Energy Harvesting with MSP430 FRAM Microcontrollers

Aguilera

Energy Harvesting with MSP430 FRAM Microcontrollers [Copy link]

For many people, their first exposure to energy harvesting was probably in the early days of solar-powered pocket calculators. While these types of calculators are no longer mainstream today, the technology and concepts used are still used in our daily lives. Today, we see energy harvesting in many applications, such as sensor nodes, wind turbines, and indoor power supply applications. However, even though the discussion of this technology has evolved greatly, when it comes to energy harvesting, developers still face the same challenges as decades ago.

In order to generate the required energy without adverse effects, it usually requires a large solar panel and a large thermal energy harvesting device, or energy is obtained by emitting vibrations in different frequency ranges from the device, all of which are determined by the system used. Therefore, in many cases, the cost of the system even exceeds the advantages of replacing traditional power sources. Of course, there are exceptions when these limitations must be ignored for certain reasons. For example, in remote areas where power lines cannot reach, wind or solar energy harvesting can provide a viable alternative energy source for battery-powered systems, although the initial cost of this approach will be higher.

Let's take a look at some of the major challenges facing current energy harvesting solutions.

Figure1—Simplified general block diagram

First, from the simplified general block diagram above, you can see that this system consists of inputs and outputs, including sensors, buttons, LEDs, displays, sounders, and increasingly wireless connectivity. The edge nodes of this typical Internet of Things (IoT) architecture can communicate via Wi-Fi, Bluetooth, NFC/RFID, or other proprietary interfaces. The power required by these wireless connections is as low as a few uA, and the maximum is only tens or hundreds of mA, which can power the related RF ICs and subsystems within tens of milliseconds.

[color=rgb(85, 85, Figure 2—Common RF power usage chart In many applications, designers want to store sensor or other data in nonvolatile memory because the acquired data can be recovered even if power is lost. Therefore, existing general-purpose memory technologies such as EEPROM or FLASH are not always the best choice in these energy-constrained situations. Fortunately, technology is moving in a direction that makes energy harvesting systems feasible. One such technology integration is TI's family of ferroelectric random access memory, or FRAM, microcontrollers (MCUs). FRAM technology combines many of the advantages of SRAM memory with the non-volatility of FLASH memory. A key advantage is the ultra-low power non-volatile FRAM writes, which, unlike FLASH, do not require a pre-erase cycle, saving time and power. Another advantage is the inherent low voltage writes to FRAM cells. Conventional Flash or EEPROM technologies require an integrated charge pump to complete the pre-erase cycle, which typically requires 5-10 mA of current and takes hundreds of milliseconds to operate. In applications that require frequent non-volatile writes, this additional power consumption can consume a considerable amount of battery power or harvested energy. The cost of purchasing disposable batteries may not be very high, but the impact they have is extremely far-reaching. Billions of new batteries are sold worldwide each year, and only a small fraction of them are recycled, which creates a large amount of landfill waste. Another drawback of disposable batteries is that both the battery itself and the entire system need to be replaced at some point in some situation, which creates potential challenges. Imagine if the battery is installed in a system deployed at the bottom of the ocean or on the top of a mountain. How should we replace it? In fact, the cost of battery replacement can be very large. Although rechargeable batteries can reduce the number of batteries to be replaced, they will not necessarily solve all the challenges of battery replacement. Rechargeable batteries do provide benefits when we use energy harvesting to charge them.

Currently, energy sources such as solar energy, thermal energy, motion energy (vibration or other dynamic effects), and RF are widely accepted. Other energy sources are also under development, such as the possibility of harvesting energy from electrochemical reactions that occur in human blood, or from such reactions inside plants and trees.

[color=rgb(85, 85,Ideally, these energy sources would be continuous, but in reality they are not. In the case of solar harvesters, for example, drifting clouds may block out the sun, and lights in indoor facilities may not always be on. Vibration-based harvesters typically operate near a resonant frequency, limiting their operating range, and thermal harvesters lose efficiency or stop operating altogether if they can't maintain the right temperature differential. Ultimately, we can't rely on this energy source for continuous 24/7 operation, so redundancy is needed. In some cases, that could be a second harvester or a rechargeable battery. Even solar-powered calculators include a CR2025 battery as a backup to the sun when the office is dark.

Dealing with power losses becomes a major consideration for designers of energy harvesting nodes. Modern microcontrollers run through a boot sequence when powered on, which often takes several milliseconds and consumes precious power. If power is lost, most microcontrollers need to reboot and run this startup code each time power is restored.

FRAM memory itself is the enabling device for a highly innovative software utility called Compute Through Power Loss (CTPL). We can even think of CTPL as a non-volatile interrupt handling routine where, when power loss is detected (usually using a comparator or ADC input), key parameters and microcontroller state are saved to non-volatile memory (NVM). In the event of a power outage, FRAM offers an advantage because designers can continue working directly from where they left off rather than starting from scratch.

A simple demonstration was enabled by the low-cost MSP430FR6989 MCU Launchpad development kit with 128KB FRAM MSP430 microcontroller.

By combining FRAM technology, Compute Through Power Loss code, and the Energy Harvesting BoosterPack plug-in module, we have laid a good foundation for many energy harvesting sensor nodes. The power-good signal provided by the bq25570 can be used as a trigger for Compute Through Power Loss activation, saving time and precious energy after a power outage.