Beruflich Dokumente
Kultur Dokumente
Filters
Filter types
Butterworth
Chebyshev I & II
Elliptic
Bessel
Lecture 2: Filters
Nomenclature
Filter Types wrt Frequency Range Selectivity
Lowpass
Highpass
Bandpass
Band-reject
(Notch)
H ( j )
H ( j )
H ( j )
H ( j )
EECS 247
Lecture 2: Filters
All-pass
H ( j )
Phase shaping
or equalization
2008 H.K. Page 2
Filter Specifications
Frequency characteristics (lowpass filter):
Phase characteristics:
Group delay
SNR (Dynamic range)
SNDR (Signal to Noise+Distortion ratio)
Linearity measures: IM3 (intermodulation distortion), HD3
(harmonic distortion), IIP3 or OIP3 (Input-referred or outputreferred third order intercept point)
Area/pole & Power/pole
EECS 247
Lecture 2: Filters
f3dB
H (0 )
3dB
Passband
Gain
H ( j )
Passband
EECS 247
Transition
Band
H ( j )
fc
f stop
Frequency (Hz)
Lecture 2: Filters
Stopband
Frequency
Stopband
Rejection
x 10
EECS 247
Lecture 2: Filters
H ( j ) =
1
R ( ) + jX ( )
Q=
EECS 247
X ( )
Energy Stored
Aver age Po w er D i s s i pa t i o n
R ( )
Lecture 2: Filters
p e r u n i t t i me
1
Rs + j L
QL = L
Rs
Rs
Capacitor Q :
1
ZC =
1 + jC
Rp
EECS 247
Rp
QC = CR p
C
Lecture 2: Filters
QP o l e =
EECS 247
Quality factor of
complex conjugate
poles are defined as:
s-Plane
p
2 x
Lecture 2: Filters
Q= fcenter /f
Magnitude [dB]
0
-3dB
f = f2 - f1
f1
0.1
EECS 247
1
fcenter
f2
Lecture 2: Filters
10
Frequency
2008 H.K. Page 9
G(j) G(j)e
j()
Lecture 2: Filters
]} +
{ (+) [ t +
+ A2 G[ j(+)] sin
Since
()
{[ t+
(+)
+
]}
then 2
0
<<1
[ ]
(+)
+
()+
()
+
EECS 247
d()
][ 1 (1 - ) ]
()
d()
d -
Lecture 2: Filters
{ [t +
+ A2 G[ j(+)]sin
()
]} +
{ (+) [ t + ()
+( d
d()
()
) ]}
Lecture 2: Filters
EECS 247
Lecture 2: Filters
[ (
)] +
+ A G[ j(+)] sin [ (+) ( t - )]
GR
EECS 247
Lecture 2: Filters
Summary
Group Delay
Phase delay is defined as:
PD -()/ [ time]
Group delay is defined as :
GR -d()/d [time]
If ()=k, k a constant, no phase distortion
For a linear phase filter GR PD =k
EECS 247
Lecture 2: Filters
H( j )
d
=0
-20
-40
-60
=0
-200
-400
0
Phase (degrees)
Magnitude (dB)
Normalized Frequency
Example: 5th Order Butterworth filter
EECS 247
Lecture 2: Filters
All poles
s-plane
pole
Example: 5th Order Butterworth Filter
EECS 247
Lecture 2: Filters
Filter Types
Chebyshev I Lowpass Filter
-20
-40
35
Phase [degrees]
-200
-400
0
Chebyshev I filter
Ripple in the passband
Sharper transition band
compared to Butterworth (for
the same number of poles)
Magnitude [dB]
Normalized Frequency
Example: 5th Order Chebyshev filter
EECS 247
Lecture 2: Filters
All poles
s-plane
Lecture 2: Filters
No ripple in passband
Nulls or notches in
stopband
Sharper transition band
compared to
Butterworth
Passband phase more
linear compared to
Chebyshev I
0
-20
-40
-60
0
Phase (deg)
Chebyshev II filter
Magnitude (dB)
Filter Types
Chebyshev II Lowpass
-90
-180
-270
-360
0
0.5
1.5
Normalized Frequency
Lecture 2: Filters
Filter Types
Chebyshev II Lowpass
s-plane
EECS 247
Example:
5th Order
Chebyshev II Filter
pole
zero
Lecture 2: Filters
Filter Types
Elliptic Lowpass Filter
0
Magnitude (dB)
Elliptic filter
Ripple in passband
