I s s u e 2 / F e b ru a r y 2 0 13

SYSTEM COMPARISON

The 5 Major Technologies

nd E d i t i o n
2
PROFINET,
POWERLINK,
EtherNet/IP,
EtherCAT,
SERCOS III
How the Systems
Work

The User
Organizations
A Look behind
the Scenes

Investment
Viability and
Performance
Everything You
Need to Know!

Safety
protocols
Learn the basics!
 Bhagath Singh Karunakaran
Stefan Schnegger

Stphane Potier
Huazhen Song
Luca Lachello

Anton Meindl
Peter Wratil

Preface Outsiders are not alone in finding the world of Industrial Ethernet somewhat confusing. Experts
who examine the matter are similarly puzzled by a broad and intransparent line-up of competing
systems. Most manufacturers provide very little information of that rare sort that captures tech-
nical characteristics and specific functionalities of a certain standard in a way that is both com-
prehensive and easy to comprehend. Users will find themselves even more out of luck if they are
seeking material that clearly compares major systems to facilitate an objective assessment.

We too have seen repeated inquiries asking for a general overview of the major systems and
wondering where the differences actually lie. We have therefore decided to dedicate an issue
of the Industrial Ethernet Facts to this very topic. In creating this, we have tried to remain as
objective as a player in this market can be. Our roundup focuses on technical and economic as
well as on strategic criteria, all of which are relevant for a consideration of the long-term viability
of investments in Industrial Ethernet equipment. The arguments made in this publication were
advanced and substantiated in numerous conversations and discussions with developers and
decision-makers in this field. We have made every attempt to verify claims whenever practically
possible.

This document must not be modified Despite all our efforts, though, we were unable to ascertain exact, verifiable information on
without prior consent of its publisher. some aspects, which prompts us to ask for your help: if you would like to propose any
Passing on the document in its entirety amendments or corrections, please send us an e-mail or simply give us a call. We look forward
is expressly encouraged. to any and all support in supplementing this overview, and we welcome all discussions that
The current version is available for download contribute to making the assessments of the various Industrial Ethernet standards as thorough
from www.ethernet-powerlink.org. and objective as possible. This second, expanded edition includes feedback submit by the
Industrial Ethernet community after publication of the first issue in November, 2011.

Contact: EPSG Office, Luca Lachello, Software Engineering Manager COMAU Robotics Italy
phone: +49 30 850885-29 Peter Wratil, Managing Director Innotec Germany
info@ethernet-powerlink.org Anton Meindl, President EPSG Germany
Stefan Schnegger, Business Unit Manager B&R Austria
Bhagath Singh Karunakaran, CEO Kalycito India
Huazhen Song, Marketing Manager POWERLINK Association China
Stphane Potier, Technology Marketing Manager EPSG France

2
 Systems Roundup:
The 5 Major Contenders
2nd Edition

INTRODUCTION 4 Selection of Systems for Review

HOW THE SYSTEMS WORK 6 Approaches to Real-Time

PROFINET Communication
POWERLINK Communication
EtherNet/IP Communication
EtherCAT Communication
SERCOS III Communication

ORGANIZATIONS 12 User Organizations and Licensing Regimes

CRITERIA FOR INVESTMENT VIABILITY 16 Compatibility / Downward Compatibility

Electromagnetic Compatibility (EMC) Electrical Contact Points
Cabling / Feasible Topologies Hot Plugging Capability
High Availability Gigabit Readiness
Availability of Safety Solutions Market Penetration

PERFORMANCE 18 Theoretically Achievable Cycle Time Communication Architecture

Direct Cross-Traffic Heavy Data Traffic
Network Load for Safety Communication
Actual Cycle Time Jitter Performance Contest

IMPLEMENTATION 22 Master Implementation

Costs for Network Components
Slave Implementation
Node Connection Costs
Operating Costs

SAFETY FUNCTIONALITY 25 Network Integrated rather than Hard Wired

THE BLACK CHANNEL PRINCIPLE 27 Safety Field Networks

Safety Data Transport via Bus or Network Lines

H O W T H E S afety S Y S T E M S W O R K 28 CIP Safety PROFIsafe openSAFETY FSoE

Certifications Technology Device Implementation
Integration Performance CRC

3
 Selection of Systems
for Review
|
This issue of Industrial Ethernet Facts compares PROFINET (RT, IRT),
POWERLINK, EtherNet/IP, EtherCAT, and SERCOS III, i.e. five out of
about 30 Industrial Ethernet systems currently in use around the
world.1 Why these five? The selection was based on technical aspects,
standardization status, and strategic market considerations. Relevant
issues include e.g. whether a user organization backs the ongoing
development of a protocol, whether a protocol is classified in the IEC
standard, and whether a system is suitable for hard real-time
requirements.

Real-time
A mechanism to resolve data collisions that is part of the IEEE 802.3
Ethernet standard causes irregular delays in data transfer. In order to
achieve real-time performance, Industrial Ethernet protocols employ
special preventive measures to avoid such collisions. For hard real-
time, signal transmission times must stick exactly to a given time
frame, or else they will trigger a failure signal. For soft real-time, some
deviation within a limited span of time is tolerable. While cycle times
of up to several hundred milliseconds may be good enough for soft
real-time applications, e.g. for temperature monitoring, digital control
systems or Motion Control applications often require cycle times
below one millisecond.

1 For a more extensive overview of systems, consult the list on

www.pdv.reutlingen-university.de/rte/ compiled by Prof. Dr.-Ing. Jrgen Schwager,
head of the Process Data Processing Lab at Reutlingen University.

4
 Systems Roundup:
The 5 Major Contenders
2nd Edition

Market Penetration
Another key aspect in selecting Industrial Ethernet systems for compa-
rison was market penetration: various IMS and ARC surveys indicate
that about three quarters of all Industrial Ethernet applications around
the world use EtherNet/IP, PROFINET, or Modbus TCP. Next in line are
POWERLINK and EtherCAT, two systems particularly suitable for hard
real-time requirements. The following roundup does not examine
Modbus TCP on its own, since its user organization ODVA has stated
that it has been integrated into EtherNet/IP. SERCOS III, however, was
included for comparison despite its marginal market share, because
this system plays a vital role for fast Motion Control solutions.

Building technology, control and

automation levels, trouble-free
processes, storage systems

Conveying systems,
simple controls,
majority of all automated systems

Machine tools,
fast processes, Real-time classes
robots and application areas
(IAONA classification)
Highly dynamic processes,
electronic drives

1 s 10 s 100 s 1 ms 10 ms 100 ms 1s 10 s

Response time / jitter

5
How the Systems Work
|
Diverse Approaches to Real-time Generation method: in each cycle, data for all network nodes is sent in one
There are three different approaches to building a real-time Ethernet telegram that travels from one node to another along the ring
solution: topology of the network, also collecting node responses on the way.
1. B
 ased on TCP/IP: Protocols are based on standard TCP/IP layers In contrast to that, the single telegram procedure used by the other
with real-time mechanisms embedded in the top layer. systems works by sending individual telegrams to the nodes, which
These solutions usually have a limited performance range. also respond individually in separate telegrams.
2. Standard Ethernet: Protocols are implemented on top of standard
Ethernet layers. These solutions benefit from Ethernet evolution The systems use three different mechanisms for network access and
without further investment. data synchronization:
3. Modified Ethernet: The standard Ethernet layer, the Ethernet A master controls the timing on the network. In POWERLINK
mechanism and infrastructure are modified. These solutions put environments, the master authorizes individual nodes to send data.
performance before standard compliance. In EtherCAT and SERCOS III networks, the transfer of summation
frame telegrams follows the masters clock.
One crucial difference of the various Industrial Ethernet systems PROFINET IRT uses synchronized switches to control communication.
compared within this publication lies in how they organize data EtherNet/IP employs CIP Sync to distribute IEEE 1588 compliant
transfer and how they manage to deliver real-time performance. time information throughout the network.
EtherCAT and SERCOS III communicate using a summation frame

