Application Information
BRIDGE CONFIGURATION EXPLANATION
As shown in Figure 1, the LM4992 has two internal opera-
tional amplifiers per channel. The first amplifier’s gain is
externally configurable , while the second amplifier is inter-
nally fixed in a unity-gain, inverting configuration. The
closed-loop gain of the first amplifier is set by selecting the
ratio of R
f
to R
i
while the second amplifier’s gain is fixed by
the two internal 20kΩresistors. Figure 1 shows that the
output of amplifier one serves as the input to amplifier two
which results in both amplifiers producing signals identical in
magnitude, but out of phase by 180˚. Consequently, the
differential gain for the IC is
A
VD
= 2 *(R
f
/R
i
)
By driving the load differentially through outputs Vo1 and
Vo2, an amplifier configuration commonly referred to as
“bridged mode” is established. Bridged mode operation is
different from the classical single-ended amplifier configura-
tion where one side of the load is connected to ground.
A bridge amplifier design has a few distinct advantages over
the single-ended configuration, as it provides differential
drive to the load, thus doubling output swing for a specified
supply voltage. Four times the output power is possible as
compared to a single-ended amplifier under the same con-
ditions. This increase in attainable output power assumes
that the amplifier is not current limited or clipped. In order to
choose an amplifier’s closed-loop gain without causing ex-
cessive clipping, please refer to the Audio Power Amplifier
Design section.
A bridge configuration, such as the one used in LM4992,
also creates a second advantage over single-ended amplifi-
ers. Since the differential outputs, Vo1 and Vo2, are biased
at half-supply, no net DC voltage exists across the load. This
eliminates the need for an output coupling capacitor which is
required in a single supply, single-ended amplifier configura-
tion. Without an output coupling capacitor, the half-supply
bias across the load would result in both increased internal
IC power dissipation and also possible loudspeaker damage.
POWER DISSIPATION
Power dissipation is a major concern when designing a
successful amplifier, whether the amplifier is bridged or
single-ended. A direct consequence of the increased power
delivered to the load by a bridge amplifier is an increase in
internal power dissipation. The maximum internal power dis-
sipation per channel is 4 times that of a single-ended ampli-
fier. The maximum power dissipation for a given application
can be derived from the power dissipation graphs or from
Equation 1.
P
DMAX
= 4*(V
DD
)
2
/(2π
2
R
L
) (1)
It is critical that the maximum junction temperature T
JMAX
of
150˚C is not exceeded. T
JMAX
is a function of P
DMAX
and the
PC board foil area. By adding copper foil, the thermal resis-
tance of the application can be reduced from the free air
value of θ
JA
, resulting in higher P
DMAX
values without ther-
mal shutdown protection circuitry being activated. Additional
copper foil can be added to any of the leads connected to the
LM4992. It is especially effective when connected to V
DD
,
GND, and the output pins. Refer to the application informa-
tion on the LM4992 reference design board for an example
of good heat sinking. If T
JMAX
still exceeds 150˚C, then
additional changes must be made. These changes can in-
clude reduced supply voltage, higher load impedance, or
reduced ambient temperature. Internal power dissipation is a
function of output power. Refer to the Typical Performance
Characteristics curves for power dissipation information for
different output powers and output loading.
EXPOSED-DAP MOUNTING CONSIDERATIONS
The LM4992’s exposed-DAP (die attach paddle) packages
(SD) provide a low thermal resistance between the die and
the PCB to which the part is mounted and soldered. This
allows rapid heat transfer from the die to the surrounding
PCB copper area heatsink, copper traces, ground plane, and
finally, surrounding air. The result is a low voltage audio
power amplifier that produces 1.07W dissipation per channel
in an 8Ωload at ≤1% THD+N. This power is achieved
through careful consideration of necessary thermal design.
Failing to optimize thermal design may compromise the
LM4992’s performance and activate unwanted, though nec-
essary, thermal shutdown protection.
The LM4992SD must have its DAP soldered to a copper pad
on the PCB. The DAP’s PCB copper pad is then, ideally,
connected to a large plane of continuous unbroken copper.
This plane forms a thermal mass, heat sink, and radiation
area. Place the heat sink area on either outside plane in the
case of a two-sided or multi-layer PCB. (The heat sink area
can also be placed on an inner layer of a multi-layer board.
The thermal resistance, however, will be higher.) Connect
the DAP copper pad to the inner layer or backside copper
heat sink area with vias. The via diameter should be 0.012in
- 0.013in with a 1.27mm pitch. Ensure efficient thermal con-
ductivity by plugging and tenting the vias with plating and
solder mask, respectively.
POWER SUPPLY BYPASSING
As with any amplifier, proper supply bypassing is critical for
low noise performance and high power supply rejection. The
capacitor location on both the bypass and power supply pins
should be as close to the device as possible. Typical appli-
cations employ a 5V regulator with 10 µF tantalum or elec-
trolytic capacitor and a ceramic bypass capacitor which aid
in supply stability. This does not eliminate the need for
bypassing the supply nodes of the LM4992. The selection of
a bypass capacitor, C
B
, is dependent upon PSRR require-
ments, click and pop performance (as explained in the sec-
tion, Proper Selection of External Components), system
cost, and size constraints.
SHUTDOWN FUNCTION
In order to reduce power consumption while not in use, the
LM4992 contains shutdown circuitry that is used to indepen-
dently turn off each channel’s bias circuitry. This shutdown
feature turns a given channel off when logic low is placed on
the corresponding shutdown pin. By switching a particular
shutdown pin to GND, the LM4992 supply current draw due
to that channel will be minimized in idle mode. Idle current is
measured with the shutdown pin connected to GND. The
trigger point for shutdown is shown as a typical value in the
Shutdown Hysteresis Voltage graphs in the Typical Perfor-
mance Characteristics section. It is best to switch between
ground and supply for maximum performance. While the
device may be disabled with shutdown voltages in between
ground and supply, the idle current may be greater than the
LM4992
www.national.com 10