For fieldbus reliability, prevention is better than a cure

01 September 2006

A planned monitoring regime to diagnose problems before the cause failure can keep productivity up and costs down.

Foundation Fieldbus (FF) has had major success across the world as a digital, two-way communications
system. Part of this success is based on its renowned reliability—a critical requirement for any process plant that has to run constantly.

But even FF is not infallible. The most common source of problems is at the physical layer. These components include the cable, wiring components, power supplies, conditioners and terminators.

Planned monitoring is one way of detecting problems before they result in failure and loss of production. However, implementing the best planned monitoring scheme is not easy, and even when it is in place, interpreting the data in order to track down the actual fault is difficult and time-consuming.

Understanding how fieldbus communicates and where failures are likely to occur makes the process less
complicated. And an informed choice of measurement parameters, test instrumentation and diagnostic
techniques further accelerates fault resolution.

Finding faults
Routine inspection of a control network can reveal problems before they turn into failures that cause expensive downtime. Unfortunately, FF is a complex system, so the list of possible faults in the physical layer uncovered during this inspection can be long.

However, adopting a methodical approach and grouping faults depending on the similarities in the symptoms simplifies the diagnosis.

For example, the fault may appear as electrical noise disrupting communication on the fieldbus. This could be due to a group of problems such as wiring running close to a frequency converter, or because the cable ground is incorrectly designed or implemented. Other common faults that don’t cause noise can be instantly discounted.

Consequently, routine inspection must take into account predetermined parameters, such as DC voltages and minimum signal levels, that give a good initial indication of the cause of any faults detected. And if the parameters are simple to measure operator training is easier (possibly to the extent that nontechnical staff can do the job).

Choosing measurement parameters While there are hundreds of possible fault modes for FF, most problems fall within five common categories: Short- circuit(s) between the shield and signal conductor; broken contacts; electrical noise; incorrect number of terminators, or a fault in the power supply and/or its conditioner. Measuring parameters that indicate one of these types of faults can narrow down the cause of the fault (see Table 1).

Detecting and interpreting a short-circuit is relatively straightforward. Progressive measurements can then lead to the actual location of the fault.

The FF specification dictates a minimum signal level. The signal level is simple to measure and any measurement below the specification indicates a fault with the particular instrument under observation, or a defective signal conditioner.

A minimum value for DC voltage is also detailed in the FF specification. By carrying out several DC measurements at different points in a segment it is possible to locate the cause of the fault. The measurements can be carried out using a simple multimeter.

Again, the FF specification indicates a maximum permissible value for noise that, if exceeded, may help
determine the source of the fault.

The number of retransmissions to a particular instrument is also a good parameter to measure. This measurement is slightly more complicated than the previous examples, but it is extremely useful for fault diagnosis. Some control systems already measure the number of message repeats and this gives a good indication of the ‘health’ of the physical layer. Lots of retransmissions to a specific instrument are an indication that the connection to that instrument is poor.

Some of these diagnostic parameters are based on the requirements laid down in the FF physical layer
specifications. Many other parameters from this specification can be measured and compared with the specified values and used as fault indicators. Examples of these include digital signal jitter, slew rate, rise-times and fall-times.

However, many of these parameters are difficult to measure, require sufficient knowledge for correct interpretation and do not clearly indicate what action should be taken to determine the cause of the problem. They are best reserved for an FF specialist.

Taking measurements
After determining which parameters should be measured, it should next be decided how, where, and with what frequency they should be carried out.

One option is to constantly monitor the parameters. This is achieved by connecting a diagnostic measuring
instrument to the fieldbus network at an appropriate point. The remote measuring instrument then reports the diagnostic parameters to a computer for storage (which is useful for later trend analysis), reproduction, and triggering of alarms. This could be sent to the computer via a separate network or via the fieldbus itself.
(Note that an FF registered instrument is needed to send the information via the fieldbus.)

One advantage of this method is that operators are immediately alerted when a problem occurs. The monitoring system is configured so that alarms are triggered at levels determined by the user. An out-ofrange signal triggers the alarm in good time for the problem to be corrected before failure occurs.

But there are disadvantages to constant monitoring. First, today’s control systems don’t have a facility to store and recover diagnostic information from test equipment. Second, if the information is being transmitted across the fieldbus network the data could well be compromised by the very fault being monitored. Third, the tester acts as a station on the network and adds to the bus traffic.

And finally, continuous monitoring is relatively expensive. Extra hardware is required, including a module to display measurements. If the communication takes place via the fieldbus network, a licence is required further increasing the price of the diagnostic module.

Periodic measurements
An alternative to continuous monitoring is periodic measurement. With this regime, measurements are carried out by means of a portable tester at selected points around the plant. If a problem is detected, immediate action can be taken. Data can be collected and stored by the instrument avoiding the need to
remotely communicate with a central computer. This allows trends to be determined by comparison with reference values taken during commissioning.

Periodic measurement has the advantage in that it allows operators to use a portable instrument. In addition, it provides the added flexibility of carrying out measurements anywhere on the network, making it easier to track down the location of the fault.

However, there are some downsides. For example, the measurements can’t be carried out remotely and continuous data aren’t available – reducing the amount of diagnostic information. And to archive the data, they must be transferred to a computer.

This last disadvantage can be overcome by using the new hand-held tester from MTL-RELCOM, the FBT-6 ‘Fieldbus Monitor’ (photo, p37). The tester boasts a built-in memory card to store measurements and a connection to interfacing with a PC, eliminating possible errors when transferring data to a PC.

Building-in fault tolerance
Selecting the correct monitoring regime for a fieldbus network is a difficult decision. The costs of monitoring must be balanced against the loss of productivity and damage that may be caused if the network fails. For example, critical network elements benefit from continuous monitoring, with immediate reporting; while in the less critical areas, regular preventive field measurements, or even reactive fault-finding after failure may be the most cost-effective methods.

Regardless of the choice of monitoring method, diagnostic problems can prove to be very difficult, time-consuming, expensive and sometimes frustrating to solve. Experience shows that the best method of minimising problems is to ensure to start with a correctly installed network. It is also sensible to choose instruments that are on the FF list of registered devices.

Moreover, good quality cable and wiring components are just as important as correct cable laying. And various tests must be carried out before commissioning to ensure the network is operating satisfactorily. Recording the measurement information during commissioning is useful as a reference for future fault-finding and trend analysis.

In the commissioning phase it is advisable to connect the instruments one by one; if a problem occurs it is then much easier to identify and fix. Ensuring the network is installed properly and fully commissioned is the best way to build-in fault tolerance and reduce the time and cost of preventative maintenance.

Enquiry Code S230

—John van Leeuwen is with MTL Instruments BV

Contact Details and Archive...

Related Articles...

Additional Information...

Print this page | E-mail this page