Batteries not included!
C.P. Ravikumar, Texas Instruments
One student who read the blog entry "The Bubblemaker" was inspired to think about energy harvesting and sent his comments. He wondered if the heat energy dissipation due to the switching activity in a CMOS circuit can be harvested to recharge the source, thereby reducing the energy drawn from the supply. This reminded me of how a seminar speaker (or a professor who teaches a class) energizes the class on the one end and harvests some energy from the participants on the other. Enthusiastic participation from the audience will boost the speaker's energy level. Ditto for the blog writer. Comments from readers will energize the blogger. And here I speak from experience - I felt so refreshed by the comments that I decided to blog about energy harvesting!
But first, let us talk about a movie!
“Batteries not included” is a “small print” that electronic product sellers have included on the boxes for many years. Steven Spielberg and Matthew Robins also made a science fiction film with the same title. The plot of this movie revolves around an old couple that is running a failing restaurant; their only son has died in an accident, but the mother is in denial. The couple gets unexpected help from an alien species that can rebuild things from scrap. What is more, these aliens derive their energy from scrap – they don’t need batteries to do their work! Perhaps we are a step closer to fiction today, with energy harvesting in electronics becoming a reality. While it would be certainly an exaggeration to say that applications such as electronic games will run without batteries and harvest energy from their surroundings, there are new applications where this would be possible. The ultra-low-power microcontrollers from Texas Instruments achieve a level of processing efficiency that many applications no longer require traditional batteries. These applications include wireless sensor networks that may involve sampling various sensors and communicating wirelessly. Since wireless sensor networks must have “infinite” life-time, energy harvesting is perhaps the only way they can be powered.
Battery Technology and Embedded Processing
Batteries such as lithium-ion cells have been used to power mobile electronics for several decades. However, the undesirable impact of batteries on the environment is well known. Innovations to improve the life time of batteries continue to take place, but the power efficiency of batteries has not kept pace with that of electronics. While Moore’s law predicts doubling of device densities every two years, battery capacity doubles every 10 years. Earth-friendly batteries, such as bionic batteries, are also being developed. For example, Prof. Angela Belcher of MIT is using viruses to grow nanowire electrodes required in bionic batteries. The weight of the batteries is another concern – to generate about 200 Watt-hours of energy, the weight of the battery must be about 1 kilogram. We certainly do not wish to carry the batteries for our mobile phones in suitcases. For applications such as wireless sensor networks, such solutions will be unfeasible.
Alternate sources of energy
Unconventional sources of energy, such as solar energy, wind, or vibration, can help generate very small amounts of energy in a given time. For example, the harvested power from outdoor Sun light is as 100 mWatt/sq-cm at 100% energy conversion efficiency. This means if you have an 8cm x 4cm solar panel that is 100% efficient, it may generate 32 Watt-hours of energy in 10 hours. Energy efficiency of panels may be lower in reality due to dust, etc. If indoor light is used, there is a 1000x reduction in this energy. When vehicles are driven on bridges, they vibrate; the energy from vibrations is about 800 microWatt/sq-cm. Radio frequency emissions such as GSM and WiFi also have energy in them, but the power that can be harvested is only 0.1 microWatt/sq-cm (GSM) or 0.001 microWatt/sq-cm (WiFi). To be able to use these, the electronic systems must be able to operate on small voltages and must be highly energy efficient. Let us consider the main sources of power in an embedded application:
- “Digital” power - Gene Frantz, Principal Fellow of Texas Instruments, has made a simple law to predict the energy efficiency of digital systems - roughly every 18 months, digital systems designed using the new CMOS technology will require half the power to deliver the same MIPS. This is possible due to better power management techniques that VLSI designers have invented (and continue to improve) over the years. The MSP430 16-bit microcontroller has active power consumption in the range ~200 µA/MIPS and standby current smaller than 1 µA. The MCU operates at speeds that are modest (< 25 MHz) and draws small peak current and simple power supplies are sufficient. It is therefore ideally suited for applications such as tamper-proof energy meters that need to be sealed and powered for tens of years.
- “Analog” power – Analog circuits such as ADC, DAC, amplifiers, etc. draw power continuously from the supplies. Mixed-signal processors such as the MSP430 are built as “SoC” and integrate analog circuitry and digital circuitry onto the same device. With the help of a low-power fabrication process and special analog design techniques, even the analog power can be brought down significantly.
As a result, applications built around low-power “mixed-signal” processors such as MSP430 can depend on “micro energy harvesting.” I recently read about a project from MIT which uses three sources (light, heat, and vibrations) to harvest energy. The energy harvester requires a means to convert ambient energy into electrical energy and a way to store the energy for later use. The storage of energy can be done on Lithium-ion batteries, thin-film batteries, or “Super Capacitors.” Researchers are looking at new applications that can use the small amount of energy available from energy harvesting. A wireless sensor network embedded into a bridge to monitor its structural properties is an example where energy harvesting can be a natural choice. Even in applications where conventional batteries are available, energy harvesting can be used to extend the life of batteries.
Applications
Wireless sensor networks can help farmers monitor the moisture content in the soil and irrigate the fields in a smarter way. Researchers have also been considering using wireless sensor networks to predict natural disasters such as a Tsunami or a volcano eruption. Such networks can be powered using solar energy harvesting. There are also medical applications where the body temperature can be used to power the implanted electronics. Consumer electronics and "smart home" applications can also benefit from energy harvesting.
More reading
Interested professors and students will find here materials for a course on Wireless Sensor Networks based on the eZ430RF2500. Information on a solar energy harvesting kit from TI is available from here. An interesting project on energy harvesting boosterpack for MSP430 launchpad is available here. Find information about several energy harvesting kits here.
All this blogging has drained my energy and I must recharge myself with coffee. Any microjoules you can send my way through comments will also help. Meanwhile, I hope you enjoy the cartoon I have included below.
Art by Ananya Ravikumar
