New in’s and out’s of I/O connections

By Control Engineering Staff August 25, 2005

I/O modules are one of the connective fabrics of automation and controls, linking sensors that touch the system with logic that makes a decision about the process. Control Engineering spoke recently with several manufacturers of I/O modules, about the fundamentals of I/O connections. For more on the topic, look for a September 2005 issue ‘Back to Basics.’

Wago Corp. , Mike Giroux, senior support engineer, explains that every I/O system has two things in common, the CPU (central processing unit or brain) and I/O (input/output). ‘Every person uses inputs and outputs every day. An example would be if you touch something hot with your hand (input), the nerves in your hand send a signal to your brain (CPU). Your brain then processes the data and tells the muscles (output) in your hand to pull away.’ For process controls and automation, I/O devices transmit data. The term ‘I/O’ is used to describe sensors, input, and actuators, output, Giroux says. ‘Without I/O devices, everything in the process control/automation world would either go out of control or come to a complete stop.’ I/O devices are then broken down into digital and analog versions. There are four categories of I/O: digital input, digital output, analog input, and analog output. Digital inputs and outputs have two states: on or off. Examples of digital input devices are photoeyes, proximity switches, and limit switches, usually used to detect the presence or absence of something. Digital outputs are devices such as solenoid valves, lights, and alarms. The term analog indicates that the device is variable in some form. Analog input devices include pressure sensors and flowmeters. Sensor signals typically are 0-20 mA, 4-20 mA, 0-10 V, orxplains.

Traditionally, sensors and actuators often have been wired to the main controller, generally a PLC, with many terminations, complicating start-up and troubleshooting, Giroux says. Industrial network communications make wiring I/O devices to the main controller much easier, allowing data transfer over one cable, via DeviceNet, Profibus, various Ethernet protocols, and others. ‘These allow for additional diagnostic information to be sent back to the controller, enabling the operator to determine what the problem may be and where the problem is located. In many cases, the sensor or actuator can tell the controller that it needs to be replaced,’ Giroux says.

Helge Hornis, intelligent systems manager, Pepperl+Fuchs , pointed out that looking at the history of I/O modules can help to understand today’s needs. First-generation PLCs used I/O cards that resided in the PLC rack. As such, the number of I/O connections was limited by the size of the rack (how many open slots were available) and the card’s I/O density (how many I/O connections could be made to each card.) Later, Hornis says, as use of PLCs increased, amount and type of I/O cards changed. At this point, it became impractical to offer larger racks (allowing more I/O to be used). This problem was solved by offering remote racks, allowing users to add many more I/O connections along a production line, reducing the length of cable needed to connect sensor and other end devices to a PLC. Remote racks communicated with the main PLC rack using proprietary and semi-open protocols. Typical protocols, Hornis says, are Allen-Bradley Remote I/O and Modicon Modbus. Distributed I/O modules represented the next step, further splitting up I/O points. Instead of having many I/O cards in a remote rack, a number of I/O points can be directly field mounted. Field blocks act like mini remote racks. Mostly open, Hornis says, but heavily pushed by specific PLC manufacturers, a main disadvantage is that ‘intelligent devices,’ such as drives, scales, vision systems, and RFID systems can be connected to the network. Since the device would normally exchange many bytes of data, a directed connection to I/O cards (possible in principle) is not practical. For more efficient communications, these devices exchange several bytes of data at a time. Many networks are available, including Profibus and DeviceNet.

Highly distributed I/O points are the latest step in distribution, defined by making I/O nodes smaller and less costly, Hornis continued. Offering a highly distributed I/O structure further limits the cable length between sensors and I/O nodes. These communicate with a PLC through scanner cards (in the PLC rack) or via a gateway to a distributed I/O network. Since I/O nodes are designed to be highly distributed, the number of data bits exchanged is kept intentionally small (usually under 1 byte). This makes them very fast. Collecting these small parts and assembling them into a larger data packet to be transported by a distributed I/O solution makes sense, since they are efficient (sending more than a few bits at a time). This can be compared to transporting people with a motorcycle (very fast but only for 2 people) or a bus (slower but for many people). Since control systems require efficiency and throughput, the ‘bus’ will probably get a lot more people through per unit of time. A typical implementation of a highly distributed I/O solution is AS-interface.

Jerry Penick, Rockwell Automation marketing manager for distributed I/O products, said today’s I/O products combine functionality, many times eliminating the need for terminal blocks for end-point connections. Distributing I/O modules can save on wiring and labor, Penick said, in the case of On-Machine applications doing without enclosures, with even more savings. On-Machine I/O’s modular design and quick connection cables and cordsets enhance ability to do repairs and allow very quick maintenance. Cost savings start in the design phase, when designing the machine or line, an excellent way to get more out of the existing workforce, with a future-proof set of technologies, he explained.

—Mark T. Hoske, editor-in-chief, Control Engineering,