Simplifying audio IC selection for smartphones

Simplifying audio IC selection for smartphones

An overview of an innovative audio architecture as well as some key characteristics to look for in audio components

BY ROBERT HATFIELD
Wolfson Microelectronics
Edinburgh, U.K.
www.wolfsonmicro.com

The pressure continues for designers of smartphones and other portable multimedia devices (PMDs) to incorporate more features into ever smaller products with longer battery life, fast time to market, and at low cost. Excellent audio capability is now a must-have in most devices, but the challenge can be how to maintain performance yet have the audio well integrated with the other features. With more complex and numerous use cases, multiple microphone inputs and multiple audio streams playing concurrently, designers need to take a more audiocentric approach to the design.

The smartphone architecture has gone through quite a metamorphosis with early designs consisting of individual ICs, each with a specific task. Improving the audio performance in a smartphone or other PMDs in these first products meant selecting a codec or audio ADCs and DACs that offered the best signal-to-noise ratio (SNR) or total harmonic distortion (THD) and low power consumption.

The second wave of designs tried to integrate all functions into the holy grail of one large system on a chip (SoC). Although, in theory, this was quite appealing from a space and perhaps a cost standpoint, functionality, time to market, and the ability to release a family of phones with varying features was severely impaired. Combining functions such as the applications processor and wireless communication into one chip proved daunting.

The feature-based innovation of the application processors has tended to diverge from the standards-driven approach to wireless communication chipset development. A Bluetooth, cellular, or Wi-Fi chipset needs to meet a standard spec with minimum cost and power consumption to be competitive, while an application processor needs powerful digital functionality and performance to compete.

Fig. 1. Typical HD audio solution for smartphones.

Adding audio functions into the application processor or modem as an afterthought with the pretense of adding value through integration has had limited success. Audio design suffered due to excessive integration with duplicated and poorly implemented audio features in two or more chipsets combined with a host of “patch” components added by system integrators to support the necessary mixing, switching, pop/click suppression and power amplification required by the end product. In addition, the different semiconductor processes required by high-speed digital circuits and high-performance audio and analog circuits make optimization of both functions impractical.

Recent smartphone designs show that there is now a “dis-integration” trend in the audio/analog design and the application processor. The processors on smartphones are being designed with very small process geometries, and they are integrating more graphics and processing capabilities, with the audio functions trending toward integration in an audio SoC. This audio hub device combines all mixed-signal audio functions, enabling lower-cost processor selection and consistency of audio signal path characteristics regardless of the source data. This enables system designers to optimally manage multiple concurrent audio uses, for example, the consumer can take a hands-free call over a Bluetooth headset, listen to music and receive navigation commands, all at the same time and on the same headset.

Figure 1 shows an idealized implementation of the architecture of a smartphone. At the core of the audio design is the audio hub. Surrounding the hub are the microphones, noise cancellation, amplifiers, and the power management system IC (PMIC).

Audio component selection

The first step in selecting the audio components is to plan the audio-use cases. For each use case, understand which chipsets need to be enabled, where the signals are coming from, where they are going, and what the most efficient way to route them is.

Where possible, replace an analog connection to the modem with a digital one. Using analog voice data paths tends to consume more power at system level, be vulnerable to crosstalk causing PCB re-spins in many cases, and have the additional cost and area overhead of passive components in the signal path.

Audio hub:

When selecting an audio hub, the designer should ensure that the inputs, outputs and features all match up to the use cases outlined. Figure 2 shows the block diagram of an audio hub from Wolfson Microelectronics specifically designed for smartphones.

Fig. 2. Wolfson WM8958 audio hub block diagram.

Amplifiers:

Using Class D mode for the speaker amplifiers can result in very large power savings, typically running into tens or even hundreds of milliwatts. These power savings can allow longer battery life for the smartphone or enable the addition of other features with this kind of power budget.

For designers unfamiliar with Class D technology, most datasheets provide recommendations on EMC, speaker selection, efficiency, and PCB design best practices. Concerns about EMC in earlier generations of phone designs have more recently given way to thermal concerns, as heat dissipation can actually limit handset capability in some cases, causing a greater focus on power efficiency.

The higher efficiency of Class D also reduces current surges into the speaker amplifier’s power supply, which cause battery voltage droop and earlier system shutdown, limiting battery life, particularly when coincident with other current surges in the system. The reduced battery droop will also help to reduce distortion at high signal levels and in some cases may also provide a little more headroom to increase the maximum speaker output signal.

For the headphone amplifiers, a device using a highly efficient topology such as Class W should be considered. Class W features stereo ground-referenced headphone amplifiers using Wolfson's amplifier techniques. Class W incorporates an innovative dual-mode charge pump architecture — to optimize efficiency and power consumption during playback. The ground-referenced outputs eliminate headphone coupling capacitors. Both headphone and line outputs include common mode feedback paths to reject ground noise.

Microphones:

Where possible, consider the benefits of digital microphones at system level rather than simply at component level. Replacing long analog microphone connections can help to reduce rework and save space. Microphone arrays are becoming more commonly employed for noise reduction and as the number of handset microphones increases, the likelihood of routing one of the many microphone signal paths close to a source of noise is high. Although differential connections and careful layout can improve noise immunity, it requires twice as many good quality capacitors in the signal path. Using a digital connection will improve noise immunity. Note that when using microphone arrays it is important to select microphones with very closely matched characteristics for best performance.

Power management:

Select an efficient power management architecture. Audio hubs usually require two external supply voltages, typically a direct battery connection and a 1.8-V supply. Since the 1.8-V supply powers the charge pump, which, in turn, powers the ground-referenced headphone amplifiers, it is preferable to use an efficient switching regulator for this supply to ensure that the low power consumption of the headphone amplifiers is not compromised by, for instance, using an LDO for this supply.

Ambient noise cancellation:

Ambient noise cancellation (ANC) can be used for both the receive side and the transmit side of the smartphone. Key specifications in an ANC device include the bandwidth of noise that is cancelled, the level of ambient noise cancellation, the number of microphones allowed and power consumption. As an example, the Wolfson WM2002 provides noise cancellation over a wide bandwidth of 40 Hz to 4 kHz with 30 dB peak and 25 dB typical ambient noise cancellation. This device supports up to 10 microphone inputs and features class-leading low power consumption of 25 mW for ear-bud headsets. ■