Analysis on choosing the right amplifier to improve speaker efficiency in portable electronic products

Over time, the use model of audio amplifier circuits in portable devices has evolved significantly. For example, when the main function of a cell phone was to simply reproduce speech from a speaker close to the ear, the earpiece required very little power. In addition, audio quality such as total harmonic distortion (THD), noise, and signal-to-noise ratio (SNR) were rarely a concern.

Voice typically consists of high crest factor, low duty cycle signals, so voice requires very low average power and efficiency is not a big concern. Since the RF and display functions dominate the total power consumption of a cell phone, most efficiency issues involve non-audio electronics.

But recently, cellular phones and other portable electronic products have integrated receivers, earpiece speakers, and near-field speakers (for hands-free operation). In addition, the reproduction of music (MP3 files) and movie soundtracks has also placed a heavy burden on the audio channel. As a result, the power consumption of the audio channel is no longer a side issue, but has become a major channel for power leakage. Moreover, low-fidelity sound reproduction is a thing of the past, and today's audio transmission requires a signal-to-noise ratio of more than 100dB and a total harmonic distortion of less than 0.1%.

Headphone Amplifier

Acoustic audio power amplifiers are generally divided into two operating types: headphone amplifiers (HPA) and speaker amplifiers (SPA). Headphone amplifiers must drive 32Ω or 16Ω speakers up to 30mW and maintain very high audio quality (typical values ​​are 105dB SNR, 0.01% THD and 20kHz bandwidth). However, for headphone applications, 30mW is a very high output power, which is high enough to be painful. Typical listening levels are between 100μW and 1mW.

Generating 30mW of power into a 32Ω load requires a peak signal swing of 1.4V, with additional margin for IR drop. Therefore, a ±1.8V supply voltage is typically used to achieve 30mW of output power.

A typical headphone cable contains three wires: two for the left and right drive signals, and one for a common return ground. In addition, other lines may be needed for volume control, mute, or microphone output. In such a configuration, the stereo headphone amplifier must use a single-ended output.

However, this can cause significant DC bias problems if the power supply is a single voltage rail. To avoid the use of large AC coupling capacitors, most headphone amplifiers are powered from a split supply, typically with an on-chip inverter charge pump to generate the negative rail.

Most headphone amplifiers use linear amplifiers (e.g., variations of the Class A/B output stage) to achieve the high-quality audio performance required of headphone amplifiers. Traditional Class A/B amplifiers consist of Class A and Class B operating modes. Such amplifiers are generally designed to operate primarily in Class A at low output power. The Class A state provides the best audio performance due to low crossover distortion.

Class B operation is effective at high output levels, where it has higher efficiency than Class A. However, Class B operation has higher crossover distortion. In summary, Class A/B amplifiers can achieve very low total harmonic distortion because the crossover distortion can be largely attenuated by closed-loop feedback.

Under constant supply conditions, the efficiency of a Class A/B amplifier is proportional to the output voltage swing. To recover the efficiency loss at low output power, the "Class G operation mode" technique can be used to reduce the voltage rail value at low-level signals.

A circuit is needed to detect the input signal level. If the level exceeds a predetermined threshold, the voltage rail can be raised to a higher value as needed. Most Class G amplifiers have two voltage rails: a high rail (VDD) for large signal swings, and a low rail that is only a fraction of VDD (such as 1/2 of VDD) for low-level signals. In this way, the efficiency of a signal at 1/4 of the full-scale output power is approximately the same as the efficiency of a full-scale power signal.

A variation of Class G operation is named "Class H operation" where the supply rails vary continuously with the peak signal requirements. This maximizes efficiency at all signal levels. However, due to circuit design and process limitations, the minimum voltage rail value for Class H operation is limited.

Some manufacturers apply the term "Class H" to headphone amplifiers that actually operate in Class G. True Class H operation is rarely seen in current IC headphone amplifiers.

