SERCOS III
Encyclopedia
| sercosIII |
>
sercos III is the third generation of the sercos interface
SERCOS interface
- Introduction :In the field of Industrial Control Systems, the interfacing of various control components must provide a means to coordinate the signals and commands sent between control modules...
, a globally standardized open digital interface for the communication between industrial controls, motion devices, and input/output devices (I/O). sercos III merges the hard real-time aspects of the sercos interface with Ethernet
Ethernet
Ethernet is a family of computer networking technologies for local area networks commercially introduced in 1980. Standardized in IEEE 802.3, Ethernet has largely replaced competing wired LAN technologies....
. It is based upon and conforms to the Ethernet standard (IEEE 802.3
IEEE 802.3
IEEE 802.3 is a working group and a collection of IEEE standards produced by the working group defining the physical layer and data link layer's media access control of wired Ethernet. This is generally a local area network technology with some wide area network applications...
& ISO/IEC 8802-3). Work began on sercos III in 2003, with vendors releasing first products supporting it in 2005. In addition to the standard sercos features cited under the sercos interface
SERCOS interface
- Introduction :In the field of Industrial Control Systems, the interfacing of various control components must provide a means to coordinate the signals and commands sent between control modules...
general description, sercos III also provides:
- Cyclic updates to devices at rates as low as 31.25 µsec
- Support for up to 511 Slave devices on one network
- Redundancy: Bump-less physical layerPhysical layerThe physical layer or layer 1 is the first and lowest layer in the seven-layer OSI model of computer networking. The implementation of this layer is often termed PHY....
single-fault recovery - Detection of a dropped physical connection within 25 µsec (less than one cycle update)
- Hot plugging: insertion & configuration of devices into network while cyclic communication is active.
General architecture
In order to achieve the throughput and jitter requirements required in the applications the interface is designed for, sercos III operates primarily in a Master/Slave arrangement exchanging cyclic data between nodes. The Master initiates all data transmission during a sercos real-time cycle. All data transmissions begin and end at the Master (circular).sercos III cycle
Communication across a sercos III network occurs in strict cyclic intervals. A cycle time is chosen by the user for a given application, ranging from 31.25 µsec. to 65 msecs. Within each cycle, data is exchanged between sercos III nodes using two types of telegrams: MDTs and ATs (see Telegram Types). After all MDTs and ATs are transmitted, sercos III nodes allow the remaining time in the cycle to be used as an NRT (Non real time) Channel, which can be used to exchange data using other formats, such as IP.The network remains available to NRT traffic until the next cycle begins, at which time the sercos III nodes close the nodes to NRT traffic again. This is an important distinction. SERCOS is purposely designed to provide open access at all ports for other protocols between cyclic real time messages. No tunneling
Tunneling protocol
Computer networks use a tunneling protocol when one network protocol encapsulates a different payload protocol...
is required. This provides the advantage that any SERCOS III node is available, whether sercos III is in cyclic mode or not, to use other protocols, such as TCP/IP, without any additional hardware to process tunneling. SERCOS nodes are specified to provide a store and forward
Store and forward
Store and forward is a telecommunications technique in which information is sent to an intermediate station where it is kept and sent at a later time to the final destination or to another intermediate station. The intermediate station, or node in a networking context, verifies the integrity of...
method of buffering non-SERCOS messages should they be received at a node while cyclic communication is active.
Telegrams
Telegram format
All sercos III telegrams conform to the IEEE 802.3 & ISO/IEC 8802-3 MAC (Media Access ControlMedia Access Control
The media access control data communication protocol sub-layer, also known as the medium access control, is a sublayer of the data link layer specified in the seven-layer OSI model , and in the four-layer TCP/IP model...
) frame format.
Destination address
- The destination address for all sercos III telegrams is always 0xFFFF FFFF FFFF (all 1s), which is defined as a broadcast addressBroadcast addressA broadcast address is a logical address at which all devices connected to a multiple-access communications network are enabled to receive datagrams...
for Ethernet telegrams. This is because all telegrams are issued by the Master, and are intended for all Slaves on the network.
Source address
- The source address for all sercos III telegrams is the MAC addressMAC addressA Media Access Control address is a unique identifier assigned to network interfaces for communications on the physical network segment. MAC addresses are used for numerous network technologies and most IEEE 802 network technologies, including Ethernet...
of the Master, as it issues all telegrams.
Ethernet type
- A unique EtherType value has been assigned via the IEEE EtherTypeEtherTypeEtherType is a two-octet field in an Ethernet frame. It is used to indicate which protocol is encapsulated in the PayLoad of an Ethernet Frame. This field was first defined by the Ethernet II framing networking standard, and later adapted for the IEEE 802.3 Ethernet networking standard.EtherType...
Field Registration Authority for sercos III (0x88CD).
sercos III header
- The beginning of the Ethernet-defined data field always begins with a sercos III header, which contains control and status information unique to sercos.
sercos III data field
- The sercos III header is followed by the sercos III data field, which contains a configurable set of variables defined for each device in the network.
Telegram types
Two main types of telegrams are used within the sercos III Cycle. The Master Data Telegram (MDT), and the Acknowledge telegram (AT). Both telegram types are issued by the Master (control). The MDT contains information provided by the Master to Slaves. It is filled by the Master, and read by Slaves. The AT is issued by the Master, but actually populated by each Slave with their appropriate response data (feedback values, input states, etc.). More than one Slave uses the same AT, filling in its pre-determined area in the AT telegram, updating checksums, and then passing the telegram to the next device. This method reduces the impact of the Ethernet frame overhead on the performance of the network without compromising IEEE 802.3 & ISO/IEC 8802-3. The amount of data sent from the Master to Slaves, as well as the sum of the data returned by the Slaves, may exceed the 802.3-specified maximum 1500-byte data field size. To comply with this limit, sercos III may use more than one MDT telegram in a cycle, as well as more than one AT telegram (up to 4 in each case).Synchronization
To achieve true hard real time characteristics, sercos III, like sercos I & II, uses a form of synchronization that depends upon a synchronization “mark” issued by the Master control at exact equidistant time intervals. All nodes in a sercos Network use this telegram to synchronize all activities in the node. To account for variations in network components, delays are measured in the node-to-node transmissions during phase-up (initialization) of a sercos network, and those values compensated for during normal operation. Unlike sercos I & II, where a separate Master Sync Telegram, or MST is used for this purpose, sercos III includes the MST in the first MDT transmitted. No separate telegram is issued. The time between two MSTs is exactly equal to the designated sercos Cycle Time, tScyc.Physical and data link layers
sercos III supports standard IEEE 802.3 & ISO/IEC 8802-3 100Base-TX or 100Base-FX (100 Mb/s baseband) Full Duplex physical layer (PHY) entities. 802.3-compliant Media-Access Controller (MAC) sub-layers are used. Autonegotiation must be enabled on each PHY, but only 100Mbit full duplex is supported. Auto (MAU [Media Attachment Unit]-Embedded) Crossover is specified between the two Physical Medium Attachment (PMA) units present with a duplex port. These two units are referred to as the Primary Channel and Secondary Channel in the sercos III specification. Dual interfaces are required (two duplex interfaces per device). Within the sercos III specification the dual interfaces are referred to as P1 and P2 (Ports 1 and 2).sercos III stack
All of the functionality required to configure a sercos III interface is contained in a stack that is available in both “hard” and “soft” versions. The hard version is widely used for embedded applications (such as drives, I/O modules and micro-controller based motion control), where:- It is important that the overhead of managing the sercos III nodes not be placed upon the device processor.
- Nanosecond jitter is required.
The hardware stack is available in a number of different forms. These currently include:
- A bit stream for XilinxXilinxXilinx, Inc. is a supplier of programmable logic devices. It is known for inventing the field programmable gate array and as the first semiconductor company with a fabless manufacturing model....
FPGAs - A bit stream for AlteraAlteraAltera Corporation is a Silicon Valley manufacturer of PLDs . The company offered its first programmable logic device in 1984. PLDs can be reprogrammed during the design cycle as well as in the field to perform multiple functions, and they support a fairly fast design process...
FPGAs - A Net list for Xilinx FPGAs
- A Net list for Altera FPGAs
- The “netX” multi-network controller chip from Hilscher, GmbH.
The maximum jitter allowed with hard-stack-based Masters and Slaves is smaller 1 µsec. Using the above stacks yields a jitter similar to sercos II (35-70 nanoseconds).
sercos III also supports a “Soft Master”, using a completely software-based stack for the master interface. Since the maximum jitter in such a configuration is dependent upon the operating system of the Master, the maximum jitter may be set by a variable for the sercos III network when a Soft Master is employed.
