Batteries Not Included: Wireless Sensor Solutions

According to the forecasts from Bosch, Texas Instruments, and Hewlett-Packard, the number of sensors installed annually between 2017 and 2025 is expected to total between 1 and 10 trillion. With this rapid growth – combined with requirements on data storage, processing, and power transmission – sustainability is an unavoidable factor.

And so, when it comes to operating sensors, any kind of energy harvesting solution is highly desirable - if not absolutely necessary.

Modern low-power sensors enable precision control across a wide variety of applications, either locally, remotely, or autonomously. They are key components, sold globally for a huge range of applications. Energy harvesting technology means that these wireless sensors can work end-to-end without a mains connection or the need to replace batteries. The flowchart in Figure 1 shows the energy harvesting element as an energy source for a typical sensor with very low power consumption.

Figure 1

Flowchart of a typical sensor with very low power consumption, powered by energy harvesting.

An autonomous sensor includes the following system elements:

• Energy harvester
• Sensor(s)
• Energy storage device
• Power management device
• Microcontroller
• Wireless connectivity element

One of the crucial challenges in designing this kind of sensor system is keeping within the energy budget. This includes the operating voltage, as well as the peak, quiescent and average operating currents.

The main types of power normally provided by nature or the environment are relatively easy to harvest. Energy harvesting techniques and sources include:

• Energy from light – sunlight and solar energy, as well as indoor and outdoor lighting, can be converted into electricity through solar cells. To work efficiently, these cells should be optimised for the typical distribution of incident light
• Mechanical energy – objects that vibrate or move with kinetic energy can generate electricity. They provide significant voltage when generated with piezoelectric materials and can easily provide hundreds of volts. However, this type of energy source has a very high internal impedance. It's so high that a piezoelectric energy harvesting device cannot generate much power. Additionally, the polarity of voltage and current – depending on the direction of the vibration change – is reversed

Table 1 shows the performance of the energy sources.

Table 1: Parameters of each energy source.

There are many different types and manufacturers of mini sensors for measuring physical quantities. Suppliers of MEMS chips include STMicroelectronics, Bosch Sensortec, Texas Instruments, NXP, Analog Devices, Seiko Epson, Infineon Technologies, Murata, Sensata and Melexis.

Wireless sensor solutions with very low power consumption require a power management device due to their variable voltage and current states. This component converts the delivered unregulated voltages and currents into regulated electrical energy that can be stored. It can also pass energy to the system load at the right voltage. Typically, a power management chip will contain circuits to protect both the load and the energy storage:

• Undervoltage protection – this protection circuit interrupts the power supply of the load and switches to the standby load of the energy storage when the output voltage of the energy harvester gets too low. This capability is important to protect the battery from excessive discharge, as it can be permanently damaged or lose part of its storage capacity
• Surge protection – this function is for monitoring the charging voltage. If the voltage gets too high, the power management chip either directs the excess charge to earth or electronically inserts a high impedance between the harvester and the energy storage. Whichever method is used, the aim is to protect energy storage
• Overcurrent protection – this protection circuit has a similar function to a household fuse. If the load draws too much power, the overflow circuit isolates it from the battery and makes sure it does not empty. When this feature is enabled, there is usually an error state in the system

An example of a power management chip is the MAX20361 from Maxim Integrated. This is a fully integrated control chip for generating energy from single or multi-cell solar sources. The device requires a quiescent current of 360nA and has a boost converter that typically starts from 225mV. To make the most of the power delivered by the source, the MAX20361 features Maximum Power Point Tracking (MPPT) technology, owned by Maxim Integrated, which efficiently processes from 15μW to over 300mW of available input power. The MAX20361 also has an integrated charging function with a protection circuit, which is optimised for Li-ion batteries. The chip can also be used to charge supercapacitors, thin-film batteries, and conventional capacitors. The charging circuit includes a programmable charging switch-off function. A description of the MAX20361 Evaluation Kit (Figure 5) can be found on the RS DesignSpark website.

The choice of MCU (microcontroller unit) is crucial for embedded systems with very low power consumption. Ideally, it will have the following features as a minimum:

• Multiple power-down, sleep, or power-off modes to maximise battery life
• Good performance for fast and efficient processing
• Very fast wake-up times from power-down modes

The latter is important because the circuits of the sensor system are designed to last as long as possible in a low consumption state before moving to an operating mode that consumes more power. The Renesas RX111 MCU, for example, offers these features.

The supply voltage requirements of the MCU are not affected by the power-controlled run modes. Operation is always permitted over the component's entire range of 1.8V to 3.6V. However, the clock frequencies that can be used in high, middle, and low-speed modes depend on the supply voltage.

Various on-chip functions are stopped or shut down in the MCU's energy-saving sleep, deep sleep, and software standby modes:

• Sleep mode – the CPU is stopped and the data is saved. This reduces the dynamic power consumption of the CPU, which contributes significantly to the total operating current of the MCU. The CPU wakes up again in 0.21μs at a 32-MHz clock rate
• Deep sleep mode – CPU, RAM, and flash memory are stopped and the data is saved. Running at 32MHz with multiple active peripherals, the typical operating current is 4.6mA. The chip requires 2.24μs for the CPU to wake up from deep sleep mode and switch to run mode
• Software standby mode – the PLL and all oscillators except Sub-Clock and IWDT (Independent Watchdog Timer) are stopped. Almost all modules of the RX111 CPU, SRAM, Flash, DTC (Data Transfer Controller) and peripheral modules are stopped and the data saved. The power-on-reset circuit keeps running. The IWDT, RTC, and LVD (Low Voltage Detection) modules can also be operated if required. The power consumption in this mode is 350nA to 790nA, depending on whether the LVD and RTC functions are used. When waking up in 4 MHz run mode, CPU operation begins after a delay of 4.8μs. When waking up in superfast 32-MHz mode, the waiting time is extended to 40μs

Although the sleep, deep sleep, and software standby modes of the RX111 MCU are very helpful in reducing power consumption, other techniques can achieve further output reductions. For example, different clock signal-frequency division ratios can be set individually. Each peripheral module in the RX111 also has a separate stop control bit.