Based on TCP/IP Standard Ethernet Modified Ethernet

IEEE 802.3 Media Access

PROFINET
EtherNet/IP EtherCAT
POWERLINK
SERCOS III
PROFINET RT
PROFINET IRT
TCP/UDP/IP

Ethernet Ethernet Modified Ethernet

Methods for
Ethernet Cabling real-time Ethernet
implementation

6
 Systems Roundup:
The 5 Major Contenders
2nd Edition

PROFINET Communication
|
PROFINET (Process Field Network) is differentiated into different
performance classes to address various timing requirements:
PROFINET RT for soft real-time, or no real-time requirements at all, Application
Engineering, Standard
business application
and PROFINET IRT for hard real-time performance. The technology was Real-time cyclical
PROFINET
Real-time acyclical
PROFINET
Standard
PROFINET
integration (FTP, HTTP etc.)

developed by Siemens and the member companies of the PROFIBUS

user organization, PNO. The Ethernet-based successor to PROFIBUS DP,
Middleware DCOM
PROFINET I/O specifies all data transfer between I/O controllers as well
as the parameterization, diagnostics, and layout of a network. Standard UDP Standard TCP

Standard IP

How It Works Standard Ethernet

In order to cover the different performance classes, PROFINET makes

free use of the producer/consumer principle and resorts to various
protocols and services. High-priority payload data sent directly via the
Ethernet protocol travels in Ethernet frames with VLAN prioritization,
whereas diagnostics and configuration data, for instance, is sent using
UDP/IP. That enables the system to achieve cycle times of around Destination Source 802.1q Type = Data CRC
address address 0x8892
10 ms for I/O applications.

Clock-synchronized cycle times below one millisecond, as required

for Motion Control applications, are provided by PROFINET IRT, which Frame ID Process data Status info

implements a time multiplex mode based on specially managed,

hardware-synchronized switches. So-called Dynamic Frame Packing
(DFP) will in the future give users a new PROFINET variant designed to
optimize cycle times making use of the summation frame principle for a
certain set of devices in the network.

7
 Device Profiles I/O Encoders Valves Drives Medical Others

POWERLINK Communication Protocol

Software
CANopen
Application Layer Object Dictionary
| Messaging (SDO and PDO)

POWERLINK Transport
CAN based
Initially developed by B&R, POWERLINK was introduced in 2001. CANopen UDP/IP
Transport
The Ethernet POWERLINK Standardization Group (EPSG), an POWERLINK Driver
independent user organization with a democratic charter, has taken CAN Driver Ethernet Driver
charge of the further development of the technology since 2003.
POWERLINK is a completely patent-free, vendor-independent and Hardware CAN Controller Ethernet Controller
purely software-based communication system that delivers hard
real-time performance. An open source version has also been made Many shared characteristics: the CANopen and POWERLINK OSI model
available free of charge in 2008. POWERLINK integrates the entire
range of CANopen mechanisms and fully complies with the IEEE 802.3 A POWERLINK cycle consists of three periods. During the Start
Ethernet standard, i.e. the protocol provides all standard Ethernet Period, the MN sends a Start of Cycle (SoC) frame to all CNs to
features including cross-traffic and hot plugging capability, and synchronize the devices. Jitter amounts to about 20 nanoseconds.
allows for deploying any network topology of choice. Cyclic isochronous data exchange takes place during the second
period (Cyclic Period). Multiplexing allows for optimized bandwidth
How It Works use in this phase. The third period marks the start of the asynchro-
POWERLINK uses a mixture of timeslot and polling procedures to nous phase, which enables the transfer of large, non-time-critical
achieve isochronous data transfer. In order to ensure co-ordination, data packets. Such data, e.g. user data or TCP/IP frames, is
a PLC or an Industrial PC is designated to be the so-called Managing scattered between the asynchronous phases of several cycles.
Node (MN). This manager enforces the cycle timing that serves to POWERLINK distinguishes between real-time and non-real-time
synchronize all devices and controls cyclical data communication. domains. Since data transfer in the asynchronous period supports
All other devices operate as Controlled Nodes (CN). In the course of standard IP frames, routers separate data safely and transparently
one clock cycle, the MN sends so-called Poll Requests to one CN from the real-time domains. POWERLINK is very well suited to all
after another in a fixed sequence. Every CN replies immediately to this sorts of automation applications including I/O, Motion Control,
request with a Poll Response on which all other nodes can listen in. robotics tasks, PLC-to-PLC communication and visualization.

Cycle Time

PReq PReq PReq PReq MN

SoC SoA
CN1 CN2 CN3 CNn

PRes PRes PRes PRes

Async Data CN
CN1 CN2 CN3 CNn

Isochronous Asynchronous
Phase Phase

SoC = Start of Cycle PReq = Poll Request MN = Managing Node

SoA = Start of Async PRes = Poll Response CN = Controlled Node

8
 Systems Roundup:
The 5 Major Contenders
2nd Edition

EtherNet/IP

Device Profiles CIP Motion Valves I/O Robots Other

EtherNet/IP Communication CIP Application Layer

Application Library
CIP
| Application CIP Data Management Services
Explicit Messages, I/O Messages

CIP Message Routing, Connection Management

Initially released in 2000, EtherNet/IP is an open industrial standard
Encapsulation
developed by Allen-Bradley (Rockwell Automation) and the ODVA Transport
TCP UDP
ControlNet DeviceNet CompoNet
Transport Transport Transport
(Open DeviceNet Vendors Association). The Ethernet Industrial Network IP
EtherNet ControlNet CAN CompoNet
Data Link
Protocol is essentially a port of the CIP application protocol CSMA/CD CTDMA CSMA/NBA Time Slot
EtherNet ControlNet DeviceNet CompoNet
(Common Industrial Protocol), which was already used by ControlNet Physical Physical Layer Phys. Layer Phys. Layer Phys. Layer
and DeviceNet, to the Ethernet data transfer protocol. EtherNet/IP is
particularly well established on the American market and is often
used with Rockwell control systems. EtherNet/IP layer model

How It Works
EtherNet/IP runs on standard Ethernet hardware and uses both acquisition. While explicit messages are embedded into TCP frames,
TCP/IP and UDP/IP for data transfer. Due to the producer/consumer real-time application data is sent via UDP due to the latter protocols
functionality supported by the CIP protocol, EtherNet/IP has various more compact format and smaller overhead. Forming the center of a
communication mechanisms at its disposal, e.g. cyclic polling, time star topology network, switches prevent collisions of data from devices
or event triggers, multicast or simple point-to-point connections. The that are connected using point-to-point connections. EtherNet/IP
CIP application protocol differentiates between implicit I/O messa- typically achieves soft real-time performance with cycle times around
ges and explicit query/reply telegrams for configuration and data 10 milliseconds. CIP Sync and CIP Motion as well as precise node
synchronization via distributed clocks as specified in the IEEE 1588
standard are used to approach cycle times and jitter values low
enough to enable servo motor control.
producer consumer consumer consumer

accept accept

prepare filter filter filter

send receive receive receive

broadcast communication

9
 Frame delay = (total byte count for header + data) x 10 ns
250 ns 115 ns

EtherCAT Communication Master

PHY

| PHY
IOS via LVDS

IOS via LVDS

PHY

EtherCAT (Ethernet for Control Automation Technology) was developed

by Beckhoff Automation. All users of this technology automatically
become members of the EtherCAT Technology Group (ETG).
of these comprises its own header, instruction data for a slave, and
How It Works a working counter. Up to 64 Kbytes configurable address space is
EtherCAT is based on the summation frame method: The EtherCAT available for each slave. Addressing proceeds by auto-increment, i.e.
master transmits an Ethernet frame containing data for all nodes on each slave counts up the 16-bit address field. Slaves can also be
the network. That frame passes through all nodes in sequence. When addressed via distributed station addresses, which are assigned by
it arrives at the last node on a trunk, the frame is turned back again. the master in the start-up phase.
The nodes process the information in the frame as it passes through in
one direction. Each node reads out data addressed to it on the fly, and EtherCAT Process Synchronization
inserts response data back into the frame. In order to support the Every slave connection provides a real-time clock that is synchronized
bandwidth of 100 Mbit/s, special hardware based on ASICs or FPGAs by the master using a technique similar to IEEE 1588. There are slave
is required for fast processing as data passes through. In effect, the devices with and without real-time mechanisms, since these are more
topology of an EtherCAT network always constitutes a logical ring. demanding on the hardware. Based on the real-time clocks, control
Even trunks branching out, which can be hooked up to nodes signals can be synchronized with high precision. In physical terms, the
especially designed for such connections, actually only add a two-way EtherCAT protocol not only runs on Ethernet, but also on LVDS (Low
junction where the summation frame telegram travels up and back Voltage Differential Signaling). This standard is used by Beckhoff as
down the branching line. an internal bus on the terminals. A PC with a standard Ethernet inter-
face is typically used to implement an EtherCAT master. In contrast
Structure of an EtherCAT frame to other protocols such as POWERLINK or PROFINET, EtherCAT solely
All EtherCAT telegrams with instructions for individual nodes are extends to Layers 1 through 3 of the seven-layer OSI model. Hence,
contained within the payload data area of a frame. Each EtherCAT in order to achieve application functionality comparable to the other
frame consists of one header and several EtherCAT commands. Each systems, an extra protocol layer (CoE, EoE) needs to be super-
imposed.

