Perry Sink, Synergetic Micro Systems, Inc.
If your hardware/software projects have anything to do with industrial automation or process control, then you know that open industrial networks -- Fieldbuses such as DeviceNet and Profibus -- are being considered and specified more and more all the time. Ethernet also is entering the fray as ERP and MIS system requirements grow. Robots, material handling systems, process control installations, assembly and packaging machines can have dozens to thousands of sensors and intelligent devices. There are three ways you can interconnect those devices: 1) hardwire each device individually, 2) use a "proprietary" network such as Modbus Plus, Remote I/O or Genius I/O, or 3) use an open architecture network.Just by using a network, you gain tremendous advantages in larger systems: faster wiring time, lower installation cost, modularity, and the potential for very powerful diagnostics. It's not necessary to go into all of the arguments for networking here, as most engineers can recognize the benefits of this.
The Advantages of Open Architecture
But using an open architecture network brings even more powerful advantages, because in nearly any product category, there are numerous vendors to choose from. The competition keeps prices low and vendors honest, and the added functionality that a networked device can provide has such potential that using a networked I/O system may truly be the only way to get best-of-class performance from your system.
Which Fieldbus Is Going To Win?
Many people are confused by the plethora of choices available -- there are at least a half dozen popular Fieldbuses, and a common question is: Which one is going to win? When is all this confusion going to end?There's never going to be a single winner, because real-world applications are too diverse. There will always be a plurality of solutions, each with its own set of tradeoffs. But if you understand the features, pros and cons of each one, you can make an informed choice. One or two of these buses will probably be an excellent solution for your particular application.Furthermore, all of the buses described here are well established, and are not going to "go away" any time soon. You can count on long term availability of devices and support.
Profibus: The World's Most Installed Open Fieldbus
- Origin: German Government in cooperation with automation manufacturers, 1989
- Implemented on ASIC chips produced by multiple vendors. Based on RS485 and the European EN50170 Electrical specification.
- Formats: Profibus DP (Master/Slave), Profibus FMS (Multi-master/Peer-to-peer), and Profibus PA (intrinsically safe). Connectors: 9-Pin D-Shell connector (impedance terminated) or 12mm IP67 quick-disconnect.
- Maximum Number of Nodes: 127
- Distance: 100M to 24 KM (with repeaters and fiber optic transmission)
- Baudrate: 9600 to 12M Bit/sec
- Message size: up to 244 bytes of data per node per message
- Messaging formats: Polling (DP/PA) and Peer-to-Peer (FMS)
- Supporting Trade Organization: Profibus Trade Organization, www.profibus.comProfibus is commonly found in process control and large assembly, and material handling machines; single-cable wiring of multi-input sensor blocks, pneumatic valves, complex intelligent devices, smaller sub-networks (such as AS-I), and operator interfaces.
Advantages: Profibus is the most widely accepted international networking standard. It is nearly universal in Europe and also very popular in North America, South America, and parts of Africa and Asia. Profibus can handle large amounts of data at high speed and serve the needs of large installations. The DP, FMS, and PA versions collectively address the majority of automation applications.
Disadvantages: High overhead to message ratio for small amounts of data; no power on the bus; slightly higher cost than some other buses; European- and Siemens-centricity is occasionally an obstacle for some North American users.Profibus' substantial speed, distance and data handling capabilities make it ideal for many process control and data intensive applications. Profibus DP, which is the most common messaging format for I/O, is a polling network, meaning that its assigned master periodically requests the status of each node. This ensures that each device on the network (which can send up to 244 bytes of data per scan) is updated consistently and reliably. Each message contains 12 bytes of overhead for a maximum message length of 256 bytes.
Multiple masters are possible with Profibus DP, in which case each slave device is assigned to one master. This means that multiple masters can read inputs from the device but only one master can write outputs to that device.
Profibus FMS is a peer-to-peer messaging format, which allows masters to communicate with one another. All can be masters if desired. FMS messages consume more overhead than DP messages."COMBI mode" is when FMS and DP are used simultaneously in the same network (Synergetic's DP/FMS masters support this). This is most commonly used in situations where a PLC is being used in conjunction with a PC, and the primary master communicates with the secondary master via FMS. DP messages are sent via the same network to I/O devices.
Profibus DP V1 Extensions
The older FMS specification is giving way to a new approach, DP with V1 extensions. This serves the needs of new devices with greater complexity. The Profibus Trade Organization has released a new specification which integrates many of the functions of Profibus FMS (multi-master, peer-to-peer communication) together with Profibus DP (master/slave I/O communication) so that the two types of messaging work together to combine synchronous scanning with on-the-fly configuration of devices. In the past, FMS and DP have been used together, but often for entirely different purposes.This integration enables Profibus to more effectively compete with some of the more advanced capabilities of its rivals, DeviceNet and Foundation Fieldbus.
Profibus DP V2 For Motion Control
A recent addition to the Profibus specification is V2, which adds 1) a synchronization feature which allows multiple devices and axes of motion to work on the same time clock, and 2) publisher / subscriber messaging which allows devices to communicate to each other on a one-to-one or one-to-many basis. This allows the coordination of synchronized axes of motion.
The Profibus PA protocol is the same as the latest Profibus DP with V1 diagnostic extensions, except that voltage and current levels are reduced to meet the requirements of intrinsic safety (Class I div. II) for the process industry. Synergetic master cards support Profibus PA, but barriers which convert between DP and PA are necessary (available from a number of companies). PA devices are powered by the network at intrinsically safe voltage and current levels.
CAN: The Low Cost, Ultra-Reliable Foundation for Several Popular Fieldbuses: DeviceNet, CANopen, SDS and Others
When the developers of DeviceNet, CANopen, SDS and other CAN based networks sought a bulletproof technology for critical factory networking purposes, they opted for a ready-made solution from the automotive industry. Bosch developed CAN in the early 1980s for eliminating large and expensive wiring harnesses in Mercedes automobiles.CAN was developed so that the primary control components in an automobile -- brake lights, airbags, sensors, lights, electric windows and door locks, etc. -- could be connected with a single cable instead of a bundle of cables 3 inches thick. (Sound familiar?)Automotive manufacturers found that if a wiring harness is faulty, it's sometimes cheaper to scrap the entire car than to troubleshoot the wiring harness. In a network, you can do wiring in software, and the added hardware cost is more than paid for by labor savings. The same applies to automated equipment in a factory.
Robustness In the Extreme
Of course in a vehicle, communication can be quite literally a life and death situation. Network errors are simply not tolerable, regardless of origin. CAN lives up to the rigorous requirements, with a statistical probability of less than one faulty message per century.
Software Application Layers On Top of CAN
CAN itself is a low-level message arbitration protocol implemented on inexpensive ( less than $1 ) chips which are available from multiple vendors and manufactured by the millions. In order to have a fully functional network protocol, an additional software layer must be added.Higher layer protocols like DeviceNet can be thought of as a sophisticated set of "macros" for CAN messages, specifically suited for automation. SDS and CAN-open are automation networks also based on CAN. (Another popular standard, J1939, was created by the Society of Automotive Engineers. It is a CAN application layer used in trucks and buses.)
DeviceNet: The Do-All Fieldbus for Low- and Mid-Level Factory Networking
- Origin: Allen-Bradley, 1994
- Based on CAN (Controller Area Network) technology, borrowed from the automotive industry, and the RS485 electrical specification
- Maximum Number of Nodes: 64
- Connectors: Popular "Mini" 18mm and "Micro" 12mm waterproof quick-disconnect plugs and receptacles, and 5 pin phoenix terminal block
- Distance: 100M to 500M
- Baudrate: 125, 250 and 500 Kbits/sec
- Maximum Message size: 8 bytes of data per node per message
- Messaging formats: Polling, Strobing, Change-of-State, Cyclic; Explicit messaging for configuration and parameter data; UCMM for peer-to-peer messaging; Producer/Consumer based model
- Supporting Trade Organization: Open DeviceNet Vendor Association, www.odva.org
- Typical Applications: Most commonly found in assembly, welding and material handling machines. Single-cable wiring of multi-input sensor blocks, smart sensors, pneumatic valves, barcode readers, drives and operator interfaces
- Especially popular in automotive and semiconductor
- Advantages: Low cost, widespread acceptance, high reliability, and efficient use of network bandwidth, power available on the network
- Disadvantages: Limited bandwidth, limited message size and maximum length.
DeviceNet: Versatile, Available, and Competitive
DeviceNet is a versatile, general purpose Fieldbus designed to satisfy 80% of the most common machine- and cell-level wiring requirements. Devices can be powered from the network so wiring is minimized. The protocol is implemented on many hundreds of different products from hundreds of manufacturers, from smart sensors to valve manifolds and operator interfaces.
Work Smarter, Not Harder
One of DeviceNet's major benefits is its multiple messaging formats, which allow the bus to "work smart" instead of working hard. They can be mixed and matched within a network to achieve the most information-rich and time-efficient information from the network at all times:
Polling: The scanner individually asks each device to send or receive an update of its status. This requires an outgoing message and incoming message for each node on the network. This is the most precise, but least time efficient way to request information from devices.
Strobing (broadcast): The scanner broadcasts a request to all devices for a status update. Each device responds in turn, with node 1 answering first, then 2, 3, 4 etc. Node numbers can be assigned to prioritize messages. Polling and strobing are the most common messaging formats used.
Cyclic: Devices are configured to automatically send messages on scheduled intervals. This is sometimes called a "heartbeat" and is often used in conjunction with Change of State messaging (see below) to indicate that the device is still functional.
Change of State: Devices only send messages to the scanner when their status changes. This occupies an absolute minimum of time on the network, and a large network using Change of State can often outperform a polling network operating at several times the speed. This is the most time efficient but (sometimes) least precise way to obtain information from devices because throughput and response time become statistical instead of deterministic.
Explicit Messaging: The explicit-messaging protocol indicates how a device should interpret a message. It is commonly used on complex devices, like drives and controllers, to download parameters that change from time to time but do not change as often as the process data itself. An explicit message supplies a generic, multipurpose communication path between two devices and provides a means for performing request/response functions such as device configuration.
Fragmented Messaging: For messages that require more than DeviceNet's maximum 8 bytes of data per node per scan, the data can be broken up into any number of 8-byte segments and reassembled at the other end. This requires multiple messages to send or receive one complete message. DeviceNet scanners typically fragment messages automatically as necessary, without intervention from the user.
