Programmable logic controllers (PLCs) are industrial computers that:
- Monitor and control industrial-automation applications
- Execute tasks related to test and measurement operations
- Perform process-type functions (including those related to HVAC systems) beyond the scope of this article.
PLCs receive data from sensors and input devices, process the data to make logic-based decisions, and output control instructions to mechanical or electrical systems. They are a type of embedded system that combines computer processor and memory with input-output (IO) devices — much like the hardwired relay-based logic as well as PC-based logic with which they compete.
In terms of physical form, PLCs today can be anything from a very simple computer having an integrated chip (IC) morphology to a large rack-mounted collection of controller subcomponents housed in multiple chassis. Simpler microcontroller-based PLCs or those taking the form of system on a chip (SoC) PLCs can be extremely reliable and operate off of very modest power input. In contrast, the most complex PLCs blur the boundaries between what constitutes a PLC and general-purpose computers for real-time industrial control … although reliability and real-time performance are still emphasized for the former.
Originally, PLCs were meant to directly replace hard-wired control logic based on relays and drum sequencers. These early PLCs only had to perform basic operations by transforming inputs into outputs. Any machine tasks necessitating proportional-integral-derivative (PID) control were outsourced to attached analog electronics. Now PID controls and even more sophisticated operations are a standard part of PLC instruction sets.
In fact, the functions expected of PLCs have proliferated over time — so that today, many PLCs are quite sophisticated and able to execute complicated and adaptive routines. The ever-increasing power and shrinking size of semiconductor chips (thanks to Moore’s law) have enabled unprecedented intelligence from smaller controllers. This trend is continuing with integrated support of motion control, vision systems, and communication protocols. At the other end of the PLC size spectrum, some programmable automation controllers (PACs) integrate a PLC with a PC to replace PLCs and proprietary control systems (run off proprietary programming languages) for certain applications. More PLCs today are also being integrated into human-machine interfaces (HMIs).
The industrial digital environment in which PLCs operate
Industrial automation today relies on machine feedback and operations data along with complex interconnections between digital devices to:
- Control digital devices.
- Run advanced capabilities — such as those related to IIoT connectivity and machine reconfigurability, for example.
- Enable human decision-making about various machine and operational conditions.
- Improve overall productivity and workpiece quality.
Such automated installations include disparate information systems to store, process, and serve this data.
Material requirements planning or manufacturing resource planning (MRP) systems provide production planning, scheduling, finance, and inventory control. In contrast, historian systems store time-series data from sensors and instruments for graphical plotting to help operators and management systems understand and process automation trends. Statistical process control (SPC) is one historian application.
Human machine interfaces (HMIs) are machine control panels (or modules that wirelessly connect to mobile devices) that let human operators view data and issue commands. Closely related to HMI functions are supervisory control and data acquisition (SCADA) systems that enable real-time control and monitoring of interactions between automated machines with their HMIs and historians. Using SCADA, an HMI can control multiple machines … and display data related to multiple devices.
Manufacturing execution systems (MESs) include functions such as operation scheduling and data collection. In some ways, it can be seen as coming between and overlapping with MRP and SCADA.
Enterprise resource planning (ERP) systems integrate manufacturing-related MRP, MES, product lifecycle management (PLM), and CRM information systems. ERP systems may be monolithic software suites which handle all of these functions … or a core ERP system that interfaces with specialized applications from multiple vendors. Typically, only the top management interacts with the ERP — and most personnel in a given organization interact with one of the component systems feeding into it.
PLCs typically operate at a level below these information systems. They send and receive information to and from machines, motors, and sensors. They may also interact with the information level above, sending data to the historian or SCADA, or receiving control inputs from the SCADA or HMI. More sophisticated PLCs can also perform SCADA and historian functions … and even HMI functions in an increasing number of cases.
Figure 1: PLCs typically operate at a level below the information systems of automation. (Image source: Jody Muelaner)
Note that PLCs aren’t just involved in automation: They’re also employed in controlling testbench (product development) and laboratory-measurement tasks.
- As described above, automation generally emphasizes diagnostics and requires deterministic real-time operation from the PLC for real effectiveness.
- In contrast, PLCs employed in measurement tasks place more emphasis on quickly and precisely executing measurement collection and other forms of data acquisition.
For machine automation tasks, PLCs rely on real-time processing in which the delay between an input and the response output is measured in milliseconds. A real-time operating system (RTOS) is required for all but the simplest of PLC functions. While many PLCs still use proprietary operating systems, there is increasing interest in open standard operating systems. Case in point: VxWorks is a proprietary RTOS which is widely licensed for use in industrial control. It is used by several leading robot manufacturers including Kuka and ABB. Or an open-source variation is FreeRTOS freely distributed under an MIT open source license. FreeRTOS includes various internet of things (IoT) libraries for a wide array of automated applications. Read more about this option at the Digi-Key article Connect Designs Quickly and Securely to the Cloud Using Amazon FreeRTOS.