Ethernet HDR FH EH Data WC CRC

EtherCAT principle of operation

10
 Systems Roundup:
The 5 Major Contenders
2nd Edition

Real-time channel Non-real-time channel

Ethernet
SERCOS III Communication I/O profile application

| Motion profile
Generic device profile S III
protocol
A freely available real-time communication standard for digital drive
interfaces, SERCOS III not only specifies the hardware architecture RT channels Safety SVC
(primary/secondary)
channel UDP/TCP
of the physical connections but also a protocol structure and an Cross-communication
extensive range of profile definitions. For SERCOS III, effectively the M/S communication IP
third generation of the Sercos Interface that was originally introduced
Synchr. Ethernet
to the market in 1985, Standard Ethernet according to IEEE 802.3
serves as the data transfer protocol. This communication system is
predominantly used in Motion Control-based automation systems. SERCON 100M/S (FPGA) netX with
+ or SERCOS III
A registered association, sercos International e.V., supports the Ethernet Dual PHY
technologys ongoing development and ensures compliance with
the standard. RT = Real Time
M/S = Master Slave
SVC = Service Channel
S III = SERCOS III
TCP = Transmission Control Protocol
FPGA = Field Progr. Gate Array
Synchr. = Synchronization UDP = User Datagram Protocol PHY = Physical Layer

How It Works
The specific master/slave communications controller for
While specific hardware is categorically needed for the slave, a SERCOS technology is known as SERCON.
software solution is also feasible for the master. The sercos user
organization provides a SERCOS III IP core to support FPGA-based
SERCOS III hardware development. SERCOS III uses a summation of communicating with each other within one communication cycle,
frame method. Network nodes must be deployed in a daisy chain or a with no need to route their data through the master.
closed ring. Data is processed while passing through a device, using
different types of telegrams for different communication types. Due Besides the real-time channel, which uses time slots with reserved
to the full-duplex capability of the Ethernet connection, a daisy chain bandwidths to ensure collision-free data transfer, SERCOS III also pro-
actually constitutes a single ring, whereas a proper ring topology will vides for an optional non-real-time channel. Nodes are synchronized
in effect provide a double ring, allowing for redundant data transfer. on the hardware level, prompted by the first real-time telegram at the
Direct cross-traffic is enabled by the two communication ports on beginning of a communication cycle. The master Synchronization
every node: in a daisy chain as well as a ring network, the real-time Telegram (MST) is embedded into the first telegram for that purpose.
telegrams pass through every node on their way back and forth, Ensuring high precision by keeping synchronization offsets below
i.e. they are processed twice per cycle. Hence, devices are capable 100 nanoseconds, a hardware-based procedure compensates
for runtime delays and variations in it resulting from the Ethernet
hardware. Various network segments may use different cycle clocks
and still achieve fully synchronized operation.
AT: Drive Telegram MDT: Master Data Telegram IP: IP Channel C

AT MDT IP

Master Slave Slave Slave Slave Slave Slave

11
The User Organizations
|
User independence is another key aspect in the overall assessment of POWERLINK EPSG
a system. Any unsettled issues regarding brand rights or patents that The Ethernet POWERLINK Standardization Group (EPSG) was founded
may limit a users own developments are crucial factors to consider in 2003 as an independent organization of companies in the drives
when making the decision for a system. Legal traps that may cause in- and automation sector. The groups goal is the standardization and
convenience later can be avoided by taking a close look at the creators ongoing development of the POWERLINK protocol introduced by B&R
and the user organizations backing the various solutions. in 2001. The EPSG cooperates with standardization organizations
such as CAN in Automation (CiA) or the IEC. The EPSG is a registered
Criteria PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III association established according to Swiss civil law.
RT | IRT
www.ethernet-powerlink.org
SERCOS
Organization PNO EPSG ODVA ETG
International
EtherNet/IP ODVA
ethernet-
www. profibus.com
powerlink.org
odva.org ethercat.org sercos.org ODVA is the union of all DeviceNet and EtherNet/IP users. The organiza-
tion attends to the continual development and further distribution of
these field buses that are predominantly used in the USA and Asia,
PROFINET PI but also in Europe. One key aspect of the organizations activities is the
PROFIBUS & PROFINET International (PI) is the international umbrella development and propagation of the CIP protocol and of other protocols
association for 25 regional PROFIBUS & PROFINET Associations based on it. Users may not only apply the technology but are also
including the user organization PNO, which spells out as PROFIBUS invited to contribute to its ongoing development by joining Special
Nutzerorganisation e. V. It runs an office that manages joint projects Interest Groups (SIG). The ODVA also actively participates in other
and disperses information to members and other interested parties. standardization bodies and industry consortia. The organizations
A certification center for PROFIBUS and PROFINET product approvals bylaws are relatively complex.
is affiliated with that office. Adopted on 24 April, 1996, the www.odva.org
organizations bylaws specify its duties and responsibilities.
Membership is open to all companies, associations, and institutions
that support the interests of PI as device vendors, users, systems
solution providers or operators of PROFIBUS or PROFINET networks.
www.profibus.com

12
 Systems Roundup:
The 5 Major Contenders
2nd Edition

EtherCAT ETG SERCOS III sercos International e. V.

The EtherCAT Technology Group is a forum jointly established by users, sercos International e.V. (SI) is an association entered into the court
OEMs, machine vendors, and other automation suppliers. The groups registry in Frankfurt am Main, Germany. The associations members
purpose is to provide support for and to propagate the benefits of are manufacturer and users of control systems, drives, and other auto-
EtherCAT as an open technology. A certification lab is affiliated with the mation components as well as machine vendors, research institutions,
organizations head office in Nuremberg. All contractual agreements for and other associations. There are subsidiary organizations in North
use of the technology must be made directly with Beckhoff Automation. America and Asia. A certification lab at the University of Stuttgart is
Based in Nuremberg, Germany, the EtherCAT Technology Group is a affiliated with the organizations head office.
nicht eingetragener Verein, i.e. a non-registered club in the sense of www.sercos.org
the German Civil Code.
www.ethercat.org

13
Status, Rights and Licensing Criteria PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III
Which is the legal status of the various user organizations? Who RT | IRT
non-registered
owns the technology? Which legal ties, depending on the licensing Type of association association association
club
association

regime, are binding for developers who use a specific technology?

organization + + + o +
The following pages provide an overview. PNO EPSG ODVA members SERCOS
Liability
+ + + o +

EtherCAT Technology Group: the non-registered club is not a legal entity, but is
effectively a hybrid between an association and a private partnership, for which
legal liabilities remain unclear.

Criteria PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III

RT | IRT

Rights members members members Beckhoff members

owners + + + o +
Brand PNO EPSG ODVA Beckhoff SERCOS
owners + + + o +

In most cases, the rights to a technology rest with the organization responsible
for it. As co-owners, members are therefore entitled to make use of it. If other
persons or companies own the rights to a technology, the prospects for future
legal use of it remain unclear.

