Reddit user sylocheed is perhaps extremely lucky to have been able to drop by the NYC Nexus Open Studio event, which was attended by representatives from all Google divisions involved with the Nexus programme. His questions revealed some greatly interesting answers from all the Google employees he asked, even if the question was mundane or controversial.
Have a look at what he asked, and what answers he got.
Original post on the Reddit sub-reddit, /r/android.
On the development/design timeline of the Nexus 5X/6P
The design and development of a smartphone typically takes around two years. For the Nexus 5X and 6P? ….a grueling seven months.
On the industrial design of the Nexus 5X
A lot of work went into the curvature and radius of the edges of the 5X and the camera “hump”. The industrial designers ensured that each of the Nexus 5X’s curves utilized “C3” continuous surfacing. It’s a lot of work for a small detail, but offers significant rewards in terms of light play around the surface and how naturally/organically the surface curves.
On the engineering challenges to nail the industrial design of the Nexus 6P
It has been a common assumption that Nexus phones tended to be based off hardware designs of existing phones (the Nexus 4 and LG Optimus G, the Nexus 5 and the LG G2, etc.), and that Google would be merely using established, solved industrial design processes that were well known by the manufacturing partner. It was interesting to find out that a lot of experimentation and testing still went into things like the bead blasting process and the diamond-cut chamfer.
It is easy to imagine that much like going to a paint store and picking from a huge wall of swatches, that the colors and finishes were a known and well-established thing for a manufacturing partner—they just merely needed to be selected. However, even something like the media used in the bead blasting process still needed constant testing and refinement. The Nexus engineers and designers tried a variety of media like coconut shells, glass beads—
“We spent lots of time over in China pulling samples fresh off the bead blasting manufacturing line, running over to a window, and scrutinizing the way the light played off the surface.”
All told, there were around 80 different manufacturing steps for the Nexus 6P (and fun fact: Google dictated that all of the aluminum leftovers in manufacturing were to be collected and reused/recycled). In fact, the circumferential chamfer that wraps around the visor is actually an industry first, using a brand-new process. In addition there were other major challenges in terms of managing the RF with an all metal phone—making sure antennas weren’t detuned by the aluminium body through careful injection molding processes.
On the relationship between the Nexus manufacturing partner and Google
It might be surprising to learn that the manufacturing partner relationship can be a mixed bag and even depends on the partner. Even with the latest Nexus 5X and 6P where Google exercised a considerable amount of design influence, there was a give and take between the manufacturing partner and Google. Oftentimes there was pushback from the partner for the sake engineering or manufacturing convenience—one designer mentioned that it was three-months of back and forth just to get the USB Type-C connector centered on the bottom of the 6P.
Another had referred to the bottom of the recently announced HTC A9 as evidence of design by cost and engineering convenience. He was sympathetic though, as he recognized that for many of Google’s OEM partners, the teams live and die by the success (or lack thereof) of a single product launch, and so cost (in both design and manufacturing) is a legitimate consideration. Google, however, can afford to take a longer view on cost and invest in trying to hit a higher bar of design.
The relationship was a lot of work (one Google engineer spent about two of the seven month development time over in China overseeing things), but it seems like both the manufacturing partners and Nexus engineers and designers were really happy with the end result.
On the Nexus family design language and the Motorola Nexus 6
Many might wonder about the Nexus 6 and the manufacturing partner selected for that generation, Motorola, as did sylocheed who asked about it. Its design seemed to share more with other phones in the Motorola family than it did with any other device in the Nexus family.
“…Somebody backed out at the last minute.”
It was added that as a manufacturing partner, Motorola exercised tighter control over the design and engineering of the phone, in a way vastly different from other Nexus partners. It sounded like Google designers and engineers (obviously) preferred the more Google-directed partnership, so this may be the last we see of a Nexus 6-ish device with a design that deviates so much from the Nexus family language.
On the lessons learned from the Nexus 6
The team there recognized and acknowledged a lot of the complaints/problems of the Nexus 6. It was too big, too bulky, hard to use one handed, and worst of all—
“…the Nexus 6 spun around when you tried to use it on a table.”
