Home » Design center » Full design center

Print - Send - -

Power management

Class G/H amplifiers: Do they deliver on their promise of high audio quality and low power consumption?

In traditional hi-fi systems, the specifications of audio amplifiers emphasised audio quality, but paid less regard to the level of power losses. As the portable hi-fi sector of the audio industry has grown, however, the shortcomings of traditional amplifier devices – in particular, their inefficiency – have come to the fore.

Traditionally, audio playback equipment has used so-called Class AB amplifiers, which generate little distortion and consequently produce high audio quality. The operation of a Class AB amplifier, however, explains its relatively low efficiency: the internal voltage of the amplifier drops as the output voltage drops. The amplifier dissipates excess power across its transistors, and so as the power output to the speakers drops, the efficiency of the system declines.

In mains-powered hi-fi equipment this is not a fundamental problem; in battery-powered audio devices such as mobile phones and MP3 players, it most certainly is, as the audio amp takes a significant portion of the system’s power budget. In the case of MP3 players, the proportion of total power consumption attributable to the audio amplifier can be as high as 80 percent.

As a result, audio equipment designers have been looking for enhancements to the Class AB topology. The question that this article addresses is, are worthwhile power savings achievable through implementing a newer technology such as Class G or Class H? And if a system designer adopts a Class G or Class H amplifier, are the differences in power consumption between different implementations of Class G or Class H large enough to affect the overall power budget?

System requirements in handheld audio equipment

The audio amplifiers used in handheld devices typically drive an impedance of 16 Ω or 32 Ω, and often consume much of the device’s power budget. This means that any improvement in the efficiency of the power amplifier has a marked effect on the efficiency of the whole device, and on battery operating time.

As we will see, the most important parameter affecting the efficiency of a conventional audio amplifier is the peak output power. This is determined by the type of headphone a device is using: earbuds have a smaller peak power requirement than full headphones, but typical output power values range from two channels consuming 4 mW each, up to 2*30 mW.

An output power of 30 mW on a headphone speaker with impedance of 32 Ω needs an amplifier output swing of ±1.38 Vpk. The amplifier stages for this application will need an additional voltage headroom of 100-200mV. So the supply voltage of the headphone amplifier will be 2*1.5 V = 3.0 V.

To avoid the use of output DC decoupling capacitors, which would be too large for the application, a charge pump is typically used to generate a negative supply rail for the headphone amplifier so that the audio output is operating around battery-ground. This configuration is known as a ‘true-ground’ headphone amplifier. It uses a positive supply of 1.5 V; the -1.5 V for the negative rail comes from the charge pump.

The most commonly used battery type is Li-Ion, which normally produces a 3.6 V output. An efficient DC-DC buck converter will turn this battery output into the positive 1.5 V supply without generating substantial losses. This is the usual configuration of a Class AB amplifier, and a typical system block diagram is shown in Figure 1.

 

A high-quality DC-DC converter will transform the Li-Ion battery voltage from 3.7 V to the fixed 1.5V output voltage at an efficiency of up to 93 percent. This is of course much preferable to burning the 2.2 V (3.7 V battery voltage – 1.5 V operating voltage) across the amplifier transistors.

But this does not mask the fact that a large amount of power will still be dissipated within the transistor, except at high output voltage levels. Solving this problem requires a change to the power supply arrangement within the amplifier itself, and this is why Class G and Class H amplifiers were developed.

Matching input voltage to output voltage

The power figures for audio amplifiers mentioned above are peak power values. In practice, this maximum supply voltage is only needed for short periods; audio signals have a wide dynamic range. Most of the time the output voltage is below 0.5 V, whereas the supply voltage to the amplifier is much higher, at 1.5 V. The difference between the output voltage and supply voltage is dissipated across the internal amplifier transistors, and so shows up as losses.

To address this defect, Class G and Class H amplifiers adapt the power supply voltage of the amplifier to some extent to fit the required output power. Class G amplifiers typically have two supply voltage levels. The higher supply level is defined by the maximum output power required. The lower supply level is defined by the minimum supply voltage the amplifier can operate at while staying above a defined threshold for Total Harmonic Distortion (THD).

