NCP2821
http://onsemi.com
13
APPLICATION INFORMATION
NCP2821 PWM Modulation Scheme
The NCP2821 uses a PWM modulation scheme with each
output switching from 0 to the supply voltage. If V
in
= 0 V
outputs OUTM and OUTP are in phase and no current is
flowing through the differential load. When a positive signal
is applied, OUTP duty cycle is greater than 50% and OUTM
is less than 50%. With this configuration, the current through
the load is 0 A most of the switching period and thus power
losses in the load are lowered.
Figure 37. Output Voltage and Current Waveforms into an Inductive Loudspeaker
DC Output Positive Voltage Configuration
OUTP
OUTM
Load Current
+Vp
0 V
−Vp
0 A
Voltage Gain
The first stage is an analog amplifier. The second stage is
a comparator: the output of the first stage is compared with
a periodic ramp signal. The output comparator gives a pulse
width modulation signal (PWM). The third and last stage is
the direct conversion of the PWM signal with MOS
transistors H−bridge into a powerful output signal with low
impedance capability.
With an 8 load, the total gain of the device is typically
set to:
− 12 dB if a low level is applied to the GS pin
− 6 dB if a high level is applied to the GS pin
Input Capacitor Selection (C
in
)
The input coupling capacitor blocks the DC voltage at the
amplifier input terminal. This capacitor creates a high−pass
filter with R
in
, the cut−off frequency is given by
Fc +
1
2 R
i
C
i
.
When a 6 dB gain is chosen the internal impedance is set
to 150 k. With a 12 dB gain, the internal resistance is 75 k
and thus an input capacitor value between 10 nF and 1 F
will give a cutoff frequency between 1 Hz and 212 Hz. The
NCP2821 also includes a built in low pass filtering function.
Its cutoff frequency is set to 20 kHz.
Optional Output Filter
This filter is optional due to the capability of the speaker
to filter by itself the high frequency signal. Nevertheless, the
high frequency is not audible and filtered by the human ear.
An optional filter can be used for filtering high frequency
signal before the speaker. In this case, the circuit consists of
two inductors (15 H) and two capacitors (2.2 F)
(Figure 38). The size of the inductors is linked to the output
power requested by the application. A simplified version of
this filter requires a 1 F capacitor in parallel with the load,
instead of two 2.2 F connected to ground (Figure 39).
Cellular phones and portable electronic devices are great
applications for Filterless Class−D as the track length
between the amplifier and the speaker is short, thus, there is
usually no need for an EMI filter. However, to lower radiated
emissions as much as possible when used in filterless mode,
a ferrite filter can often be used. Select a ferrite bead with the
high impedance around 100 MHz and a very low DCR value
in the audio frequency range is the best choice. The
MPZ1608S221A1 from TDK is a good choice. The package
size is 0603.
