The idea of Class D amplification has been around for a long time (remember the kits that appeared briefly in the 1960s?) but only recently has been applied for use in amplifiers that can be taken seriously. The principle is to use fast-switching transistors with pulse waveforms that have been modulated with an audio signal, and one enormous advantage is that the dissipation in the transistors can be low even for outputs of several hundred watts. This allows higher output power for an IC chip to become a reality, particularly when fast-switching MOSFETs with very low forward resistance can be used. The modulation system is usually a form of pulse time coding, in which the time of switching transitions (between minimum and maximum) is varied according to the input amplitude. The conversion to audio is carried out using a low-pass filter. A simple circuit that illustrates the principles is shown in Figure 7.19. The differential operational amplifier has inputs consisting of an audio wave and a fast pulse waveform, and the combined signal, the pulses modulated by the audio, is fed to high-power switching MOSFETs. At the output, a low-pass LC filter removes the high-frequency switching signals leaving the audio. Audio feedback can be applied from the switching output to an earlier audio stage. Practical IC implementations usually are of ‘H’ (or bridge) configuration, so the loudspeaker is fed from two switching circuits operated in anti phase. Hybrid circuits use an IC driver to feed the power FETs, as illustrated by the AudioMax LX1710 (Figure 7.20). Circuits of this type can be used for very large power outputs, as exemplified by the IR2011S assembly from International Rectifier, providing a maximum stereo output of 500W + 500W.

The use of Class D can also allow the construction of single-chip power amplifiers of fairly high power requiring the minimum of heat sinking. The Texas Instruments TPA3100D2 is a modern example, delivering 20 W per channel into an 8 ohm load (with a supply voltage of 18 V) using bridge connected speakers of 20 W/ch into an 8 ohm load from a 18 V supply. Power output into 8 ohms for a 12 V supply is 10Wper channel, and for a 4 ohm load using a 12 V supply is 15Wper channel. The permitted supply

Figure 7.20 A circuit using the Audio Max LX1710 Class D driver chip.

voltage range is 10 V to 26 V, and the efficiency is 92%, emphasizing the advantages of Class D operation so that the chip does not require a heatsink. The circuitry incorporates thermal and short-circuit protection with autorecovery, and two pins can be used to provide four gain settings of 20, 26, 32 and 36 dB. Differential inputs are used and the chip is packaged in SM format with 48 pins. A typical application circuit is shown in Figure 7.21. For full details of this chip, see the website http://focus.ti.com/lit/ds/symlink/tpa3100d2.pdf. Another single-chip solution is the STA5150 from STMicroelectronics with a maximum mono output of 200 W. Other chips are in course of development, particularly with digital rather than analogue inputs.

Figure 7.21 Typical circuitry around the T.I. TPA3100D2 chip as used in TV applications.