SparkFun will be closed Nov 26th and 27th for the Thanksgiving holiday. Orders placed after 2:00pm MT on the 25th will ship out Monday the 30th.

Google Ara Modular Smart Phone Developers Convention

The Ara Modular Smartphone Developers Conference was on April 15-16. We were there. This is what we learned.

Favorited Favorite 0

alt text

Image Credit: idownloadblog.com

On April 15th and 16th, Google held its first Ara Modular Smart Phone developer conference at the Computer History Museum in Mountainview, California. Ara is a project that is being spearheaded by Google ATAP, their “Advanced Technology and Projects” group, which is a stay-over from Motorola that Google kept after the Motorola sale. The Ara Project is their concept for a modular cell phone. They were forced to push the announcement of the project forward with the sudden popularity of PhoneBloks, a modular phone concept conceived by Dave Hakkens that went viral recently. The ATAP team had been working on the Ara modular phone for a while at that point, and decided to announce the phone as an answer to the popularity of the PhoneBloks idea.

Since the project’s announcement, the reception has been mixed. The reactions seem to fall into two camps: the “I want that gimme that right now” camp, and the “That will never work don’t even try and here’s why” camp. That reaction notwithstanding, the DevCon this weekend was full of about 400 people who had formed an “Okay, let’s see if we can make this work” camp.

Google’s message for the developers at the convention was clear; they need manufacturers and the developers-that-be to start building modules. They want to reach what they call a “demonstration at convincing scale,” i.e. have enough developers and companies on board and time-invested by 2015 that the project has enough outside investment to demonstrate value. They’ve released a v0.10 Development Kit complete with reference designs, a high-level description of the platform, and an industrial design language to insure that developers build modules to their form-factor specifications. The conference highlighted all of the major systems and components of the phone, and gave developers a good starting point for module design. What follows is what I took from the conference, from the hacker/maker perspective.

Ara Base Platform (The Endoskeleton)

alt text

The base platform, or “Endo” as the developers refer to it. is what Google plans to manufacture and sell. They have a vision of users purchasing a bare-bones endo from stores that includes only the screen and wifi module. That’s it. All the other modules will be purchased from their module store or other retailers. This way they can sell platforms, and not phone plans. Users will be responsible for purchasing a separate module from a cellular provider if cellular service is desired. This may be a frustrating step for consumers looking for a new phone, but for electronics geeks it’s great news; Ara will be an off-the-shelf expandable hardware platform with a touch screen and wifi module that runs Android out-of-the-box. To me that’s more exciting than the Ara’s viability as a mass-market cell phone.

The Endo will house all the electronics that run Android and connect to the modules. The central controller will be the Texas Instruments OMAP4460, the chip on the PandaBoard. The central controller will connect to the modules via the MIPI Unipro network protocol. The MIPI protocol was developed by a consortium of cell phone developers to create a fast, universal protocol for connecting peripheral hardware in cellular designs. Modules will connect to this protocol using a special chip that the Android will recognize. The chips require a special driver in Android, that is currently being compiled into the kernel.

Right now, the only method the Ara group has offered to connect a user-designed module to the backbone is an FPGA with an instantiation that talks to MIPI. This FPGA then tunnels protocols like I2C and GPIO to the module. The problem with this FPGA solution is that it is expensive, high-power, and large. For the future development, Toshiba has promised to manufacture a specific ASIC that will solve these problems and connect the module to MIPI natively. Google will provide a driver that will automatically recognize this ASIC, making module hardware development much more streamlined.

Electro-Permanent Magnets: How Do They Work?

Another interesting feature of the Ara phone is the way the modules are held in the Endo. Mechanically, the modules slide into the slots on the back of the phone and are held in the x-direction by dovetails. In the y-direction, however, they are held by electro-permanent magnets, or EPMs. Magnets that can be magnetized or demagnetized with a burst of electrical current. They draw no power unless they are changing state. The idea is that the user inserts the module, and the EPMs automatically magnetize to hold the unit in place.

alt text

Before the conference, I couldn’t find these magnets for sale anywhere. I cornered one of the Ara team members at the conference and asked where I might purchase hundreds of these, and I was told “No one is actually building them yet. The EPMs we have we had to build by hand.” Frustrating.