14
 Systems Roundup:
The 5 Major Contenders
2nd Edition

Criteria PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III Criteria PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III
RT | IRT RT | IRT
Free source
Funding
obligations
membership
fees
o
membership
fees
o
membership no membership membership
fees
o
fees
+
fees
o
code for
master
- + - o +
Free source
code for
slave
- + + - o
Membership in the ETG is free of charge. A fee is due for memberships in
all other organizations. Annual contributions usually vary with the size of a
corporate member. POWERLINKs and sercos user organizations also allow
PROFINET: The PROFIBUS user organization (PNO) provides source code and
non-members to develop products and put them on the market.
documentations for PROFINET implementations (PROFINET runtime software)
to its members. Clause 1.5 of the license agreement for this software gives
PNO members the right to use five patents.
Criteria PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III
RT | IRT POWERLINK: POWERLINK master and slave code is freely available under a
BSD open source license. The software stack is available on SourceForge.net.
Master and PNO EPSG ODVA Beckhoff SERCOS
EtherNet/IP: Stacks are available for purchase from various service providers.
slave
specification + + + o o
An open source variant has been developed by a university.
EtherCAT: Slave implementations necessarily require an ASIC or an FPGA.
The VHDL or IP code for the FPGA must be purchased from Beckhoff; no source
While communication mechanisms are specified for SERCOS III and EtherCAT, code for it is available. The ETG provides sample source code for the master
the inner workings of a slave remain undisclosed. Users must resort to an ASIC side. Since the patent holder has not agreed to an open source licensing
or an FPGA. FPGA IP code from Beckhoff is available for EtherCAT as object regime, that source code does not qualify as open source.*
code, the source code is not disclosed.
SERCOS III: Software master is provided free of charge under an LGPL
license. ASICs or FPGA code must be purchased for the slave.

* Source: Open Source Automation Development Lab (www.osadl.org)

15
Investment Viability
|
Openness as one issue with a bearing on the long-term viability of Flexible Cabling Topology
investments in a system has already been mentioned. In addition, EtherCAT and SERCOS III networks always constitute a logical ring.
a number of technical and strategic considerations also play crucial That ring can be physically closed at the master, or, in the case of a
roles in making a safe investment decision for the long term. daisy chain, closed internally at the last node in the physical line.
EtherCAT does provide for trunks to branch out via special junctions,
Compatibility to Existing Application Profiles but the entire frame travels up and back down such lateral network
lines, i.e. the network as a whole still represents a logical ring.
Criteria PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III
RT | IRT
Criteria PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III
PROFIBUS CANopen DeviceNet CANopen SERCOS II RT | IRT

Downward Tree
compatibility topology + + + o o
+ + + + +
Star
topology + + + o o

Ring
EMC Susceptibility/Transmission Reliability topology + + + + +
Summation frame protocols are more susceptible to interference
Daisy-chain
than single frame protocols. If a frame is destroyed, summation frame topology + + + + +
protocols always lose an entire cycle.

Criteria PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III

RT | IRT
High Availability
EMC
susceptibility + + + o o Only in the case of POWERLINK have master and cable redundancy been
included in the specifications, and have been implemented in actual
projects. For PROFINET and EtherNet/IP, application implementations
Since it uses two telegrams, in this comparison SERCOS III actually provides
50 % better performance than EtherCAT. based on special switches are feasible.

Criteria PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III

Electrical Contact Points RT | IRT
One special EtherCAT feature is the option to route all communication
Ring
through the internal I/O terminal bus as well. However, the superior per- redundancy o + o + +
formance often cited in connection with this feature is offset by the safety Master
risk due to increased susceptibility for interference (contacts and EMC). and cable
redundancy
o + o o -
Criteria PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III
RT | IRT
Electrical
contact
points
+ + + o +

16
 Systems Roundup:
The 5 Major Contenders
2nd Edition

Hot Plugging Capability Support of International Standards

POWERLINK, EtherNet/IP, and PROFINET give users hot plugging The IEC 61158 international standard standardizes protocols (called
capability. Some restrictions apply for SERCOS III and EtherCAT due to Types) for use in industrial control systems. IEC 61784-2 standardizes
the compulsory ring topology. In a physical ring topology, SERCOS III communication profile families (called CPF). GB standards are National
does allow for taking a single node off a network. In this event, the two Chinese Standards written and issued by the Standardization Authority in
neighboring nodes close the TX and RX lines. Nodes can then be China (SAC). They are valid across all industries and nationwide. GB/Z
reached from either side of the master. EtherCAT provides some hot stands for national technical guidelines. These are primarily informative
plugging capability: In the EtherCAT Slave Controller, open ports are in nature and in no way binding. The highest authorized standardization
automatically closed if no link is detected. EtherCATs distributed level for communication technologies is GB/T. As a Chinese recommended
clocks, however, requires re-synchronization, which may affect certain industrial standard, GB/T must meet several requirements: It must be
applications. fully open technology, widely used and standard technology in the world.
It must not be subject to any country or company.
Criteria PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III
RT | IRT
PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III
Hot plugging + + + o o
IEC 61158 Type 10 Type 13 Type 2 Type 12 Type 19

For technologies based on a logical ring (EtherCAT and SERCOS III),

the limitations of the network topology also limit hot plugging capability. IEC 61784-2 CPF 3 CPF 13 CPF 2 CPF 12 CPF 16
Hot pluggable modules can only be connected to one end of a daisy chain
(SERCOS III) and distributed clocks require re-synchronization after node GB National GB/Z GB/T GB/Z
failure, which may impose restrictions on applications. Chinese
Standard 25105-2010 27960-2011 26157-2010

Gigabit Readiness
As EtherNet/IP and POWERLINK are entirely software-based techno-
logies, these protocols can also be used with Gigabit hardware. Products on the Market
EtherCAT can be scaled to Gigabit but requires an ASIC redesign. IRT products based on ERTEC technology are generally available on
PROFINET IRT also requires some redesign of the hardware, which con- the market. However, the introduction of the DFP feature and the new
cerns switches in particular. FPGA solutions can be ported to Gigabit. generation of ASICs in conjunction with it (e.g. the Tiger Chip supplied
by Phoenix) has raised doubts concerning the future compatibility of
Criteria PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III
RT | IRT current IRT solutions.

Gigabit
readiness + - + + - o
Criteria PROFINET
RT | IRT
POWERLINK EtherNet/IP EtherCAT SERCOS III

Products on
the market + o + + + +
sercos International has stated that their IP core is basically Gigabit-ready.

17
 Performance
|
Theoretically Achievable Cycle Time
The performance of the systems has been the subject of intense
debate, which has focused on the theoretical cycle times achievable
by Industrial Ethernet systems. The briefest possible cycle time in
theory is calculated as follows:

Number of bytes: 7 1 6 6 2 38 1500 4

10100101..011 Destination Source

Preamble
Starting frame delimiter
Destination MAC address
Source MAC address
Length (if <1501*) / EtherType (if >1535*)
Payload data
Check sequence (cycling redundancy check)
*Decimal values

Number of bytes Number of bits Duration at 100 Mbit/s

Minimum length 26 + 38 = 64 512 5.1 s
Maximum length 26 + 1500 = 1526 12208 122 s

Source: frame makeup as defined in IEEE 802.3

(The interframe gap of .96 s must be added on top of the 5.1 s cited above.)

