Examining field instruments under the ‘fieldbus microscope’

01 September 2007

New field monitoring capabilities in Profibus provide more exact field tests. They help end users verify vendor claims and ensure a new instrument performs well.

How do you know if a new field instrument is working correctly? How can you be sure it will perform properly
in your application? End users and instrument vendors use field trials to test new instruments; however, the choices of test methods are limited. For example, you can install a new instrument alongside existing
technology and compare the readings from the two devices. Or you can do periodic manual checks. If you have a chart recorder, you can verify the instrument is operating correctly at that moment.

These periodic checks provide some data, but there is no way to really closely and continuously monitor the
instrument’s performance. There is no way to tell what is happening between the check points.

With the introduction of fieldbus technology, and Profibus in particular, this has changed. It is possible to
monitor field instruments like never before.

Millisecond response
When an instrument uses a traditional 4-20 mA or even HART interface, there is a built-in time filter because of the speed of the response of the interface. The response is relatively slow (on the order of seconds). If something happens in the interval between checks, it is not detected and you simply do not see it.

Much depends on the duration of the error. A loss of signal would be detected if it lasted for a few seconds or more, but momentary lapses would not be detected. With these devices, the status information is sparse with long periods between the data readings. Just because you can’t see the errors doesn’t mean they aren’t occurring.

By contrast, Profibus reports errors fast: the status byte is read every bus scan. While the typical scan is every 200 milliseconds, the status byte can be updated as often as every 20 milliseconds. You can monitor the instrument’s, performance and see what is going on like never before. It’s like the difference between normal vision and looking at something through a powerful microscope. Everything is bigger and sharper, with far more details in evidence.

Monitoring a device
The output from a Profibus PA device is five bytes. The first four bytes, bytes 0 to 3, are a floating point representation of the process variable. Byte 4 is the status byte, using status codes defined by the Profibus standard (see figure 2).

This output information is sent to the master device (PLC, DCS, or controller) every bus scan and is referred to as cyclic data since they are updated every cycle. In addition to these data, the master device has a chance every cycle to have a ‘special’ conversation with one of its slaves. This special conversation can take
several bus scans to complete and the data the master receives are referred to as acyclic data. This type of
communication is typically used to transmit configuration data and advanced diagnostics.

Profibus instruments have always supported acyclic data communications from a ‘Class 2 master.’ Class 2 Profibus term for an engineering station— configuration software such as Simatic Process Device Manager, for example. Traditional ‘Class 1’ masters such as a PLC or DCS now also support this function.

The impact
At the outset, for full Profibus compatibility, instruments require additional intelligence in the initial design phases so they will have the diagnostic ‘intelligence’ to handle the increased update speeds. They need to
determine the type of error and assess its relevance or importance. This also needs to be configurable by the end user.

The ability to assess instrument performance to such a fine degree has big implications for vendors and for
product development. Being able to detect even minute lapses in performance means researchers and
developers have the data they need to fine-tune devices and their software to a higher level. This promises to enhance overall product quality and performance and will, of course, ultimately benefit end users.

As for the end users, being able to look at instruments through this ‘microscope’ has its advantages and its
challenges. The main advantage is you now can know the instant your instrument is having a problem. You
have a far more accurate picture of the accuracy of your data. You also now have to manage this information—this is the challenge!

With 4-20mA, an end user was blissfully ignorant of what was going on inside the instrument. Now, with
Profibus, you are going to see things that never came up before. It is up to the end user to decide how much they want to know. During setup, you can decide when an error is a problem or not, and adjust the
‘magnification’ of your ‘microscope’ to the appropriate level of detail.

In some processes, you want to know if an instrument has even a momentary problem such as loss of signal. In other processes, momentary lapses do not matter. You can now adjust the settings at the instrument end and monitor it to that level at the system end. As an end user, you really have a lot of power to
configure the instrument to match your application and gather meaningful data.

It should be noted that the ability for the end user to decide when an error is a problem or not was added to the Profibus Profile specification in 2005 (Profile specification Version 3.01). The Sitrans Probe LU was one of the first instruments to have this functionality. Siemens is in the process of adding this functionality to most of its Profibus PA instruments.

A field test case study
Researchers at Siemens Milltronics recently employed this aspect during the field testing of their new instrument, the Sitrans Probe LU Profibus ultrasonic transmitter. The testing process serves as a good example of what is now possible with rapid fieldbus communications.

This transmitter had successfully passed and exceeded all lab tests, and was ready for field trials that would
particularly focus on checking the instrument’s new diagnostic capabilities.

The test sites we used varied in size. The smallest one had two instruments and an update rate of 50ms and the largest had 12 instruments and an update rate of 240ms. With speeds this fast, even instantaneous glitches can be detected.

The main requirement was to monitor the status byte every bus scan and log any changes in value. In addition, when it changed, we initiated an acyclic read to determine why it changed. The acyclic read gathered up this additional information.

We monitored all communication errors or power cycles. We recorded the main process variable and internal temperature every five minutes, using an acyclic read.

To conduct the test, we used a Siemens Simatic S7-300 PLC as the Profibus master because it supports acyclic data and it can handle lots of data.

The fact that both the PLC and Probe LU support acyclic reads was critical in this project. For an end user who has a Simatic PCS7 DCS with the Maintenance Station option, he could readily monitor all these data without writing new application code.

The test program
The program was designed to sequence the acyclic reads one at a time. This ensured we would never exceed the maximum number of acyclic connections that the CPU can handle.

For the data logging, the memory of the S7-300 is divided up into working and loaded memory. Working memory is for the program and current data. Loaded memory is for data storage. We moved the data blocks from working memory to loaded memory as soon as they were full to avoid overloading the working memory. The S7-300 with a 12 megabyte memory card was easily able to handle this.

We tested the program extensively inhouse before conducting field validation, verifying that each function of the program worked correctly. In the field test, we were able to use the PLC fault table as a backup monitor as a cross-check to ensure nothing was missed.

Both the program and Probe LU performed extremely well during the trials. We observed several ‘errors’ but
the acyclic data showed they could be traced back to a process event.

An example is one test unit that was mounted over a spillway. During a large thunderstorm, this spillway overflowed from two directions, creating splashing. During this event, the device reported a rate of level change that exceeded the normal process rate of change. This was the correct thing to do. With the date
stamp and trend information, we were able to detect and verify this.

Our experience
The experience of this field trial shows in practical terms the value of close monitoring through Profibus PA acyclic communications. It provided very detailed additional information about the behaviour of a field device. In this case, it verified the reliability of the Sitrans Probe LU Profibus.

With modern fieldbus technology, a user can see what is happening in a field device like never before. How far you want to go in monitoring a device is really now up to you. This added monitoring capability will help
vendors test their products more closely before release and should, ultimately, encourage higher product quality. For end users, it provides an opportunity to put new instruments through a rigorous test to verify vendor claims and really help you judge if the new device is appropriate for your application.

—James Powell, Siemens Milltronics Process Instruments Inc., Peterborough, Ontario, Canada

Contact Details and Archive...

Related Articles...

Additional Information...

Print this page | E-mail this page