AN48
Application Note
DESIGNNOTESFORA2-POLEFILTERWITH
DIFFERENTIAL INPUT
by Steven Green
AIN-
AIN+
R1
R4
R1
R3
C2
C2
Figure 1. 2-Pole Low-Pass Filter with Differential Input
R4
R3
C5
_
+
C5
1. Introduction
Many of today’s Digital-to-Analog Converters
(DACs) require a circuit that has a differential input
and will remove common-mode errors, reduce the
out-of-band noise produced by the delta-sigma
modulator and produce a single-ended output. The
circuit in Figure 1 includes a differential input and a
two-pole analog filter to achieve these design requirements. This application note outlines the design steps required to select component values.
www.cirrus.com
Copyright Cirrus Logic, Inc. 2003
(All Rights Reserved)
Notice the similarities between Figure 1 and the
multiple-feedback low-pass filter shown in
Figure 2. The 2-Pole Low-Pass Filter with Differential Input is easily designed using the design equations for the multiple-feedback low-pass filter. Also,
notice the similarities between Figure 1 and
Figure 3. The differential input function is accomplished by simply duplicating the component values generated in the filter design
MAR ‘03
AN48REV2
1
AN48
R4
C5
R3R1
C2
Figure 2. Multiple-Feedback Low-Pass
Filter
_
+
2. Design Steps
Step 1: Determine the required pass band gain,
H
. The circuit parameters require that the magni-
o
tude of H
also negative due to the inverting op-amp configuration.
be greater than or equal to one. Hois
o
R4
R1
R1
Figure 3. Differential Input Circuit
_
+
R4
Step 4: Select convenient values for C5 and C2.
Notice in Step 5 that K and H
such that is real.
Step 5: Given F
ζ2K 1 Ho–()–
,C2,C5, alpha and beta, cal-
c,Ho
must be selected
o
culate R1,R2 and R3 using the following equa-
tions.
Step 2: Determine the minimum input impedance.
Step 3: Select the desired filter type, Butterworth,
Bessel, etc. and the corner frequency, F
,forthefi-
c
nal design. The filter response and corner frequency determine the pass band phase and amplitude
response. The filter type determines the pole-locations and therefore alpha and beta. Table 1 lists
the normalized pole locations for several filter
types.
Table 1: Normalized Pole Locations
FILTER TYPE αβ
Butterworth 0.7071 0.7071
Bessel 1.1030 0.6368
0.01 dB Chebyshev 0.6743 0.7075
0.1 dB Chebyshev 0.6104 0.7106
----------------------=
α
2πFcα2β2+=
C
5
-----=
C
2
R
--------------=
--------------------------------------------------------------------=
ω
oC2
α
2β2
+
4
Ho–()
1
ζζ2K 1 Ho–()–±[]
ζ
ω
o
K
R
1
R
3
ζζ2K 1 Ho–()–±
R
-------------------------------------------------=
4
ω
oC5
⋅
2