Industrial Design Language

alt text

Google spent a good deal of time at the conference discussing the Ara’s industrial design. They’ve developed a very specific set of rules for module developers. The modules can only be so big, so hot, and so high-power. The also talked at length for their plans for manufacturing the coverings for the Endo. Stated simply, they plan to 3D print everything. Consumers will be directed to a site where they can design a fully custom enclosure for their phone that will be 3D printed on a special printer that is being designed by 3D Systems. It’s certainly exciting to see a company planning to manufacture a mass-produced product on technology as young (and open) as 3D printers. Be sure to check out the MDK for a full run-down of their industrial design language.

Module Design Tools (Metamorphosis Software)

Along with hardware guidelines for developers, Google has contracted MetaMorph Software to build an interesting design tool for module developers. The tool is called CyPhyML. It is built upon the open-source CyPhy software developed by DARPA for rapid prototyping and virtual environment testing.

alt text

CyPhyML is not a CAD program, but rather an “umbrella” program that integrates several CAD programs and facilitates the design process by translating a CAD model designed in one program (like a layout in Eagle) to a model in another program (like a 3D model in Creo). CyPhyML takes the design from the part-level to the module level, and facilitates tests like Spice models and heat dissipation. The designers tried to keep the “utility programs” as open and available as possible. The tool is a bit too comprehensive to describe here, but check out metamorphsoftware.com for a basic run-down. The software has yet to be released, but the tool has a lot of potential for wide use in the prototyper market.

Licensing and Distribution (Play Store for Hardware)

To sell the modules and make them available and understandable, Google has plans for a “Play Store for Hardware.” In other words, the modules that they approve will be sold through a special repository, like the Play Store for applications, where users can search through tested-and-proven hardware modules to build a custom phone. There was not a lot of talk about the fees and processes associated with this “approval process”, but there wasn’t anything to suggest that a developer couldn’t develop and sell a “non-approved” module without going through an official Google approval process. How Google is going to deal with these renegade modules remains to be seen, but the promised availability of the Toshiba ASIC and openness of the CyPhyML software is a good sign.

What this means for DIY'ers

As I stated before, the most exciting part of the Ara project is that Google plans to mass-produce cheap Android platforms that include a touch screen and wifi module in stores. That in itself should get a geek’s blood flowing. As far as hackability, there is nothing to suggest that the Ara won’t be super-hackable. The ASIC that Toshiba has promised and the EPMs (that no one has promised, yet) seem to be the only pieces of specialized hardware that designers will require to build modules. The seeming “openness” of the CyPhyML software is promising, and may turn out to be a useful tool for hobbyists even if the Ara platform doesn’t succeed as a mass-market cell phone. To keep an eye on the project, bookmark the Ara Developers Google Group for the latest.

Comments 16 comments

  • A couple of things: is the core module (CPU, RAM and flash) part of the display? It would be nice to have interchangeable displays, as well; you could switch to your 7" to 10" display for watching movies and such, but carry the 4" in your pocket for normal use. Also, it would be nice if there were different core modules to choose from, like the OMAP mentioned, or the power Exynos 5 Octa. Let’s just hope the various vendors will all make their modules play nice with everything else.

    • No. The CPU/RAM is in a 2x2 module in the back, while the display is on the front. Note that the OMAP was used because they found it well-supported and easy to program on (Does not require NDAs, like many Snapdragon and Exynos chips). (Note: I wasn’t at the conference, but I livestreamed it)

  • It’s definitely a case where the intended cell phone user is just for marketing & the real consumer is going to be someone who wants to connect phone modules to robot microcontrollers. A standalone LTE transceiver that can plug into a microcontroller would be huge. There’s never been any standalone module beyond GSM. A standalone camera with the quality of a phone cam would be huge. A standalone Linux board for a phone would instantly blow past any current single board computer in miniaturization. The devil is in how much interfacing the unipro bus will cost & how much those phone breakout boards are going to add to the size.

    The only disappointment is for all the years we’ve had standalone GSM modules, no-one has ever gotten one to talk to another over its 2 mile range without requiring a cell phone plan. Any LTE module would be equally crippled.

    • It’s not true that there aren’t any LTE modules out there, have a look at this: http://www.telit.com/telit/Pulsar/en_US.Store.display.891./4g

  • Unrelated to my previous comment: I hope that they provide for “reconfiguring” – I can sure see times when a consumer might have a module that they need for a while, but then a few months later, not need it and want to remove it and (maybe) install something new (or just get rid of the battery load).

