Internet-of-Things Radio Chip Consumes a Little Power to Save a Lot

MIT engineers build a 100-fold more efficient transmitter by stopping up leaky transistors

2 min read
Internet-of-Things Radio Chip Consumes a Little Power to Save a Lot
Illustration: Jose-Luis Olivares/MIT

Even as electronics become more efficient overall, many gadgets are also requiring power for a new use: to connect to the Internet. Engineers at MIT presented new research this week at the IEEE International Solid-State Circuits Conference that could help keep the power draw of connected devices in check.

Connected devices, which are seemingly everywhere these days, all require power to send data wirelessly. The need of each thing on the Internet of Things might be small, but the number of devices is expected to more than double between now and 2020 to more than 30 million, according to ABI Research.

Anantha Chandrakasan, professor of electrical engineering at MIT, presented a new transmitter design that reduces power leakage when a radio is in the off state by 100-fold. Even though it has ultra-low power needs, the system can still provide enough power for communication across different standards, including Bluetooth and 802.15.4.

“A key challenge is designing these circuits with extremely low standby power, because most of these devices are just sitting idling, waiting for some event to trigger a communication,” Chandrakasan told MIT News. “When it’s on, you want to be as efficient as possible, and when it’s off, you want to really cut off the off-state power, the leakage power.”

Chandrakasan said the key was to reduce the leakage of power in the transistor. Even when there is no charge applied to the transistor’s gate, it leaks some current. For devices that mostly sit idle waiting for a signal to power up, the slow leak can take a toll on battery life. (Limiting leakage was a main factor in two fundamental redesigns of transistors in computer processors.)

Arun Paidimarri, an MIT graduate student in electrical engineering and computer science and Nathan Ickes, a research scientist in Chandrakasan’s lab, applied a negative charge to the gate when the transmitter is idle, making the transistor a better insulator. Just a small negative charge, consuming just 20 picowatts of power, was able to save 10,000 picowatts in leakage.

“Ultralow leakage energy is critical for future sensor nodes that need the transmitter to be on only a very small percentage of time,” Baher Haroun, director of the Embedded Processing Systems Labs at Texas Instruments, said in a statement. Texas Instruments and Shell helped fund the work by Chandrakasan’s team.

The Conversation (0)

The Future of Deep Learning Is Photonic

Computing with light could slash the energy needs of neural networks

10 min read

This computer rendering depicts the pattern on a photonic chip that the author and his colleagues have devised for performing neural-network calculations using light.

Alexander Sludds
DarkBlue1

Think of the many tasks to which computers are being applied that in the not-so-distant past required human intuition. Computers routinely identify objects in images, transcribe speech, translate between languages, diagnose medical conditions, play complex games, and drive cars.

The technique that has empowered these stunning developments is called deep learning, a term that refers to mathematical models known as artificial neural networks. Deep learning is a subfield of machine learning, a branch of computer science based on fitting complex models to data.

Keep Reading ↓ Show less