By now it is clear that manufacturing is getting digitized – meaning that an increasing amount of data is being collected during any manufacturing process and this data is increasingly being analyzed to optimize efficiency, utilization, and even to generate new streams of business.
As market research firm ARC says, “The move to digitization has been sold, at least in part, based on the possibility of significantly improving or transforming the business. About half of the respondents see opportunities for new business models and revenue streams, as well as opportunities for improving business responsiveness and agility.” 1
This digitization goes by many names – industrial IoT, Industrie 4.0 (a term coined in Germany), and industrial internet (favored by GE). But, in large part, these refer to the same promise of making manufacturing smarter and more agile by imbibing within the flow the digital smarts that we so take for granted in other spheres. In fact, just recently, Industrial Internet Consortium and Platform Industrie 4.0 have started working collaboratively. In late 2017 they published a joint whitepaper that details how their reference architectures are aligned.2
Far more important than the name is the fact that there is a massive amount of data in a manufacturing process that can be harnessed to achieve very substantial goals – predict faults, optimize equipment lifetimes, derive new revenue streams, and even optimize the production process to better align with market needs. But, first, all of this data needs to be captured and then it must be brought out to be analyzed. Finally, there must be a feedback path to optimize the edge devices and the controllers to fine-tune the manufacturing process in response to the data analysis.
Collecting, formatting, and using this vast amount of data is driving a sea change in industrial automation system design. At the most basic level we are witnessing dramatic miniaturization coupled with increasing levels of edge processing across the spectrum of industrial automation systems. This has implications on system architectures as well as the component specifications within next-generation industrial automation systems. This article looks at some of these system design trends and also shows examples of how some of this miniaturization coupled with increasing intelligent processing can be achieved.
Development and availability of the next generation of high-performance, yet small and rugged automation systems is the key to building out or upgrading to a digital manufacturing facility – or, in other words, realizing the industrial IoT vision.
Industrial IoT Demands Tiny, Connected Sensors
Data collected from the edge sensors and from the controllers is the lifeblood of any digital factory. No existing assembly lines will be replaced – instead, edge sensors that are small, smart, and connected will have to be developed to fit within the existing assembly line.
As shown in Figure 1, small sensor systems are required that will have to not only collect the information but also do some processing in real time to clean the data that must be delivered via standard communication links. Also, these sensors will have to be tiny so that they can unobtrusively fit within the existing manufacturing flow.3
Small sensor systems collect data and also perform real-time data processing (Photo courtesy of SICK AG).
This steady downward mark of a system’s form factor has been going on for many years now and is accelerating to build out the vision of the industrial IoT. In fact, we find that industrial sensors used in various manufacturing facilities have steadily shrunk in size even as the functionality has gone up. Figure 2 shows how even industrial light curtains have evolved over the last 50 years.
Evolution of industrial light curtains (Photo courtesy of SICK AG).
This importance of shrinking industrial sensors to be able to accommodate narrow assemblies or even so that they can be integrated into tiny valves and actuators is echoed by different manufacturers. Balluff highlights the advantage of their small sensor size plus high performance here: “The increasing miniaturization of assemblies demands the smallest possible yet still high-performance components. Balluff mini sensors meet these requirements. With small dimensions and top performance, they offer a great degree of freedom in design and make possible considerably more applications.” 4 Figure 3 shows photoelectric mini-sensors from Balluff.
Photoelectric mini-sensors from Balluff (Photo courtesy of Balluff).
Industrial IoT Drives Distributed Control and Shrinks Controller Size
Given the proliferation of intelligent, connected sensors in a modern manufacturing facility, distributed data collection and processing are becoming the norm. Instead of a large central PLC, we are seeing distributed controllers sprinkled across the manufacturing flow. At the Siemens Amberg Electronics Plant – which is a showcase for Industrie 4.0 – multiple distributed PLCs control each step of the highly automated production flow. As Siemens notes in their press release: “ …Some one thousand of these controllers are in action from one end of the production line to the other…”5
These distributed controllers are getting smaller and more compact while also incorporating more features. If we take a longer historical view, then PLCs have shrunk from the size of a small room/cabinet (circa 1970s) to something that can fit in the palm of your hand (circa 2000). Figure 4 shows a desktop PLC circa 1990s; today’s PLCs are even more compact. At the same time the newer and significantly smaller PLCs feature a stunningly higher processing and interfacing capability.
Example of a desktop PLC, circa 1990s. (Photo courtesy of Siemens Press)
Design Challenges with Smaller Systems – Heat Dissipation and Form Factor
Industrial systems are rugged – they have to be able to work in 55o C to 75o C ambient operating environments. In most cases, these are passively cooled, i.e. there is no fan for active cooling. So, it is crucial that the system architecture takes into consideration the power consumed by each section.
Let’s take an example of architecting a controller with multiple digital inputs – very common in industrial PLCs. A traditional design using discretes and an opto-coupler for isolation per input channel would look something like shown in Figure 5. While this design is not the most integrated design, what is important to consider is the amount of power consumed by each channel. With a 2.2Kohm input resistor and 24V Vin, the input current is 11mA, which equates to power consumption of 264mW per input channel. This may not seem like a lot unless you stop and realize that it is not uncommon for a modern PLC to have 8, 16, or even 32 digital IO channels.
Traditional controller design with discrete components
As shown in Table 1, the power consumed by the digital IO module/section of your PLC can go up dramatically. And as your PLC size shrinks and, with only passive cooling allowed, this kind of power consumption by just the digital input portion of your design can be problematic.
The other option is to use an integrated IC to replace the discrete logic – an integrated approach that allows for configurable input-current limiting. Many solutions allow one to set the max input current to something like 2.5mA. This would cut the power consumed by 1/4. In the example above for 32 channels, your digital input section would now have a sub-2W maximum power consumption. Also, eight channels of discretes can now be replaced by a single IC whose dimensions are shown in Figure 6.
Controller architecture based on integrated IC
One of the key aspects of building a smaller and yet more feature-rich industrial system is to increasingly make use of integrated solutions where applicable. Not only does this approach reduce the solution size, but it can also dramatically limit the power consumption and, thereby, the heat dissipation. To do this we need to use integrated chips that incorporate functionality specifically to eliminate the need for external components.
Let’s take an example of a digital out module in a control system. Digital out modules generally drive actuators like a motor, a solenoid, or a relay. Many of these loads are inductive in nature. As shown in Figure 7, this means that when the switch is opened, the inductor “tries” to keep current flowing. So, a free-wheeling diode has to be used across the inductor for back EMF suppression and/or to protect the MOSFET.
Digital out module in a control system
Some of the newer digital out drivers will incorporate a FET switch that eliminates the need for external free-wheeling diodes. Internal clamping diodes limit the negative excursion to (VDD – VCL) and allow free-wheeling currents to demagnetize the inductive loads quickly. In practice this frees up a large amount of board space, as shown in Figure 8.
Examples of an 8-channel digital out module with and without the free-wheeling diodes
Designing the Ever-Shrinking Sensors
Sensors are continually shrinking in size while incorporating more signal conditioning and sophisticated communications capability. So, the new sensor designs have to carefully select components based on their size and power dissipation. Figure 9 shows a Himalaya switching regulator IC and the companion inductor housed in a small, M8-sized proximity sensor. Clearly the DC-DC switching regulator along with inductor size that it requires are an important consideration in the design of this ultra-small sensor.
Himalaya switching regulator IC (MAX17552) and the companion inductor housed in a small, M8-sized proximity sensor
Many sensors are getting equipped to support IO-Link, a new open communication protocol that allows the sensor to communicate digitally with the controller. An IO-Link-enabled sensor does not just tell you that the proximity limit has been violated – it tells you the precise distance by which the proximity limits have been crossed. IO-Link transceivers are getting smaller and more power-/heat-efficient.
As sensors become more ubiquitous and more pervasive, there is a corresponding need to make them smaller and more power efficient. Building these sensors requires a judicious selection of components that enable the small form-factor, and which display the absolute minimum power dissipation.
This article has just scratched the surface of some of the decisions that must be made by system architects as they design for the next generation of automation systems that will enable the industrial IoT. As more intelligence is squeezed into traditional manufacturing processes, system designers must carefully consider criteria like solution size and power efficiency.
Even though power is not a traditional metric evaluated in industrial systems, it directly affects the heat generated by a system that must survive for decades in a rugged environment with high ambient temperature and no active cooling. Form factor, another metric once rarely considered for industrial systems, now matters since it directly affects how and where you can place your edge sensors and, sometimes, also your edge controllers.
As the industrial IoT vision is built out, it will profoundly change how industrial systems are architected and how component selection is done.
- Digital Transformation of Manufacturing Industries – ARC Advisory Group
- Architecture Alignment and Interoperability: An Industrial Internet Consortium and Platform Industrie 4.0 Joint Whitepaper
- Smart Sensors
- Miniature Sensors: Compact sensors for factory automation by Balluff
- Siemens showcase “digital factory” celebrates 25th anniversary