
TAS5112
SLES048C  JULY 2003  REVISED MARCH 2004
www.ti.com
14
Another parameter to be considered is the idle current loss
in the inductor. This can be measured or specified as
inductor dissipation (D). The target specification for
dissipation is less than 0.05.
In general, 10-
H inductors suffice for most applications.
The frequency response of the amplifier is slightly altered
by the change in output load resistance; however, unless
tight control of frequency response is necessary (better
than 0.5 dB), it is not necessary to deviate from 10
H.
The graphs in Figure 12 display the inductance vs current
characteristics of two inductors that are recommended for
use with the TAS5112.
Figure 12. Inductance Saturation
I  Current  A
4
5
6
7
8
9
10
11
0
5
10
15
L
Inductance
H
INDUCTANCE
vs
CURRENT
DFB1310A
DASL983XX1023
The selection of the capacitor that is placed across the
output of each inductor (C2 in Figure 11) is simple. To
complete the output filter, use a 0.47-
F capacitor with a
voltage rating at least twice the voltage applied to the
output stage (PVDD).
This capacitor should be a good quality polyester dielectric
such as a Wima MKS2-047ufd/100/10 or equivalent.
In order to minimize the EMI effect of unbalanced ripple
loss in the inductors, 0.1-
F 50-V SMD capacitors (X7R or
better) (C1A and C1B in Figure 11) should be added from
the output of each inductor to ground.
THERMAL INFORMATION
The thermally augmented package provided with the
TAS5112 is designed to be interfaced directly to heatsinks
using a thermal interface compound (for example,
Wakefield Engineering type 126 thermal grease.) The
heatsink then absorbs heat from the ICs and couples it to
the local air. If the heatsink is carefully designed, this
process can reach equilibrium and heat can be continually
removed from the ICs. Because of the efficiency of the
TAS5112, heatsinks can be smaller than those required for
linear amplifiers of equivalent performance.
RθJA is a system thermal resistance from junction to
ambient air. As such, it is a system parameter with roughly
the following components:
D RθJC (the thermal resistance from junction to
case, or in this case the metal pad)
D Thermal grease thermal resistance
D Heatsink thermal resistance
RθJC has been provided in the General Information
section.
The thermal grease thermal resistance can be calculated
from the exposed pad area and the thermal grease
manufacturer’s area thermal resistance (expressed in
°C-in2/W). The area thermal resistance of the example
thermal grease with a 0.002-inch thick layer is about 0.1
°C-in2/W. The approximate exposed pad area is as
follows:
56-pin HTSSOP
0.045 in2
Dividing the example thermal grease area resistance by
the surface area gives the actual resistance through the
thermal grease for both ICs inside the package:
56-pin HTSSOP
2.27
°C/W
The thermal resistance of thermal pads is generally
considerably higher than a thin thermal grease layer.
Thermal tape has an even higher thermal resistance.
Neither pads nor tape should be used with either of these
two packages. A thin layer of thermal grease with careful
clamping of the heatsink is recommended. It may be
difficult to achieve a layer 0.001-inch thick or less, so the
modeling below is done with a 0.002-inch thick layer,
which may be more representative of production thermal
grease thickness.
Heatsink thermal resistance is generally predicted by the
heatsink vendor, modeled using a continuous flow
dynamics (CFD) model, or measured.
Thus, for a single monaural IC, the system RθJA = RθJC +
thermal-grease resistance + heatsink resistance.
Table 4,
Table 5, and
Table 6 indicate modeled
parameters for one or two TAS5112 ICs on a single
heatsink. The final junction temperature is set at 110
°C in