Speaker Amplifier

Speaker amplifiers in portable electronics (for near-field applications such as hands-free and speakerphone operation) are often required to drive 8Ω or 4Ω speakers. Typical listening levels fall in the 100 to 300mW range, but IC amplifiers are typically able to deliver 1 to 2.7W average output power, with peak outputs approaching twice that level.

To produce 1.7W into an 8Ω load, the speaker amplifier must deliver 5.2V peak, or about 3.7V rms, to the speaker load. To allow for margin in IR drop, a 1.7W speaker amplifier typically uses a 5.5V rail. If a lower IR drop can be achieved with a larger switch, then slightly more than 1.8W is possible. These output power values ​​are with 1% total harmonic distortion. Even higher output power can be produced with 10% total harmonic distortion.

Generally speaking, in portable audio products, near-field speakers do not reproduce high-quality audio. Therefore, the speaker amplifier does not usually need to achieve the audio performance of a headphone amplifier. Typical audio performance is 1% total harmonic distortion at full power, 10kHz bandwidth and 94dB signal-to-noise ratio.

Efficiency is a more important factor for speaker amplifiers than headphone amplifiers, because the power levels of speaker amplifiers are much higher. Headphone amplifiers are typically less than 50% efficient - not great, but a small power consumption compared to a battery with a capacity of 4.7Wh (about 0.01% of the battery capacity for normal listening levels). However, the same 50% power consumption of a speaker amplifier operating at 1W equals 0.5W, or about 10% of the battery capacity.

Class D Speaker Amplifier

The importance of headphone amplifier versus speaker amplifier operating efficiency is a function of the time spent in one listening mode or the other. For example, a cell phone consumes more power when in speaker mode, so efficiency becomes very important. Linear amplifiers (such as Class A/B) can be used to drive speakers (and often have been in the past), but today the preferred speaker driver is a Class D amplifier. Class D speaker amplifiers maintain high efficiency over a wide range of output power levels, with efficiency starting to drop only at power levels below 1% to 2% of full power.

Class D amplifiers are not linear, but rather switching amplifiers. In a switching amplifier, a high-frequency carrier (relative to the audio band) modulates the audio input signal, typically from 100kHz to 1MHz. As a result, the output stage can be "digitally" switched (rail-to-rail) to place the output power device in the on or off state, which is the point of highest efficiency.

Switching amplifiers are often configured in bridge mode to drive speaker loads differentially, thus avoiding the need for output AC coupling capacitors. Because bridge mode amplifiers use four power switches per channel, they are twice as large as single-ended output stage amplifiers. However, for a given voltage rail, the output power of a bridge mode output stage is four times that of a single-ended amplifier.

Class D amplifiers can achieve very high efficiencies, typically over 90%. However, there are disadvantages to using this type of amplifier. Because the audio content is now a modulated signal, it must be demodulated by some kind of low-pass filter (LPF) before it can drive the speaker load. High-power LPFs that do not cause efficiency losses or distortion issues are not only large in size, but also expensive, so they cannot be used in portable devices.

However, the speaker in a portable device is itself an LPF that presents a high impedance to typical carrier frequencies. In portable devices such as cell phones, the speaker is often used as an LPF to demodulate the output signal of a switching amplifier. Sometimes, some ferrite beads are connected in series with the output of the Class D to reduce the electromagnetic interference (EMI) generated by the high-power switching output. Since the speaker has a high impedance, its modulated signal dissipates very little energy, so it can maintain good efficiency.

However, using a switching amplifier can cause serious EMI problems when long wires are used between the speaker amplifier output and the speaker load without a separate low-pass filter. For this reason, Class D amplifiers are not used for headphone amplifiers if the headphones are at the end of long wires. Therefore, Class D amplifiers should be placed close to the speaker load to avoid excessive EMI radiation.

Other types of speaker amplifiers are commonly used in the industry, but most are variations of the linear and switch-mode amplifier designs described in this article. In modern portable electronics, the demand for higher battery energy is increasing. Large, high-resolution color displays for video content, high-resolution cameras and flash memory, and high-power audio output all affect battery life. Improving the efficiency of audio speaker amplifiers has become an important design consideration to extend battery run time.