TPA005D14
2-W STEREO CLASS-D AUDIO POWER AMPLIFIER
SLOS240A – AUGUST 1999 – REVISED MARCH 2000
22
POST OFFICE BOX 655303
DALLAS, TEXAS 75265
APPLICATION INFORMATION
losses in a real-world class-D amplifier (continued)
Losses due to rise and fall times are called switching losses. A plot of the output, showing switching losses, is
shown in Figure 26.
tSWon
+
tSWoff
=
tSW
1
fSW
Figure 26. Output Switching Losses
Rise and fall times are greater than zero for several reasons. One is that the output transistors cannot switch
instantaneously because (assuming a MOSFET) the channel from drain to source requires a specific period
of time to form. Another is that transistor gate-source capacitance and parasitic resistance in traces form RC
time constants that also increase rise and fall times.
Switching losses are constant at all output power levels, which means that switching losses can be ignored at
high power levels in most cases. At low power levels, however, switching losses must be taken into account
when calculating efficiency. Switching losses are dominated by conduction losses at the high output powers,
but should be considered at low powers. The switching losses are automatically taken into account if you
consider the quiescent current with the output filter and load.
class-D effect on power supply
Efficiency calculations are an important factor for proper power supply design in amplifier systems. Table 2
shows class-D efficiency at a range of output power levels (per channel) with a 1-kHz sine wave input. The
maximum power supply draw from a stereo 1-W per channel audio system with 8-
loads and a 5-V supply is
almost 2.7 W. A similar linear amplifier such as the TPA005D14 has a maximum draw of 3.25 W under the same
circumstances.
Table 2. Efficiency vs Output Power in 5-V 8-
H-Bridge Systems
Output Power (W)
Efficiency (%)
Peak Voltage (V)
Internal Dissipation (W)
0.25
63.4
2
0.145
0.5
73
2.83
0.183
0.75
77.1
3.46
0.222
1
79.3
4
0.314
1.25
80.6
4.47
0.3
High peak voltages cause the THD to increase