    If so, this should give the Ara a big leg up on more conventional phones. (Hmmm… the “nay sayers” are kinda sounding like the folks who poo-pooed digital camera technology many years ago – like Kodak.)

    Taking it to extremes, we might see folks swapping modules on a frequent basis – configuring their phone for a particular activity.

    • They are absolutely providing for reconfiguring. That’s one of the main features of the phone.

      The goal is to make everything hot-swappable. Even the battery! They stated that you’ll be able to swap out the battery without turning off the phone. They didn’t say how, but I’m assuming some sort of high-capacity cap that can run the main controller for a short period of time. As for the other modules, the drivers for most of those should be identical, with the applications taking care of the details for each module.

      The example they used was having one configuration for work and one for play. You come home from work and you’re going out for the night. You can remove the bluetooth module and replace it with a camera, and remove the wifi module and replace it with another battery module for extra power (the power system allows the user to have as many batteries as they want).

      • The thing I like is you can have companies create hardware and software for the phone. On thing I heard about is being able to swap out your camera for a thermal camera. Given that FLIR has made a thermal imaging camera for iphone, the technology might be there to shrink it enough to fit in one of those slots.

        But what I find most amazing is the time line for the Ara phone. A lot of this technology used to make the Ara possible is being worked on as we speak, including the production 3D printer for the cases. They want to have all of that done by 2015. http://img.pandawhale.com/82057-thats-a-bold-strategy-cotton-g-0hV3.gif

      • Why aren’t they just using a screw or equivalent mechanical system instead of the magnets? Or just put some sort of case on top. That seems so much simpler.

        • Screws get lost, and who has a screwdriver on demand (we’re talking general public).

          The issue is, you need something secure, and has the proper platform (hot-swapping components should use the same size parts), A case, although nice, won’t provide the connectivity pressure to make anything work well for long periods of time.

          The idea behind the electropermanent magnets (which are used in scrapyards and shipyards) provides extremely good connectivity and strength. However, my only concern is that, these are magnets. Electromagnetism tells us that you can induce a current from a magnet, and on a device that small with that much conductive material, you can either 1) destroy the device. 2) Make it inoperable from the noise that the induction can produce.

          Now that is the scientific answer to your question, but let’s look at a manufacturing aspect. The longer the BOM, the higher the cost. This is why in electronics you want to reduce BOM as much as possible. And screws add a lot to that. A case, great for protecting edges, is not gonna be all that helpful either. The basic wear and tear on them will cause issues. Besides having to develop a special case (higher BOM) and making sure it works, they have to do some large endurance tests to make sure that it will always be functional when we all know it won’t.

          The magnets provide a strong solution because 1) they are very hard to destroy/damage. 2) They have great strength. 3) They are “programmable” because of the electropermanent aspect.

  • This sounds like a fantastic opportunity. I’m sure that if they don’t price the “tinkerers” out, they’ll get a LOT of new applications of cell phone technology that they haven’t thought of – just as there are things that the Internet has enabled that none of us saw coming back 20 or so years ago.

    Thanks for sharing the info! Hopefully in a year or two, we’ll see some “Ara” “DIY” modules on SparkFun.com…

    Now I’ve got to go clean the drool off my computer…

    • This is where I’m concerned. If Google wants this to be a “lego” phone people can put together, how are they going to control what modules people can make and put to it? If I were to make one that does something, how can I prove that it works with the phone flawlessly? The answer might have to be in special licensing.

      I’m all for DIY'ers and tinkerers to do whatever they want to it, but I want to know that what someone creates will not break the phone or use it maliciously.

  • Ara will be an off-the-shelf expandable hardware platform with a touch screen and wifi module that runs Android out-of-the-box.

    It seems like there’s some real promise here.

    I want somebody to build me the Just Runs a Real Linux Already For Crying Out Loud version, but I’ll take what I can get.

    • You can already make a Linux ROM for your device. You can base it off of your kernel (Possibly the CM ones). Note that the baseband and some drivers need proprietary software. If you have a Nexus, all drivers are downloadable from Google’s website.

  • Awesome information! They should make a version that is all lego blocks, so you can build your very own Lego Phone!

  • The link to PhoneBloks in the article is wrong (links to Project Ara page). PhoneBloks page address is: