At I/O 2014, Google has publicly turned on its modular Project Ara smartphone for the first time. Considering Project Ara nothing more than some pretty concept art back in October 2013, it’s impressive how quickly Google’s Advanced Technology and Projects (ATAP) group has produced a working prototype. If all goes to plan, Google hopes to sell the modular Ara smartphone in 2015, starting at $50 for a bare-bones endoskeleton that you plug other modules into. During the ATAP tech session at I/O 2014, Ara chief Paul Eremenko also gave us lots of technical details about how to actually build a modular phone — no mean feat, as you can probably imagine.
The Project Ara prototype, called Spiral 1, was given its first public outing at Google I/O yesterday. As you can see in the photos throughout this story, Spiral 1 actually looks surprisingly like a finished device — except, of course, for the fact that the power switch is a jumper that needs to be physically shorted. If you want to see the device being powered on, skip through to around the 26:10 mark in the video below. The manual jumper method is shown at around the 27:40 point. Sadly, the device actually crashes and fails to reach the Android desktop — but hey, it is just a prototype.
Eremenko also gave us lots of juicy technical details about Ara. The current prototype uses FPGAs to implement the industry standard UniPro MIPI packet-switched network protocol to provide an interconnect between the modules. The antenna modules use the metal endoskeleton (phone chassis) to boost reception.
Android, which will power the Ara smartphone, will eventually be updated to support hotplugging of third-party modules — and later, Eremenko expounded on that theme with this interesting tidbit: “Think of [Ara] as an analog of the Android app ecosystem, but in hardware.” To kickstart development of these third-party modules, Google also announced a competition to design and implement a novel module, with the winner walking away with a grand prize of $100,000.
A lot of time was spent discussing Project Ara’s most significant barrier: Overhead. “The principle challenge to modularity is overhead,” Eremenko said. “What we found was that Moore’s law, the miniaturization of electromechanical components, and a modern data protocol, could get the modularity penalty at a system level down to around 25% across the board; PCB area, device weight, and overall power consumption.” Eremenko specifically discussed the electropermanent magnets that will keep the modules in place, and the socket connectors (they have to be small and thin, but still effective at carrying different signal frequencies).
Interestingly, Eremenko also discussed battery technology. He noted that there are now battery chemistries with three times the energy density of standard lithium-ion batteries (LIBs), but with much reduced cycle life (i.e. they don’t last for 300+ recharges). To get around the power consumption overhead of a modular phone, you could use these high-energy batteries and switch them out more often. These batteries could also be smaller than current LIBs, giving you more space for other modules.