Acoustic System Design
of an Apple AirPod
What is inside? How does it work? Let's have a look!
Probably everyone of us is familiar with AirPods and the product does not need an introduction anymore. But it is a more complex device than its unpretentious black and white surface might indicate. We will have a look at the system design of the AirPods as an example for a True Wireless Speaker (TWS). We will also do a little deep dive on how the acoustics inside an AirPod work and learn what functional elements are responsible for the sound of the AirPod 1 and 2.
But before digging into the details of the engineering inside, let's have a short look on some main requirements as they set the foundation on what was implemented by the engineers. It is also worth to remember that the final design of the AirPod 1 dates back to 2016 (8 years ago from the day of writing this article).
Background and Requirements
Obvious requirements
- Play sound in a good quality - Speaker and acoustic tuning
- Wireless connectivity - Bluetooth transceiver + antenna
- Runtime - Battery capacity and strict power management
- Make phone calls - Microphones, beamforming, echo cancellation
- User input by tapping - Accelerometer for tap detection
- Active Noise Cancellation - Microphones and capable DSP
(not for AirPod 1/2)
A few additional considerations
- Production - Cheap, easy and reliable (and possible!)
- Cost - It is a mass product, every cent counts
- Supply Chain management - For many external components
- Safety standards - Battery failure, maximum sound output
- Acoustic seeling - Housing (except acoustic ports) must be air tight
- Ear fit - Sit comfortable in most ears
- Ear presence detection - Proximity sensors
- IPX4-Rating (repel splashing water) - Meshes on all openings
- Sweat and chemical resistance - Earwax is surprisingly nasty
- Temperature, humidity - E.g. no condensation within the housing
- Drop, Pressure, Shock resistance - Soft glue, robust sensors, Flex-PCB-layout
- Certifications - Such as FCC, CE, ...
- Test - Characterization and qualification must be possible
- End of line test - No product without Quality Management
- IP - Freedom to operate
- Software & Apps - Compatibility to several end devices
- Design - Last but not least: look&feel
- Many more
Unfortunately, the list of requirements above are far from complete, but they show the complexity of the system and the care and love for each single component that goes into a final product.
It is also worth mentioning that 3rd party products (such as the MEMS microphones, accelerometers, ASICs, ...) are often customized by their respective manufacturer. This is necessary to meet the specifications of Apple, which are again derived from the product requirements of the AirPods.
Functional Acoustic Requirements
We start by having a look at one of the main product requirements: “The AirPod shall play sound in a good quality”. To work on a technical level, the product requirement is transferred into a number of functional requirements.
With this work done, we can set ourselves in the position of an engineer, who is designing the electrical system. They have to choose a digital to analog converter and output amplifier, which is able to deliver a dynamic range of 65dB (from 40dB to 105dB). So their task is to deliver a clean signal (THD<<0.01%) for output voltages between 850µV to 1.5V (while having an eye on power consumption and cost, which are defined in other requirements).
In the same way, this affects all other part selections and design decisions: Speaker selection, acoustic volume size, outer shape, etc. So let's see how the Apple engineers solved this puzzle.
The Main Components of an AirPod 1
When having a look inside an AirPod 1 (A1722), all the mentioned requirements are represented by functional units.
What we find are the acoustic system (a speaker, meshes and ports, acoustic volumes), microphones for phone calls, proximity sensors for presence detection, one accelerometer for motion and tap detection, another one for speech detection, a class-D audio amplifier, a main processor, a custom bluetooth IC, a bluetooth antenna and of course a battery with power management.
All of this is packed together using two connected rigid-flex-PCBs and folded into the housing we all know. If you want to learn more, it is always interesting to have a look into the patents that Apple filed on the AirPods. For example US10397682 - “Earbuds with acoustic insert” is a good place to start reading.
Since it would fill (at least) one book to go through every single one of the many remarkable details (such as how and why the PCB is folded multiple times at certain positions to provide strain relive, manufacturability, design fit and electrical connectivity at the same time) we would like to focus on the acoustic subsystem of the headphone in this article.
The Acoustic System of an AirPod
Overview on the Design
An acoustic speaker system consists of three main components. A driver, a front volume, and a back volume.
- The driver in AirPods is an electrodynamic speaker with 14mm diameter and 32 ohms impedance. This is a relatively large and flat driver unit, which suits the semi-open design of the AirPod because it is able to deal with a (in comparison) big volume flow (high excursion driver).
- The front volume is the combination of a small volume inside the AirPod and the volume inside the users ear. The sound pressure created in this volume (by deflecting the membrane of the driver and therefore compressing or expanding the volume) forms the audio signal at the listeners perceives.
- The back volume is the air on the backside of the membrane, similar to a speaker box. Making the back volume very small, the membrane is not able to move anymore, because the counter pressure from compressing the volume becomes very high. In contrast, opening the volume to the outer world (infinite volume) is also not a good idea: Then the in comparison smaller front volume dictates the load of the driver. The front volume depends on the seat and anatomy the ear of the user, so the listening experience would differ a lot between different listeners especially for mid and high frequencies. Thus, the designer chooses a small back volume and a well defined load for the driver.
The AirPod is also no exception here, with implementations of the driver, a back volume, a front volume, and ports between these volumes.
