Mixed-Signal Audio Subsystems Enable High-Performance Audio in Portable Products

Feature phones, smartphones, PDAs, and many other cell phone derivatives are taking over many portable electronic devices. This convergence of functionality, while reducing the number of devices consumers carry, expands the audio requirements of the system and increases the burden on designers to solve audio challenges.

As audio requirements increase, system designers may choose to use discrete audio functional blocks. However, taking this approach in a mixed-signal system is a multi-front battle. In the digital domain, providing multiple sampling rates, formats, and digital levels increases complexity exponentially. In the analog domain, signals are biased at different levels, need to be mixed and switched, amplified and attenuated, and are prone to picking up noise. In fact, it is common for portable media devices today to have 10 to 20 different audio signal paths. Finding a path through this maze is a daunting task. Mixed-signal subsystems help solve this problem by integrating multiple effective elements.

Signal Routing

The most notable feature of a mixed-signal subsystem is its ability to route many signals to multiple locations. By using routed signals, a portable media device or cell phone can perform many tasks. An example of a mixed-signal audio subsystem is shown in Figure 1.

For example, consider a system that functions as both a cell phone and a digital audio player. The pulse code modulation (PCM) digital signal from the cell phone baseband needs to be connected to a digital-to-analog converter (DAC) and then to a headphone amplifier for use with the headphones. The same headphone amplifier also works for the digital audio player, which is an I2S data stream that is played through the DAC and then connected to the headphones. A mixed signal subsystem with dual digital audio ports can easily accomplish this task.

Another benefit of a mixed-signal audio subsystem with multiplexing capability is the ability to process analog FM radio signals. Although FM radio signal levels are usually controlled, they often exceed specifications. These out-of-spec levels are often much greater than expected, which can cause speaker damage. A mixed-signal audio subsystem can digitize the FM signal, use DSP to provide automatic level control (ALC) and equalization, and then convert it back to an analog signal for amplification to speakers or headphones. In addition, the mixed-signal subsystem can pass the digitized signal to a baseband processor for more DSP processing.

In addition to audio routing and processing, the mixed signal subsystem can also mix multiple audio streams. Sidetone can be generated by mixing the signal from the microphone into the headset. Similarly, a ringtone can be played while listening to music without muting the music.

Having two digital audio ports makes the mixed-signal audio subsystem a powerful tool for connecting digital audio within a system. For example, an I2S digital audio stream can be converted to PCM and sent to baseband. Alternatively, a 48kHz I2S interface data stream can be converted to a 44.1kHz signal using the same method.

One application that benefits from dual digital audio ports and sample rate conversion is a Bluetooth bridge. The mixed-signal audio subsystem provides a connection bridge from the Bluetooth transceiver to the baseband. Sample rate conversion can be performed if required, as well as digital equalization. An example of such a connection is shown in Figure 2.

Connecting to a Bluetooth transceiver through a mixed-signal audio subsystem enables many use cases. Obviously, a telephone can handle two-way voice. Audio signals received by Bluetooth can be sent to speakers or headphones. FM radio signals are digitized in the mixed-signal subsystem and sent to a Bluetooth headset. The baseband processor can send digital audio from flash memory through the mixed-signal subsystem to headphones or an amplifier, such as a Bluetooth-enabled docking station or headphones that can help achieve stereo effects in a car.

Class D output power

Class D speaker amplifiers are becoming the industry standard for smartphones and feature phones due to their high efficiency. The advantage of Class D amplifiers is their output power. High output power Class D amplifiers enable mobile phone speakers to be loud and clear. In areas with high ambient noise, such as train stations and airports, it is often necessary to quickly distinguish ringtones.

Feature phones or smartphones are also often used to share media resources, such as sharing a song with a friend or sharing information with a colleague.

The mixed-signal audio subsystem features a high-power Class-D amplifier. For example, the LM49352 can typically deliver 970mW into an 8Ω load with a 4.2V signal and only 1% total harmonic distortion and noise (THD+N). This superior output power ensures clear messaging at higher volume levels.

One of the latest features to be incorporated into mobile phones is the micro-projector, which, at high output power levels, enables video sharing with a group of people.

PSRR

Mobile phones rely on switch mode power supplies (SMPS) to efficiently provide multiple supply voltages. In addition to the high frequency noise generated by the SMPS power supply, the phone itself cycles power from the RF power amplifier (PA). This PA cycling occurs in the audio band, typically 217Hz.

All of these noise sources can degrade the audio quality of a mobile phone, sometimes significantly. One of the most important characteristics in a mixed-signal audio subsystem is high immunity to these noises. The power supply rejection ratio (PSRR) of a mixed-signal audio subsystem can be 90dB or higher, minimizing any noise caused by these sources. For example, the PSRR test results of the headphone amplifier of the mixed-signal audio subsystem LM49350 show that the device has a PSRR of 95dB at 217Hz and maintains high audio quality in the higher frequency region.

High PSRR is of great value to the system. The analog power supply of the mixed-signal audio subsystem can be directly connected to the battery, and the digital power supply derived from the SMPS can be used to generate other digital core voltages. Since the mixed-signal audio subsystem is inherently noise-rejecting, no additional low-dropout regulator (LDO) or passive filter is required to eliminate the noise.

Separate headphone power supply

A common feature found in almost all portable media devices is their stereo headphone connection. The connection to the headphones is generally a standard 3.5mm jack, a proprietary connector, or a variation of the mini-USB port. In all of these cases, the headphone impedance is usually around 32Ω. A true grounded headphone amplifier with a charge pump generating a negative voltage can deliver 16mW of power by simply applying 1V to a 32Ω load. For most users, 16mW is quite loud, so the actual voltage required is much lower.

Because the headphone amplifier is Class AB, separate and lower supply voltage headphone needs offer significant power advantages. In Figure 3, the two curves show a single-channel ideal amplifier with Class AB output. Simply reducing the headphone supply from 3.3V to 1.8V can save 45% of energy. Although a Class D amplifier will theoretically save more energy, it requires a larger and more expensive LC output filter. In addition, the unknown headphone cable length and load impedance will make the filter design very difficult.

High SNR Data Converters

High-performance data converters are a factor in the ever-decreasing state of geometric processing technology. Unfortunately, the baseband ICs in mobile phones have been able to provide high performance in the smallest size and lowest power levels thanks to advanced processing technology. While it has achieved these advantages, it has become increasingly difficult to maintain a high signal-to-noise ratio (SNR) in the baseband DAC and ADC.

This degradation is exacerbated by the multifunctional integration of mobile phones. If they were just used as mobile phones, this would not be a big problem. However, for many people, the mobile phone is also their portable music player. This raises the signal-to-noise ratio requirements from telecommunication quality to high fidelity, especially when using high-quality headphones.

Some might object that any SNR above 90dB would be wasteful, but this is not true. It is true that the vast majority of audio portable media devices originate from CD quality (44.1kHz sampling, 16-bit resolution) and are compressed to lower resolution and fidelity using algorithms such as MP3. However, for normal listening levels, most headphones are sensitive enough for around 2mW of power. The standard set for SNR is 40mW or more at full load output, so designers only lose about 26dB of SNR.

Another advantage of moving the digital-to-analog conversion out of baseband is that it allows the DAC to be closer to the load. Digital signals are more immune to noise than analog signals. The mixed-signal subsystem eliminates this source of noise by eliminating the routing from the baseband DAC to the external amplifier.