Control signals – which to choose?

It is about 20 years since industry ‘experts’ stated that 4-20mA would disappear being replaced by digital networks or fieldbus solutions.  While there is some truth that networked systems have now taken a portion of market share, 4-20mA is still in use. There are two main reasons for this; firstly that higher-priced intelligent sensors, made for specific networks are not always the most cost-effective option, nor are the ‘smart’ features on offer from intelligent field devices always wanted or required.  

The process control industry has also changed significantly, with much more automation being installed in smaller plants, like food processing, water treatment, remote monitoring, etc.  It could be said that 4-20mA was the last truly universal signal, accepted by every system, every vendor, and every control component and this means that single control loops, small plants and transportable skid-type equipment will always have a need for the simple, low-cost solution that 4-20mA offers. 

Using just two wires to transmit the signal, 4-20mA can power the field transmitter with its own supply voltage and can run for long distances on standard cable. Using the live zero (4mA instead of 0mA) it senses a cable break or instrument failure, it is not overly susceptible to noise and is highly accurate and low-cost.  The downside is that two wires can only carry one variable or instrument.  

However, as the number of control loops increases, the costs start to multiply, as you need individual cables, glands, terminals and inputs for every element.  As well as the hardware, the space it requires and most importantly, labour to design, install and commission, which makes it uncompetitive versus fieldbus networks in applications with many sensors and feedback devices.  It is different for every installation, but there is a point where the change from an analogue to a digital network becomes more cost effective.

Fieldbus

With fieldbus networks, it is possible to have thousands of inputs and outputs on a single cable.  This reduces the infrastructure needed for a traditional system and can save up to 50% of project costs, even though the actual hardware becomes more expensive. By using smart field equipment, the entire transmission of the signal is digitally accurate and generally there is configuration and often diagnostic data available. For large plant design, especially for heavy industry like petrochemical for example, this methodology is ideal.  

The downside is the number of choices that a process designer has to make when designing these systems; the first being which network to use as it is often not easy to mix networks on site. While the most popular networks used are not proprietary, they are still not available from all vendors, which starts to limit choices in components, even though they may be better suited to the application. For these reasons, it is rare to see a site that utilises only digital networks, however almost every larger site does use fieldbus technology. This leads us to the most common installation, the hybrid system.

The hybrid system incorporates both technologies. The 4-20mA is used in the field, which keeps the components generic and lower cost, and then the signal is fed into a field I/O panel where these signals are clustered and then transmitted to the central control via fieldbus. Without doubt, this system is the lowest cost to implement and simple to understand and maintain. The only downside for a network that is not fully digital is that it is difficult to get configuration or diagnostic information from the field instrument. The benefit however, is that the instrument itself will be considerably lower cost.

Imagine a single flow control loop; for example, is the water feed to another process that needs to be maintained at the same rate. A flow meter based on the application requirements, with 4-20mA output would be chosen. To control the flow a modulating control valve with 4-20mA input would be specified. But to close the loop a Proportional Integral Derivative (PID) controller between these two instruments is needed that can be configured to the application; this will need a 4-20mA input and output.  

With, for example, 15 of these loops, all identical, and the plant manager wants to monitor and record the flows in the system (assume a PLC & SCADA layer).  Using 4-20mA, it would be necessary to send 15 signals to the PLC, plus 15 signals out, needing 6-8 input cards and as many as 90 terminals, 30 glands, cable tray, panel space, power supplies, signal isolation, etc. This option now starts to become more expensive.
 

With a fully digital solution – for example with a fieldbus protocol such as Profibus, it is necessary to install a Profibus master at the PLC (unless it is incorporated) and then choose a Profibus-enabled  flow meter and control valve.  So now we have one slot in the PLC rack, one gland and one cable to all the devices. The PLC can now read the process signal and control the loop as well as log the data, via the 30 fieldbus nodes.  

The only downside is that the field equipment is more expensive, sometimes considerably, and specific to Profibus, so if the plant changed to Ethernet in the future, the field equipment would have to be discarded or a customised network-network gateway would be required. This design is far more complex and may require highly qualified maintenance engineers.

The hybrid solution would be to use the original lower-cost field equipment, but to wire them directly to a local (field) I/O panel that converts them to Profibus and transmits them to the PLC. This offers the cost savings of the installation, with lower cost field instrumentation, even allowing for the I/O cluster hardware.  For this example the hybrid design provides the most cost effective solution.

If a customer came to Bürkert with this application at the design stage we would choose flow meters with 4-20mA output and feed these directly into our modulating valves with Profibus enabled positioner and optional inbuilt PID controller. This would control the flows directly as well as transmit process, output, configuration and even diagnostic data to the PLC, via just 15 fieldbus nodes. This ‘solution, although being partly hybrid, would give all the functionality of the fully digital system, yet is around 30% lower cost than the cheapest of the other options. 

There is no perfect fit for all situations. There are many considerations that weigh into the correct selection of the site signal wiring, with cost being just one. It is advisable to always seek competent advice and develop a tool that will estimate the total cost of the installation, including hardware, software and labour. Then consider whether it will be easy to commission and maintain, and whether there is available competence on site for fieldbus technology.