AN48

Application Note
Design Notes for a 2-Pole Filter
with Differential Input
by
Steven Green

R4 C5

R1 R3
AIN- _

C2 +

R1
AIN+
C2 R4 R3 C5

Figure 1. 2-Pole Low-Pass Filter with Differential Input

Introduction

The CS4329 evaluation board required an analog Notice the similarities between Figure 1 and the
circuit capable of removing common-mode multiple-feedback low-pass filter shown in
errors and lowering the out-of -band noise Figure 2. The 2-Pole Low-Pass Filter with
produced by the delta-sigma modulator. The Differential Input is easily designed using the
circuit in Figure 1 includes a differential input design equations for the multiple-feedback
and a two-pole analog filter to achieve these low-pass filter. Also, notice the similarities
design requirements. This application note between Figure 1 and Figure 3. The differential
outlines the design steps required to select input function is accomplished by simply
component values. duplicating the component values generated in
the filter design.
Cirrus Logic, Inc.
Crystal Semiconductor Products Division Copyright  Cirrus Logic, Inc. 1996 OCT ’96
P.O. Box 17847, Austin, TX 78760 (All Rights Reserved) AN48REV1
(512) 445-7222 FAX 445-7581
http://www.crystal.com 1
 Design Notes for a 2-Pole Filter with Differential Input

R4 R4

C5 R1
_

+
R1 R3
_ R1

+ R4
C2

Figure 2. Multiple-Feedback Low-Pass Filter Figure 3. Differential Input Circuit

Design Steps

Step 1: Determine the required pass band gain, Step 3: Select convenient values for C5 and C2.
Ho. The circuit parameters require that the Notice in Step 4 that K (C5/C2) and Ho must be
magnitude of Ho be greater than or equal to one. selected such that √

ζ2 − K (1 − Ho) is real.
Ho is also negative due to the inverting op-amp
configuration. Step 4: Given Fc, Ho, C2, C5, alpha and beta,
Note: For the CS4329, the single-ended output calculate R1, R2 and R3 using the following
will be 2Vrms for Ho = -1. equations.

Step 2: Select the desired filter type, α

Butterworth, Bessel, etc. and the corner ζ=
frequency, Fc, for the final design. The filter 
√ α2 + β2
response and corner frequency determine the
pass band phase and amplitude response. The ωo = 2πFc√

α2 + β2
filter type determines the pole-locations and
therefore alpha and beta. Table 1 lists the C5
normalized pole locations for several filter types. K=
C2

R4
R1 =
(−Ho)
FILTER TYPE α β
1
Butterworth Second Order 0.7071 0.7071 R3 =
Bessel 1.1030 0.6368 ωo C2 [ ζ ± √

ζ 2 − K(1 − Ho)]
0.01 dB Chebyshev 0.6743 0.7075


0.1 dB Chebyshev 0.6104 0.7106
ζ±√ ζ2 − K(1 − Ho)
R4 =
Table 1. Normalized Pole Locations ωo C5

2 AN48REV1
 Design Notes for a 2-Pole Filter with Differential Input

Step 5: It is recommended that R1 be a Design Example

minimum of 10 kohm to meet the CS4329 load
requirements. Values much larger than 10 kohm The following example shows the steps required
can lead to small capacitor values where it is to duplicate the CDB4329 filter.
desireable to keep the capacitor values large to
minimize the effects of stray capacitance. It may Step 1: The required output level is 2 Vrms,
be necessary to adjust the capacitor values therefore the required pass band gain, Ho, is -1.
chosen in Step 3 to achieve this requirement.
Step 2: A two-pole Butterworth with a corner
Step 6: The resistor values calculated in Step 4 frequency of 50 kHz attenuates the signal at
are generally not standard values. Select standard 20 kHz by approximately 0.1 dB and has nearly
values which are nearest the calculated values. ideal phase linearity within the audio band. A
This should not create a large change in the filter Bessel response would achieve negligible phase
characteristics since metal film resistors are improvement at the expense of degraded
available in approximately 2.5% increments amplitude response.
which allows for component selection near the Fc = 50 kHz
calculated values. alpha = .7071
beta = .7071
Step 7: The conversion from the singled-ended
circuit to the differential circuit requires Step 3: Select convenient values for C5 and C2.
duplicating the values of R1, C2, R3, R4 and C5 C5 = 220 pF
in the non-inverting input as shown in Figure 1. C2 = 1000 pF

Step 4: Given Fc, Ho, C2, C5, alpha and beta,

calculate R1, R3 and R4.
R1 = 13.77 kΩ
R3 = 3.343 kΩ
R4 = 13.77 kΩ

Step 5: Verify that R1 is greater than 10 kohm.

R1 = 13.77 kΩ

Step 6: Select standard values which are nearest

the calculated values.
R1 = 13.7 kΩ
R3 = 3.32 kΩ
R4 = 13.7 kΩ

Step 7: The conversion from the singled-ended

to the differential circuit requires duplicating the
values of R1, C2, R3, R4 and C5 in the
non-inverting input as shown in Figures 4 and 5.

AN48REV1 3
 Design Notes for a 2-Pole Filter with Differential Input

220 pF
13.7 kΩ

13.7 kΩ 3.32 kΩ
_

+
1000 pF

Figure 4. Filter Design

13.7 kΩ

220 pF

13.7 kΩ 3.32 kΩ
AIN- _

1000 pF +
13.7 kΩ 3.32 kΩ

AIN+
1000 pF 13.7 kΩ 220 pF

Figure 5. CDB4329 2-Pole Filter with Differential Input

References

[1] C. L. Lindquist, "Active Network Design with Signal Filtering Applications", Steward & Sons

[2] A. B. Williams, "Electronic Filter Design Handbook", McGraw-Hill

[3] "Reference Data for Radio Engineers" (Fourth Edition), International Telephone and Telegraph
Corporation

4 AN48REV1
• Notes •