18
 Systems Roundup:
The 5 Major Contenders
2nd Edition

Hence, if a master sends out a frame addressed to itself that does Communication Architecture of the Systems
not pass through any other nodes, that frame will be available to the
Criteria PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III
master again after 122 microseconds have elapsed (in the case of a RT | IRT
single, maximum-length Ethernet frame). Supports
central
control
+ + + + +
In theory, it would be possible to process parts of a frame as soon as Supports
they are received. However, the CRC bytes that confirm the validity of decentral
control
+ + + - o
the data received are last to arrive at the end of a frame. This scenario
does not factor in delays affected by PHYs, cables, and Ethernet ports,
times for internal data transfer in the master, etc. Moreover, once a
signal leaves the master, the time it takes to travel along network lines Direct Cross-Traffic
(5 ns/m) and the processing time inside a slave have to be taken into Direct cross-traffic provides crucial benefits particularly in case of very
account as well. demanding real-time requirements: for fast drive controllers, axes can
be synchronized easily and with extreme precision, since all position
Prospective extensions of a system and possible future requirements values can be distributed directly without having to go through a
need to be carefully considered for selecting either a centralized or master. That results in lower network load and also ensures that data
a decentralized architecture. One advantage of the decentralized (e.g. actual angle positions of axes) is available to all relevant nodes
processing of various control loops is that it allows for adding within the current cycle. If data needs to pass through a master first,
nodes without any noticeable effect on the basic cycle time, i.e. it is not only delayed by one cycle, but overall data traffic on the
no fundamental changes to the overall concept must be made. network is increased as well.
Moreover, additional functionality such as condition monitoring or
Criteria PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III
integrated safety technology will have less impact on the control RT | IRT
concept than in central architectures, which depend significantly
on a low volume of data.
Direct
cross-traffic + + + - +

In order to select a solution that is viable for future use as well,

wherever possible preference should be given to a decentralized With POWERLINK and SERCOS III, direct cross-traffic is a feature even for
modules that only have slave functionality, while EtherNet/IP requires a module
handling of control loops for cycle times below 500 microseconds, with scanner functionality.
especially in drive applications.

19
Heavy Data Traffic Actual Cycle Time
In applications involving a large volume of process data, the time In solutions using the summation frame method, data must pass
required for passing through the nodes greatly impacts the overall twice through each controller. If a signal has to go through many
cycle time. Data prioritization, on the other hand, enables lower cycle nodes, total transfer time will rise considerably as it makes its way.
times. Systems that support prioritization mechanisms allow for Raw performance data cited by the organizations supporting such
reading high-priority data once every cycle and polling for data with solutions has to be adjusted to account for this effect. Another
a lower priority only every n-th cycle. aspect to consider is that performance depends on implementation
specifics, e.g. task classes, in the actual control systems used for
Criteria PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III
RT | IRT an application.

Prioritization + + + o + Criteria PROFINET

RT | IRT
POWERLINK EtherNet/IP EtherCAT SERCOS III

Performance o + + o + +
For POWERLINK, EtherNet/IP, and PROFINET, variable cycle times have been
firmly established in the protocols specifications. SERCOS III has only recently
added this feature. For EtherCAT, solutions for this requirement can be
implemented as part of a specific application.
Jitter
It is crucial for control quality on a network to ensure minimal jitter
Network Load for Safety Communication (clock deviation) and to determine signal delays very precisely.
Safety over Ethernet is based on a cyclic exchange of protected data To this end, network nodes must be synchronized as precisely as
between safety nodes (emergency stop switches, drives with Safety possible. Competing Ethernet variants employ different mechanisms
controllers). The safeguard procedures in this process involve data to achieve that goal. While EtherCAT uses the principle of distributed
duplication and wrapping data in safe containers. This increases clocks solved by a proprietary algorithm within the ESC (EtherCAT
data rates on the network. Solutions using the summation frame Slave Controller), synchronization is accomplished via a simple sync
method will see the frame count go up, whereas the single frame signal (SoC) in POWERLINK networks.
method will increase the volume of data in each of the frames that
Criteria PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III
are due to be sent anyway. All in all, the theoretically superior RT | IRT
performance of the summation frame method is neutralized.
Jitter o + + o + +

EtherCAT, POWERLINK, and SERCOS III give users a system with almost no jitter
(< 100 ns) at all times. On EtherNet/IP networks, jitter can be considerably
reduced with special IEEE 1588 extensions in all components. Reduced jitter
can also be achieved in PROFINET IRT applications.

20
 Systems Roundup:
The 5 Major Contenders
2nd Edition

Performance Contest In practice, POWERLINK is faster than EtherCAT in most applications.

In practice, comparing system performance proves to be a difficult EtherCAT is optimized for applications with only very low network traffic
endeavor due to the specific characteristics of the various systems: volume. In systems with a heavier data load, there is a disproportionate
EtherNet/IP and PROFINET RT are excluded from the start because rise in cycle times in EtherCAT environments. Where decentralized
these systems are only suitable for soft real-time requirements. architectures (e.g. for decentralized Motion control) are implemented,
PROFINET IRT poses problems due to the indispensable switches, EtherCAT suffers greatly from the lack of direct cross-traffic (in both
which lead to a different application architecture that makes a direct directions), which sharply reduces the performance that can theoretically
comparison of measurements complicated. The values below were be achieved. A direct I/O integration of EtherCAT also results in lower
determined based on published calculation schemes. sampling rates (I/O system), since the time the signal takes to pass
through the I/O has a direct impact on the cycle time within reach.
Test scenarios were For POWERLINK and SERCOS III, there are no such effects.
1. a small machine comprising a master and 33 I/O modules The publication by Prytz (2008)1 was used as a reference for the calcu-
(64 analog and 136 digital channels); lations concerning EtherCAT. Delays for signals passing through the
2. an I/O system with a master and twelve Ethernet slaves with EtherCAT ASIC were verified again by measurements. For POWERLINK,
33 modules each (in total, 2000 digital and 500 analog channels applications with actual products were set up for practical measure-
were taken into account in this application); ments, leaving no room for doubt and reconfirming the cited figures.
3. a Motion Control network with 24 axes and one I/O station
with 110 digital and 30 analog I/Os. No tests and calculations were conducted for SERCOS III. However,
SERCOS III can be expected to provide a performance level similar to
POWERLINK, making it faster than EtherCAT in many applications.

Better

Motion decentralized: EtherCAT 542.88

Motion decentralized: POWERLINK 269.98

I/O System: EtherCAT 363.48

I/O System: POWERLINK 325.25

Motion centralized: EtherCAT 271.44

1 G. Prytz, EFTA Conference 2008, A Performance
analysis of EtherCAT and PROFINET IRT. Referenced
Motion centralized: POWERLINK 269.98 on the EtherCAT Technology Groups website, www.
ethercat.org. Last accessed: 14 September, 2011.
Small I/O: EtherCAT 53.4

Small I/O: POWERLINK 81.21

0 100 200 300 400 500
Cycle time [s]

21
Implementation
|
Implementation costs include development expenses, license costs,
and hardware costs. Code availability (program or VHDL in case of
a hardware implementation) must be taken into consideration here
as well.

Master Implementation
Master designs PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III
RT | IRT

Master access + o +

openPOWERLINK common SERCOS III master API

no open source master available no open source master available patent-protected 1
(open source) (open source)

Implementation costs o + o + o

requires special
pricey software typically with
hardware with runs on standard hardware pricey software stack runs on standard hardware
stack coprocessor support
coprocessor

1 No open source master, only sample code that does not warrant applicability.

All protocols allow for a software implementation of the master

on a standard Ethernet chip.

22
 Systems Roundup:
The 5 Major Contenders
2nd Edition

Costs for Potentially Required

Network Components
External devices = external switches or hubs
Internal multiports = ports that are directly integrated into the devices, primarily for daisy chain and ring topologies

Network components PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III

costs RT | IRT

External devices + o + o o o

designated for future use of

special switch, managed switch with complex
standard special network external infrastructure devices,
IRT support standard hubs or switches functionality required (IGMP
switch components required 1 but no such use at the time
required snooping, port mirroring, etc.)
of writing

