Class-D amplifiers

The output transistors in a Class-A amplifier remain in conduction for all of the output waveform cycle; in a Class-B amplifier, each transistor remains in conduction for exactly half the cycle, as explained above. Extending this nomenclature, an amplifier in which the transistors remain cut-off for a period of the cycle (like those in the Current-Dumping amplifier described above) can be termed as operating in Class-C. This regime is only used where either output waveform purity is not required or where purity is achieved by another means (for instance by an output filter or resonant circuit). This latter situation is sometimes found in radio transmitters where a Class-C amplifier energises a resonant tank circuit. Continuing the nomenclature even further is the Class-D type amplifier, in which the output devices act simply as switches; producing a pulse-width modulated version of the signal for subsequent reconstruction by a filter. Some commercial audio amplifier products exist using this technology.

The first stage of a Class-D amplifier (after an initial amplification stage) is the audio modulation stage which changes the analogue input signal to a constant frequency, varying duty cycle, PWM signal, as shown below.

This seemingly complex operation is accomplished with a ramp generator and comparator. First, the ramp generator is set to a fixed frequency, 250kHz, for example. The resulting 250-kHz triangle waveform, from the ramp generator, and the analogue input signal are connected to a comparator. The comparator's output switches each time the analogue input signal and triangle waveform cross. The resulting output from the comparator is a 250kHz PWM waveform, which contains the analogue input signal, as shown. The input modulation stage drives the DMOS transistors in the output stage which transfer power from the power supply to the low pass filter as "packets of energy". Transistors in Class-D amplifiers operate in the cut-off or saturation regions (i.e. "on" or "off", like a switch). This mode of operation is very similar to switch-mode power supplies. Operating the DMOS transistors in this way minimises power losses commonly associated with transistors in linear power amplifiers. The final stage of a Class-D amplifier is a low pass filter comprised of an inductor and a capacitor. Much of the gains in efficiency will be lost in the filtering stage if it is not properly designed. A 2nd order LC, Butterworth low-pass filter is typical.

The Class-D design approach can produce an amplifier with better than 90% efficiency but the resulting amplifier is more complex than its linear counterpart and there is an important tradeoff: the higher the frequency of operation, the easier it is to design the filter that removes the carrier frequency; but the higher the switching losses in the output transistors which eliminate the Class-D amplifier's efficiency advantages. In the 1980s, MOSFETs became available that could meet both the switching-speed and conduction-loss requirements to effectively implement Class-D amplifiers. In a typical, practical Class-D amplifier, the configuration is the bridge type in which two output stages and two Butterworth filters supply the amplified audio signal to a floating load, as shown in the figure below.

When driving 100W into a 4(ohm) load, typical, commercial first-generation amplifiers had a THD+N across the audio spectrum of a few percent. These figures were roughly maintained at low power levels too meaning that this type of design was not yet really of interest for high-quality monitoring and listening applications. However, this level of performance - coupled with high efficiency - was immediately of very considerable interest in audio amplifiers for cars and public address. However, it has been discovered that THD may be practically reduced by (amongst other things); filter selection, higher grade op-amps in the circuit, careful design of the feedback compensation network, and improving the linearity of the comparative triangle waveform. In particular, matching the output filter to the speaker's impedance reduces peaking in the amplifier's closed-loop response and improves the distortion characteristics of the amplifier. These gradual improvements have resulted in the development of the latest generation of integrated IC switching amplifiers which feature a power output range of 20 to 150 W and up to 95% efficiency with only 0.01% THD; a performance that may be said to rival that of a conventional amplifier, but with the advantage of much greater efficiency.

Digital amplifiers

It is clearly possible to derive a PWM modulated waveform, like that derived from the analogue modulation stage described above, directly from a PCM digital input in a PCM to PWM converter. This waveform thereby might subsequently be used to modulate directly the output-stage devices as illustrated above. An amplifier so devised might (somewhat extravagantly) be described as a digital amplifier. (In fact it's just that the power amplifier has become a powerful DAC topology.) Several commercial amplifiers have been developed along these lines.


Links

Back to home page

Address all mail to richard@richardbrice.net


© Richard Brice 2002. All rights reserved. Newnes, CYP other information by permission