We help our customers develop breakthrough products and services within their market.

How did Apple do it?

While there are products out there that claim to have truly wireless ear buds, reviews for them have not been so great. Now we have Apple's AirPods. I'm hoping that Apple engineers have solved all the problems that others have experienced and that true wireless earbuds will soon reach mainstream consumer adoption.

Until they do, let’s take an unashamedly geeky look at:

  • Design challenges for wireless earbuds
  • How those challenges can be overcome
  • My thoughts on the most likely tech approach that Apple took

At the risk of stating the obvious, the main design challenge in any wireless system is getting rid of the cable. Many things that are incredibly easy to do over a wired connection become much more involved when a radio link is used instead. Apple AirPods are no exception. To make them truly wireless, Apple haven’t just cut the cord between your phone and the earbuds, but also between the two buds.

This presents the interesting design challenge of getting radio waves around or through a person’s head, safely and without undue power consumption. Unfortunately, radio waves don’t propagate well through the human body in general, and especially through the bone and water of the skull. Another effect of losing the wired connection between earbuds is that synchronising playback between them, something essential for any stereo audio playback, now becomes a design challenge all of its own. Finally, Apple is advertising a full five hours of continuous playback on a single charge. This is bold. Whatever solutions are used to tackle these design challenges, they need to be incredibly low power, and then some.

It is probably worth mentioning that although Apple are well known for disregarding interoperability, complex system solutions such as AirPods really should be interoperable. Solving a particular problem while maintaining standards compliance and performance targets is the really juicy stuff – often one of the largest design challenges on any project.

So how did Apple do it?

With some reports on the internet claiming the AirPods are compatible with Android devices, it seems likely that Apple have deployed standards compliant Bluetooth. It is also likely that they have adopted a standards plus approach – adding in proprietary extensions to improve the user experience with all Apple devices, such as the pairing on case opening. Whether true or not, using a proven technology with such broad market penetration is a smart choice. This would allow Apple engineers to concentrate on the specific challenges of their use case - through-the-head wireless and very low power consumption - while inheriting or borrowing system level solutions that exist today.

It would also allow for interoperability, which may prevent the vendor lock in that Apple would enjoy, but does mean that customers are likely to be much happier. And who knows, an Android user may be so impressed by the performance of her new truly wireless Apple AirPods that she may chose the iPhone for her next smartphone.

How did Apple achieve a five hour playback time in such a constrained form factor?

Apple has already demystified this for us, the Apple W1 chip is the heart and soul of the AirPods. The W1 is claimed to be a custom wireless chip. Based on Apple publicity pictures, official AirPod dimension/weight specs, and Apple iPhone 6 Plus battery volumetric power density, I estimate that each AirPod has a 70mAh battery. The usable capacity of each battery is, of course, smaller than the datasheet value, but even at 90% utilisation, achieving five hours of playback is no mean feat. This comes down to an average current draw of 12.6 mA or less when streaming AAC audio over Bluetooth, decoding and playing it back. Current highly integrated Bluetooth chips solutions drain about 50% more than that.

A quick look at available Bluetooth radio and controller chips shows us that existing Bluetooth classic technology is run on design and processes that are often 5 to 7 years old. Much has changed in the semiconductor and ASIC RF business since then, allowing for smaller sizes, lower power consumption and increased performance.

It is highly likely that in order to achieve their five hour playtime requirement, Apple chose to design its own single chip, highly integrated, Bluetooth audio solution. This is a time consuming and sophisticated engineering challenge, so no small decision by Apple’s leadership. Here at Cambridge Consultants we help clients hit similar targets by developing custom ASIC designs that fit their needs. In fact, 16 years ago, Cambridge Consultants created the low-power integrated RF ASIC design that served as the basis of the Cambridge Silicon Radio (CSR) Bluetooth Audio chips that enabled Bluetooth wireless headsets in the first place. The added value in investing in a custom ASIC design, especially when high performance and low power consumption are needed, often justifies the engineering costs of developing one.

But low power consumption is useless if Apple cannot send music from one AirPod to the other. This is another area of interest for us at Cambridge Consultants, where we create breakthrough in-body antenna technology for leading-edge connected medical devices. But forcing radio signals through the head is not practical, and in some cases may not be safe. There are two main approaches to getting the radio waves across, and each has benefits and drawbacks.

The easiest way to get the audio to your second AirPod is to use Bluetooth forwarding. This is a tried and tested approach between two Bluetooth devices. It requires no extra hardware, no additional stack or software support for a new protocol. However, Bluetooth cannot pass through the head, and Bluetooth waves cannot be made to go around the head. This does not mean that this approach doesn’t work. In fact for indoor radio use, radio waves reflect off objects, walls and ceilings, creating a path for Bluetooth to travel from one AirPod to the other. Performance outdoors is a different matter. To put it mildly, performance may vary, and is likely to be extremely poor in the middle of a large grassy lawn for example.

A second very popular approach is to use Near Field Magnetic Induction (NFMI), an approach similar to induction loops for hearing aids. This approach works in all environments, as it uses magnetic signals that go through the user’s head with minimal loss. Unfortunately, they operate at completely different frequencies to Bluetooth and require very sophisticated RF design to fit into such a small form factor. The additional RF front end, software requirements, and antenna design make this a less attractive option and less likely to hit the aggressive power targets that Apple has set out.

In fact, fitting a Bluetooth and NFMI radio on such a small form factor as the W1 chip, and ensuring antenna performance in such as small form factor as the AirPods, makes NFMI an unlikely candidate for Apple’s end design.

We don’t yet know how AirPods perform for people in open spaces, and we will have to wait and see if Apple has managed to tackle this issue. A Wall Street Journal writer, writing a generally positive review of preproduction units said “…I experienced a few drop outs on my AirPods outside”. However, we can speculate as to what Apple could have done to make a pure Bluetooth solution work for them.

I suspect that Apple combined a fully standards compliant and existing solution based on Bluetooth forwarding (with all of its interoperability benefits and performance drawbacks) with a proprietary approach when using the latest iOS devices to solve the existing solutions problems.

Apple have probably done the smartest thing an engineering team can do when faced with a very hard challenge – step back and find another way round it. We do this constantly here at Cambridge Consultants. Sometimes the hardest problems can be overcome by our interdisciplinary teams by focusing on the end goal and finding all the different ways it could be achieved instead of focusing on that massive obstacle that one of the ways has presented us with.

So how do you send music from one side of your head to the other, through your head in the simplest, fastest to market, and most robust way?

Perhaps you don't.

The task Apple engineers have is to get synchronised audio to both AirPods, not to get it through your head. Solving the through-the-head problem may be interesting, but is not actually required. The best engineering solution, when compatibility isn’t a factor, is not to send anything between the two AirPods. At the expense of a more complicated software solution on the iOS device side, the device is sending data to both AirPods and handling synchronisation between them.

Is this Bluetooth compliant? Probably entirely so, even if Apple is using some custom control profiles, or maybe even the iAP 2 protocol, to sync devices. Such an approach will only work with iOS, but using standards compliant Bluetooth and keeping existing solutions for Bluetooth forwarding will allow the AirPods to work with any Bluetooth enabled device.

Regardless of the specific design solution that Apple has used, it is clear that they have their sights set on the emerging hearables market, and are aiming to stir up the personal audio market.

We’ll learn more about how the AirPods have been developed in the weeks to come, but for now, what do you think of my educated guesswork? Let me know in the comments.

Author
Kalin Dimitrov