UCMM (Unconnected Message Manager): DeviceNet UCMM interfaces are capable of peer-to-peer communication. Unlike the plain-vanilla master/slave configuration, each UCMM capable device can communicate with another directly, without having to go through a master first.UCMM devices must accept all generic CAN messages, then perform filtering of irrelevant or undesired message types in the upper software layer. This requires more RAM and ROM than ordinary master/slave messaging.
Synergetic repeatedly heard complaints that electricians and maintenance people were having a difficult time troubleshooting routine DeviceNet problems -- it was too time consuming to hook up a laptop computer to the network and find the source of a problem -- so we created a simple portable network analyzer for DeviceNet.
This new, Do-All hand-held configuration and diagnostic tool is a fast way to troubleshoot a network -- indispensable wherever real people (not just technical wizards) need to make a network go and keep it going! The DeviceNet Detective monitors important signal levels, displays complete device and configuration information for each node, and has an "error log book" which time stamps problems which occur over a period of minutes or hours. It shows all of the most important network information without excessive complexity.
CANopen: The European CAN Bus
- Origin: CAN in Automation, 1993
- Based on CAN (Controller Area Network) technology, borrowed from the automotive industry, and the RS485 electrical specification
- Maximum Number of Nodes: 64
- Connectors: Popular "Mini" 18mm and "Micro" 12mm waterproof quick-disconnect plugs and receptacles, and 9 pin D-shell
- Distance: 100M to 500M
- Baudrate: 125, 250 500 and 1000 Kbits/sec
- Maximum Message size: 8 bytes of data per node per message
- Messaging formats: Polling, Strobing, Change-of-State, Cyclic, and others
- Supporting Trade Organization: CAN In Automation, http://www.can-cia.de
- Typical Applications: Commonly found in motion control systems, assembly, welding, and material handling machines. Single-cable wiring of multi-input sensor blocks, smart sensors, pneumatic valves, barcode readers, drives and operator interfaces.
- Advantages: Better suited for high-speed motion control and feedback loop closure than other CAN based networks. High reliability, efficient use of network bandwidth, power available on the network.
- Disadvantages: Acceptance limited outside of Europe. Protocol is complex and involved from developers' point of view. Same general limitations as other CAN based networks (limited bandwidth, limited message size, and maximum network length).
CANopen is a family of profiles based on CAN, a higher-layer protocol (CAN application layer or CAL, and communication profile) providing additional functionality such as standardized communication objects for process data, service data, network management, synchronization, time-stamping, and emergency messages. CANopen was developed within the CAN in Automation (CiA) international users and manufacturers group, over 300 companies strong.
CANopen networks provide multi-master functionality and broadcast communication, because they are based on CAN. In CANopen different methods are specified to achieve real-time communications. CANopen also specifies a predefined master/slave connection set to unburden the newcomer from the task of distributing identifiers as required in CAN data link layer solutions. One of the important functions is the fragmentation of data blocks larger than 8 bytes. The transport protocol uses confirmed services to guarantee that the communication profile layer correctly receives configuration data. The CANopen communication profile makes products from different vendors interoperable.
The general difference of CANopen to master/slave-oriented Fieldbuses is the capability that each node can access the bus and communicate directly to each other node without any master. Because CANopen is based on CAN, the communication profile provides event-driven process data transmission, which reduces communication as much as possible. For motion control applications there is also synchronous operation (cyclic and acyclic).
CANopen, DeviceNet and Smart Distributed System are all object-oriented and provide a similar functionality regarding the transmission of real-time data, configuration data, and network management information. However, DeviceNet and Smart Distributed System are more connection-oriented, and CANopen is more message-oriented. There are also some minor differences in the fragmentation of larger data blocks. The three mentioned CAN-based higher-layer protocols are supported by the CiA organization and specify also the behavior and functionality of standard devices such as I/O modules, drives etc. to achieve interchangeability of devices produced by different manufacturers.
Applications in Multiple Industries
Besides the use in machine control systems (e.g. textile, printing, packaging, injection Molding, and other applications), CANopen is implemented by robotics manufacturers and in medical equipment such as computer tomographs and X-ray apparatus. CANopen networks are also used in forklifts and cranes. Other applications are public transportation (passenger and driver information systems) and ship control systems. But the main application area is decentralized machine control.
Interbus: The High Speed, Deterministic, European Fieldbus
- Origin: Phoenix Contact, 1984
- High Speed Shift Register topology
- Maximum Number of Nodes: 256
- Connectors: 9 Pin D-Shell and 23mm circular DIN; Cabling options allow for twisted pair, fiber optic, slipring, infrared or SMG connections
- Distance: 400M Per Segment, 12.8 KM Total
- Baudrate: 500 Kbits/sec (2Mbit also available)
- Message size: 512 bytes of data per node, unlimited block transfers
- Messaging formats: I/O scanning and PCP channel for data transfer
- Supporting Trade Organization: The Interbus Club, www.interbusclub.com
- Typical Applications: Commonly found in assembly, welding and material handling machines. Single-cable wiring of multi-input sensor blocks, pneumatic valves, barcode readers, drives, and operator interfaces. Also can be used with SensorLoop and AS-I sub-networks.