Internal multiports o o + o + +

integrated Siemens ASIC standard hub integrated switch,

Beckhoff ASIC required 2 FPGA-based technology
switch required very complex

1 With EtherCAT, special network components are required for star or tree topologies.

2 Beckhoff ET1100.

23
Slave Implementation
For EtherCAT, SERCOS III, and PROFINET IRT, bus protocol implemen- attractive technology for Industrial Ethernet solutions primarily due
tations into a slave require hardware solutions (ASICs or FPGAs). For to these lower expenses, their high performance, and multi-protocol
POWERLINK, EtherNet/IP, and PROFINET RT, microcontroller-based capability, but also because they allow for using pre-assembled
software solutions are also feasible. Expenses for software solutions components to integrate Layer 2 functionality (hubs, switches).
comprise license costs for the stack, possibly complemented by extra However, users need to be aware that the complexity of a protocol has
costs for more powerful and therefore more expensive controllers. an impact on the volume of code and, by extension, the required
For hardware solutions, users may opt for either FPGA- or ASIC-based number of logic blocks. L2 functionality can also have a substantial
communication interfaces. In principle, FPGAs may also be used for bearing on this number. Switches need more blocks than hubs, and
software solutions. complex managed switches require an excessive number of logic
blocks. POWERLINK is the least complicated real-time Ethernet
An FPGA (Field-Programmable Gate Array) is an integrated circuit that solution. Moreover, since POWERLINK only resorts to hubs in its
hardware developers can configure themselves. It consists of pro- network layout, this protocol requires only a small number of logic
grammable logic components, so-called logic blocks, and a hierarchy blocks, and is suitable for small FPGAs.
for the reconfigurable component circuitry. All logic functions ASICs
are able to execute can be implemented with FPGAs as well. Functionality EtherCAT and SERCOS III, on the other hand, are more demanding
can be customized before commissioning. The one-off development and therefore need many more logic blocks.
costs for FPGAs are lower than that of ASICs. FPGAs comprise an

Node Connection Costs in Different Real-time

Ethernet Environments
The connection costs per node below refer to the running expenses for
the hardware. Potentially owed license costs for software stacks etc.
have not been taken into account.

24
 Systems Roundup:
The 5 Major Contenders
2nd Edition

25 $
Minimum Hardware Costs

20 $

Operating Costs
15 $ Operating costs largely consist of maintenance and network
administration expenses. Some technologies such as EtherNet/IP
10 $ with CIP Sync and PROFINET IRT are highly complex and may therefore
entail considerable network administration costs. Moreover, any use
5$
of managed switches requires network expertise. In many cases,
22.2 $ 9.2 $ 15.2 $ 11.0 $ 15.2 $ a network engineer will be needed on location for commissioning
PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III and maintenance.
RT | IRT

The synchronization technology used by a solution plays a key role

The figures in this diagram have been derived from feedback from various
for real-time communication. POWERLINK and SERCOS III ensure
manufacturers with implementation experience covering different Industrial
Ethernet solutions. Several figures have also been quoted by manufacturers synchronization via a master-managed mechanism that is very precise
in automation industry magazines. Costs for the PHY (2 1.1 $) have been and very rarely disturbed by faults. PROFINET IRT and EtherNet/IP
factored in equal measure for all protocols. Connectors are excluded.
Cost estimates per node are made for an annual volume of 1000 units. with CIP Sync depend on an IEEE 1588 compliant synchronization
Profinet: The calculation reflects a solution with an ERTEC200 ASIC. Future
mechanism. That results in significantly more complex network
implementations may also use devices equipped with a TPS1 chip developed administration, especially if devices must be isolated because they
by Phoenix Contact. In that case, costs are likely to drop to a level comparable trigger synchronization faults due to hardware or software failures.
to EtherCAT. POWERLINKs price level will not be met.
Powerlink: The calculation applies to an FPGA-based solution. RAM and flash
memory costs have been taken into account. Functions such as hot plugging i.e. the option to swap devices on a
live network can also help greatly to bring down maintenance costs:
EtherNet/IP: The EtherNet/IP figure applies to a typical FPGA solution.
replacement devices can be updated and configured without any
EtherCAT: The calculation is based on the least expensive EtherCAT ASIC
solution with two Ethernet ports (ET1100). EtherCAT solutions for FPGAs are impairment to the real-time function of the system as a whole.
much more costly; the difference is most striking for synchronous solutions
with real-time clocks. Costs PROFINET POWERLINK EtherNet/IP EtherCAT SERCOS III
RT | IRT
Sercos III: The SERCOS III figure applies to a typical FPGA solution.
Purchase
costs o - + o + o

Operating
costs
o + o + +

25
Safety Functionality
|
Requirements toward safety in production environments have become The new standards led to the necessity for new machinery to undergo
considerably more demanding over the past decade. Introduction of strict certification procedures and to elevated performance require-
the 2006/42/EC Machinery Directive by the European Union made ments for the safety components used. Supported by a multitude of
machine and plant manufacturers focus their attention on this issue. new and innovative safety products, they also facilitated changes in
They are required to design comprehensive solutions to ensure the approach toward the conceptual design of safety solutions. No
protection of workers against injuries and the machinery itself against longer is an emergency stop immediately halting all parts of a machine
damage while maintaining high levels of productivity. the only safe reaction to violations of the machines safe boundaries.
Smart safe reactions such as continued operation at a safe limited
speed can in many cases deliver the required level of protection while
providing better productivity by reducing the time to resuming full
speed. In many instances, it enables a more direct interaction
between worker and machine, particularly in teaching and adjustment
scenarios.

Network Integrated rather than Hard Wired

Traditionally, safety equipment used to be hard wired with dedicated
switching circuitry, often logic cast in pure hardware. Although with
some effort it is theoretically possible to cover many cases using
this method, more and more machine manufacturers have been
recognizing the benefits of integrated safety. It is based on safe
programmable control hardware and I/O modules using the existing
field bus to exchange safety-related data.

26
 Systems Roundup:
The 5 Major Contenders
2nd Edition

At first glance, the older methods may appear less costly. Due to the How It Works
lower purchasing costs of their hardware components, this may in Safety applications based on certified software are programmed
many cases be true, but not if safety solutions are viewed in their using function blocks such as counters, timers or speed monitors.
entirety. Wherever the complexity of such systems goes beyond a Running on dedicated safety controllers, this replaces the traditionally
single emergency stop button, network-integrated safety systems have hard-wired safety circuitry. Implementation of the safety application in
become the preferred choice. They lower the number of components software reduces the number of both safety components and standard
as well as required cabling and provide more flexibility of safe logic I/O modules. Along with the replacement of discrete cabling by safety
design by replacing hard wiring with configuration and parameter data transfer via the existing network connections, this greatly
setting. Also, error diagnostics are greatly simplified. Combined with minimizes both costs and complexity of safety installations. Due to
centralized data storage, this results in faster recovery. Maximum the use of existing network connections, varying machine layouts
availability of plants and machines is provided by network-integrated and options do not require dedicated safety connections. This also
safety technology through: improves flexibility and freedom of safety application design as well as
modifications of existing plants and machines during their lifecycle.
Safety sensors directly attached to the network Also, diagnostic signals can be transferred without any additional
Direct read-out of component information hardware. All in all, using integrated safety speeds up engineering and
Simplified maintenance due to automated component substantially shortens time to market.
parameter setting across the network
Safer operating mode switching due to parameter setting
during runtime
Decreased response time, as latency induced by relays is eliminated
Modular design supported by network structure and safe software
Increased availability as a result of comprehensive diagnosis
Reduction of component count and wiring
Greater variety of safety functions
(safe operating stop, safely limited speed)

27
The Black Channel Principle
|
Safety Field Networks Safety Data Transport via Standard Bus
Safety-oriented field buses simplify placement of components within or Network Lines
a plant or machine. In most cases, two cables, one for power and one The Black Channel Principle allows transmission of failsafe and
for communication, are all that is required. Sensors can be attached standard data via the same network or bus line. Independent of
directly to the safety network. They do not require separate cables for the regular data transport mechanism used on that line, safety
the return of diagnostic signals. This results in a reduction of required components can transmit data using an isolated safe protocol
hardware components. tunneling the underlying network channel. As safe fieldbuses are pure
application protocols without physical characteristics of their own,
Using the Black Channel Principle, safety-relevant data as well as available bandwidth and cycle times depend on the data transport
diagnostic information is exchanged via the existing network protocol used. Possible transmission errors are known and listed in
connections. This enables faster responses. Sensors receive their the relevant standards IEC 61784-3 and IEC 61508. Their prevention
parameter setting and configuration via the network. This allows the needs to be implemented as a crucial part of the safety data trans-
download of optimized parameter settings to the sensors in case of mission protocols. The required quality of transmission error
changes to the mode of operation and also eliminates the need to detection depends on the safety level that needs to be achieved.
adjust parameters at the device itself in case of component replace-
ment after failure. All this adds to a maximized productivity and
reduced down time.

