Wanted for the Internet of Things: Ant-Sized Computers
If the Internet is to reach everywhere—from the pills you swallow to the shoes on your feet—then computers will need to get a whole lot smaller. A new microchip that is two millimeters square and contains almost all the components of a tiny functioning computer is a promising start.
The KL02 chip, made by Freescale, is shorter on each side than most ants are long and crams in memory, RAM, a processor, and more. The genesis of the chip was a customer asking for help creating a wireless device small enough to be easily swallowed and cheap enough to be considered “digestible.” Freescale is now offering the chip for general sale, and also embarking on an R&D push to create more tiny computers that also include sensors and wireless data connections.
“The Internet of things is ultimately about services, like your thermostat connecting to the Internet and knowing when you’re coming home,” says Kaivan Karimi, director of global strategy for microcontrollers at Freescale, “but the technology those [services] are based on is embedded processing and sensors.”
If connected sensors are to be spread throughout the world around us, those technologies need to shrink in size, power consumption, and price, says Karimi. Freescale is betting that one of the best ways to do that is to integrate, into a single chip, components such as processors, memory, sensors, radios, and antennas that would usually be laid out across a circuit board.
Freescale will start offering the KL02, and some slightly larger microcontrollers, with Zigbee or low-power Bluetooth wireless integrated later this year. Wireless connectivity is added by adding the guts of a radio chip to the current designs. The company is also working to refine technology for packaging chips and other components together to enable many more millimeter-scale computers.
“All these heterogeneous things need to come together and be integrated,” says Karimi, “but we have to figure out how these components can coexist without degrading their performance.”
Bringing sensors and other components close together creates challenges because they each produce their own kind of electronic noise that can interfere with the workings of other components. It’s an area of chip engineering that is suddenly more important than processing power, says Karimi. “It is a packaging exercise, not a Moore’s Law problem, and the ultimate miniaturized version requires different sorts of packaging technology from what we’ve used in the past.”
One challenge in designing compact chips is that flash memory, the kind found in smartphones, creates interference for radio chips. In chips too small to avoid the problem, Freescale’s engineers design tiny Faraday cages around memory to shut in this electronic noise.
Freescale is betting that a technology called redistributive chip packaging (RCP), largely developed in-house, is going to make it possible to overcome similar problems. It has been used for some years in defense systems that needed very compact electronics able to withstand extremes of heat and pressure. “It isn’t a futuristic technology, because parts of it have been around in very specific applications,” says Karimi. “It does allow for packages with a very small footprint area, and we will see this technology migrate to consumer, industrial, and auto [uses].”
RCP also makes it possible to stack up components and chips; one possibility being explored is using the process to integrate antennas, as well as radio circuits, into chips.
“Such wafer-scale packaging is getting close to ‘smart dust’ design points,” says Prabal Dutta, an assistant professor at the University of Michigan, referring to the idea that very cheap, tiny sensors could eventually be scattered like dust to gather data.
However, Dutta points out that packaging alone can’t solve all the problems of making tiny, multifunctional computers, and says work will be needed on the components being packed together. “All of the system components—sensing, computing, wireless communications, data storage, and power conversion, will need careful attention,” says Dutta, with power consumption a particular challenge. “As a [miniature computer]’s length shrinks, its volume falls cubically, meaning battery capacity falls quickly,” he points out—a problem Dutta faces in his project designing radio-equipped sensors just one cubic millimeter in size.
Karimi agrees that batteries are a problem, saying that Freescale is working with partners developing energy harvesting components—of heat, radio waves, or light—that could power very small devices.
Freescale is not the only chip company that sees potential profit in supplying chips for a new wave of sensor and other tiny computers that feed data to the Internet. The company’s KL02 chip is based on a design announced by ARM last year as the world’s most energy-efficient microprocessor, which other companies are also licensing. However, Karimi says, the RCP packaging technology crucial to Freescale’s ambitions is protected by patents.
The original version of this article stated that Freescale would add Wi-Fi to its compact chips later this year. In fact, other wireless technologies that use the same frequency band but with lower power consumption are to be added.
Geoffrey Hinton tells us why he’s now scared of the tech he helped build
“I have suddenly switched my views on whether these things are going to be more intelligent than us.”
ChatGPT is going to change education, not destroy it
The narrative around cheating students doesn’t tell the whole story. Meet the teachers who think generative AI could actually make learning better.
Meet the people who use Notion to plan their whole lives
The workplace tool’s appeal extends far beyond organizing work projects. Many users find it’s just as useful for managing their free time.
Learning to code isn’t enough
Historically, learn-to-code efforts have provided opportunities for the few, but new efforts are aiming to be inclusive.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.