- Advantages: Auto-addressing capability makes configuration very simple; extensive diagnostic capability, widespread acceptance (especially in Europe), low overhead, fast response time and efficient use of bandwidth, power (for input devices) available on the network.
- Disadvantages: One failed connection disables entire network; limited ability to transfer large amounts of data.
A Different Twist on Moving Data
Interbus was one of the very first Fieldbuses to achieve widespread popularity. It continues to be popular because of its versatility, speed, diagnostic and auto-addressing capabilities. Physically it has the appearance of being a typical line-and-drop based network, but in reality it is a serial ring shift register. Each slave node has two connectors, one which receives data, and one which passes data on to the next slave.Address information is not contained in the protocol; data is pushed through the network in a circular fashion and the master is able to determine which node is being read or written to by its position in the circle, so to speak. Therefore the protocol has minimal overhead, and for typical installations which might incorporate a few dozen nodes and perhaps a dozen I/O per node, few buses are faster than Interbus.
Because of the unusual network topology, Interbus has two other advantages. First, a master can configure itself because of the ring topology. Synergetic's Master can configure itself without intervention from the user. Interbus has the potential to be "idiot proof" (at least as much as any network can be). Second, precise information regarding network faults and where they have occurred can drastically simplify troubleshooting.
Interbus handles both analog and digital I/O with ease, and the PCP channel is a mechanism by which block transfers of configuration data or downloads can be packed within the Interbus protocol without interfering with the transmission of normal I/O data.
AS-I (Actuator Sensor Interface): The Simplest (and Least Expensive?) Fieldbus
- Origin: AS-I Consortium, 1993
- Maximum Number of Nodes: 31 slaves, 1 master
- Connectors: Insulation displacement connectors on flat yellow cable, 2 position terminal block or 12mm "micro" quick-disconnect connectors
- Distance: 100M, 300M with repeaters
- Baudrate: 167 Kbits/sec
- Message size: 8 bits (4 inputs, 4 outputs) per node per message
- Messaging formats: Strobing
- Supporting Trade Organization: AS-I Trade Organization, http://www.as-interface. com/ Typical Applications: Commonly found in assembly, packaging and material handling machines. Single-cable wiring of multi-input sensor blocks, smart sensors, pneumatic valves, switches and indicators.
- Advantages: Extreme simplicity, low cost, widespread acceptance, high speed, power available on the network. Excellent network for wiring discrete I/O devices.
- Disadvantages: Poorly suited to connecting analog I/O; limited network size; limited visibility in North America (at present).
NOTE: A new specification has been released for AS-I which allows for 62 slave devices (instead of 31), each supporting 4 inputs and 3 outputs. Although there are few manufacturers who support this new spec at this time, old devices and new are interoperable.
Simplicity In Networking
ASI was developed by a consortium of European automation and sensor companies, which saw a need for networking the simplest devices at the lowest level. ASI is easy to configure and low in cost. It is most often used for proximity sensors, photoeyes, limit switches, valves and indicators in applications like packaging machines and material handling systems.ASI is designed for small systems employing discrete I/O. It allows for up to 31 slaves, which can provide up to 4 inputs and 4 outputs each for a total of 248 I/O.
"If you can program a VCR, you can configure an ASI network."
ASI is arguably the simplest Fieldbus to use. ASI uses a number of sophisticated and clever mechanisms to ensure fast and reliable data transmission and user friendliness. The only configuration issues are choosing the address of each node and assigning individual inputs and outputs within those nodes.
The Flat Yellow Cable
ASI is best known for its flat yellow cable, which is pierced by insulation displacement connectors so that the expense of tees and complex connectors is avoided. Devices are simply clamped onto the cable and a connection is made. In addition to the popular flat cable, ordinary lamp cord can be used and normally no adverse effects will be experienced.
Power On the Bus
The signal cable also carries 30VDC at low current to power input devices; supplemental power for outputs can be provided via an additional flat (black) cable. Most output devices have provisions for this extra cable.
A common question is the lack of shielding and the obvious concerns about noise immunity. Digital signals are encoded on the cable in a sinusoidal signal, which has a very narrow frequency bandwidth. Filtering which is distributed through the network rejects all extraneous frequencies, and in this way ASI can be operated in electrically noisy environments (even on robotic welders) without experiencing transmission errors.
Analog signals can be transmitted on ASI, but a node can represent only one analog device, and fragmented messaging must be used to transmit signals requiring more than 4 bits.
Determinism and Scan Time
ASI is deterministic; meaning that one can know with certainty how long it will take for status changes to be reported to the master. To calculate scan time, multiply the number of nodes (including the master) by 150 microseconds. The maximum network delay is 4.7mS, which is certainly speedy enough for most applications (most PLC's have a scan time of 20mS or more!).
ControlNet: The High Level, Mission Critical Fieldbus
- Origin: Allen-Bradley, 1995
- Based on RG6/U cabling (popular in cable TV applications) and Rockwell ASIC chip
- Maximum Number of Nodes: 99
- Connectors: Twin redundant BNC
- Maximum Distance: 250 to 5000M (with repeaters)
- Baud rate: 5M bit/Sec
- Message Size: 0-510 bytes
- Messaging Formats: Based on Producer/Consumer model; multi-master, peer-to-peer, fragmented, prioritized and deterministically scheduled repeatable messages; dual transmission paths for built-in redundancy
- Supporting Trade Organization: ControlNet International, http://www.controlnet.org/
- Typical Applications: Mission critical, plant-wide networking between multiple PCs, PLCs and sub-networks (i.e. DeviceNet, Foundation FieldBus H1, etc.) and process control, and situations requiring high-speed transport of both time-critical I/O and messaging data, including upload/download of programming and configuration data and peer-to-peer messaging.
