Design of wireless sensor nodes for the Internet of Things

Wireless sensor nodes (WSN) play a key role in promoting the development of the Internet of Things (IoT). The advantages of WSN are that it has extremely low power consumption, extremely small size, and easy installation. For many IoT applications, such as those installed outdoors, WSN can be powered by solar energy. When there is light indoors, the system is powered by sunlight and charges a small button battery or supercapacitor to power the system when there is no light.

In general, wireless sensor nodes are sensor-based devices that monitor conditions such as temperature, humidity, or pressure. The nodes collect data from any type of sensor and then wirelessly transmit the data to a control unit, such as a computer or mobile device, where the data is processed, evaluated, and action is taken. Ideally, the nodes can obtain operating power from energy harvesting mechanisms and become standalone devices. In general, the process of energy harvesting is the process of capturing and converting very small amounts of energy from sources such as light, vibration, or heat into electrical energy.

Figure 1: Example of energy harvesting system design

Figure 1 shows a block diagram of an energy harvesting system. Energy is collected by an energy harvesting system (e.g., a solar panel) and converted into stable energy by a power management integrated circuit (IC) (PMIC) and stored using low-leakage, low-impedance capacitors. This energy can be supplied to a sensor interface load (e.g., a microcontroller MCU), which is a sensor that transmits data wirelessly. In this figure, the energy harvesting sensor (EHS) is the wireless sensor node. 

Figure 2: Example of a wireless sensor node system

Figure 2 shows a block diagram of a wireless sensor node. Here, processed sensor data is transmitted wirelessly via Bluetooth Low Energy (BLE). BLE is a standard for short-range, low-power wireless applications to communicate status or control information. BLE operates in the 2.4 GHz ISM band and uses binary frequency shift keying (GFSK) modulation, which supports a data rate of 1 Mbps.

The power management IC is used to stabilize the power required by the energy harvesting device to support its ultra-low power operation. For example, the current consumption of the Cypress S6AE103A PMIC device is as low as 280 nA and the startup power is 1.2uW (see Figure 3). Therefore, in a low-brightness environment of about 100 lux (lx), a compact solar cell can still obtain a small amount of energy.

Figure 3: Block diagram of the S6AE103A PMIC device for energy harvesting

Efficient design of wireless sensor nodes

Let us consider the steps involved in designing a wireless sensor node:

Step 1: Select Hardware:

On the hardware side, you need the appropriate sensors, an MCU and PMIC that can ultimately be powered by the energy harvesting device. You may need additional passive components, depending on the design. 

Sensors can be analog or digital. Many sensors available today are digital sensors based on the Inter-Integrated Circuit (I2C), Serial Peripheral Interface (SPI) or UART interfaces. Sensors with very low power consumption are also available. To keep device costs low and form factors small, an MCU with integrated BLE can simplify design and reduce time to market. To further speed up design, many manufacturers use fully integrated, fully certified programmable modules such as Cypress EZ-BLE Modules. Modules consist of a main MCU, two crystals, chip or trace antennas, expansion boards, and passive components. Since these modules already have the necessary BLE certifications, products can be quickly brought to market.

Figure 4: BLE module example: Solar-powered Bluetooth Low Energy sensor beacon CYALKIT-E02

Step 2: Design firmware and estimate power consumption 

After choosing a programmable MCU, the next step is to write the appropriate firmware. The basic functions that the firmware needs to have are an interface to collect sensor data, a BLE component or stack to transmit data wirelessly, and a CPU that can handle firmware processing.  

Since ultra-low power operation is key, the total current consumption needs to be considered from the beginning. The total current consumption is the sum of the current consumed by the sensor and the current consumed by the MCU. Since sensors usually do not consume much total current, the focus should be on how to minimize the current consumed by the MCU. Before optimizing the current, consider the three main components that consume current within the MCU: the CPU, the sensor interface module (such as I2C, SPI, etc.), and the BLE subsystem. Here, when the radio is turned on (such as BLE Tx and Rx), the main consumer of current will be the BLE radio. 

