FPGAs Take on the 4th Revolution

The industrial revolution, powered by the Industrial Internet of Things (IIoT) will see embedded systems adapt to connectivity and data gathering in the new wave of productivity. FPGAs are well-placed to meet demands for flexibility, time-to-market and efficiency.

Development tools need to take into account hardware and software resources. This not only impacts how developers can explore architectural trade-offs and product development, but how they can handle debug.

The IIoT is dramatically altering manufacturing, energy, transportation, cities, medical, and other industrial sectors, says Giles Peckham, marketing director, Xilinx. Most experts believe that IIoT is happening now and with very tangible, measurable business impact,” he says. The impact shows itself most in savings or productivity gains, with some figures suggesting that small savings can translate to large financial gains. For example, in aviation it could save $30 billion in fuel, and the rail industry could realize $27 billion in productivity gains.

Figure 1: LabVIEW FPGA bridges the FPGA design gap, says National Instruments. (Picture credit: National Instruments)

Figure 1: LabVIEW FPGA bridges the FPGA design gap, says National Instruments. (Picture credit: National Instruments)

“The IIoT enables companies to collect, aggregate, and analyze data from sensors to maximize the efficiency of machines and the throughput of an entire operation,” Peckham continues, citing applications including smart energy and smart grid, motion control, machine-to-machine (M2M), predictive maintenance, and connected medical systems.

These efficiencies will be brought about by improved connectivity and autonomous control, believes John Johnson, Strategic Marketing Manager, Industrial Control and Manufacturing Segment, Intel Programmable Solutions Group. Connectivity is driven by security, interoperability, and service level guarantee demands, whereas autonomous control will drive functional safety, due to the increased automation, and also edge analytics, with much higher levels of performance, particularly at the edge.

Integrated Solutions

Robotics, power electronics and ruggedized control are placing demands on the pipeline, resulting in more processing being pushed to the edge, rather than being pushed back to the server, agrees Kyle Voosen, Director of Marketing, National Instruments.

National Instruments launched its high synthesis level tool, LabVIEW in 1999, and it is now the dominant language for industrial design. With the introduction of LabVIEW FPGA it brought the data flow language and bridges to FPGAs, explains Voosen. It extends the LabVIEW graphical development platform to FPGAs on the company’s Reconfigurable I/O (RIO) hardware and provides an ecosystem of IP libraries, simulator, and debugging features for the FPGA developer.

Energy Efficiency

Today’s factory systems include a considerable number of motors, and efficient control can significantly lower the overall power consumed. Peckham points out that the company’s All Programmable FPGAs and SoCs can be used on the lowest power consumption controllers. He adds that embedded controllers, such as the MicroBlaze or ARM Cortex-A9 processor cores, have software layers that guarantee a common Application Programming Interface (API) for different actuator hardware, to enable modular development. There are also interfaces to design environments such as Simulink, Scilab and a Graphical User Interface (GUI) with NI LabVIEW.

Another characteristic of the industrial environment is communication, and the versatility of FPGAs is evident here. “For industrial networking,” says Peckham, “All Programmable solutions . . . provide scalability and flexibility for factories. The computing capabilities match the needs of powerful Ethernet-based protocol for real-time communication. PROFINET, EtherCAT, EtherNet/IP, Powerlink, and many more are supported by IP cores from industry-leading Xilinx Alliance Program partners,” he says. “Acceleration in hardware allows the strictest real-time and latency requirements to be fulfilled, and follows the evolution of standards easily,” he adds.

Addressing industrial Ethernet and fieldbus connectivity, Intel FPGA has partnered with German industrial Ethernet IP protocol provider Softing Industrial Automation to make Ethernet connectivity easier. Johnson maintains that combining Intel FPGAs and a security Complex Programmable Logic Device (CPLD) makes it easy for developers to implement industrial Ethernet and Fieldbus connectivity without obstacles such as license negotiation, or per-unit royalty reporting.