- Advantages: Deterministic, repeatable, efficient use of network bandwidth, provides redundancy at lower cost than most other available networks including Ethernet. Can be transmitted on any IP transport protocol via Ethernet, Firewire or USB.
- Disadvantages: Limited multi-vendor support and expensive Rockwell ASICs.
ControlNet was conceived as the ultimate high-level Fieldbus network, and was designed to meet several high-performance automation and process control criteria. Of primary importance is the ability of devices to communicate to each other with 100% determinism while achieving faster response than traditional master/slave poll/strobe networks. (Determinism means knowing absolute worst-case response times with 100% certainty.) This is made possible by the Producer/Consumer communication model and the scheduler, which rigorously prioritizes messages.
Multi-Master and Repeatability
ControlNet allows multiple masters to control the same I/O points. Repeatability ensures that transmit times are constant and unaffected by devices connecting to, or leaving, the network. These features are further enhanced with user selectable I/O and controller interlocking update times to match application requirements.
Large Quantities of Data and Complex Devices
ControlNet is specifically designed to accommodate the high-level information and control needs of literally dozens of sub-networks and controllers. In process control situations where hazardous materials are involved and absolute certainty with respect to control processes is required, the deterministic capabilities of ControlNet are extremely important.
The ControlNet architecture has redundant connectivity as an integral feature. Redundancy is rather difficult to achieve with other networks, but each ControlNet node has dual connections for this very purpose.Synergetic provides a number of useful ControlNet connectivity solutions:
The PC Adapter Cards. We have PC/ISA and PC104 Adapters for ControlNet. They use the same configuration tools (the SYCON configuration tool for ControlNet is available; these cards can also be configured with Allen-Bradley RSNetworX Software), hardware and software interface as all of the other FieldBus interfaces (DeviceNet, ProfiBus, InterBus, etc.).
Embedded Slave Modules. The COM OEM ControlNet module allows you to design instant ControlNet, DeviceNet, Profibus, Interbus, and other Fieldbus connectivity into your product in a matter of weeks. Now you don't have to be a ControlNet expert to manufacture a ControlNet-compatible product!
After designed in, all COM cards in the same family can be interchanged as different Fieldbuses require.
- Foundation Fieldbus: The Open International Standard for Mission Critical, Process Control and Intrinsically Safe Environments
- Origin: ISA, 1998
- Implemented on chips produced by multiple vendors
- "H1" Intrinsically Safe, 31.25Kbit/sec; "HSE" High Speed Ethernet, 100Mbit/sec.
- Based on ISA SP50/IEC 61158
- Maximum Number of Nodes: 240 per segment; 65,000 possible segments.
- Distance: 1900M for H1
- Baudrate: 31.25K and 100M Bit/sec
- Message size: 128 Octets
- Messaging format: Client/Server, Publisher/Subscriber, Event Notification
- Supporting Trade Organization: Fieldbus Foundation (www.fieldbus.org)
- Typical Applications: Distributed Control Systems; Continuous process control, Batching, Oil and Gas
- Advantages: Flexible, sophisticated protocol with many capabilities; Intrinsically safe; Integrated device level/plant level approach; Very strong contender as future process industry standard.
- Disadvantages: "Process Industry" centric; limited availability of compatible devices; slow process of standardization and industry adoption.
Foundation Fieldbus has finally come into its own, and is rapidly establishing itself as the future standard for process industry networking.Since its official introduction in 1997, many DCS vendors have been embracing Foundation Fieldbus, developing and certifying devices. Foundation Fieldbus contends with Modbus, HART, and Profibus PA as a standard.
The Unique Requirements of the Process Industry
Process industry installations tend to be "campus wide": much larger than the cell-level and "production line" sized applications more typical of the automation industry. Upgrades and changes tend to be done much less often and are more expensive. (When process vendors find a customer, they usually put out press releases and have big parties, because these projects are typically very large.)
The process industry is very, very cautious about new technology. These installations can be extremely hazardous and there's no room for error. So the Fieldbus Foundation's objective of developing the "ultimate" networking standard, that would please everybody, was an ambitious one and it did not come about easily.
What was produced, however, is a sophisticated, object-oriented protocol, which uses multiple messaging formats and allows a controller to recognize a rich set of configuration and parameter information ("Device Description") from devices which have been plugged into the bus. Foundation Fieldbus even allows a device to transmit parameters relating to the estimated reliability of a particular piece of data.Foundation Fieldbus uses a scheduler to guarantee the delivery of messages, so issues of determinism and repeatability are solidly addressed. Each segment of the network contains one scheduler. HSE, High Speed Ethernet, is a 100Mbit Ethernet standard which uses the same protocol and objects as FF H1, on TCP/IP. HSE is not commercially available yet, but there are currently some beta sites.