Nulls in the stopband
-20
-40
-60
Phase (degrees)
-200
-400
0
Normalized Frequency
EECS 247
Lecture 2: Filters
Filter Types
Elliptic Lowpass Filter
j
s-plane
Sharp cut-off
Narrower transition
band
Pole Q higher
compared to the
previous filter types
Pole
Zero
Example: 5th Order Elliptic Filter
EECS 247
Lecture 2: Filters
Filter Types
Bessel Lowpass Filter
Bessel
s-plane
All poles
Poles outside unit circle
Pole
Example: 5th Order Bessel filter
EECS 247
Lecture 2: Filters
Magnitude [dB]
-20
-30
-40
lte
Fi
-50
er
rd
O
r
d
se
ea
r
c
n=1
In
n Filter order
-60
-70
-80
-90
-100
0.1
10
100
Normalized Frequency
EECS 247
Lecture 2: Filters
Filter Types
Comparison of Various Type LPF Magnitude Response
Magnitude (dB)
-20
-40
Magnitude (dB)
-60
2
Normalized Frequency
EECS 247
Lecture 2: Filters
Bessel
Butterworth
Chebyshev I
Chebyshev II
Elliptic
2008 H.K. Page 26
Filter Types
Comparison of Various LPF Singularities
j
Poles Bessel
Poles Butterworth
Poles Elliptic
Zeros Elliptic
Poles Chebyshev I 0.1dB
s-plane
EECS 247
Lecture 2: Filters
28
Bessel
Chebyshev I
0.5dB Passband Ripple
1
1
12
10
Butterworth
4
1
1
EECS 247
EECS 247
Lecture 2: Filters
Magnitude Response
4th
Magnitude (dB)
-20
-40
-60
10
10
10
Frequency [Hz]
EECS 247
Lecture 2: Filters
Phase Response
4th Order Chebyshev I versus Bessel
0
Phase [degrees]
-50
-100
-150
-200
-250
-300
-350 0
50
100
150
200
Frequency [kHz]
EECS 247
Lecture 2: Filters
Group Delay
4th Order Chebyshev I versus Bessel
14
12
10
8
4th order
Chebyshev 1
6
4
2
0
10
EECS 247
100
Frequency [kHz]
Lecture 2: Filters
1000
Step Response
4th Order Chebyshev I versus Bessel
1.4
Amplitude
1.2
1
4th order
Chebyshev 1
0.8
0.6
0.4
0.2
00
EECS 247
10
Time (usec)
Lecture 2: Filters
15
20
EECS 247
Lecture 2: Filters
Pulse Impairment
Bessel versus Chebyshev
1.5
1.5
Input
Output
0.5
0.5
-0.5
-0.5
-1
-1
-1.5
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
x 10
-1.5
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
-4
1.9
2
-4
x 10
Note that in the case of the Chebyshev filter not only the pulse has broadened but it
also has a long tail
More ISI for Chebyshev compared to Bessel
EECS 247
Lecture 2: Filters
Input Signal:
Symbol rate 1/130kHz
1111011111001010000100010111101110001001
0.5
-0.5
-1
-1.5
0
0.2
0.4
0.6
0.8
1.2
1.4
x 10
1.5
1.5
0.5
0.5
1111011111001010000100010111101110001001
-0.5
-0.5
-1
-1
-1.5
0
-4
0.2
0.4
0.6
0.8
1.2
-1.5
0
1.4
x 10
-4
1111011111001010000100010111101110001001
0.2
0.4
0.6
0.8
1.2
1.4
x 10
-4
Summary
Filter Types
Filters with high signal attenuation per pole
poor
phase response
For a given signal attenuation, requirement of
preserving constant groupdelay Higher order filter
In the case of passive filters
For integrated active filters
Lecture 2: Filters
RLC Filters
Vin
Vo =
RC
Vin s2 + o s + o2
Vo
o = 1 LC
Q = o RC = R
L
s-Plane
Lecture 2: Filters
RLC Filters
Example
Design a bandpass filter with:
Vo
Vin
Lecture 2: Filters
RLC Filters
Effect of Finite Component Q
Qfilt.=20 (ideal L)
= 1 + 1
Q filt Qid ea l Qin d .
filt
Lecture 2: Filters
RLC Filters
Vo
Vin
Question:
Can RLC filters be integrated on-chip?
