Fault-Free Fieldbus

To avoid any implementation headaches with that new fieldbus system, consider all options and specify the correct hardware at the beginning, recommends Desmond Ho.

Many new oil refineries, chemical plants and other major process plants in Asia are being controlled with fieldbus-based distributed control systems (DCS) from Emerson Process Management, Yokogawa, Siemens, Honeywell, Invensys and other control system companies. This is because fieldbus promises many benefits for operations and maintenance. Processes run at higher efficiencies because of more information, fewer shutdowns occur because of predictive process diagnostics, and instruments tell you when they need to be recalibrated or repaired.

In a conventional plant, about 65 percent of the time when an instrument technician is called for, it is not an instrument problem at all. With fieldbus, you can tell from an HMI display where the problems are, and what to do about them.

Although fieldbus systems provide many advantages to end users, there are key decisions made early in a project that affect maintenance and operations during construction and long after the project team has turned over the plant. These issues are not typically brought up by the control system supplier. These include:

• Installation and startup
• Surviving short circuits
• Removing and replacing instruments
• Installing intrinsically safe systems
• Setting up a redundant system

Many of these problems can be avoided through proper specification of fieldbus equipment at the beginning of the project. End users should become aware of these problems, realize that newer hardware solutions are available, and ask the DCS vendors to specify hardware that will minimize maintenance costs.

A little research on the Internet will find several articles that note the importance of considering maintenance, start-up and installation issues when specifying a fieldbus system. To quote from one article, “It's been estimated that about 90 percent of fieldbus problems are caused by wiring problems: the biggest causes are bad termination, interference and inadequate earthing.”

Such problems are especially critical in China. Chinese plant operators are not familiar with fieldbus, so two big projects – Tianjin ethylene and Zenghai ethylene—were originally specified for FOUNDATION Fieldbus (FF), but now may revert to traditional 4-20mA systems.


Starting up
One of the most baffling problems occurs during startup: Why don’t the instruments work? In many cases, it’s an installation problem. Simply put, the technicians didn’t set the segment terminators properly. All fieldbus segments need proper termination to prevent communications errors through uncontrolled signal reflections. Illustrated in Figure 1, the square “T” boxes indicate a properly terminated fieldbus segment.
It is unfortunate that many installation subcontractors pay little heed to the terminators and either forget them completely or simply install them everywhere, neither of which allows the segment to operate properly. Often, physical inspection of junction boxes and field enclosures is the only way to locate and correct the terminator position, which is a significant delay to the commissioning process.

Delays in commissioning cost money. One estimate, by a major oil company, says that time lost to a refinery unit being “down” can cost $400,000 per hour.

This problem can be avoided from the beginning, simply by specifying newer device couplers with automatic segment terminators, such as MooreHawke’s TRUNKGUARD. The automatic segment terminator greatly assists in segment commissioning by eliminating the issue of over or under terminating. If multiple field device couplers are used, the auto-terminator is always activated at the farthest unit, and automatically migrates up the segment if that line is disconnected.


Surviving shorts
Short circuits are a common problem in any fieldbus installation. Maintenance technicians can jostle cables; corrosion can weaken connections; and vibration from pumps and motors can loosen cables and connectors. Maintenance people must be concerned about what might happen to the entire fieldbus segment if any single instrument shorts out.

Older fieldbus designs rely upon a “current limiting” approach to short circuit protection. This technique limits the amount of power the short circuit can draw to between 40 and 90 mA, but it also holds that fault on the segment continuously. The additional current draw can deprive other instruments on the segment of the minimum 9V power they need to operate. When devices have less than 9V, they “drop off” the segment.

A typical fieldbus segment has 350mA at 25V available, enough to theoretically power up to 16 x 20 mA fieldbus devices. However, the system must also take into account voltage drop due to cable runs. For example, in a system with 10 devices and 4,000 ft of cable, each 20 mA device receives 10.52 V, which is fine. If a short occurs and a 60 mA load is locked in because of current limiting short circuit protection, this takes away enough power so the remaining devices receive less than 9V (8.39V), and some will drop off the segment. If two shorts occur, all the devices could drop off, and an entire process unit might go down.

Two methods are available to prevent this. The first is to allow a safety margin. That is, do not install as many instruments as the segment can theoretically power; instead, leave a certain number of spurs empty. This, of course, limits the number of devices that can be connected to every segment, but allows the segment to continue performing if a single fault occurs. Considering the high cost of fieldbus equipment, this is an expensive solution.

The other solution is to select device couplers that use “fold back” short circuit protection, where the fold-back circuit disconnects a shorted spur from the coupler. The fold-back technique has a logic circuit on each spur (see Figure 2) that detects a short in an instrument or spur, disconnects that spur from the segment, illuminates a red LED associated with that spur, and uses a trickle current to determine when the short goes away.
Total current consumed during a shorted condition is 4-5 mA. So, during a short, the voltage will actually go up. Using device couplers with this capability allows the maximum number of devices to be used in a segment, and allows users to make full use of fieldbus capabilities.

Cost is a major issue. A typical FOUNDATION fieldbus segment, consisting of an H1 card, power supply, device couplers and cables, can cost about $5,000. A large process plant may have hundreds if not thousands of devices. If the “safety margin” approach is used, where the entire capability of fieldbus is not used, the cost of all the extra fieldbus segments can become substantial.

For example, assuming that a typical fieldbus segment with modern fold-back protection can accommodate 16 x 20mA fieldbus devices, it requires 63 fieldbus segments to support 1,000 devices, at an approximate cost of $312,500. If a safety margin approach must be used because of current limiting protection, and each segment can now only accommodate 10 instruments, then 100 segments are needed, at an approximate cost of $500,000. Simply by specifying fold-back short circuit protection, an end user can save $188,000.



