We are currently poised at the beginning of an exponential increase in the deployment of sensors and microcontrollers (MCUs) for a wide range of applications, including smart utility metering, wearable electronics, data acquisition systems, and industrial and remote control systems.
A modern factory, for example, contains numerous building automation systems and data acquisition systems, all involving sensors and MCUs as illustrated below:
Similarly, a modern home features an ever-increasing number of security and automation systems, and -- once again -- these all involve sensors and MCUs as illustrated below:
A major consideration for all these systems is reducing power consumption and increasing reliability as much as possible. Designers are driven to add functionality while simultaneously reducing the system's power budget -- at the same time, embedded software is becoming larger and more complex, requiring more memory and further stressing the power budget.
The majority of today's MCUs come equipped with two sorts of memory: Flash and SRAM. The Flash is relatively slow and supports a limited number of write cycles, but it is non-volatile and so is used to hold the code. By comparison, the SRAM is fast and has unlimited write cycle endurance, but it volatile and can only hold temporary data.
All of this furthers the complexity of the designer's choice of MCU -- how much Flash is required (how little will suffice)? How much SRAM will be needed (how little can one get by with)?
In order to address all of these issues, Texas Instruments (TI) has developed a line of FRAM-based MCUs. The original family was the MSP430FR57x series of devices. Now, TI has introduced two new families: the MSP430FR59x series and the MSP430FR69x series.
FRAM conveys many advantages, including the fact that it is non-volatile, which means it can maintain both code and data when power is removed from the system. Unlike Flash, FRAM has an extremely high write speed, it's bit-wise programmable, and it can be written to an unlimited amount of times. Unlike SRAM, FRAM is non-volatile and it's not susceptible to soft (radiation-induced) errors.
Another key consideration is that the FRAM simplifies the software architecture by allowing flexible code and data partitioning coupled with fast writes and unlimited endurance (FRAM does not require pre-erase and is the only non-volatile embedded memory that can be written at 8MBps).
This means that the same FRAM-based MCU can be used to address both "code-heavy data-lite" and "data-heavy code-lite" applications.
One more piece to the puzzle is TI's EnergyTrace++ technology. Developers of embedded systems with low power budgets typically spend a huge amount of time figuring out where power is "leaking" out of their system. As illustrated in the image below, traditional debug systems -- which are expensive and incompatible -- are difficult to use, are limited by a one-way data flow, and don't provide visibility into real-world MCU activity.
In order to address this, TI's EnergyTrace++ technology enables developers to analyze power consumption in real-time for each peripheral, thereby allowing them to make intelligent decisions about code execution and to optimize peripheral use based on the power profile.
Costing less than US$30, EnergyTrace++ technology provides a closed-loop seamless integration of the IDE, tools, and devices, thereby resulting in embedded systems that are optimized for the lowest-possible power consumption.
Please visit the TI website for more information on EnergyTrace++ technology, the MSP430FR59x and MSP430FR69x MCU families, and the new low-cost MSP430FR5969 LaunchPad development kit.
— Max Maxfield, Editor of All Things Fun & Interesting