For basic Slaves, such as I/O devices, a license-free core is available. The Easy-I/O core can be downloaded and loaded onto Xilinx Spartan-3 FPGA devices.
Data consistency
A term usually associated with the IT enterprise, data consistencyData consistency
Data consistency summarizes the validity, accuracy, usability and integrity of related data between applications and across an IT enterprise. This ensures that each user observes a consistent view of the data, including visible changes made by the user's own transactions and transactions of other...
can also apply to real-time control (see for example Peer to Peer Communication). For this reason, sercos III specifies that no data be overwritten (destroyed) during a transmission. Every slave on a network may access input and output data for every other slave on the network.
Addressing
Devices must support Ethernet’s MAC addressing, plus the sercos III addressing. Other addressing schemes are optional.sercos III address
- Each sercos III device contains a numeric address used by other devices on the sercos III network to exchange data. The address may be any whole integer from 1 to 511.
IP address
- sercos III does not use an IP address for its own operation. Whether a device contains an IP address or not is dependent on its support of other specifications, either independent (exclusive) of sercos III operation, or via the NRT portion of the cycle.
Network topologies
The sercos III specification defines two possible network topologies; Ring and Line. To those familiar with other networks, they may appear to both be configured as a Ring. All telegrams begin and end at the Master. The Full Duplex feature of the physical layer is used to achieve this.Line topology
- A line topology is the simpler of the two possible arrangements, and provides no redundancy. However, this configuration saves the cost of one cable. In it, only one of the two interfaces on the Master is used. Telegrams are issued out of the transmit PMA on the Master’s active port. Either port on the Master may be the active one. sercos III determines this during phase-up (initialization). The first Slave receives the telegrams on the connected interface’s receive PMA, modifies them as required, and issues them out on the transmit PMA of the second interface. Each cascading Slave does likewise until the last Slave in the Line is reached. That Slave, detecting no sercos III connection on its second port, folds the telegram back on the receiving interface’s transmit port. The telegram then makes it way through each Slave back to the Master. Note the last Slave also emits all sercos III telegrams on its second port, even though no sercos III connection is detected. This is for snooping, ring closures (see below), as well as hot-plugging. Keep in mind that since the Ethernet destination field in all sercos III telegrams is the broadcast addressBroadcast addressA broadcast address is a logical address at which all devices connected to a multiple-access communications network are enabled to receive datagrams...
of 0xFFFF FFFF FFFF (all 1s), all telegrams issued from this open port will be seen by other devices as broadcast telegrams. This behavior is by design, and cannot be disabled. To avoid taxing networks attached to an open sercos port, an NRT-Plug can be used, or alternately a managed Ethernet switch programmed to block broadcast telegrams received from the sercos port can be used.
Ring topology
- A ring topology simply closes the network by attaching the unused port on the last device in a ring back to the unused port on the Master. When the sercos III Master senses that a ring exists, it sets up two counter-rotating telegrams. The same data is issued simultaneously out of the transmit PMAs of both ports on the Master. From there both telegrams are managed essentially identically as they make their way through each Slave, ending back at the opposite port on the Master they were emitted from. Advantages to this topology include tighter synchronization, as well as automatic infrastructure redundancy (see below).
Other network topologies
- With both the line or ring structure, sercos III operates in a “circular” approach. All telegrams leave the Master, and return there. As with any network that operates in this manner, modified structures can be constructed to appear as a tree or star network, utilizing hardware that manages the branches, but the structure is still circular in nature.
Infrastructure hardware
sercos III is designed in such a way that no additional network infrastructure (standard Ethernet switches, HubsEthernet hub
An Ethernet hub, active hub, network hub, repeater hub or hub is a device for connecting multiple Ethernet devices together and making them act as a single network segment. A hub works at the physical layer of the OSI model. The device is a form of multiport repeater...
, etc.) is required to operate. In fact, no additional standard Ethernet (non-sercos III capable) components may be placed within a sercos III network, as their presence will adversely affect the timing and synchronization of the network.