The front volume inside the AirPod is coupled to the ear channel and the outside world through three acoustic meshes. The connection to the outside stems from the semi-open design of the AirPod as it is not tightly sealed in the ear channel. The same is true for the backside of the driver. In addition, acoustic ports from the back volume to the outside and back volume to front volume give the designer the possibility for an acoustic feedback loop. Aside from static pressure equalization, this is used to tune the frequency response of the whole assembly.
It has to be mentioned that the (acoustic) meshes and the plastic port 1 to the backside are complex parts by itself, integrate multiple functions, have to be selected very carefully, and add a lot to the bill of materials.
A look at Ports and Meshes
Since ports and meshes are expensive, hard to design, and there are so many of them, we should have a little closer look at them.
The picture below shows the influence on the sound pressure level when closing the ports or opening them further (by removing the mesh).
Backside Mesh and Ports
The backside volume is connected to the outer world via two acoustic paths. One is a direct opening that is covered by a mesh (marked in green in the figures above) and another path conducting a small channel (orange) and a mesh (yellow). This is used in design to change the frequency response at low frequencies and to control the damping of the resonance of the driver (as mentioned before: the back volume is used to control the load of the driver).
To analyze the influence of the backside mesh and port, acoustic measurements have been performed, where the ports where blocked and where the mesh was removed. During all measurements, there was no leakage path between front volume and outside world to better insulate the effects from each other.
Blocking the port on the backside completely closes the back-volume. Since the air cannot escape, the counter pressure in the back volume on the driver rises. Thus the membrane movement is reduced (for the same input voltage) and the sound pressure level drops on the front side drops. Blocking just one of the ports or removing the mesh, it becomes clear, that the resonance of the system (around 2kHz) is affected and that the two ports have slightly different tuning. Removing the mesh also affects the mid and high frequencies (above 300Hz). However, it can not be assured, that the change is purely by removing the mesh, as it is not easy to cleanly take it apart.
Front Side Mesh
On the front side, there is an opening not directly facing towards the ear channel (blue). Inserted into the ear, it is blocked by the pinna, but still opens a transfer path to the outside world but as well as into the ear channel. Opening this port adds lot of leakage (connection to the outside) to the system which can be modelled as a virtual increase of the front volume. With the small movement of the tiny membrane and the -now big- volume, the sound pressure level at low frequencies drops. In opposition to this, closing the port increases the sound pressure level especially at low frequencies. Therefore it becomes clear, that the port is used to tune the frequency response of the AirPod, while also allowing a fast static pressure exchange to the outside (for example when inserting the AirPods into your ear). The mesh is quite (acoustic) dense. This means, it is difficult for sound waves to escape through the mesh. Therefore, closing the port completely, changes the sound pressure level "only" about 5dB. We can find the opposite at the output port towards the ear, which we will have a look at next.
Output Port
Through this opening (purple), the sound is supposed to leave the AirPot and find the way into your ear. Still, we find a mesh covering the output there. It allows to shape the sound response especially at high frequencies. Therefore not only its size and mesh density, but also position is important. However, people do try out crazy things with their AirPods like cleaning them with air guns or throwing them into the washing machine (yes, it may survive!). Also earwax is quite a nasty chemical. To protect the AirPod against all of this, a mesh is also placed here. In the measurements, it shows a damping of the output by up to 6dB at 3kHz (the airpod was used a few years, so it might have been less when it was new).
Overall acoustics
We have covered the design of the AirPod with a driver, a front side volume and a backside volume, the outside air, and ports covered with meshes between each of these domains. There are many points for criticism, such as the lack of SPL at low frequencies (bass) due to the AirPods open design choice, and the SPL should be more pronounced in the 500Hz-2kHz range to improve quality for speech (phone calls) (Apple went this path towards the AirPod 3). But we also do not have all background information and insight to understand the individual engineering design decisions and constrains for the acoustic system.
And there is so much more!
The acoustic system is much more than we just covered here and we just scratched the surface. We have not talked about phone calls. Yet. Okay, sounds easy - just add a microphone and that's it? Unfortunately not. We have to go back to the early 2010s, when the development of the AirPod started. MEMS microphones were on the market, but their performance was far off from what we are used to today. Below is an image of one of the two microphones that are inside an AirPod 1.
The design of a MEMS microphone itself might be as complex as the design of the entire AirPod acoustic system. This also indicates the incredible amount of work going into their specification, quality control, and implementation (beam forming, echo cancelling, ...), which could be a topic for a whole new article (or book).
There is also an accelerometer input used for “Hey Siri”. Why not just rely on the microphone? In the end, it comes down to the probability of false detection and energy consumption.
But lets take a break here and ...
Apply this Knowledge
Thank you so much for taking the time and making it this far! We hope you learned something or at least enjoyed the read.
Luckily (or unfortunately), we are still just scratching the surface. There is so much more to discover.
The system design of an AirPod is just one example out of many for the complexity of a supposedly simple product. However, this systemic approach can (and should) be adopted to each product during development. Analyzing main KPIs (Key Performance Indicators), the requirements behind them, and how everything comes together allows an analytical, transparent, structured, and cost efficient product development.
Still, it is a very complex topic and hard to get right (even on a second try). Many cross-relations and external factors require a deep understanding of the product, the costumer, and available technology. We are happy to support you with experience and knowledge on demand, provide lectures, and also directly help you in the development of your own fantastic product: pragmatic, analytical, fast, and avoiding traps on your way!
If you enjoyed this article, have feedback or questions about AirPods or any other sensor system or just want to say hello and connect: we would love to hear from you. You find our contact details below.
Thank you so much!