements. As the physical layer technology is already in use in some different flavor in the automotive industry, the industrial users can rely on an already field proven physical technology when they start implementations. However, there is still a distance to go to embed SPE into the actual Ethernet standards specifically in the factory automation environment. From system installation view-point, an IP network into each sensor allows to configure and maintain sensors in the field by a vendor-independent tool environment. From system installation point of view, an IP network into each sensor allows to configure and maintain sensors in the field by a vendor-independent tool environment.
Thank you for reading our second part. Next up: Part 3 - System adoption in different scenarios
Read the first part here: Part 1 - Single-pair Ethernet: An Introduction