Application layer (profiles)
The sercos III specification defines a broad range of variables developed by a consortium of product suppliers to provide interoperability between components (motion controls, drives, etc.). All Traffic across a sercos III network consists of Idents (parameters) with attributes. This method was first defined in sercos I, as an essentially flat set of Idents. They were later grouped into application sets to aid in selection of pertinent Idents required for a given industry, such as the “Pack Profile” for use with packaging machinery. During the development of the sercos III specification, this methodology was further refined to group the Idents logically by device class. The definition of the legacy Idents has remained largely untouched; rather their grouping has been re-evaluated for a more understandable architecture. This has also enabled the separation of communication Idents into a logical subset, simplifying migration from sercos I/II to sercos III, and providing a clear overview to users.Redundancy
When a ring network is employed, sercos III provides for automatic infrastructure redundancy. If any interconnection point in the ring ceases to function, the associated sercos III nodes will detect a “ring break” and “loop back” the end nodes, effectively operating as two lines rather than one ring.The operation is “bump-less”, as the detection & recovery time to such a break is less than 25 µsecs, which is less than the minimum sercos III cycle time. sercos III can also recover from ring breaks and “heal” with no interruption in operation. Since sercos III telegrams continue to be emitted by transmit PMAs on unconnected ports, and receive PMAs on unconnected ports continue to monitor for incoming data, when a sercos III port recognizes that a ring has by physically re-closed, it will re-activate the counter-rotating telegrams to functionally close the rings again. This operation is also bump-less.
Peer communications
To ensure the determinism required, most Real-time Ethernet standards enforce a master-to-slave-only method of communications. This can conflict with the need for a node in the system to exchange data efficiently with a node other than the network master. The conventional method to achieve this in a master-slave network is to pass data from one slave node to the master, where it is reissued to one or more different slaves. For example, if several servo drives on a network are to be synchronized to a signal from another drive on the network, the master must fetch the signal from this drive and reissue it to all other drives on the network. Disadvantages to this method are that delays are induced due to the multiple cycles required, and the master’s processing load is increased as it must actively participate in the function, although it contributes nothing. Since no data is destroyed in a sercos III telegram, data to and from any slave can be accessed by another node on the network without any additional cycle delay or master intervention. Additionally, as telegrams pass each node twice in a cycle (for both topology types), a node can even have the opportunity to access data supplied by a subsequent node. Two peer communication methods are defined in the sercos III specification: Controller to Controller (C2C) for multiple masters to communicate with one another, and Cross Communication (CC) for multiple slaves.Hot-plugging
Another feature of sercos III is hot-plugging, which is the ability to add devices to an active network. Using the features described for redundancy, a network can detect when a new device is attached to an active network. Processes exist that configure the new device, and announce it’s availability to the Master control. After that, the Master control can select to make use of the new device based on the application currently running.Non-real-time (NRT) channel
The time between the end of the transmission of all sercos III Real Time (RT) cyclic telegrams, and the beginning of the next communication cycle is defined as the “sercos III Non Real Time Channel” (NRT Channel). During this time period, the sercos Network is opened to allow transmission of Ethernet-compliant frames for other services and protocols. For example:- Web serverWeb serverWeb server can refer to either the hardware or the software that helps to deliver content that can be accessed through the Internet....
s can be embedded in sercos III-compliant devices to respond to standard Hypertext Transfer ProtocolHypertext Transfer ProtocolThe Hypertext Transfer Protocol is a networking protocol for distributed, collaborative, hypermedia information systems. HTTP is the foundation of data communication for the World Wide Web....
(HTTP) messages received via the NRT Channel. - Frames from other FieldbusFieldbusFieldbus is the name of a family of industrial computer network protocols used for real-time distributed control, now standardized as IEC 61158....
standards that conform to Ethernet frame formatting may be transmitted across a sercos III network.
Every sercos III-compliant node must support the passing of NRT frames through its sercos III interface. Whether a sercos III node actively makes use of the NRT feature is determined by the feature set of the product. If, for example, the device has an embedded web server, it could make available its IP address for access by other devices.
A sercos III network will always pass NRT frames, even when cyclic operation has not been initialized. This means that devices always have access to the network for NRT messages, as long as the ports are powered.
sercos III does not define whether a port should operate in cut-through switching
Cut-through switching
In computer networking, cut-through switching is a method for packet switching systems, wherein the switch starts forwarding a frame before the whole frame has been received, normally as soon as the destination address is processed...
or store-and-forward mode when handling NRT frames. There are sercos III products currently on the market that support both modes. Likewise, sercos III does not define whether a port should intelligently process NRT telegrams, such as learn the network topology.