Safety Standard Safety Standard

Application Application Application Application

Safety layer Safety layer Safety

layer

Communication Protocol Communication Protocol

Black Channel

Industrial Ethernet, Fieldbus, Backplanes...

Black-Channel mechanism

28
 Systems Roundup:
The 5 Major Contenders
2nd Edition

How the Safety Systems Work

|
CIP Safety
The CIP Safety protocol was specified for safety data transmission For CRC (cyclic redundancy check) calculation, the CIP Safety protocol
via EtherNet/IP or DeviceNet. Using the pre-existing CIP (Common makes use of five different formats, ranging from 8 bit to 32 bit CRC.
Industrial Protocol) Services as its foundation, the CIP Safety protocol These depend on whether data size is one or two Bytes or between
makes use of the producer-consumer mechanism for the exchange of three and 254 bytes and of the data range to be covered by checksum
data between safe nodes. In this context, a consumer is designated calculation. A Unique Node Identifier (UNID) is used for the unique
originator and a producer is called target. Safe time synchronization identification of the safe nodes. It is a combination of a network ID and
between producers and consumers relies on chronological monitoring. the node address, which is equivalent to the MAC address. It can be
Synchronicity between all nodes throughout the network provided, the set either manually using DIP switches or via software configuration.
time of origination of safe messages can be determined using a time During the ramp-up phase, the originator checks the presence of the
stamp. These methods guarantee that processed data is still up to configured UNIDs in the network. Further parameters such as timeout
date. For the transfer of the safe data, Safety Validator Objects are delays, ping intervals or the maximum number of nodes are configured
used. They organize and guarantee the integrity of messages in a CIP using a Safety Configuration Tool (SNCT).
Safety network. These objects also constitute the bridge between the
safety communication and the field bus or network used. For data
transmission, the protocol provides single transfer or multicast
connections. Their use depends on the capability of the channel used
to support either of these connections.

Pneumatic AC Drives Semi Devices Other Profiles Safety Other

Valves I/O Block Safety Profiles

Object Library Safety-Specific CIP- Common

Object Library
Safety Industrial
Protocol
Data Management Services Safety Layer (CIP)
Explicit Messages, I/O Messages
Connection Management, Routing

TCP UDP
DeviceNet ControlNet
Transport Transport
Internet Protocol (IP)

Ethernet CAN ControlNet Network

CSMA/CD CSMA/NBA CTDMA Adaptions
of CIP

Ethernet DeviceNet ControlNet

Physical Layer Physical Layer Physical Layer

EtherNet/IP DeviceNet ControlNet

29
PROFIsafe
PROFIsafe uses the Master-Slave mechanism for transmission of Although the addresses (Unique Codename) are automatically passed
safety telegrams. The master, typically called the F-Host, cyclically on to the F-Devices, the target addresses need to be adjusted directly
exchanges safety-relevant data with all its configured slaves called at the device via DIP switches. The F-Devices receive their configuration
F-Devices. Each F-Device has an F-Driver organizing the co-ordination through transfer of the F-Parameter via GSD (General Station
of safe messages called Safety PDUs (Protocol Data Unit) between Description) and of the I-Parameter (individual F-Device Parameter).
the F-Host and F-Slave. CRC calculation of the PDUs depends on the These parameters are managed within the iPar server, from where
message size to be transferred, the distinction being between slim they can be transferred to a PROFIsafe Device using standardized
PDU up to 12 bytes and long PDU up to 123 bytes. CRC 24 is used interfaces. Usually, the iPar Server comes integrated in a CPD-Tool
for slim PDU calculation, while for long PDUs, CRC 32 is used. As a (collaborative product design) engineering tool. For product designers,
means for message recipients to determine whether telegrams arrive this means that for the ability to completely configure an F-Device,
in the right sequence, PROFIsafe uses consecutive numbers for the a GSD file needs to be created and an interface to the CPD-Tool must
safety telegrams. Additionally, monitoring of the tolerance time be provided for each product.
(F-Watchdog Time) that is reset upon receipt of a telegram ensures
that always the currently valid telegrams are read. The so-called
F-Parameters (PROFIsafe parameters) supply a unique identifier
between F-Host and F-Device.

User Program F-Device Technology

(Logic Operations) iParameter (e.g. Laser Scanner)

Services Services

F-Host Diver Instance F-Parameter F-Device Driver

State Machine State Machine

PROFIsafe Message

CRC Control Byte Output Data

Input Data Status Byte CRC

30
 Systems Roundup:
The 5 Major Contenders
2nd Edition

openSAFETY
openSAFETY was designed aiming at transmission of safety relevant The openSAFETY Object Dictionary (SOD) manages all parameters,
data over any field bus or network. It can be used on all field buses, which are then transferred to the safety nodes using Safety Service
Ethernet-based or not. Data Objects (SSDO). Upon completion of Node configuration and
the booting phase, the cyclic data transfer between producer and
For the transmission of safety data, the producer-consumer model Consumer commences. For the transfer of safety-critical process data,
is used. The advantage of this model is that all consumers in an this uses Safety Process Data Objects (SPDO). The openSAFETY frame
openSAFETY network can receive and subsequently process the consists of two sub frames. It can transport a maximum of 254 bytes
messages sent by the producer. Each openSAFETY node has a unique of safety data, using CRC 8 for payloads from 1 to 8 bytes and CRC 16
UDID (openSAFETY Unique Device Identification) number. This is a for payloads from 9 to 254 bytes.
combination of the MAC address and the manufacturers device
number. During the booting process, the Safety Network Management With openSAFETY, very large networks can be created. For each
(SNMT) checks the device type and the UDID, so that it automatically openSAFETY Domain (SD), up to 1023 safe nodes can be connected.
detects replaced devices. In such a case, the required parameters are As they are addressed by the SCM, no additional hardware switches
automatically transferred to the Safety Nodes (SN). The SNMT can are required. The maximum total configuration of an openSAFETY
optionally be integrated with the Safety Configuration Manager (SCM). network has 1023 openSAFETY Domains with a total of more than a
In analogy to other communication protocols, the SCM can be viewed Million safe nodes. Communication between the individual Domains
as an openSAFETY master using services to manage the network. is performed by the openSAFETY Domain Gateway (SDG).

Safety Related Application

openSAFETY

Industrial Ethernet
TCP/IP

IO-Link

RS485
LVDS
USB
CAN

...

POWERLINK PROFINET EtherNet/IP Modbus/TCP SERCOS III

31
FSoE
Fail Safe over EtherCAT (FSoE) is a transmission scheme for safety the master and the slave generate a Sequence Number ranging
data over EtherCAT using an FSoE master and FSoE slaves. In each from 0 to 65535 for every message. This ensures that only currently
FSoE cycle, the master sends its Safety PDU (Protocol Data Unit) to valid messages are processed. Addressing of the individual devices
the slave, concurrently starting a watchdog timer. The slave verifies requires designation of unique Node numbers by hardware setting
and calculates the data received prior to returning it to the master. In using DIP switches. Each FSoE master includes an FSoE master
this case, the slave also starts a watchdog timer. The master receives Handler, which communicates with a slave through an FSoE slave
and processes the data as described for the slave, stopping the Handler. Optionally, an additional FSoE slave handler that can be
watchdog timer. Only when this cycle is completed, the master implemented in the master allows communication between different
generates a new Safety PDU. Due to this mechanism, safe communi- masters within a network. For safeguarding the PDUs to transfer, for
cation always depends on the hardware and topology used. every 2 bytes of Safety Data, a CRC 16 is used. This implies that for a
10 Byte transfer, a CRC 16 is applied five times.
The address relation between a master and a slave is called FSoE-
Connection. It is characterized by a unique Connection ID. The 16-bit Parameter setting as such is not specified. The parameterization
Connection-ID is transferred by the master to the individual slaves. process needs to be part of the user-programmed application
Users need to take measures to ensure providing each slave with a software. While the FSoE specification does detail the required
unique ID. For correct identification of the ramp-up sequence, both parameters, users need to ensure for the individual FSoE slaves
to receive their correct parameters.