Removing & replacing
Maintenance people want to be able to remove devices from fieldbus segments in hazardous areas without turning off the whole segment, and without going through complex disconnection procedures and mechanical interlocks, if they can be avoided. Once again, these headaches can be avoided by specifying fieldbus device couplers properly.

In Division 1 (Zone 1) applications, simply specify a device coupler that has a magnetic interlock on each spur. The technician puts the key in the slot, which isolates the spur, and makes it accessible for rewiring without shutting down the segment. This works if IEC/AEx standards are being followed, since that particular device coupler can fit inside an Exe/AExe enclosure and spurs are fully accessible in Zone 1. Some device couplers are designed and approved for use in Zone 1 and Zone 2 with flameproof Exd devices.

For flameproof Division 1 applications, live de-mateable plug/socket combinations are available from many manufacturers. If an application demands live exposure in Division 1 or connection into Zone 0, then field barriers can be used which allow intrinsically-safe (IS) spurs to be attached to the non-intrinsically safe trunk.

Cost issues involve the amount of time a maintenance technician must spend removing and replacing instruments. If the process is laborious, it might take hours to follow all the safety procedures. If the process simply requires a key, then an instrument can be disconnected in a few seconds.

Intrinsically safe systems
Intrinsically safe systems are one way to allow access to instruments in hazardous areas. Intrinsic safety restricts the amount of available energy and remains safe even if the connecting wires are shorted accidentally in the hazardous location. IS instruments can be opened and adjusted as required, wiring connections can be made and unmade and even bare sensors can be exposed to explosive hazards, such as inside closed vessels and pipes.

IS systems such as FISCO are popular in Europe, primarily because of German influence on fieldbus installations, but have not caught on elsewhere in the world, primarily because of the extremely high cost. An IS system requires very expensive barriers and power supplies, and supports only a small number of fieldbus devices on a segment. It is not unusual to see an IS segment with only four or six fieldbus devices. FISCO also dramatically reduces the MTBF of the system and halves the allowable cable length.

An alternative split-architecture concept has become available in recent years, which offers all the benefits of IS at a fraction of the cost. Essentially, it splits the barrier into two pieces, one piece at the safe/hazardous interface and the other matching piece in the field in the device coupler (see Figure 3). This very efficient design allows a full 350 mA per segment and still maintains standard IS approvals at the device connections.

Cost savings are dramatic. Boehringer Ingelheim Chemicals, a pharmaceutical plant in Virginia, USA, converted its plant from IS to a split-architecture system, and found that it was able to install 17 fieldbus devices on a segment compared to only four devices on an IS segment. Not only that, the savings in enclosure space from eliminating barriers allowed it to install four DeltaV DCS controllers in cabinets that previously held only one with the IS system.


Delivering redundancy
Most plants have critical components whose failure will shut down whole processes. The classic solution is to install redundant instrumentation and controls. With most fieldbus products, the only way to build a redundant system is to duplicate the entire segment completely, including H1 interfaces, power supplies, device couplers, fieldbus instruments, wiring from instruments to the couplers, and the fieldbus cables. This is illustrated in Figure 4.

But duplicating every component, from H1 cards to field instruments, can be a very expensive solution. Not only does such a system require twice as much fieldbus hardware, but the cost of duplicating complex field instruments, such as flow and pressure transmitters, is very high.

Also, conventional redundant systems don’t always work very well. This is because control system vendors have no way to determine if a segment has failed. Typically, a fieldbus host can only detect the failure of an H1 card. Therefore, conventional redundant fieldbus systems have to use complex software voting schemes that determine – by analyzing instrument signals (or the lack thereof) – that the segment has failed. Needless to say, such redundancy schemes are expensive, complex, and can be hard to maintain. A catastrophic process failure could result while the control system is determining what’s wrong.

Newer fieldbus systems, such as TRUNKSAFE, consist of dual, redundant Fieldbus Power Conditioners (one for each leg of the segment), two fieldbus cables, and a fault-tolerant Device Coupler (Figure 5). There is no need to duplicate the field instruments unless the user is concerned that a particular instrument is prone to failure.

TRUNKSAFE provides two legs, both of which are normally active in the segment. In the event of a failure in one leg, TRUNKSAFE makes it possible for the control system to maintain normal communications through the alternate path, with no special software or DCS reconfiguration required. This helps prevent the loss of a complete segment, the shutdown of associated plant or equipment, and potential catastrophic process failures.

As far as cost is concerned, a system like TRUNKSAFE costs about half or a third of a conventional redundant system (not counting the extra instrumentation). However, it is only about 10 percent more than a non-redundant fieldbus segment. This makes it economically possible for end users to install redundant systems on many more segments than with a conventional redundant approach.

No more headaches
Fieldbus is an exciting technology. Maintenance headaches, which have irritated users for the last decade, now have solutions. To avoid maintenance problems with your new fieldbus system, consider all the options and specify the correct hardware at the beginning.

It is important for end users to realize that the DCS vendors do not provide most of the fieldbus hardware. In a FOUNDATION Fieldbus system, for example, the DCS vendors usually make only the H1 cards. The rest of the fieldbus hardware, including power supplies and device couplers, are supplied by approved vendors, such as MooreHawke, Relcom and P+F.

Therefore, end users have the right to ask major DCS vendors to consider modern hardware and alternate solutions, instead of accepting ten-year-old hardware that creates maintenance and operational headaches.


Based in Shanghai, Desmond Ho is Fieldbus Applications Consultant for MooreHawke, a division of Moore Industries-Internationa (www.miinet.com).

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5