Ethernet: The Worldwide De Facto Standard for Business and PC Networking
- Origin: Digital Equipment Corporation, Intel and Xerox, 1976
- Implemented on multitudes of chips produced by many vendors. Based on IEEE 802.3
- Formats: 10 Base 2, 10 Base T and 100 Base T, 100 Base FX, 1 Gigabit; Copper (Twisted Pair / Thin Coax) and Fiber
- Connectors: RJ45 or Coaxial
- Maximum Number of Nodes: 1024, Expandable with Routers
- Distance: 100M (10 Base T) to 50 KM (Mono mode, Fiber with Switches)
- Baudrate: 10M to 100M Bit/sec
- Message size: 46 to 1500 bytes
- Messaging format: Peer-to-Peer
- Supporting Trade Organization: Industrial Ethernet Association (www.IndustrialEthernet.com) and IAONA (www.iaona.com).
- Typical Applications: Nearly universal in office / business Local Area Networks. Widely used also in PC to PC, PLC to PLC and supervisory control applications. Gradually working its way toward the "sensor level" in plant floor applications.
- Advantages: Ethernet is the most widely accepted international networking standard. Nearly universal worldwide. Ethernet can handle large amounts of data at high speed and serve the needs of large installations.
- Disadvantages: High overhead to message ratio for small amounts of data; No power on the bus; Physically vulnerable connectors and greater susceptibility to EMI/RFI than most Fieldbuses; Confusion based on multiple open and proprietary standards for process data.
The networking of millions of PCs in offices and the proliferation of the Internet across the world has made Ethernet a universal networking standard. Ethernet hardware and related software has evolved to the point where even inexperienced users can build simple networks and connect computers together.
Ethernet hardware is "dirt cheap" and can be purchased in office supply stores, computer stores, and e-commerce sites everywhere.
In contrast to the perception that the popular Fieldbuses are expensive, difficult to use, and that there are too many of them vying for market dominance, Ethernet appears to be a panacea. Furthermore, a study by a Big Three automotive manufacturer showed that Ethernet could potentially serve up to 70% of plant floor networking applications.
Reality may be somewhat different, however -- at least right now. There are at least four major issues, which must be addressed satisfactorily for Ethernet to become a viable, popular, plant-floor Fieldbus:
1. A common "Application Layer" must be established. When your device receives a packet of data, what format is that data in? Is it a string of I/O values, a text document or a spreadsheet? Is it a series of parameters for a Variable Frequency Drive? How is that data arranged? There are several competing standards; more about this below.2. Industrial grade connectors will be necessary for many applications. Cheap plastic "Telephone Connectors" don't cut it on the plant floor, and the RJ45 connectors aren't up to the task. An industrial strength connector will be a great benefit.3. Many users desire 24 Volt power on the Bus. This is advantageous from a practical standpoint -- it reduces wiring and power supply problems -- but it adds cost and introduces noise and other technical problems.4. Some applications require determinism. Ethernet as it is typically used is not deterministic or repeatable; in other words, throughput rates are not guaranteed. However, methods exist for architecting deterministic Ethernet systems. Note that, in reality, most people who think they need determinism really just need speed.
Ethernet Doesn't Guarantee Interoperability
Ethernet is just a physical layer standard, in much the same way an RS232, or for that matter, a telephone line, is. Having a physical connection means that messages can be transmitted, but it does not assure successful communication. Just because you can make a telephone ring in Shanghai doesn't mean you can speak Mandarin.
TCP/IP Doesn't Guarantee Interoperability, Either
Many transmission protocols can be used on Ethernet; the most popular, and the one used on the World Wide Web, is TCP/IP, which stands for Transmission Control Protocol / Internet Protocol.
When you download a file from the web, you can see the speed of the transmission speed up and slow down as network traffic levels change. TCP/IP is the mechanism that breaks the downloaded file into any number of bits and pieces and reassembles them at the other side. TCP/IP was developed at Stanford University in the 1970s as a "handshaking" mechanism that would assure that "the message would always eventually get through." It does seem a bit miraculous that those files usually come across perfectly, with every single bit intact, wouldn't you say?
To carry the Web example a bit further, we've all had the experience of downloading a large file, only to discover that our PC "cannot find an associated application for this file type." So you end up downloading a plug-in like Shockwave or RealAudio or Winamp or Adobe Acrobat Reader so you can open the file.
The exact same problem applies to industrial controls. You can send any file or piece of process data you want to over Ethernet or the Internet, but the receiving end has to know what to do with the data. TCP/IP doesn't assure you of opening the file; it just guarantees that it will arrive.
Fieldbuses, School Buses, and Flatbed Trucks
To an extent, you can think of the current popular Fieldbuses -- DeviceNet, Profibus, Interbus -- as being like a school bus. The school children (bits or bytes of data) file into the bus and sit down in their respective seats, in a predefined arrangement. The popular Fieldbuses have profiles (arrived at over years and years of laborious committee meetings) for the most common devices: analog and digital I/O, drives, controllers, etc. -- and the data comes in prescribed formats.
Well, if Profibus is like an orderly school bus, then Ethernet and TCP/IP are like a flatbed truck. You can put anything you want on the back of the truck and send it. But it may be a mess for the guy on the other side who's trying to unload it. Back to our school children: if you put a bunch of kids on the back of a truck, you'd have chaos. The approach that is used, then, is "strapping the school bus on the back of the flatbed truck." You put the children, lunchboxes in hand, on the bus, put the bus on the truck, transport it at high speed, then take the bus off the truck. The children can exit the same way they got in.
Similarly, data can be packaged into an existing FieldBus format, and then transported on TCP/IP. This is the likely direction the industry will take.
Existing Fieldbuses On Ethernet
The next frontier for the established FieldBus organizations is to produce Ethernet TCP/IP application layers of their protocols. Presently, there are four major contenders: Modbus/TCP (Modbus protocol on TCP/IP), EtherNet/IP (the ControlNet/DeviceNet objects on TCP/IP), Foundation Fieldbus High Speed Ethernet, and Profinet (Profibus on Ethernet).
One could propose an infinite number of potential application layer protocols, and in fact right now there are, in addition to the above protocols, a myriad of other, proprietary standards from various vendors. But there are several significant advantages to employing the existing bus architectures:
- Profiles for many devices have already been defined, and can be transferred to Ethernet with relatively little effort.
- In systems which use, for example, Profibus as an I/O level network, and Profibus on Ethernet at the supervisory level, the relationship between the two networks is relatively transparent. Data can be transferred between the upper and lower network fairly easily.
- Many developers and users are familiar with these existing protocols, and this speeds the process of product development and adoption.
To go into the details of Modbus/TCP, EtherNet/IP and Profibus/Ethernet is beyond the scope of this article. But for more information, see the following links:
If there is currently a defacto standard for Ethernet on the plant floor, this is it. Takes advantage of the simplicity and availability of Modbus. However, the limitations of Modbus also apply; it has limited ability to transmit complex sets of parameter data between devices.
Ethernet/IP is essentially the ControlNet / DeviceNet objects on TCP/IP and UDP. This specification has strong possibilities but is just now being released. Example code and the specification itself will be available from ODVA, ControlNet International and Industrial Ethernet Association.
Profibus On Ethernet
This specification is, as of this writing, still unreleased. It combines the existing Profibus protocol with open source OPC / XML services. With the worldwide acceptance of Profibus, it has strong likelihood of commercial success.
Foundation Fieldbus High Speed Ethernet (HSE)
HSE puts the Foundation Fieldbus H1 protocol on TCP/IP and also adds OPC, XML, and Simple Object Access Protocol on TCP/IP. Definitely geared towards the process industry, and will be a very strong contender there.
A final note about "Application Layer" protocols: They can co-exist on the same network. You could possibly have all four of the above protocols on the same wire, running simultaneously, just like you pass Word documents, HTML files and .exe's all on the same office LAN. However, this is bound to create added confusion and expense as well. Furthermore, no standard which defines a relationship between dissimilar application layers, operating on the same wire, has been proposed to date.
Practical Applications and the Future of Ethernet On the Plant Floor
This was written without the help of seers, soothsayers, astrologers or palm readers. But here's what Synergetic sees in its crystal ball of industrial networking:
- It will be difficult and expensive to get Ethernet to the sensor level. Using Ethernet to turn a valve on or off, or to connect a node to a photo eye or prox switch, is kind of like putting five-foot monster tires on a Ford Escort. However, it will do well for "Racks" and clusters of I/O tying into a single node.
- Ethernet will not necessarily be "dirt cheap" in industrial applications, and overall may prove to be more expensive.-- Ethernet PC cards cost 1/10 as much as, for example, DeviceNet cards, for several reasons. Obviously, they're made by the millions. They don't have a processor: they're passive, meaning that the PC does most of the work. DeviceNet cards usually have a processor on board that handles all communication. Also, long term availability is a real problem with consumer computer products. Product life cycles are measured in days, not years! Finally, the quality of cards you buy at an office supply store may be lacking, certainly not for professional, industrial use. (This is why Synergetic makes industrial grade Ethernet NIC cards for PC/ISA and PC/104.)-- Ethernet in embedded applications is much more expensive than CAN. CAN chips cost $1 or less; Ethernet chips are much more. So devices themselves with Ethernet built in will definitely have a cost factor. You can expect Ethernet to have costs similar to Profibus, which has more expensive ASICs.-- Industrial grade cables and connectors will drive the cost up as well.
- Not only are Ethernet ASICs more expensive than CAN chips, running a TCP/IP stack takes more horsepower than your usual 8051 can supply. The processors will cost more, too.
- A TCP/IP packet has 68 bytes of overhead. For short messages, corresponding to typical industrial I/O products, that's a lot of overhead. So 10Mbits may not be as fast as it sounds.
- The promise of fiber optic, 100M and 1000Mbit Ethernet is certainly exciting, and potentially overcomes most people's speed issues. However, the cost associated with these enhancements may be substantial.
The Bottom Line
Ethernet will establish itself among the popular Fieldbuses as a legitimate and attractive option. It will not replace them, but for some applications it will be the clear winner.
The Fieldbus Wars and You
With eight different Fieldbuses described here -- with multiple variations within several of them -- it's easy to be indecisive and just wait for one standard to "shake out." You can wait until your teeth fall out, but the emergence of one single Fieldbus for everybody is not going to happen. No single bus will dominate, because no single bus solves all of the problems.
You could make the same arguments for standardizing all automobile engines to a single mounting size, or making all wooden doors the same height and width. But the fact is, the creativity of vendors in the constant race for newer and better technology, and the diverse and growing needs of customers, make it impossible for this to happen.
What you must do is narrow your choices to a few likely options (that's pretty easy) and then research the details and the available vendors for all of the devices you wish to use. Then do something. The savings in wiring, the additional diagnostics and functionality that's available in devices, and the lower long term cost of maintenance are well worth the extra hardware expense.
It's also important to recognize that the established Fieldbuses described in this series -- particularly Profibus, DeviceNet, Foundation Fieldbus, Interbus, CANopen and AS-I -- will be around for a long time. If you choose one of these, the odds are good that you'll still be able to buy products in five years. Industrial Ethernet, however, is still young enough that questions may exist as to whether a standard you buy into now will be valid two or three years from now.
The Real Question for OEMs: One Network or Many?
If you are an "end user," i.e. a factory or manufacturer of finished goods, then your decision is, "Which bus (or which two buses -- one for the sensor level, one for the control level) should I choose?" If that's the case, the above descriptions will be useful to you. You may wish to team up with an experienced high tech distributor or integrator who has experience with Fieldbus installations, and with proper planning, you'll be pleased with the results.However, if you are any kind of OEM, system integrator, software developer, equipment or device manufacturer, the reality of supporting multiple buses should definitely be part of your plan. Your customers may ask you for any or all of the above buses, and a strategy for supporting a plurality of solutions at minimum cost is very important! Synergetic's approach to this problem is a universal software interface and single configuration tool.
PC/Embedded Interfaces: Multiple Networks, but Only One Driver
Visit www.synergetic.com for ISA, PCI, PC104, PCMCIA, STD32, and snap-in OEM daughterboards for the above networks. They use the same hardware and software interface as all of our Fieldbus interfaces (DeviceNet, Profibus, AS-I, Interbus, ControlNet, etc.), which means that you can write a single software / firmware / controls program, in whatever operating system you choose, and use the same drivers and code for all of these networks.
Multiple Networks, One Configuration Tool
Whenever you commission a network, you must assign node addresses and attributes and map the data to your controls program. The SYCON configuration tool is a single software package that configures all of these buses with the same look and feel. Once you've used SYCON to configure one network, doing any of the others is fairly easy.
Making Sure It All Works
"End User" plants who want to install new Fieldbus technology should seriously consider using a qualified, experienced system integrator who can guarantee successful operation of the system. This can drastically reduce the risk in exchange for a known up-front cost.
...And Some Advice for OEMs
If you produce or integrate controls equipment, or develop hardware or software applications, there's no question about it: You need to be Fieldbus savvy. When your customers want to buy your product with a Fieldbus option, that's not the time to start designing new stuff. Be ready beforehand.
This is a difficult challenge, because developing a FieldBus interface is an expensive and time-consuming proposition. Most of all, you don't want to do it five times for five different buses. Synergetic's approach to this problem is PC cards and embedded daughterboards, which all have a common hardware and software interface, regardless of the bus. This makes them essentially interchangeable with little or no changes in your software or firmware. A detailed explanation of this common architecture concept is available at www.synergetic.com/oem.
OEMs can use Fieldbus connectivity to differentiate their products in the marketplace. First of all, just having the capability is somewhat distinctive. But taking advantage of the diagnostic and information capabilities is a significant advantage that customers will pay more for. You win both ways.
Networking: The Center of New Technology Development
Networking is obviously a hot topic with the Internet and all of the business and ERP applications out there. And although the Wall Street Journal isn't necessarily doing stories on plant floor connectivity, keep in mind that the general connectivity that's available to citizens everywhere is changing people's expectations. They expect the instant availability of information and results.
Fieldbuses meet this requirement, and quite honestly, most Fieldbus users have only begun to tap the true potential of the technology. Adopt a Fieldbus architecture and squeeze the most performance from it! If you need education, or training, then seek it out. (Hands-on "Boot Camp" classes for Profibus and DeviceNet are available at www.synergetic.com/education.)
As you embrace the technology and scale the learning curve, your efforts will be amply rewarded with more productive equipment, more distinctive products, and the ability to attract leading-edge customers.
The Value of Information
When you buy connectivity, you are really buying information. Let's say you make products for the factory floor, and your control system is installed on a packaging line that seals 500 boxes of cereal per minute.Using simple calculations, if each box sells for $3, then down time costs $1500 per minute. Fifteen minutes of down time costs $22,500.
If you design diagnostic features into your controller, which, via a Fieldbus, transmit critical data that alerts a maintenance electrician to a potential failure before it happens, this is going to save your customer a lot of money -- $90,000 per hour, to be exact. Of course automotive plant down time is much more expensive than that -- by an order of magnitude. Semiconductor industry down time is yet an order of magnitude more expensive than that. So the cost of what you sell, versus the potential benefits to the customer, is miniscule. Never forget that!
Now of course theory is different than reality. Just because you say it will prevent a problem doesn't make it true. But if you can prove to your customer that your enhanced device prevents such problems in advance, on a regular, predictable basis, then your customer will have little resistance to paying a few hundred extra dollars for your device.
Remember, when you're selling Fieldbus, you're selling information, not wires. The information your devices provide can be very, very valuable. And that's what's so great about Fieldbus technology: Delivering the right information to the right place at the right time, for maximum productivity.