Micro-Harvesters Powering Wireless Devices

Today, sensors like thermometers and fire detectors are connected by wires. In the future, energy-efficient smart buildings will need sensors capable of measuring multiple parameters: humidity, CO, CO2, O3, fire, illumination, pressure, presence and more. Installing sensors with wires has limitations, as cable installations might not be possible in old buildings, and significant design effort is required in new buildings for wire installation. In addition, wire installation is costly. Any change to existing wiring needs to be properly planned, and in many cases, further changes are not possible. Today, wiring is one factor limiting widespread sensor deployment.

Wireless sensors and wireless remote controllers can overcome the limitations of wired installations. In place of wires, radio frequency transmissions are used to send information. But how do you supply power to these devices? The first solution that comes to mind is to use a battery. Powering a remote wireless device with a battery works well, but the battery may need to be replaced every year or two. Changing batteries is not much of a problem if you have just four or five sensor units installed in your home. However, changing batteries is a bigger problem when many sensors and remote controllers are in place in a larger building. In an industrial building, maintaining large numbers of battery-operated devices is prohibitively costly.

Harvesting power from the environment

Instead of power from batteries, energy harvesting devices collect energy from the environment to drive wireless sensors and transmitters. You can now choose from multiple micro/nano harvesters based on electromagnetic technology, thermogenerators, vibration, piezoelectric systems, or dye-sensitised solar cells.

A typical wireless device includes an energy source, an energy management unit, a sensor, a microcontroller and a transmitter. Let's look at the steps involved in designing a wireless sensor powered with energy harvesting technology.


Figure 1: Dynamic Current Drain

Step 1: Determine and minimise power requirements

The first step is understanding the power requirements of your wireless sensor design and finding ways to reduce them.

To keep energy consumption to a minimum, remote wireless sensors typically operate in three modes with different levels of activity: sleep, active, and transmission. Engineers commonly minimise power consumption by adopting discontinuous operation, whereby most of the time the wireless sensor is operating with its energy drain minimised in sleep mode.

Periodically, the device is activated to perform internal functions and to transmit data. This activation may happen every minute for a temperature sensor or by a key press on a remote controller. The microcontroller periodically activates the sensor, takes a measurement, converts it to a message, and then activates the transmitter to send information to a receiver. Information is sent by RF transmission in short bursts, generally in the ISM (industrial scientific and medical) frequency bands. Only a few volts are required to drive the transmitter, but the current generally is in the range of 10mA to 100mA. After the transmission, the device is placed back in sleep mode. Common current consumption in sleep mode is on the order of 100s of nA.

To quantify the amount of energy required for your device, it’s important to understand the power consumed in each mode and the amount of time the sensor spends at each energy level.

Typically power states:

Sleep: This state is the lowest level current, and this is where the device spends most of its time. Thus the energy required needs to be the lowest possible. You need to know the duration of this mode and maximise time spent in this state.

Active: The microcontroller is actively acquiring data from the sensor element, converting a measurement and passing the information to the RF transmitter. You need to ensure the microcontroller activates for the shortest time possible while performing these activities and promptly returns the device to the sleep condition. Energy drain during the active state is 100 to 1000 times the sleep current. One of the most common causes of high energy consumption in wireless sensors is a lack of attention to power management in the control software. Without vigilant power management, the device is kept in active mode for longer than necessary, resulting in significant energy draining off in a short time. When the energy source (either a battery or a micro-harvester energy cell) is prematurely depleted, the device is useless.

Transmission: After the measurement computation is performed during active mode, the information is transmitted. Transmission drains the most power because it has to drive the RF transmitter, and in some cases, a power amplifier. Current is typically 10 to 100mA. As current drain at this level is high, it’s critical to operate for only the time required and avoid unnecessary energy usage. Energy at this level is 100,000 to 1,000,000 times the sleep current. This means that 1 second of activity at this level drains the same energy as about a day to over a week of operation in sleep state.

Choosing the right communication strategy plays an important role in the energy consumption. Proprietary communication protocols are often more energy efficient because they employ short messages, which require shorter TX times. In contrast, standard protocols need to use the defined format with overhead resulting in long coded messages that require long transmit times.

Making Measurement

The most common measurement approach is to use an oscilloscope with a current probe. While this setup allows you to correlate current peaks in the order of tens of mA with transmitter timings for the peaks, the measurements needed are very close to the limits of the current probe’s sensitivity. Furthermore, when using current probes, you need to periodically zero the probe to achieve measurement repeatability. Oscilloscope vertical resolution is also a big limiting factor. But the biggest limit is that you don’t have visibility on what happens below mA levels where the device spends the majority of its time.

As an alternative to using current probes, you can use shunts combined with operational amplifiers. In that case, the voltage drop across the shunt (known as the burden voltage) reduces the voltage that reaches the device, so less voltage is delivered. The lower voltage may cause improper device operation when the reduced voltage is close to the minimum operating voltage of the device. Also, the operational amp's offset and gain errors do not allow you to reach the required accuracy across the device’s full operating range from 100s of nA (sleep current) to 100s of mA (transmit current).