In contrast to Class G amplifiers, Class H amplifiers can regulate the output voltage smoothly according to the requirements of the output signal. So unlike a Class G amplifier, a Class H amplifier is not limited to two or three fixed output voltages. It can create smooth transitions from the lowest permissible supply voltage to any other discrete output voltage available from the buck converter.

A Class H amplifier can therefore operate with a supply voltage that is closely matched to the actual output voltage; reducing the difference between the supply voltage and the output reduces the amount of power that needs to be dissipated (see Figure 2).





In fact, since audio equipment operates for most of the time at the minimum permissible supply voltage, power consumption in Class G and Class H amplifiers, and therefore the contribution of the headphones to total system power consumption, is more or less determined by this THD threshold-limited lower level of supply. Let’s give this important parameter, the minimum supply voltage, the name V_SUPMIN.

The diagrams in Figure 2 show the differences in power dissipation for the different types of amplifiers. The positive and negative supply voltage for each amplifier is indicated with the black dashed line. The Class G amplifier supports two different output voltages: 1.5 V, and 1.2 V (V_SUPMIN). The Class H amplifier on the other hand supports additional voltage levels between 1.5 V and 1.2 V. Class G and Class H amplifiers clearly reduce power dissipation.

 

Moreover, the implementation of a Class G or Class H topology is not greatly more complex or costly than a Class AB amplifier. (A simplified block diagram of a class G/H amplifier is shown in Figure 3.) The main difference from a Class AB amplifier is that the DC-DC converter no longer has a fixed output voltage. The variable voltage output requires the addition of a feedback signal from the amplifier output stage to the DC-DC converter to adjust the output voltage in relation to the audio signal.

The scale of the difference between the best and worst Class H or Class G amplifiers

For reasons of cost and availability, headphone amplifiers are generally produced in CMOS technology. While the ideal value of VSUPMIN would be determined by the design of the amplifier circuit, in practice it is more often determined by the minimum thresholds of the CMOS technology used to manufacture the amplifier.

In today’s amplifier designs, V_SUPMIN is defined by 2*VTH+Vdsat. Much better results are achievable if the device is manufactured in a CMOS process that can support a lower value of V_SUPMIN by achieving lower transistor threshold voltages. For instance, if the amplifier stage operates with a minimum supply voltage of ±1.2 V, it consumes 30 percent more power than an amplifier playing the same audio signals and operating down to ±0.9 V.

This is the promise of special LowVT CMOS processes available from austriamicrosystems. These processes have been used in the manufacture of the new AS3561, a Class H amplifier that provides a low ±0.9 V supply voltage to headphone amplifiers. Together with a highly efficient DC-DC converter and an adaptive charge pump, it yields extremely low power consumption: typical current draw of 1.7 mA at 2 x 0.1 mW playback and 3.6 V battery supply voltage. The difference in power consumption between such a highly efficient architecture on the one hand, and Class G and Class AB architectures on the other, is shown in Figure 4.

 

Conclusion

The development of Class G and Class H audio amplifiers with dynamically adjustable supply voltage has led to a significant improvement in power efficiency over the widely used Class AB alternative. But the analysis of typical operation, which for most of the time is at the amplifier’s THD threshold, shows that an additional 30 percent power saving can typically be achieved by using a Class H amplifier manufactured using low voltage threshold process technology.

About the authors

Horst Gether is austriamicrosystems' Product Manager - Consumer and Communications Business Unit. 

Herbert Lenhard is austriamicrosystems' Design Engineer in CON MS Design Department.

Helmut Theiler is austriamicrosystems' Design Engineer in CON Backlight Design.

Microcontrollers

 

Batteries

 

Power

 

Energy Harvesting

 

IMS Research

 

Diodes

 

NXP Semiconductors

 

Photovoltaic

 

MOSFETs

 

Solar

 

GaN

 

Power Supply

 

Power Management

 

International Rectifier

 

Fairchild Semiconductor

 

MOSFET

 

Texas Instruments

 

Battery

 

Linear Technology

 

Intersil

 

Smart Grid

 

Maxim Integrated Products

 

UPS

 

STMicroelectronics

 

Analog

 

Microcontroller

 

Analog Devices

 

Vishay Intertechnology

 

National Semiconductor

 

Power Supplies