Along with the fact that the Nexus 5 was Google’s best selling Nexus phone, this seemed to set the stage for what we see with the Nexus 5X and 6P today. First, a return to the tried-and-true formula of the Nexus 5, but with updated hardware and a modernized design for 2015.
On the Nexus 6P, there were obviously more design changes. For one thing, the screen shrinkage from 6″ to 5.7″ seems small, but vastly improved one-handed use. Also, they were proud that they were able to fit a 5.7″ size screen into a device with the same phone size as competitors with a smaller 5.5″ screen (clearly a reference to the iPhone 6 Plus). In addition, the visor was a key design component addition. For one thing, an all-metal frame did need RF windows for sensors, cellular reception, etc. But moreover, the full-width visor allowed for a stable base… and no more spinning phones on tables.
On the just-barely “off center” placement of the camera module in the Nexus 6P visor
Astute owners of the Nexus 6P might notice that the camera lens is just-oh-so-slightly above the center axis of the oval camera “surround” it sits within. Well, it turns out it wasn’t just some design oversight.
Thanks to the all-aluminum design of the Nexus 6P, all of the RF (wireless) “stuff” and sensors needs to fit within either the visor or the bottom plastic RF “windows”. This includes the NFC sensor, which was even of a special design that allowed for a small part of the sensor to sit “above the fold” in the RF window, while the rest was below.
When it came to QC test the NFC sensor in the Nexus 6P’s first iteration… it failed. Not enough of the sensor had been exposed in the visor, and they had to painfully widen the visor just a little. Well, the camera had originally been symmetrically aligned in the visor, but when they widened it, the camera was now off center (and unable to be moved around anymore because of design and timing limitations). To help conceal the now slightly-off center nature of the camera within the visor, the design team threw in the oval camera surround.
On bucking the trend of thinner phones with smaller batteries
When discussing the Nexus 6P “visor hump”, it was mentioned that one of the frequent suggestions that have been seen in terms of eliminating humps altogether was to just maintain thickness on a phone and use the added volume for more battery capacity. The engineer just shrugged and replied:
“Apple really sets the bar on thinness there.”
On the Nexus 6P and 5X placement of the Nexus Imprint fingerprint sensor
The Nexus design team considered and spent a lot of time researching three different fingerprint sensor locations before settling on the rear placement we know today: the bottom front (e.g., Samsung/iPhone), the side power button (e.g., Sony Xperia), and the center rear. Designers liked the discoverability and accessibility of the Samsung/iPhone model, but felt that the bottom edge placement placed all the weight of the phone aft of the sensor and there was a cantilevering effect that was both awkward and increased the risk of dropping.
The side positioning had limitations based upon the thinness of the phone and those limitations would reduce the surface area available to read a fingerprint. The backside approach allowed for a large surface area to capture a fingerprint, was where the index finger naturally fell, and also was suitable for both right and left-handed users.
“As you scale the phone, the fingerprint sensor does not move—because it’s not based on [the size of] the phone, it’s based on the human hand.”
The backside sensor approach also had the benefit of being scalable and independent of phone size, which was particularly relevant for the two-sized Nexus launch. It was pointed out that the fingerprint sensor was in an almost identical location on both the Nexus 5X and 6P, despite the differences in the size of the phones. The hardware engineer conceded that the industrial designers were so passionate about perfecting the fingerprint sensor location that…
“In fact, [the industrial designers] were so adamant about basing this on the human hand we had to compromise on the battery size , to make sure he got his way—to make sure it fit the human hand.“
On how Google tested and benchmarked the Nexus Imprint fingerprint sensor
Google wanted a quantitative way to measure the speed of the fingerprint sensor while they were optimizing its speed. So, Google engineers built a rig with a robot “hand” and highspeed camera for testing and benchmarking the fingerprint sensor. However, the missing link was what to use for the actual fleshy finger stand-in.
“The product manager for fingerprint… she was having lunch one day in the cafe and she was having chicken drumsticks. And she realized that the skin of the chicken drumstick is really similar to the human hand. So she took a batch of chicken drumsticks, took it upstairs to the lab, washed it, picked off a bit and tied it to the robot, and used that to measure the latency of the fingerprint sensor.”
