In audio applications there is a trend towards Class D audio amplifiers as replacements for traditional Class AB electronics. The main driving forces are improved efficiency and space savings. Converting to Class D designs promises higher output power within the same system volume or miniaturization at existing power levelsboth desirable for automotive audio and infotainment applications.
Converting to Class D for traditional Class AB amplifier designers can be quite daunting as the modes of operation are significantly different. Moreover, the circuit protection schemes must also be adapted for the different topologies. In an attempt to simplify these issues, the basic design procedure for a 500W, 2Ω automotive Class D subwoofer amplifier is presented here.
The Class D amplifier
The major differences between a Class AB amplifier and a Class D amplifier are tabulated in Table 1 below. It can be seen that the main advantages of Class D amplifiers are efficiency, stability, and inherently low output impedance (voltage source), which are all beneficial for driving speaker loads. (For a more complete discussion on the differences between the Class AB and Class D amplifiers, please refer to Application Note AN-1071 [Reference 1].)
View a full-size image
Figure 1: A half bridge Class D amplifier with voltage feedback can be cost effective if higher voltage requirements can be met.
View a full-size image
Design of a Class D amplifier
For the Class D amplifier design shown in Figure 1 above, consider the following requirements, given in Table 2 below. For this design a half bridge topology was chosen, as it tends to be the most cost effective solution if the proportionally higher bus voltage requirements can be met. As can be seen in Figure 1, the supply voltages, active devices, output filter, gate drive, and protection circuits have to be designed.
View a full-size image
Selection of power devices
The first step for selecting the switching devices is to estimate the minimum bus voltage, which is dependent on output voltage swing. In this respect, the design task starts out identical to that for a Class AB amplifier. Considering the power requirement is for 1% THD (Total Harmonic Distortion), the required peak-to-peak voltage swing (without clipping) at full load is given by Equation 1:
where M is the maximum modulation index. For this design, a modulation index of 100% is possible. Thus neglecting supply regulation and device on-resistance, the minimum bus voltages (+B or "B) for a half bridge topology will each be half of the peak-to-peak voltage swing in Equation 1 above.
As with other hard switching converters, the voltage rating of the switching devices should be on the order of 50% higher than the overall bus voltage to allow for power supply fluctuations and peak turn-off voltage. The peak load current (thus peak switching device current) is given by Equation 2:
To determine the design parameters of the switching device, the switching frequency also needs to be known. Establishing the optimum frequency can be a complicated task. First, the switching (carrier) frequency should be at least 10 times higher than the required bandwidth of the amplifier or performance (THD) deterioration at the high end of the audible frequency spectrum will occur. For a full bandwidth (20 kHz and above) amplifier, this typically means a switching frequency around 400 kHz. The higher the switching frequency, the smaller the low pass output filter, because the filter corner frequency is shifted higher.
As the switching frequency increases, the amplifier efficiency will drop due to the switching loss increases. Furthermore, the finite switching time will become a larger portion of the overall switching period. This will affect the "square-ness" of the carrier, which will impact the THD of the amplifier. Thus the switching frequency is limited at both the upper and lower frequency range. Because the bandwidth for the subwoofer amplifier in particular is very low, the "10 times bandwidth" requirement can easily be met. Therefore, the output filter size will dominate the switching frequency selection.
For determining the die size of the switching devices, a trade-off between switching loss (increasing with die size) and conduction loss (decreasing with die size) has to be made. For Class D applications in particular, where fast switching transients are preferred for better THD, the optimum die size tends to be larger than for switching-mode power supplies (SMPS) of similar power levels.
For the subwoofer design, the switching frequency is lower than for full bandwidth amplifiers and the optimum die size tend to be even larger. (For a more complete discussion on choosing the right switching devices, please refer to Application Note AN-1070 [Reference 2].)
For this subwoofer design, the following MOSFET devices were selected for each switching device: three IRFB41N15Ds in parallel, rated at 150V(BR)DSS.