Embedded MCUs offer various low-power modes to reduce current consumption. Firmware designers need to consider these low-power modes and design code so that average current consumption is minimized. For example, sensor data does not change instantaneously, and the firmware needs to scan the sensor data occasionally (for example, every 5 to 10 seconds, depending on the sensor). The sensor's read data is transmitted wirelessly via BLE. 

As far as the BLE firmware is concerned, the sensor can send data along with the BLE advertising packet. We recommend not to forward too much other data along with the advertising packet, as this will further increase the current. Between the advertising interval and the sensor scanning interval, the MCU needs to enter a low-power mode, such as the "sleep function". The low-power timer is like a watchdog timer, which can wake up the device when the timer counts down. In order to use low-power operation, the MCU is optimized to provide a BLE internal timer that can wake up the device that has entered the sleep function when the advertising interval ends. Figure 5 shows the firmware flow of the operation.

Figure 5: Firmware flow for efficient wireless sensor node design

Once the firmware is designed, you can measure the current. You can use a prototype board to measure the current. Please note that the current of the MCU's startup and low-power modes needs to be measured separately. As long as you know the time the MCU is operating in startup and low-power modes, the average current consumption is: 

(Iactivex Tactive) + (ILowPower x TLowPower)

Tactive+TLowPower

Tactive+TLowPower

With the average current figure in hand, you can then multiply it by the PMIC voltage to find the average power. 

Step 3: Optimize firmware to minimize average current consumption

It may be the case that the initially calculated design power is too high for the solar PMIC to support. If this is the case, you will need to optimize the firmware. Here are a few effective ways to do this:

Execute optimized MCU boot code: When the MCU is booting, you do not need to use a high-frequency external clock such as a 24MHz crystal clock to operate BLE. By turning this clock off initially, energy can be saved. Furthermore, the clock crystal can use this time to stabilize, which is also one of the components of the boot. As these clocks stabilize, the MCU can be adjusted to a low-power mode again, and the internal low-frequency clock can wake up the device when the clock is ready. In short, the execution time of the boot code can be long, and the firmware designer needs to minimize the startup current consumption.

a. Reduce the main CPU operating frequency

b. Before entering low power mode, control the drive mode to prevent MCU pin leakage current.

c. If the MCU supports any debug interfaces, disable them.

These steps help reduce the average current consumption.

Step 4: Design the Hardware

With the power-optimized firmware in place, it is time to design the hardware based on the PMIC. Figure 6 shows a simple energy harvesting based PMIC design.

Figure 6: Simple energy harvesting design

The PMIC first stores solar energy in a storage device VSTORE1 (VST1), in this case a 300-μF ceramic capacitor. When VST1 reaches VOUTH V, the energy can be sent to the MCU. But this simple energy harvesting design cannot operate all day because there is no backup capacitor. Let's take a look at how backup capacitors are added to PMIC devices and how capacitors can help the MCU.

Figure 7: Energy harvesting and backup capacitors

The energy required to operate the WSN is first stored in VST 1, and the remaining energy is used to charge VST 2. The energy stored in VST 2 can be continuously provided to the WSN in the absence of light. In addition, an additional button battery can be connected to the PMIC to increase reliability, as shown in Figure 8.

Figure 8: Energy harvesting from multiple power inputs

The PMIC converts the two power sources so that the WSN can operate in all conditions (even when there is no light). The conversion occurs automatically so that energy is supplied to the WSN when needed. Therefore, this may be the most appropriate hardware design for WSN.

Step 5: Design the User Interface

The user interface design connected to the wireless sensor node can be a mobile phone application that receives data transmitted by WSN. It is as simple as that. Since the sensor data may appear at fixed locations in the broadcast packet, the BLE application can be designed to extract the relevant data from these locations and display the data on your mobile phone. This technology can be used to manage complex networks composed of multiple WSNs.