On sourcing the new Sony camera module
The camera was clearly the major highlight of the event (other than the Nexus Imprint, both of which had key presentations), and many members of the team were incredibly proud of it. One product manager proudly noted that the sensor sourced was an industry-first and that it was pulled from Sony’s digital camera parts bin (and not mobile camera menu).
One of the things revealed by the team was that picking out a sensor is not merely just the picking of a sensor from a parts menu. In every implementation, Sony needs to attach a team to support the integration of the camera sensor. Sony was initially concerned that the sales volumes wouldn’t be sufficient on the Nexus phones to be worth the investment for Google to reach so deep into their premium parts menu. It took a lot of convincing to get Sony onboard to provide Google this particular camera module. Thankfully it turns out Sony was really satisfied about the press and acclaim it received in connection with the new Nexus phones and their stellar camera.
On the absence of camera OIS
Sony did not offer this particular camera module with an OIS option. Sorry folks!
On the Nexus 6P display, and whether it is a “current generation” Samsung AMOLED display
“…The Nexus 6P actually uses the same display panel as the Note 5”
Knowing that this is a hot topic on many forums, sylocheed brought up the fact that testing has shown that on the Nexus 6P, its display has a lower maximum brightness and higher display power consumption than seen on the Note 5—key indicators of the AMOLED panel generation. It turns out that the display panel is the same, but that much like the practice of CPU/SoC binning, there is variation inherent in AMOLED panel fabrication. Here, Google was able to procure the lower binned, yet current-generation displays from Samsung for the Nexus 6P.
On the thermals of the Snapdragon 810 SoC
The thermals of the Snapdragon 810 were known to Google and were of some concern. Qualcomm attached a team of engineers to help control thermals and supplied Google with a third version of the Snapdragon 810 (which seems to match public, well-established reports that the Nexus 6P uses the v2.1 version of the 810).
On the removal of wireless charging from the Nexus 5X and 6P
It sounded like the intention to go with a never-been-done-before-on-Nexus all-aluminum design had a big role to play. The Qi wireless charging was not going to work on the metal back, and so it needed to be dropped. For the Nexus 5X, even though it had a plastic design, the product team consciously wanted to maintain a certain feature parity/consistency between the two models (like in the same way they kept the same camera, flash, IR sensor, etc.) Furthermore, the reduced pain point of the universal direction USB Type C connector added onto the decision to drop wireless Qi charging.
On the Nexus 6P “Bendgate” controversy
All of the Nexus phones go through rigorous strength and wear testing. When the videos first broke out about the “Nexus 6P Bendgate”, the engineers rushed back into the labs to double-check and rerun the tests to make sure they hadn’t missed anything. The engineers indicated that nearly all phones are strength tested to similar specifications and all have their first “bend” at a similar weight pressure—he quoted the iPhone 6 Plus and Nexus 6P to be around 40-45 kgs on the particular bending bench. Apparently, where phones vary is the ultimate failure strength (the amount of pressure for the phone to catastrophically fail after a first bend), but the pressure-to-first-bend was apparently consistent across most phones.
On the “twist to launch” camera action that was apparently pulled from the Nexus 6P
According to one product manager, the feature didn’t test well and was never formally implemented as opposed to being dropped/pulled. It was an accident it was included on websites (this is why we don’t report rumours, ha!).
On controversial topics like internal storage and integrated batteries
Sylocheed asked some of the product managers what they thought about the controversies over internal storage and integrated batteries—particularly whether there was any truth to the theories that the limited internal storage was really an initiative to push users to cloud storage or that integrated batteries were there for planned obsolescence.
The product managers were quite genuine and adamant that there were no ulterior motives and that things really boiled down to usability and design tradeoffs. On the internal storage side, they noted that this was an interesting and complex issue, and first pointed out that when they balance tradeoffs for a feature like expandable microSD storage, the same balance isn’t made the same way for every market. For example, on Android One phones, the usability issues with microSD was outweighed by the need for price sensitivity and the greater challenges of single-up-front payment—and this use case is reflected in the inclusion of microSD in Android One designs and in Marshmallow’s stronger adoption of external storage.
