Wireless sensors are becoming more common and will continue to be developed and deployed in the near term. The technological challenges of wireless communications are steadily being overcome, permitting sensors to be interconnected into intelligent networks. The increased density of systems-on-a-chip (SoCs) is leading to increased functionality and smaller sizes for wireless sensors. But one key challenge still remains: power.
Wireless Sensors Emerge
Today, wires for power and communications are one of the limiting factors of widespread sensor deployment. Let’s look at an example in building controls. Smart buildings will use wireless sensors to monitor humidity, carbon monoxide, carbon dioxide, fire, and lighting. Wired installations might be not possible in old buildings and too costly in new buildings.
Batteries have been an enabling technology for wireless sensors. As the average operating power consumption of the sensor is driven down, it becomes practical to power wireless sensors from long-lived batteries. For example, I have home security sensors that are wirelessly connected to the alarm system. These sensors are running on lithium batteries that will last 10 years, so I will have to go through the trouble to replace them once every decade. This is a workable solution for these wireless sensors.
But for some wireless sensors, battery replacement isn’t practical. The sensors could be in remote locations where access isn’t easy. Or, the sensors could be so small and in such large numbers that a technician just can’t cost-effectively get to all of them to replace their batteries. For these applications, some form of energy harvesting would be ideal.
Energy Harvesting And Power Management
Energy harvesting devices can collect energy from the environment and then use that energy to power wireless sensors. Today, many energy harvesters use electromagnetic, Peltier cell, piezoelectric, and solar dye cell energy sources. A typical energy harvesting wireless sensor includes an energy harvesting source, an energy management unit, a sensor, a microcontroller, and a transmitter.
The key to a successful wireless sensor is power management. Typically, the energy is collected at some extremely low rate and accumulated in some form of storage device, like a supercapacitor or rechargeable battery. The sensor will then need to very sparingly use its stored energy for its primary task.
Often, the sensor sleeps for a long time consuming almost no power while it is harvesting energy. Current consumption in sleep mode can be as low as hundreds of nano-amperes. Then, the sensor will wake up and perform its task, sending information in a short burst of a few milliseconds (Fig. 1).
1. The waveform shows the wireless temperature sensor waking up three times and transmitting a pulse each time. The sensor sleeps for approximately 40 seconds between transmissions. The sleep mode current, between the pulses, is about 7.5 µA as shown by the markers.
The RF transmission is generally in the industrial, scientific, and medical (ISM) frequency bands. This active state power can be 1000 to 1 million times the sleep mode power. This means that 10 ms of activity can drain the same energy as 2.7 hours in sleep mode. Thus, choosing the right communication strategy (interval and transmission scheme) plays an important role in managing energy consumption.
Proprietary communication protocols are often more efficient because they employ short messages so short transmit times can be used (Fig. 2). Finally, after sending its message, the sensor will go back to sleep and collect energy until the next wakeup interval. The resultant power consumption profile is an extremely low-frequency waveform, with perhaps seconds, minutes, or hours of sleep time broken up by a few milliseconds of operation.
2. This close-up view of one of the transmit pulses actually is a burst of around 200 ms in duration with a peak current of 14 mA, or 2000 times more current than sleep mode.
Design And Test Challenges
Engineers must carefully design their wireless sensors to be as energy efficient as possible. From a test equipment perspective, they need tools that can measure micro-amperes or nano-amperes of current flow (during harvesting operations) for long periods of time and accurately accumulate that information to determine the total amount of energy harvested and total amount of energy consumed during the sleep operation. Those same tools will have to be able to accurately measure the brief higher currents consumed during sensor operations.
Normally, an engineer might turn to an accurate ammeter to make these measurements. But ammeters that can accurately measure micro-amperes or nano-amperes will typically have long integration times, making it impossible to measure the short high-current active state while the sensor does its task. If the ammeter is set up to accurately measure the high current (tens to thousands of milli-amperes), it will not be able to accurately measure the sleep mode/harvesting currents.
An oscilloscope or digitizer is another possibility, as either of these instruments can capture the narrow pulses (low duty cycle) of the wireless sensor as it goes in and out of sleep and active modes. However, using an oscilloscope or digitizer requires a current transducer, such as a current shunt. It is nearly impossible to select a current shunt that will allow the oscilloscope or digitizer to accurately measure micro-amperes or nano-amperes and also accurately measure tens to thousands of milli-amperes.
Summary
So the challenges of successfully designing and deploying a wireless sensor that gathers its power from energy harvesting are more than just designing a small, highly functional sensor and wireless communications system. The sensor must also be an extremely energy efficient system.
Testing the designs and validating the operation of energy harvesting wireless sensors involves specialized tools. Thankfully, a few test and measurement equipment providers have developed a set of tools for sourcing and measuring dynamic current and logging power consumption. These tools are well suited for this application (Fig. 3).
