User requirements are changing quickly for next generation appliances. This trend is driven by the need to design products that are more energy efficient and environmentally friendly, while at the same time can be produced cost-effectively. Added to that is the goal of meeting consumer demands for improved performance and reliability, while satisfying the desire for the latest features such as color displays and touch control panels.

A major limitation for many manufacturers is the tendency to think about new product design decisions in isolation, or to think only about a single product in the line instead of looking at the entire process of designing new products sequentially. For optimal long-term results, a systems approach to design is required.

New requirements translate into new features and new implementation approaches for manufacturers. In order to deliver these new features while considering the process of designing product lines, many manufacturers are migrating to lean product development or platform-based development, which allows them to spread costs over entire lines and multiple designs.

Using various research and development programs, the lean product development approach focuses on the development of a basic platform for each line of products and targets specific components for incremental enhancement. As enhancements are made, they are introduced into the various products. Low-end items get the older, inexpensive proven technologies with cost minimization, while higher end models get innovative new features.

Feature R&D

Enlarge this picture Fig. 2. Architecture of the DSPnano and Unison RTOS products. The layered nano-kernel approach ensures easy understanding and modular growth.

Using lean development, designers first consider what each of the upgrades means in terms of changes to individual components across multiple units in the line, then considering the effects that this might have on the aggregation of all modules.

For example, upgrades for a new, environmentally friendly refrigerator product might require:

An alternative refrigerant.

A more efficient compressor.

Better insulation and/or alternative blowing agent.

Quieter operation.

These features, some of which may help against global warming, are things that certain consumers are willing to pay for. In research and development terms, the materials and components investigated would include: refrigerants, efficient compressors compatible with alternative refrigerant, different insulation, and improved motor control to achieve variable speed operation of the compressor.

A leading contender for quiet, efficient operation is a permanent-magnet synchronous motor (PMSM) with field oriented control (FOC). Taken together, these new technologies would be introduced in high-end models initially, and then migrate to the lower end over time as costs decreased with volume, allowing margins to be preserved.

For appliances in general, the display and interface components represent another opportunity for product enhancement. Color touch panels give the perception of quality, but also allow appliances to expand their information exchange. Mechanical switches and knobs can be completely eliminated if desired, and improved branding options are available at low or no cost using these new features.

Another feature likely to be added to appliances in the future is a networking capability that enables remote monitoring and energy rate negotiation with a smart electrical grid.

As designers plan for improvements, they must realize that these changes can require a fundamental overhaul of the overall design approach. A network connection is required for energy-rate negotiation and remote monitoring. PMSM motors with FOC require more computing power. And color touch panel displays can offer additional features, such as family message boards. The Fig. 1 block diagrams show a system using two MCUs to provide new features.

New technology can also enable field upgrades, reduce service costs, improve reliability, and provide downstream revenue through the sale of add-on components. Adding diagnostics can provide fast and simple service using network connections to diagnose problems so the service person shows up with the right parts. Flash updates can improve algorithms for increased reliability, as well as add completely new features. A power-on self-test feature can ensure that the unit is safe and independent of failed components.

A unified approach

Enlarge this picture Fig. 3. The hardware and software architectures show greater simplicity with a single, more powerful MCU than with multiple MCUs.

The optimal means to implement a lean development program that enables easy adding or swapping of features at minimal cost is to use system on chip (SoC) microcontrollers (MCU) and an open-standards-based real-time operating system (RTOS). While SoC MCUs are already widely used, their software is typically handcrafted and difficult to maintain. It is also difficult to upgrade single loops of control on absolutely minimal MCUs.

Appliance designers always strive to provide new features while controlling costs. By using open industry standard software interfaces or APIs, they can maximize software life and reuse while minimizing training and risk. These open standard APIs support the use of standard I/O modules and interfaces, substantially reducing development effort and training on sophisticated algorithms, while also slashing development cost. Fig. 2 shows the architectures of DSPnano RTOS and the Unison RTOS, both of which implement a modular, size-constrained approach to providing these APIs.

The globally accepted open standards for embedded operating systems are limited to Linux and POSIX. (See Sidebar below.) Using Linux-compatible and POSIX-compatible operating systems is also the ideal approach for lean, platform-based product development. Proprietary solutions are available as well; however, the cost of development delays, rework and training required to support proprietary solutions far outweighs their perceived benefits. Examining the total cost of ownership is the better approach to analyze alternatives. Such an analysis will show that lean product development maximizes profits while minimizing time and risk.

The DSPnano RTOS for 8/16-bit processors and Unison for 32/64-bit processors offers Linux and POSIX open standards and a broad set of off-the-shelf components to maximize profit and minimize time and risk. DSPnano and Unison also offer the only native implementation of Linux-compatible and POSIX-compatible application programming interfaces (API) for 8/16/32 bit MCUs. DSPnano and Unison are fully API compatible with each other, and applications port in minutes, supporting the full range of 8-bit through 32-bit MCUs with a single platform. Larger microprocessors with open-standards-based Linux and POSIX operating systems are also upward compatible.

Consider an example in which one or more MCU is connected via asynchronous serial lines for communication. If new features need to be added on a separate MCU, the master controller requires a single serial port and some flash to implement the feature. DSPnano and Unison fully support this kind of incremental growth and offer solutions to grow the operating system. In this way, costs and features can track the position in the product line. This is one approach to managing BOM cost.

Also consider the fact that a single, powerful MCU is less expensive than a number of smaller MCUs. With either approach, if the operating system applications programming interface (API) is standardized, the effort to swap features in and out is minimal given underlying hardware support. With a more powerful MCU, products can also use the time-based use of features to minimize overall MCU processing requirements, making the system much more responsive. Of course, overall BOM costs must always be minimized to analyze these trade-offs.

Software transformation

To minimize research and development effort, a platform-based approach requires an open-standards-based embedded or real-time operating system with a set of components to minimize development effort and speed time to market. The DSPnano RTOS (8/16 bit) and Unison RTOS (32 bit) offer Linux and POSIX compatibility on tiny MCUs along with a full complement of I/O, diagnostics, data logging, and testing. Using pre-integrated, tested, and documented components drastically cuts development time.

Sensorless FOC motor control, touch control panels, networking capability, on-board diagnostics, data logging, the capability for field upgrades with USB sticks, and many other features are coming in next-generation appliances. From a software and hardware point of view, this means more computation and more complex design with higher processing requirements. However, these software components can now come off the shelf.

Using an operating system like Unison or DSPnano allows manufacturers to quickly realize new lines and new models by implementing a lean product development approach based on platform-based design. Both are Linux and POSIX compatible to maximize software life and minimize training and risk.

For more information, visit: www.rowebots.com

What is POSIX?

The globally accepted open standards for embedded operating systems today are POSIX and Linux. POSIX is an acronym for “Portable Operating System Interface for Unix” and Linux is the world's second most popular operating system (second only to Windows). POSIX was modeled after many Unix variants with the idea of standardization.

How did Linux end up POSIX compatible?
All prominent operating system vendors adopted these open POSIX standards to some degree, including Windows, Linux, Solaris and AIX. Real-time POSIX standards were broadly adopted from 1995 through 2009 in MMU-based embedded and real-time operating systems. These standards are used in VxWorks, QNX, Integrity, and Monta Vista Linux. The early Linux emulation of Unix, and the POSIX standardization of Unix, led to the POSIX compatibility in Linux.