ARMed and ready

3While the rest of the engineering world has been developing embedded systems using the ARM microcontroller unit (MCU) architecture for many years, the high-reliability (hi-rel) market has been slow to adopt it. That is a pity because the biggest benefit of using an ARM-based microcontroller is the ease of development that ARM offers, by virtue of the large supporting ecosystem of tools.

One reason for this slow adoption of -based microcontrollers is that suppliers of these MCUs have been even slower to embrace the requirements of the military and aerospace design community. High-reliability (hi-rel) integrated circuits (ICs) need to be guaranteed to operate in extreme environments such as extreme radiation or temperature. It consumes considerable resources and budget to develop and qualify chips that are guaranteed to operate in such environments – frankly, it is easier for the semiconductor suppliers to play it safe by developing chips that will sell in higher volume in consumer applications.

Engineers are trained to solve problems: The lack of availability of hi-rel ARM microcontrollers has not stopped these engineers from doing their jobs and creating hi-rel anyway. Today there is a limited pool of components, often very mature products, that will get the job done, although it may not always be pretty. Field-programmable gate arrays (), (), and legacy microprocessors are often used in for the simple reason that they have already been qualified to military spec and that they are trusted and proven.

What do MCUs bring to the party?

Now that there are options available for rad-hard and extreme temperature ARM Cortex-M based microcontrollers, embedded systems designers can choose the most efficient solution to the problem. There are quite a few benefits of adopting an ARM-based microcontroller, the biggest one being the supporting ecosystem.

The ecosystem for a microcontroller covers many aspects, the most obvious being hardware and software development tools. There exist lots of options from freeware, from cheap and cheerful all the way up to professional-grade tools that cost real money. The ecosystem also encompasses the community – the places to go for help to ask a question, find a software driver or a communications stack that is already available, or get technical support. Training is also included in the ecosystem – either bringing a professional trainer into your facility for a day or two, or going to YouTube to watch a video.

The ARM CPU and instruction set were designed for embedded-control applications. The core is efficient for monitoring incoming data, processing it with math-intensive routines (such as digital filtering on noisy data from sensors), and managing on-chip peripherals that interface to external chips, sensors, and actuators. Most embedded designers that are starting a new design with a blank sheet of paper would probably be expected to start with an ARM microcontroller. The benefit is not just realized in the initial design, but it actually greater in subsequent designs, because the code is reusable. Often a few tweaks to the known good code and some additional routines are all that is required for spin-off or development of next-generation systems. Contrast that ease of use with FPGAs, which are more expensive, are more difficult to design with, have a smaller existing ecosystem than the , and do not offer the same level of reusability as the MCU.

Power consumption is another major benefit of using microcontrollers, particularly ARM. Because battery-powered systems have been a big driver in the embedded market for many years now, microcontroller architectures have evolved in response to the low-power requirement. Operating current consumption with ARM is typically less than a tenth of that of an , , or application processor. Moreover, standby current (when the CPU is not executing but non-volatile memory contents are maintained and the core will wake up instantaneously in response to an interrupt) is significantly lower.

Hundreds of different ARM-based microcontrollers are available today, covering almost every conceivable combination of on-chip peripherals, pin count, and memory size option. There are also now many different ARM CPUs in the Cortex-M class (the range that are optimized for embedded applications) to choose from. Fortunately, these CPUs have the same look and feel as well as a high degree of compatibility. The higher performance cores typically have more functionality but retain backward compatibility. Not all of these products are a great fit for military embedded systems; in fact, not many of them are. The good news: A growing number of MCUs have been designed for hi-rel applications and it is now possible to take advantage of the ARM ecosystem.

What makes a hi-rel ARM MCU?

Figure 1 is a block diagram of a hi-rel microcontroller based on ARM Cortex-M0. The IC has a combination of “standard” ARM peripherals such as serial communication interfaces (SPIs, UARTs, I2Cs) and counter/timers (for PWMs, etc.) as well as “hi-rel application”-type features.

21
Figure 1: Block diagram of ARM microcontroller intended for hi-rel embedded applications.

The hi-rel features that are now available on this class of chip start with enhanced wafer fabrication processing to immunize against latch-up in the presence of radiation or extreme temperature. In addition, all internal registers have been implemented with triple modular redundancy (TMR). Both of these attributes are invisible to the firmware programmer but are critical to the system design.

On-chip peripheral hardware enhancements include an on-chip error detection and correction system with a scrub engine that can automatically correct flipped memory bits. This feature has been included to ensure that the memory operates reliably in the presence of ionizing radiation particles. Note that while the TMR also protects the chip against ionizing radiation, the chip is designed to protect the logic and circuit routing rather than the memory cells.

Along with integrating processing enhancements and hi-rel type peripherals, the packaging options are optimized for extreme environments. Figure 2 shows a PCB that has been developed for a space application – the PCB is located on the International Space Station. Right in the center of the board is an ARM Cortex-M0 based microcontroller in a hi-rel ceramic package.

22
Figure 2: Hi-rel embedded system using ceramic packaged ARM-based MCU. Image courtesy .

A robust MCU that doesn’t cost an ARM and a leg

The good news is that after years of watching the ARM microcontroller world blossom into a treasure trove for embedded designers, there are finally products being developed for hi-rel environments such as military end uses. It is now possible to take advantage of the benefits that these offerings bring to embedded designs while conforming to the hi-rel specification. It is still true that ICs that are designed for USB sticks are not suitable for defense systems; in this spirit, designers must still choose carefully.

Why are these chips just now being developed? There is a demand for a higher volume of reasonably priced processor products such as ARM Cortex-M microcontrollers to support new growth applications such as megaconstellations of small and picosatellites. These applications need components that are sure to work in extreme environments but cannot have the high price of traditional space/mil FPGAs. These demands are driving the development of this new class of hi-rel ARM microcontrollers.

Ross Bannatyne is director of market-ing for VORAGO Tech-nologies, based in Austin, Texas. He was educated at the University of Edinburgh and the University of Texas at Austin. Ross has published a college text called “Using Microprocessors and Microcomputers” and a book on automotive electronics called “Electronic Control Systems” (published by the Society of Automotive Engineers); he also holds patents in failsafe electronic systems and microcontroller development tools. Ross can be reached at rbannatyne@voragotech.com.

VORAGO Technologies www.voragotech.com