Safety-over-EtherCAT Software Architecture

Safety Application
Application

Safety Objects Safety Data

Safety Management

Application Layer (AL)

EtherCAT Data Link Layer (DL)

EtherCAT Physical Layer

32
 Systems Roundup:
The 5 Major Contenders
2nd Edition

Integrated Safety System Comparison

|
Certifications Technology
Criteria CIP Safety PROFIsafe openSAFETY FSoE Criteria CIP Safety PROFIsafe openSAFETY FSoE

Black- Payload Data

Channel
based
+ + + + duplication
support
+ +
Multicast
IEC 61784-3 + + + + messaging
support
+ +
Certification TV Rheinland TV Sd TV Sd Safety device
body IFA IFA TV Rheinland
TV Sd
configuration + o + o

Safe motion
control
o + + +
Generally speaking, all the various integrated safety technologies equally
fulfill safety requirements. They are all based on the Black Channel principle,
listed in the IEC 61784-3 standard and certified up to SIL 3. Hidden behind
the raw safety aspects, however, are relevant criteria that define whether a In the design of safety devices, technology considerations have great signifi-
technology will be adopted by component manufacturers or end customers. cance. Depending on the complexity of Safety frames, their composition can
The distinguishing differences are the ease of integration of the technologies require undesired extra implementation efforts.
in the application serving the problem-solving purpose at hand. Support of multicast messaging helps achieve fast response times. These in
turn can be influential for the overall plant or machine design, for instance
reducing machine footprints and required floor space.
After maintenance or device replacement, safety slaves should automatically
Criteria CIP Safety PROFIsafe openSAFETY FSoE be configured by the safety master. Configuration interfaces need to be spe-
cified and unique so that devices can be configured by different masters. For
SIL3 certified
PROFIsafe, the iPar-Server has been developed to cover this requirement. Its
(IEC 61508) + + + + interoperability status in the market is unclear, because in the past, configu-
ration data came from the manufacturer of the master used rather than from
Ready within the system.
for SIL4
o o + o In April 2012, the sercos International (SI) user organization announced the
development of a Safe Motion Profile for CIP Safety on SERCOS III. By the time
of publication of this brochure, however, a Safe Motion Profile based on CIP
Safety does not seem to be available.
openSAFETY technology is certified up to SIL3. Though it has not been qualified FsOE offers a safe parameterization channel to transfer safe encapsulated
yet, the core principle of this technology, including probability of failure on data to the safe application, but addressing scheme of the safe application
demand (PFD), is ready for SIL4. parameters does not exist.

33
Device Implementation Integration
Criteria CIP Safety PROFIsafe openSAFETY FSoE Criteria CIP Safety PROFIsafe openSAFETY FSoE

Legal
limitation + + Stack
compatibility
o o +
Investment
security
o o + o Performance o o + o

Time-To-
Market + + + o Addressing + +
Safe Reaction
Imple-
mentation o + + Time
o o + o

Current
market share
o + o To ensure compatibility between safety products from different manufacturers,
compatibility of all stacks on the market is essential.
certified
stack available + + +
Open source strategy of openSAFETY guarantees stack compatibility.
In a safety network, all nodes must have unique IDs. To avoid parameter setting
For device manufacturers, independence and implementation costs are the
errors, addressing should be automated. The PROFIsafe and FSoE protocols,
most significant considerations. For implementation, all costs for license fees,
however, require manual address setting for each safety device using DIP
software stack, conformance test and certification were taken into account in
switches. This makes installation of safety devices in the control cabinet more
the comparison. So was the complexity of each technology and its impact on
complicated. Human error, particularly in maintenance scenarios, could result
the required resources and costs for implementation.
in faulty parameter setting. It is also very difficult to create modular machine
ProfiSAFE and FSoE are limited to the protocols of their user organizations.
concepts using hardware switches, as this form of addressing is always rigid
This could lead to the requirement to implement several safety protocols if
and components cannot be configured automatically.
equipment using different automation systems and fieldbuses is combined.
CIP Safety requires the implementation of a dedicated CIP Abstraction Layer Following the producer consumer principle, openSAFETY supports direct cross
within the Black Channel, thus increasing engineering efforts. Acccording to communication, which results in singularly fast reactions. Routing all safety
our researches a FSoE slave is under development, but currently there is no messages via the master, as in PROFIsafe and FSoE, extends cycle times.
certified FSoE slave stack available on the market. This may pose a potential Consequently, valuable time for safe reactions is lost.
risk for safety technology implementations on the device level.
CIP Safety requires originator functions to support cross-traffic, thus a cross-
Criteria CIP Safety PROFIsafe openSAFETY FSoE
traffic between slaves (targets) is not possible.

PROFINET
EtherCAT
Supported EtherNet/IP
Industrial EhterNet/IP PROFINET Modbus EtherCAT
Ethernet SERCOSIII POWERLINK
protocols PROFINET
SERCOSIII

Open-Source
Implementation
available
+

The openSAFETY stack is currently the only open source software for safety
communication. Technically as well as from a legal point of view, openSAFETY is
entirely technology-independent.

34
 Systems Roundup:
The 5 Major Contenders
2nd Edition

Performance
Since Safety protocols are application protocols, a safety networks the square of the fault response time and negative acceleration,
performance depends on the underlying data transfer protocol. The quadrupling the signal transfer time will result in a 16-fold extension
base protocol selection determines the available bandwidth and the of the emergency stopping distance.
cycle times, but also functional features such as hot plugging
capability or data communication via cross-traffic.
Criteria CIP Safety PROFIsafe openSAFETY FSoE
Cross-traffic plays a crucial role in the performance of safety-oriented
systems. In networks supporting cross-traffic, safety nodes can
CRC range 8-32 bits 24-32 bits 8-16 bits 16 bits
transmit signals directly to each other without routing them through
a master. This provides for optimized reaction times in hazardous Required CRC
situations. On a network not supporting cross-traffic, safety nodes computations for 2 1 2 10
20 bytes net data
send their signals to a fieldbus master node, which relays it to the
Number of
networks safety master for acknowledgement. It is then handed back different CRC
5 2 2 1

to the fieldbus master node which forwards it to the receiving safety

node. Compared to direct data transfer via cross-traffic, this process
The required number of different checksums increases implementation
causes four times the signal delay valuable reaction time elapses.
complexity, resulting in higher development costs. Additionally, computation of
Since the emergency stopping distance of an axis increases with multiple CRCs may result in significantly slower reaction to safety violations.

openSAFETY Failsafe over EtherCAT

POWERLINK EtherCAT
Master Master

Safe Safe 3
PLC PLC 2
4

Safe Safe
1 Motion 1 Motion
Safe Safe
Sensor X Sensor X
Example for shorter signal transfer times
due to cross-traffic: Cross-traffic enables
safety nodes to directly communicate
Task: Task: with each other (left), whereas signal
(X) Safe Sensor has to send data to Safe Motion (X) Safe Sensor has to send data to Safe Motion paths are quadrupled in a system that
does not support cross-traffic (right).
Solution: Solution:
(1) Safe Sensor sends data to Safe Motion (1) Safe Sensor sends data to EtherCAT Master
(2) EtherCAT Master relays data to Safety Master
(3) Safety Master sends data to EtherCAT master
(4) EtherCAT master relays data to Safe Motion

35
Masthead

MM-E01306.009 INDUSTRIALETHERNETFACTS FEB. 2013 English

INDUSTRIALETHERNETFACTS is an information
service provided by the EPSG ETHERNET POWERLINK
STANDARDIZATION GROUP, POWERLINK Office,
Schaperstr. 18, 10719 Berlin.
http://www.ethernet-powerlink.org/
Conception, Layout, Project Marketing,
and Coordination:
FR&P Werbeagentur GmbH,
Schaperstrae 18, 10719 Berlin,
Phone:+49(0)30-85 08 85-0, Fax: -86
Publication Management:
A.-Christian Broddack
Coordination Editorial Office/Production Team:
Heide Rennemann-Ihlenburg
Copyright Notice
The name and layout of
INDUSTRIALETHERNETFACTS
are protected by copyright laws.
Republication in full or in excerpts
requires advance permission
from the editorial office.