For test and measurement tasks, PLCs rely on real-time processing in which the delay between field-device measurements and its collection is measured in milliseconds. Gone are the days when engineers had no choice but to employ interface converters and systems of transfer channels. Now smart peripheral devices and I/O assemblies have advanced and simplified signal collection via digital and analog inputs.
Today’s engineers also have more options based on standardized interfaces and cross-manufacturer compatibility of components that can serve as interoperable components.
Just consider I/O components with integrated PLC functionality. These are compatible with configurable HMIs running Windows or Linux operating systems and having Ethernet connectivity — but lacking easy recalibration options or analog I/O for field devices that generate low-voltage analog signals. Such I/O components also work with PLCs set up to collect data from remote I/O devices … and directly from sensors via their own onboard I/O.
Figure 2: T7 multifunction data-acquisition devices (DAQs) include Ethernet, USB, wifi, and Modbus connectivity to work with a wide array of field devices as well as industrial HMIs and PLCs. Modbus/TCP connectivity, in particular, imparts controllability via various third-party software and hardware options for openness and flexibility — which in turn gives industrial-system architects as well as research and development (R&D) engineers vendor-neutral choices for their data collection and automation applications. (Image source: LabJack)
Of course, PLCs aren’t the only option for machine automation or test and measurement. As all industrial controls have become more complex, some vendors have come to differentiate certain hardware as programmable automation controllers (PACs) to signify enhanced capabilities … and in many cases, multiple processors on a single piece of hardware. In reality, PLCs have also seen increasing sophistication — so there’s no hard-and-fast rule about when some hardware that executes PLC functions constitutes a PAC. Most PACs integrate PLC and PC aspects to serve as complex automation systems featuring multiple PC-based applications as well as an HMI and historian. One clear differentiation is that PACs are easier for developers to employ, as PACs have a more open architecture than traditional controls.
Yet another option today are modular PLCs. These consist of modules that perform different functions. All PLCs must include a CPU module which includes the processor and memory for the operating system and program. There may be a separate power supply module and additional input/output (I/O) modules. A PLC may include both digital and analog I/O modules. Another module may be required for network communications.
The PLC may be either integrated — with all modules in a single enclosure — or modular. Integrated PLCs are more compact but modular PLCs are more versatile, typically allowing multiple modules to be easily connected together either by plugging directly into one another or by using a common rack as a bus. Modules are addressed according to their position on the bus. Although the physical support aspect of the rack may conform to a standard such as DIN, the data bus is typically proprietary to the PLC manufacturer.
The role of PLCs in IoT
As interest in Industry 4.0 (also called the IIoT) grows, industrial users increasingly expect to have the option to connect their industrial controllers to company networks using internet protocols. This means communication using the Transmission Control Protocol (TCP) and the Internet Protocol (IP) or simply TCP/IP. However, the IIoT trend isn’t just about using internet protocols … it’s also about machine learning and big data. As PLCs become more powerful (and more advanced controls render PLC functions a feature), more host functions such as vision systems. Internet connectivity also allows engineers (via system PLCs) to leverage cloud-based algorithms for the processing of extremely large datasets — also called big data — for machine learning.
In practical applications, Ethernet for Control Automation Technology (EtherCAT) excels for such IIoT PLC functionality. This is a communications protocol based on Ethernet suitable for real-time control applications with cycle times of less than 0.1 msec — the fastest industrial Ethernet technology with the ability to synchronize with nanosecond accuracy. Another important advantage is the flexibility of the EtherCAT network topology that works sans network hubs and switches. Devices can simply chain together in a ring, line, star, or tree configuration. PROFINET is a competing standard offering similar capabilities.
The current trend towards increasingly sophisticated data collection and industrial control will continue. That means PLCs for industrial automation and test and measurement will increasingly resemble PACs … and integrate with SCADA and historians. Internet protocols and open standards such as EtherCAT are also seeing steady adoption for PLC communications. Such connectivity will in turn spur greater use of Industry 4.0 technologies such as big data analytics and machine learning partly facilitated by the ability to distribute required processing power and memory to:
- Cloud-based computing
- Edge devices capable of data processing
In addition to these trends, there will remain a need for more traditional PLCs to execute relatively simple test and measurement as well as control functions with maximum reliability and energy efficiency.