Rapid prototyping paves the way for more responsive projects

The prototype has always been an early part of product design, as it lets stakeholders see what design choices deliver the best results in a real-world context. As modern design becomes more agile, development teams can deliver results to key customers regularly for feedback and feature requests, to ensure the product meets needs and expectations.

One strategy to build upon this foundation and limit the number of rounds of reviews is rapid prototyping. This strategy for hardware-based systems identifies design flaws early, reducing iterations and costs associated with rework and waste. Cost efficiency is improved using small batches of prototypes to test different deployment scenarios before full-scale production.

The Avnet Zynq UltraScale+ RFSoC development kit with Qorvo RF Front-End supports rapid prototyping of LTE systems. Source: Tria Americas

Data from this product cycle is now referred to as the digital thread. This virtual chain supports teams working on current and future product iterations. The result is a more agile and flexible continuous development process, with rapid prototyping as the foundation of successful product introduction.

Flexibility is the key

Rapid prototyping works best when there is a flexible approach to design, such as that enabled by a digital thread. Ideally, the system can adapt to changes in the specification or unforeseen problems with the original concept without impacting the underlying hardware. The key to rapid prototyping success is to select platforms and development techniques to support these changes.

Following a software-defined approach requires good programmable hardware.  Advanced 32-bit multicore microcontrollers now combine high-speed general-purpose processing with application-specific peripherals and accelerators. Common industrial functions, like motor control, can run alongside responsive user interfaces on the same platform.

Manufacturers are making these MCUs available through development boards. Many of these boards are supplied as a system-on-module (SoM), providing I/O expansion through header connectors. These standard pinouts provide developers with many expansion options, enabling them to construct an initial prototype in a matter of days.

The field programmable gate array (FPGA) has long been a favorite prototyping platform. After decades of evolution, the FPGA has turned into the programmable system-on-chip (SoC). With increasing performance available, SoCs now include heterogeneous processing subsystems supporting application and real-time processors, dedicated GPUs and video codecs, along with programmable mixed-signal functionality able to run at extremely high frequencies. These powerful platforms are well suited to developing software-defined RF front ends and other applications that need high-performance interfaces.

Effective rapid prototyping extends to developing hardware capable of field trials. Developers can now use 3D printing to create custom enclosures and can even create custom actuators for integration into mechanical systems using a metal-powder system. The result is a high-functioning prototype without the lead times that are often needed to create custom tooling at a subcontractor for volume manufacture.

Tools for rapid design

The computer-aided design (CAD) tool industry is also evolving. CAD provides essential support for 3D modeling, creating enclosures and supporting integration of PCBs into the target system. Engineering design automation (EDA) tools cover logic and software development environments with high-level system-design tools, all of which support rapid prototyping processes.

System-level tools can be helpful in the earliest stages of a project. Engineers can evaluate algorithms in a virtual environment using data collected from the field before committing to hardware. The virtual environment and hardware development can run in parallel throughout the product’s lifecycle. A growing list of hardware platforms can be modeled in the software domain, on cloud servers or desktop computers. These platforms can support multiple engineers working in parallel more easily than sharing access to a limited number of hardware platforms.

Source: Getty Images

Software modeling also supports virtual prototyping, enabling greater collaboration across teams. Software developers get more accurate feedback on the changes they make. It also provides manufacturing with a better view of how a platform is expected to operate. System engineers can suggest changes to improve production flow once the hardware is complete. Using integrated platforms that allow for real-time communication and data sharing helps ensure all the stakeholders remain aligned throughout the design cycle.

Once practices such as virtual prototyping are embedded in the design process, teams are in a good position to take a further step and move towards a continuous deployment paradigm. Software changes to improve performance made after commercial launch can be tested on a combination of virtual and physical platforms before being sent to customers’ systems.

Rise of AI in prototyping

As artificial intelligence (AI) continues to play a growing role across technology and product development, its’ effects can also be seen across rapid prototyping tools and strategies.

Rapid prototyping fits well with the increasing use of different sensor modalities in a single device. For example, modern sensors, coupled with innovations like machine learning, are enabling new applications to deliver real benefits in combination with predictive maintenance in industrial systems.

Creating these new and complex systems requires a different approach to interpreting data and determining how best to react to that data. Combining different sensor data gives a more accurate picture of the system and can provide the foundation for creating a digital twin.

Machine learning and AI represent key elements of the underlying system design. AI, in particular, is useful in sensor-driven applications because it can link inputs to outputs in new ways. The development of predictive systems benefits from a process built on rapid prototyping, and retraining AI models is part of this process. By focusing on techniques that support easy iteration, design teams can achieve production-worthy implementation with less risk of post-launch failures.

Like many other strategies in the industry, rapid prototyping is evolving with new technologies that will further strengthen its benefit. Overall, rapid prototyping is crucial in an agile design world as it can help designers save time and costs as a product moves through various stages of development.

—————————

Jim Beneke is vice-president of Tria Americas, which is the new name for embedded compute boards, systems and associated design and manufacturing services at Avnet. He has more than 39 years of experience in technology management, business and strategy development, technical marketing, research and development and design engineering.

https://www.tria-technologies.com/