The time allotted for NRT is dictated by the amount of data transmitted during the RT portion of the cycle. In real-world applications, there is a significant amount of bandwidth available for NRT frames. For example, in a typical application with 8 axes of motion and a cycle rate of 250 microseconds, the equivalent of 85 Mbps is available for NRT use. This amount of time means the NRT frames in this example can be as long as the maximum defined for Ethernet (Maximum Transmission Unit
Maximum transmission unit
In computer networking, the maximum transmission unit of a communications protocol of a layer is the size of the largest protocol data unit that the layer can pass onwards. MTU parameters usually appear in association with a communications interface...
[MTU] =1500). Using the same example of 8 axes, but with a cycle time of 62.5 microseconds, the effective bandwidth available for NRT frames would be 40 Mbps, and the MTU would be reduced to 325. As with any network where time on the bus is shared, MTU values should be configured to ensure reliable communication. Properly configured sercos networks will set the sercos parameter “Requested MTU” (S-0-1027.0.1) to the recommended MTU value, which can then be read by other devices to match their MTU settings. Regardless of the value of this parameter, a sercos node will allow non-sercos traffic to pass for the entire NRT channel time period (i.e., telegrams longer than the MTU setting are not discarded by the sercos stack). sercos parameter S-0-1027.0.1 is set by default to 576, the minimum value called out in RFC 791.
NRT access
NRT frames may only enter a sercos III network through a sercos III-compliant port. This can be achieved two different ways. One is to employ the unused sercos III port at the end of a sercos III network configured in line topology, as shown to the right.In a network configured in ring topology, the ring can be temporarily broken at any point to also attach a device. Since the redundancy feature of sercos III will reconfigure the network in a bump-less manner (responding in less than one cycle), no disruption of network transmission will occur. The ring can again be closed after the access is no longer required.
If access is desired in the middle of a line topology (where no free ports are available), or it is undesirable to break a ring topology for extended periods of time, the sercos III specification permits a device called an “NRT-plug” that can be used to provide access to the NRT channel anywhere along the network. NRT-plugs supply two sercos III-compliant ports, and one or more ports for NRT access.
Commercially available NRT Switches block the transmission of sercos III broadcast telegrams out their non-sercos III port(s), to prevent flooding of non-sercos III networks with sercos III cyclic data.
Functional safety support
"Functional safety" is a general term referring to the design of a system that reduces the risk that a hazardous event harmful to humans can occur with a system. The main definition is contained in the international standard IEC 61508IEC 61508
IEC 61508 is an international standard of rules applied in industry. It is titled "Functional Safety of Electrical/Electronic/Programmable Electronic Safety-related Systems"....
. Most industrial networks contain some type of features to conform to functional safety requirements. Rather than define a unique specification for this functional safety, sercos III Safety is based upon the CIP Safety safety protocol developed by the Open DeviceNet Vendors Association
Open DeviceNet Vendors Association
Open DeviceNet Vendors Association is an international organization that supports computing network technologies based upon the Common Industrial Protocol . These include DeviceNet, EtherNet/IP, CIP Safety and CIP Sync. Members of ODVA include representatives of leading automation companies...
(ODVA). This provides interoperability at the safety level with all networks based upon the Common Industry Protocol (CIP), including DeviceNet and EtherNet/IP.
Development tools
A number of different manufacturers provide tools to aid in the design of sercos III compliant network nodes. Since sercos III uses standard Ethernet frames, most off-the-shelf network tools can be used on sercos III networks. For example, a driver is available for the free Wireshark network protocol analyzer.Open source driver
On April 21, 2009, sercos International and sercos North America announced a cooperation with the Open Source Automation Development Lab eG (OSADL)OSADL
The Open Source Automation Development Lab eG was founded on December 8, 2005 and assigned the cooperative register number 440085 at the Stuttgart, Germany, District Court in August 2006...
to provide an open source software driver library for sercos III. The library simplifies the development of a sercos Master node. Released under the GNU General Public License
GNU General Public License
The GNU General Public License is the most widely used free software license, originally written by Richard Stallman for the GNU Project....
(GPL), the driver is available to download from either OSADL or SourceForge
SourceForge
SourceForge Enterprise Edition is a collaborative revision control and software development management system. It provides a front-end to a range of software development lifecycle services and integrates with a number of free software / open source software applications .While originally itself...
at http://sourceforge.net/projects/cosema.
Further information
Additional information on sercos III, including sources of sercos III compliant products and specifications for developing sercos III products, can be obtained from the governing organizations (sercos North America or sercos International e.V.).See also
- SERCOS interfaceSERCOS interface- Introduction :In the field of Industrial Control Systems, the interfacing of various control components must provide a means to coordinate the signals and commands sent between control modules...