You can argue that current levels can be measured using a special version of the device’s software that is capable of setting the device to a fixed state. While this is possible, using this approach means that final operating software will be different from the version used for test, and it may introduce differences or include bugs that keep the device operating in a high-consumption state for longer than necessary and drains more energy than expected. Clearly, making current measurements involves more than just testing device hardware. You also have to evaluate the effects of the software that drives the device. Low energy consumption is the combined results of optimisation in both areas.

The new Agilent patented seamless ranging technology available on a source measure unit enables you to acquire dynamic current drain with the high accuracy required. As the current drawn by the DUT changes, the seamless ranging available on the new N6781A and N6782A 2-quadrant SMUs automatically detects which current measurement range is needed, changes to that range instantaneously, and returns the most precise measurement. This gives you an unprecedented vertical resolution. It is like having a 28-bit vertical resolution with a timing resolution of 5µs.

When used with the N6705B DC power analyser and 14585A control and analysis software, the N6781A allows you to acquire current, voltage and power measurements and to view them in an oscilloscope- or data-logger-like display. You can place graphical markers on acquired traces to read off multiple values, including energy (Watt-hours or Joules) over user-defined time windows.

Step 2: Engineer the energy source

Once the device power requirements are defined and minimised, you can move to the next step and properly engineer the energy source. Energy harvesting in a wireless, wall-mounted, remote light switch is a perfect example.

A wall-mounted wireless switch is activated by pressing a mechanical element that is connected to an electromagnetic power generator. No wires are needed for switches using this technology, so the switch can be placed in a location where wiring is not possible or would be extremely complicated.


Figure 2a: Schematic diagram: energy drain measurement setup

The measurements challenges associated with this technology include:

-Power is only generated during physical movement;
-Peak power is required for the whole device operation, from pressing the key, to activating the device, to completing the transmission;
-The device must store energy and enable the device to activate over the required time, and;
-The whole process takes a few milliseconds to complete.

An advantage of this particular design is that it operates only during the time when the key is pressed, causing energy to be available; there is no a sleep state. Here, the challenge is to wake up the device in a short time.

Measuring the energy required is made simple with the N6781A SMU and N6715B DC power analyser. The SMU is used as a source in combination with the built-in arbitrary generator to generate a voltage pulse to activate the device. Seamless ranging selects the most appropriate current range and returns the most accurate current measurements.

Using the 14585A software set to its power measurement mode, you can position markers on the scope-like display and directly read how many µJ are required from activation to completion of the transmitting process. In the example shown in Figure 2, the device requires about 80µJ. This parameter is the key to define how much energy the harvester must provide.


Figure 2b: Energy drain by wireless transmitter

Once you know how much energy you need the harvester to provide, you can select the most appropriate harvester to provide the energy in a single narrow pulse. Next, you need to select a means to store energy in a storage element such as a capacitor. After selecting a capacitor, you can test it by connecting the harvester, rectifier and capacitor circuit and using the SMU as an electronic load set to drain constant current (CC load). In the measured example shown in Figure 3, the 20mA setting is used. The SMU is triggered on a rising voltage edge with a single acquisition. Power measurement is displayed.


Figure 3: Energy measurement result with markers showing result in µJ

Testing the whole device

With a suitable harvester and wireless transmitter identified, the final step is to connect the energy source, storage element, and user device together and to see how they operate. Voltage and current measurements are required to quantify its performance.

Measuring the voltage is easily accomplished. However, measuring the low dynamic current in the circuit can be problematic. You might use a shunt, but the voltage drop across it reduces the voltage that reaches the device and may impact the operation. Potentially, it might not operate if the voltage at the device falls below the minimum required for operation.

The Agilent N6781A SMU can operate in ammeter mode with zero voltage drop across its connection points. This enables it to measure the current drain with high dynamic range and present the data on a scope-like and/or data logger display. The voltage is measured with the auxiliary DVM available on the N6781A SMU. Both current and voltage waveforms are captured and displayed on the same screen. Note that V and I are synchronously sampled, so a power trace can also be displayed. You can use markers to see the energy consumed during the whole operating cycle.


Figure 4: Schematic diagram - N6781A SMU used as virtual shunt

As device activation comes from the mechanical movement of the electromagnetic energy harvester, it’s a single-shot event. Hence the rising voltage edge can be used as a trigger point for the N6705B DC power analyser. In addition, its trigger output connector provides a signal that can be used to start other measurements with instruments such as a scope or spectrum analyser.

Large deployment of wireless sensors and remote controllers is limited today by battery life. Adoption of low-power microcontrollers, efficient transmitters, storage devices, and power management designs will drive deployment only if these devices are optimised for low power consumption. Low-power devices can then be combined with micro-harvester energy sources.

Agilent provides the tools needed for wireless sensor testing. These powerful test tools accelerate power management device development with an easy-to-use, high-resolution SMU that can help you achieve new device designs in a short amount of time.

The Agilent N6781A source measurement unit for battery drain is the highest performance module in the N6700 modular power supply family, which includes more than 30 modules and 4 mainframes that offer basic power supply performance to autoranging and precision modules. In January 2012 Agilent added seven new higher power modules in the power range up to 500W per channel to meet your application needs. Learn more at www.agilent.com/find/N6700.