However, on the Nexus side, the usability issues with microSD won the day. The product managers were concerned about things like slower, poorer quality microSD cards and microSD failure that were challenging for users to self-diagnose when there were multiple kinds of storage. A single, high quality internal storage solved this issue (at the cost of expandability). They also noted that in addition to considering users, a single unified storage was easier and more predictable for developers—a single fast and reliable internal storage meant better predictability on app performance, less problems with defining where to put files, and reduced troubleshooting/customer service.
The product managers also noted that they really cared about the products they build. If they could have it their way, their phones would be timeless in a way similar to “like a Gucci bag,” with great longevity and resale value evidencing the quality of their product. So they too are disappointed with the inherent problems of lithium ion batteries. That said, the design benefits to an integrated battery were so great that the balance of tradeoffs really went in favor of non-removable batteries.
On feedback on the Nexus products and Android UI
Some of the representatives noted that they definitely have heard the requests for a smaller, more premium phone instead of the premium 6P phablet and economical 5X phone. They said they are always looking into the decisions around this and are always rethinking those decisions for the next generation.
Another noted that in addition to the outside user feedback, they hear a lot of the same sentiment directly after the launch of a new feature or device. The engineers, designers, and other Googlers are a very vocal and strongly opinionated group and are frequently dogfooding and using their own products.
When asking one of the UX designers what one of the most controversial things was that she had worked on, she confessed that she worked on and helped to implement the bolded hour digit in the Jelly Bean clock app. She noted with a smile that the reaction of Google staff on this change alone was swift and unequivocal.
On the functionality and evolution of Android volume control
A presentation was devoted to the user experience design and evolution of Android volume control from pre-Lollipop, to Lollipop, to Marshmallow. The main takeaway was that there are a serious amount of different permutations and use cases of alarms and volume control.
These challenges were compounded by the fact that unlike many of Google’s apps that are now decoupled from the OS and are updated routinely on the Play Store, volume control is still tightly woven into the OS. The UX team needs to finalize their decisions well before the launch of the OS and only has an annual window for which to make changes.
On the sale of Motorola to Lenovo
Google’s OEM partners were really concerned that despite the firewalls between the two organizations, that there was too much threat that a Google-Motorola would turn into a fully vertical entity like Apple (or Microsoft today with its Surface tablets), squeezing out its partners in the Android ecosystem. When asked why the OEM partners were okay with the Nexus program, it sounded like the Nexus sales were still quite relatively small compared to the bigger OEM players (Samsung, LG, etc.) and were not threatening.
On the Nexus 6P three-microphone noise cancellation
On the iFixit teardown, it can be seen that two of the advertised “3 microphones (2 front, 1 rear) with noise cancellation” were located within each of the front-facing speaker grills, which could have caused interference. It turns out the microphones are in fact in the speaker grills, and that they can be spaced in a way that minimizes interference from the speakers.
On the rumors of the next iPhone to pull the 3.5mm headphone jack in favor of the Lightning connector
“If anyone can do it, it’s Apple and Beats.”
The engineer noted that we’ve been stuck with the same form factor and technology on the 3.5mm analog jack for decades and figuring out the placement of it is a consistent design challenge and design limitation. There are a lot of new design opportunities afforded by moving to a new connector and opportunities around digital audio output (headphones with audiophile DACs and amps instead of relying on those in your phone!) and although Google doesn’t really have the leverage to initiate a change like this, Apple definitely does—especially through its ownership of Beats.
On knowing the name of the next Android version
The dessert naming decision of the next version of Android is not known to most Googlers, even those in the Android UX design team. So they are as surprised as everyone else when the statue is finally unveiled.
However, while they don’t often know the what the right choice will be for a name, they often know what it won’t be, so internal and code references to wordy names like “Lemon Meringue Pie” were placed with tongues firmly in cheeks. The team at Google also had a lot of fun when David Burke trolled the community with his infamous “milkshake” watch stunt.