Meet Beckhoff at embedded world
Industrial Ethernet Book Issue 49 / 40
Request Further Info   Print this Page   Send to a Friend  

Working with industrial networks: Introduction to fieldbus

Industrial networks take in much more than just IP-based Ethernet and there are tens of millions of fieldbus nodes which are not disappearing anytime soon! The characteristics of fieldbus will keep it as the bedrock of Process Automation for the foreseeable future, and almost always with a connection to enterprise management systems. In this first introductory article about industrial network types, Moore Industries' Mike O'Neil introduces the basics of two major architectures: Profibus and Foundation Fieldbus H1. Fieldbus is simple... so simple, you'll appreciate how this simplicity is a major strength, but with the odd weakness which can be addressed with new hardware implementations. In this article, we'll stick to two basic fieldbus types commonly used in process control: Profibus PA and Foundation Fieldbus H1. We'll cover how fieldbus works, show how to connect instruments, and explain why 每 in most cases 每 you can't connect all 32 instruments on a single fieldbus segment as all the advertising claims! We will also talk about the differences between Profibus and Foundation Fieldbus, FISCO vs. Entity intrinsically safe fieldbus systems, installing redundant segments, and EDDL vs. FDT. How fieldbus works
In analogue controls systems, instruments produce a 4-20mA output signal that travels all the way from the remote column, tank or process unit to the control room, marshalling rack, remote I/O concentrator or RTU over twisted pair cables. Similarly, 4-20mA control signals travel from the control system to valve actuators and other control devices. Hundreds, sometimes thousands, of cables snake their way through cable trays, termination racks, cabinets, enclosures and conduit. Instead of running individual cables, fieldbus allows multiple instruments to use a single cable, called a 'trunk' or a 'segment,' each instrument connects to the cable as a 'drop.' Instruments, of course, must have a fieldbus interface to connect to the segment, and some sort of software running to provide the fieldbus communications. A fieldbus trunk or segment 每 either FF H1 or Profibus PA 每 is a single twisted pair wire carrying both a digital signal and DC power that connects up to 32 fieldbus devices (temperature, flow, level and pressure transmitters, smart valves, actuators, etc.) to a DCS or similar control system. Most devices are two wire bus powered units requiring 10 to 20mA, but it is also possible to have 4-wire fieldbus devices, typically where a device has a particularly high current draw. The fieldbus segment begins at an interface device at the control system. On a FF H1 (FF) system, the interface is called an H1 card; on a Profibus PA system (PA), it is a Profibus DP/PA segment coupler. In terms of signal wiring and power requirements for the segment, FF and PA are identical:
  • Minimum device operating voltage of 9V;
  • Maximum bus voltage of 32V;
  • Maximum cable length of 1900m (shielded twisted pair).
The DC power required by the bus is normally sourced through a fieldbus power supply or 'power conditioner' which prevents the high frequency communications signal from being shorted out by the DC voltage regulators. Typical power conditioners make 350 to 500mA available on the bus and usually incorporate isolation to prevent segment to segment crosstalk. For PA, the segment coupler usually incorporates the power conditioning component. In FF segments, the power conditioners are separate from the H1 interface card and are often installed in redundant pairs to improve the overall reliability. Figure 1 shows a typical fieldbus segment. When calculating how many devices can fit on a fieldbus segment, a user must take into account the maximum current requirement of each device, the length of the segment (because of resistive voltage drop along the cable), and other factors. The calculation itself is a simple Ohm's law problem, with the aim of showing that at least 9V can be delivered at the farthest end of the segment, after taking into account all the voltage drops from the total segment current. For example, driving 16 devices at 20mA each would require 320mA, so if the segment is based on 18AWG cable (50次/km/loop) with a 25V source power conditioner, the maximum cable length guaranteed to deliver 9V at the furtherest end is 1000m. Note that many users also specify a safety margin on top of the 9V minimum operating voltage, to allow for unexpected current loads and the adding of additional devices in the future.
Connecting instruments
As noted, each fieldbus device connects to the segment in parallel, via a drop on the fieldbus segment called a 'spur'. The simplest spur connection is as a T form. The problem with simple T connections is that if any one of the devices or cables short out, it takes down the entire segment. A short can occur during field maintenance of an instrument, from an accident in the field, corrosion causing electrical problems, or a host of other possibilities. Short circuit protection is therefore a requirement for proper fieldbus implementation.
Fig. 1. A fieldbus segment starts with an H1 interface card and a power supply for Foundation Fieldbus or a segment coupler for Profibus. Up to 32 devices can be supported on a single segment. The boxes with a T in them indicate the location of the segment terminators
Fig. 2. 'T' configurations are the simplest fieldbus connection. However, if one device fails or short-circuits, it takes down the entire segment (inset). A device coupler permits multiple instruments to be connected to a fieldbus segment. Each spur has shortcircuit protection, so it cannot harm the entire segment. More than one device coupler can be used in a segment. Another way to connect fieldbus devices is via junction boxes specifically designed for fieldbus, often referred to as device couplers (Fig. 2) 每 that allow multiple fieldbus devices to connect at one location. Typically, users will install a device coupler in a field enclosure, and connect nearby instruments to it. The fieldbus cable may continue onwards to another device coupler. A multi instrument segment may have several device couplers. Two basic types of electronic spur short circuit protection are used in device couplers: current limiting and foldback. Both prevent a spur fault from shorting out the segment and both auto-reset back to normal on removal of the fault. The current limiting technique limits the amount of power the short circuit can draw to between 40 and 60mA (vendor-dependent) but it also sustains that fault on the segment continuously. Although this design protects the segment from the initial short, the additional current load can deprive other instruments on the segment of power, overload the segment power supply, and possibly cause other failures on the segment. When a short circuit deprives other instruments of power, some may drop off the segment because they do not have enough power to operate properly. Consequently, when current limiting protection is used in a device coupler, system designers would allow a safety margin: they do not install as many instruments as the segment can theoretically provide power. In effect, a number of spurs are left empty to reduce fault condition power consumption to a level which guarantees operation of the remaining devices. For example, if a user wants the segment to be able to keep working with two failures 每 which can draw up to 120mA of current in worst case 每 the segment calculations must assume a maximum current availability of 350mA minus the 120mA taken by the faults, or 230mA. Instead of theoretically being able to power 32 devices that draw 10mA each, the segment is now only able to support 23 such devices. In practice, some users are wary of relying on current limiting couplers, and most would limit each segment to only 16 devices to prevent large scale segment failures. The foldback technique, when implemented in a device coupler, disconnects the shorted spur from the segment thus preventing loss of an entire segment. The foldback technique has a logic circuit on each spur (Fig. 3) that detects a short in an instrument or spur, disconnects that spur from the segment, and illuminates a fault condition light. With foldback device couplers, users no longer have to worry about spur failures and can have confidence about placing more devices on fieldbus segments. Since the cost of H1 cards (typically in the region $2k+) and other segment hardware can be expensive, being able to place more devices on a segment can save on system installation cost.
Fig. 3. Fold-back short circuit protection has logic that detects a short, removes the shorted circuit from the segment Segment termination
Every fieldbus segment must be terminated at both ends for proper communication. If a segment is not terminated properly, communication errors may occur from signal [transmission line mismatch] reflections. Most device couplers use manual on/off DIP switches to terminate couplers. In a segment, the last device coupler should contain the terminator, and all couplers between the last coupler and the H1 card should have their terminator switches set to off. The boxes with a 'T' in the previous diagrams illustrate where a typical segment is terminated properly. A frequent commissioning problem during startup arises from correctly locating the terminators. During installation of the fieldbus system, the DIP termination switches may sometimes be set incorrectly, creating problems. The instruments can behave erratically, drop off the segment mysteriously, and generally raise havoc 每 all because the terminations are not set properly. Diagnosing the problem often requires physically examining each device coupler to determine if the switches are set correctly throughout the segment. Automatic segment termination as found in some device couplers simplifies commissioning and startup. This facility automatically activates when the device coupler determines that it is the last fieldbus device coupler in the segment; if it is, it terminates the segment correctly. If it is not the last device, it does not terminate the segment, since the downstream device coupler will assume that responsibility. This obviates the setting of DIP switches by installation personnel for correct segment termination. If a device coupler is disconnected from the segment accidentally, or for maintenance purposes, automatic segment termination detects the change and terminates the segment at the proper place. This allows the remaining devices on the segment to continue operation. Redundant fieldbuses
Fieldbus has one major problem: all communications and power are dependent upon a single twisted pair trunk cable. If the trunk cable fails, it can take down all the devices on the segment simultaneously. Not only is the fieldbus segment lost to the control system, devices on the segment can no longer talk to each other. Although fieldbus instruments can continue to operate if the control system fails, any cable fault (open or short circuit) could render the entire segment inoperable. This problem is particularly serious when it affects plant-critical segments, where the failure of a segment may adversely affect plant or process applications, lead to costly process shutdowns, cause a hazardous condition, or release materials into the environment. No provision is made within either fieldbus standard for redundant segment communications. Various fieldbus vendors, including major process control companies, have developed redundant fieldbus schemes that involve complete duplication of all equipment 每 including H1 or DP/PA interfaces, power supplies, fieldbus cables, device couplers, and some critical fieldbus instruments (Fig. 4) 每 plus complex software voting schemes. A 'voting scheme' is needed because most control systems cannot tell when a fieldbus segment fails. They can only detect if the H1 or DP/PA interface itself fails, or if a particular fieldbus device fails. If the interface remains powered, the control system has to determine that the entire segment has failed by analyzing instrument signals (or the lack of them). Needless to say, such redundancy schemes are expensive. Our own company has looked at this problem as it applies to FF systems dealing with any single point failure 每 such as open or short circuit 每 on a FF segment. Our particular approach would be to use dual redundant foldback power conditioners (one for each leg of the segment), two fieldbus cables, and a specialised device coupler (Fig. 8). The power conditioners are located on DIN carriers for up to four segments at a time and the device couplers accommodate six or 10 fieldbus instruments. In a typical application, two redundant H1 interface cards are connected to two legs of a fieldbus segment, and wired out into the field. The power on each leg is properly conditioned by the power conditioner, and run to the device coupler, which are connected to multiple fieldbus devices. If a fault occurs on either cable, the power conditioner on the affected leg immediately cuts power to that leg, automatically switching the power draw to the remaining trunk while applying a segment terminator. This switchover takes just a few milliseconds. The cost of such architecture is only slightly more than a standard fieldbus system 每 and far less than that of a fully duplicated system.
Fig. 4. Some fieldbus redundancy techniques require complete duplication of the segment; in some cases, this means duplicate field instruments. When a segment fails, logic in the DCS determines that a failure has occurred, and switches from one H1 card to the other
Fig. 5. This fault tolerant fieldbus system only requires an extra power conditioner and a trunk cable. If one segment fails, it instantly switches over to the backup segment FF H1 vs Profibus PA
From a field wiring perspective, FF and Profibus are physically identical. They use the same twisted pair cables and device couplers, and require the same segment terminators. Both handle up to 32 devices per segment. One primary difference is that Profibus is a polling system, while FF uses cyclic transmission. Other differences include: FF devices have scheduled times at which they transmit their information, whereas PA devices submit their data at random times. In PA, slave polling involves the bus master asking for information from the devices. The link active scheduler in FF has a timetable, so it determines when devices communicate on the segment. Address allocation in PA has to be done by communication with each device individually, whereas FF devices will announce themselves to the bus master. Each PA device will go back to the bus master with its information, which will then transmit to other devices the relevant data. FF devices can talk to each other and bypass the bus master, hence providing peer to peer communication. FF devices can have built-in function blocks that allow them to talk to each other in peer-to-peer fashion, perform control functions, and continue to operate in the event that communications are lost to the control system.
Profibus systems do not have function block capability. Profibus instrumentation reports to, and takes directions from the PA master; if communications to the PA master(s) are lost, the instruments must go to a failsafe position or maintain their last settings until directed otherwise. FF and PA differ in the way that the segment control cards connect to the DCS or control system. FF uses an HSE (High Speed Ethernet) network to connect remote H1 cards to the DCS; PA uses Profibus DP, which is an RS485 network, or Ethernet-based Profinet architecture to connect its PA devices to the DCS. DD, EDDL and FDT
Both FF and PA systems require that instruments and controllers are 'mapped' into the control system. That is, the DCS or control system must be told what devices are on the segment, what variables are to be recognized as input and output data types, and what functions are available. This is usually done with Device Description (DD) files, which are text based files that can be downloaded from a web site into either an FF or PA based control system. For example, DD files from all FF devices that have received FF certification can be downloaded from the Fieldbus Foundation web site. In general, DD files are universal; that is, they can be used in either a FF or PA system. During control system commissioning, DD files are downloaded into the control system's equipment configuration. DD provides a standardized representation for the device that allows the host to interact with the device. This in turn permits the host to provide a consistent user interface for devices which is independent of the actual device type. EDDL (Electronic Device Description Language) and FDT/ DTM (Field Device Tool and Device Type Management) are more advanced ways to describe devices. Typically, FF systems use EDDL, while FDT/DTM is used by PA systems, but the two technologies are growing closer. EDDL starts with DD files. The DDL (Device Description Language) is the method for writing DD files. The next generation is EDDL (Electronic DDL), which provides additional information for GUI displays, configuration and calibration procedures, alarms and interfaces to higher-level software, such as MES, SCADA and plant historians. EDDL is independent of devices and operating systems. DTMs are vendor specific drivers that interact with the device. DTMs allow more sophisticated programs to be written, so configuration and management tools such as PACTware can be used to configure and view vendor specific data. They are also dependent on the host operating system, while EDDL is not. To process the EDDL or FDT files, each control system vendor must have software capable of reading and understanding the files. For a while, a controversy existed in the fieldbus world, where some FF vendors would not accommodate FDT/DTM files, and vice versa. Fortunately, two of the biggest control system vendors recently agreed to patch up their differences and support both systems, so the EDDL vs. FDT controversy appears to be ending. Soon, most control system vendors will support both EDDL and FDT. An end user must ensure that all the instruments and field devices planned for a control system conform to the chosen fieldbus. If, for example, a FF system with EDDL is chosen, then all the instruments must have EDDL or DD files; conversely, a PB system with FDT/DTM must use instruments that have those files. Fortunately, it is relatively easy to check, simply by consulting the web sites for FF or PA. Mike O'Neil is director, MooreHawke Fieldbus, a division of Moore Industries-International, Inc Mike O'Neil is director, MooreHawke Fieldbus,
a division of Moore Industries-International, Inc
Source: Industrial Ethernet Book Issue 49 / 40
Request Further Info    Print this Page    Send to a Friend  


IoT World 2020
Accelerate your HART data at the speed of Ethernet
embedded world 2020
Industry of Things World

Get Social with us:

© 2010-2020 Published by IEB Media GbR · Last Update: 18.02.2020 · 28 User online · Privacy Policy · Contact Us