Can Ara create a hardware revolution?
The Ara phones will require significant innovations in antennas, interconnects, and mechanical structures—and a new layer of Android code. Most importantly, Google needs to enable and inspire a new community to start creating compelling modules that let users do things they can't do with today's smartphones.
"If it's just today's phone and the pieces come off, it will be a big yawn," said Kaigham Gabriel, deputy director of the Google R&D lab working on Ara.
To enable new modules, Google will have at least two contractors make a blank module platform it designed currently based on an FPGA. A follow-on version with an ASIC being developed by Toshiba will give engineers twice as much space to pack in their innovations.
To prime the pump, Google recently announced a $100,000 challenge for the best module design. It showed a working prototype phone that booted Android two weeks ago.
Taiwan ODM Quanta is working on the basic skeletons that will carry the modules. The 9mm thick base includes a flexible circuit board and 400mAh battery for system functions.
In terms of its core technical challenges, Google and partners are working on new kinds of antenna designs integrated into the module packages. They require new antenna modelling tools and deposition techniques.
Designers are still experimenting with different capacitive, inductive and hybrid contactless interfaces to link the modules and skeletons. The approach keeps the interconnects thin while supporting data rates in tens of GHz. On the electrical level, Google chose the MIPI M-Phy and Unipro interfaces because they are widely used in smartphones and do not use the master/slave approach of alternatives like USB and PCI Express.
Mechanically, Ara is using electro-magnets to hold modules with a 30 Newton force when turned on. Turned off, they exert a 5 Newton force, enough not to fall off but still be easily removed by users. The technology is well known on factory floors where it is used to lift cars but less well used in miniature consumer forms.
Finally, adding a new driver communications layer to Android is also a big challenge given the breadth and pace of Android development. But the dynamic software interface is needed to make the OS one that can deal with hardware components coming and going.
In the end, Gabriel expects adding modularity will increase costs about 25 per cent. The phones will also be larger and consume more power than integrated models, and their value must rise above all those costs.
There are management issues, too. All this work is coordinated by just four Google researchers working with third parties in some 22 universities and mainly small companies. Some large companies have shown interest.
|Related Articles||Editor's Choice|