EECS 247
Lecture 2: Filters
Monolithic Inductors
Feasible Quality Factor & Value
Typically, on-chip
inductors built as
spiral structures out
of metal/s layers
QL= ( L/R)
QL measured at
frequencies of
operation ( >1GHz)
Lecture 2: Filters
Integrated Filters
Implementation of RLC filters in CMOS technologies requires
on-chip inductors
Integrated L<10nH with Q<10
Combined with max. cap. 20pF
LC filters in the monolithic form feasible: freq>350MHz
(Learn more in EE242 & RF circuit courses)
Analog/Digital interface circuitry require fully integrated filters
with critical frequencies << 350MHz
Hence:
Lecture 2: Filters
Filters
2nd
Biquadratic
(2nd
H( s ) =
1+
H( j ) =
s2
PQP P2
H ( j ) =0 = 1
1
2
2 2
1 2 +
PQP
Bi quad pole s @: s =
Note : for QP 1
2
EECS 247
H ( j ) = 0
H ( j ) = = QP
P
P
2
1 1 4QP
2QP
1
2
s=
P
2
1 j 4QP 1
2QP
S-plane
poles
d
Distance from origin in s-plane:
d 2 = P (1 + 4QP2 1)
2QP
= P2
EECS 247
Lecture 2: Filters
s-Plane
j
radius = P
arccos
1
2QP
real part = -
EECS 247
poles
P
2Q P
s=
Lecture 2: Filters
P
2
1 j 4QP 1
2QP
Implementation of Biquads
Passive RC: only real poles cant implement complex conjugate
poles
Terminated LC
Low power, since it is passive
Only fundamental noise sources load and source resistance
As previously analyzed, not feasible in the monolithic form for
f <350MHz
Active Biquads
Many topologies can be found in filter textbooks!
Widely used topologies:
Single-opamp biquad: Sallen-Key
Multi-opamp biquad: Tow-Thomas
Integrator based biquads
EECS 247
Lecture 2: Filters
Active Biquad
Sallen-Key Low-Pass Filter
R1
Vin
R2
C2
1+
Vout
P =
QP =
H ( s) =
C1
s
s2
+
PQP P2
1
R1C1R2C2
1
1
1 G
+
+
R1C1 R2C1 R2C2
Single gain element
Can be implemented both in discrete & monolithic form
Parasitic sensitive
Versions for LPF, HPF, BP,
Advantage: Only one opamp used
Disadvantage: Sensitive to parasitic all pole no zeros
EECS 247
Lecture 2: Filters
s
1+
H( s ) = K
1+
s
+
PQP P
H( j ) = K P
Z
EECS 247
Lecture 2: Filters
Imaginary Zeros
f P = 100kHz
QP = 2
fZ = 3 fP
6
x 10
2
Pole-Zero Map
1.5
With zeros
No zeros
Imag Axis
Magnitude [dB]
10
0.5
-10
-20
-0.5
-30
-1
-40
-50
4
10
EECS 247
-1.5
5
10
6
10
Frequency [Hz]
7
10
-2
-2
-1.5
-1
-0.5
0.5
Real Axis
Lecture 2: Filters
1.5
2
6
x 10
5
x10
fZ = fP
6
4
Imag Axis
20
Magnitude [dB]
Pole-Zero Map
10
0
-10
2
0
-2
-20
-4
-30
-6
-40
-50 4
10
105
106
Frequency [Hz]
EECS 247
107
Lecture 2: Filters
-6
-4
-2
Real Axis
6
5
x10
Parasitic insensitive
Multiple outputs
Ref:
P. E. Fleischer and J. Tow, Design Formulas for biquad active filters using three
operational amplifiers, Proc. IEEE, vol. 61, pp. 662-3, May 1973.
EECS 247
Lecture 2: Filters
Frequency Response
(b a b )s + (b2a0 b0 )
Vo1
= k2 2 1 2 1
Vin
s + a1s + a0
Vo 2 b2 s 2 + b1s + b0
= 2
Vin
s + a1s + a0
Vo 3
1 (b0 b2a0 )s + (a1b0 a0b1 )
=
Vin
s 2 + a1s + a0
k1 a0
Vo2 implements a general biquad section with arbitrary poles and zeros
Vo1 and Vo3 realize the same poles but are limited to at most one finite zero
EECS 247
Lecture 2: Filters
Component Values
given ai , bi , ki , C1 , C2 and R8
b0 =
R8
R3 R5 R7C1C2
b1 =
1 R8 R1R8
R1C1 R6 R4 R7
R2 =
b2 =
R8
R6
R3 =
a0 =
R8
R2 R3 R7C1C2
1
k1k2
a1 =
1
R1C1
R4 =
1
1
1
k2 a1b2 b1 C1
R5 =
k1 a0
b0C2
R6 =
R8
b2
k1 =
k2 =
R2 R8C2
R3 R7C1
R7
R8
EECS 247
R1 =
1
a1C1
k1
a0 C2
1
a0 C1
it follows that
P =
R8
R2 R3 R7C1C2
QP = P R1C1
R7 = k2 R8
Lecture 2: Filters