Johnson has a warning for those assessing energy efficiency in industrial automation. When viewed at the component level, it may be misleading, he cautions. “To be an effective figure of merit, energy efficiency must be normalized and measured on the basis of machine throughput,” he says. “While energy requirements of a piece of machinery may or may not increase to improve machine throughput, machines that demonstrate higher production levels can also contribute significantly to energy efficiency,” he clarifies.

He maintains that Intel FPGAs can redefine performance in industrial designs, as portions of the system that are computationally intensive can be accelerated in hardware. “This helps motors run more efficiently, robots operate more quickly and precisely (often compensating for variations of the mechanical system over time), and systems detect and compensate for inefficiencies caused by machine/tool wear,” he adds.

All agree that FPGAs will drive the IIoT market. For Johnson, this is because of intrinsic strengths that meet tough market requirements better than conventional solutions. “Markets that employ FPGAs typically exhibit one or more of several common characteristics,” he says. “First, FPGAs are used in markets in which standards are fluid and will likely change at least once over the life of a product family,” he notes. As a result, the ASIC approach is usually eliminates and software-centric solutions may or may not be able to address updates and changes effectively, he argues. “Intel FPGAs are particularly suitable for applications that require hardware acceleration,” he says. A final factor is market demands. The scalability of Intel FPGAs, says Johnson, provide scalability, enabling manufacturers to develop multiple product families “by equipping engineers with tools and technologies that produce designs that are portable across entire product families, from end to end, including cloud, fog, gateway and edge,” he says.

Peckham agrees. “A sophisticated architecture and design process for functional safety applications integrated in FPGA and SoC FPGA can reduce customers’ risks and increase time to revenue and profits significantly,” he says. The company provides tool support with the functional certified Vivado Design Suite. For Peckham, this combination of FPGA products, pre-approved architectures and design methods, enables customers to “quickly incorporate functional safety in their products with fast development times.”

There is also the UltraFast Design Methodology for accelerated and predictable design cycles. This is used by Xilinx Alliance Program Partners in tools and IP. The Vivado IP Integrator communicates between blocks to ease the design of IP subsystems.

Figure 2: Xilinx’s Zynq-7000 All Programmable SoCs are a scalable, single chip option with analytic functionality and cloud connectivity.

Figure 2: Xilinx’s Zynq-7000 All Programmable SoCs are a scalable, single chip option with analytic functionality and cloud connectivity.

The company’s Zynq-7000 All Programmable SoC single core device is optimized for industrial applications, says Peckham. It offers a range of options—from entry point to fully scalable ARM processor-based platform. The single chip offers analytic functions and cloud connectivity, with multi-layered security, namely processor-driven secure boot and bitstream encryption and authentication.

Tool support is important, agrees Johnson. He points out that development tools need to take into account hardware and software resources. “When development tools take hardware and software resources into account, it not only affects how developers explore architectural trade-offs and product development, but also how they handle debug,” he says. The company’s Motor Control Development Framework comprises hardware platforms (including motors), reference designs where the demarcation lines between hardware and software are controllable, and graphical, system-level debug tools.

IIoT Security

Johnson believes that implementations for security, safety and reliability need more than link encryption. He explains that Intel FPGAs offer the capability to secure ‘green field’ environments, where there are no legacy systems, as well as ‘brown field’ ones, where the new solutions and components have to co-exist and interoperate with legacy solutions. Creating secure gateways makes co-existence and interoperation with legacy buses possible and adds the required security. “FPGA implementations are well suited to accelerating software encryption functions, providing white listing, and to firewall functionality,” he asserts.

Figure 3: Intel FPGAs store, manage, and control integrity within the IIoT.

Figure 3: Intel FPGAs store, manage, and control integrity within the IIoT.

Studies by BarcoSilex have shown that FPGA acceleration of secure IPsec connections reduce usage by a factor of four, and accelerate the throughput by greater than 7x, adds Johnson.

Last month’s SPS Drives show in Nuremberg, Germany, showed the level of activity in manufacturing, automation and the growth of the Industrial IoT. It was clear that the rise of the machines will be heavily dependent on the development and support of the FPGA community.


hayes_caroline_115Caroline Hayes has been a journalist covering the electronics sector for more than 20 years. She has worked on many titles, most recently the pan-Euro