Active Filters: Conventional Passive Filters Consist of LCR Networks. Inductors Are Undesirable Components

Download as ppt, pdf, or txt
Download as ppt, pdf, or txt
You are on page 1of 18

Active Filters

 Conventional passive filters consist of LCR


networks.
 Inductors are undesirable components:
 They are particularly non-ideal (lossy)
 They are bulky and expensive
 Active filters replace inductors using op-amp
based equivalent circuits.
Active Filter Designs
Three active filter design techniques will be covered:

 Synthesis by Sections
 Cascade of second order sections.
 Component Simulation
 Replace inductors with op-amp inductor simulations.
 Operational Simulation
 Simulate all currents and voltages in the LCR ladder using
an analogue computer.
Analogue Filter Responses

H(f) H(f)

0 f 0 f
fc fc

Ideal “brick wall” filter Practical filter


Standard Transfer Functions
 Butterworth
 Flat Pass-band.
 20n dB per decade roll-off.
 Chebyshev
 Pass-band ripple.
 Sharper cut-off than Butterworth.
 Elliptic
 Pass-band and stop-band ripple.
 Even sharper cut-off.
 Bessel
 Linear phase response – i.e. no signal distortion in pass-
band.
Analogue Transfer Functions
The transfer function of any analogue filter (active
or passive) can be expressed as the ratio of two
polynomials : M 1

N s   s  z m 
H s   K  K mN 01
Ds 
  s  pn 
n 0

Special case when M=0, all-pole response :


K K
H s    N 1
Ds 
  s  pn 
n 0
Poles and Zeros
 Poles
 Complex values of s where the transfer function is infinite.
 i.e. the denominator of the transfer function is zero.
 Zeros
 Complex values of s where the transfer function is zero.
 An N-th order filter will have N poles and up to N
zeros.
 Some poles may be in the same place (as may some
zeros).
Example – Two Pole Bessel Filter

Low pass, cut-off frequency = 1 rad/s, from tables :

1 1
H s   
s 2  1.73s  1 s  p0 s  p1 
where,
p0  0.866  j 0.5
p1  0.866  j 0.5
Operational Amplifiers
 All the active filters we shall study are based
on operational amplifiers (op-amps).
 Analysis of linear op-amp circuits is usually
based on simplifying assumptions :
 The difference between the non-inverting and
inverting inputs is zero.
 The input current is zero.
 The output voltage and current is arbitrary.
Op-Amp Assumptions
I+
V+ + Iout
Vout
I-
V- -

V  V
I  0
I  0
Vout  anything
I out  anything
Inverting Amplifier
Z2
Z1
VIN -
VOUT
+

0V

VIN V
V  V  0  I Z 1  & I Z 2  OUT
Z1 Z2
VIN  VOUT VOUT Z2
I   0 I Z1  I Z 2    
Z1 Z2 VIN Z1
Non-Inverting Amplifier

VIN +
VOUT
-
I  0
Z1 Z2
Z2 V  VOUT  V  VIN
Z1  Z 2
VOUT Z1  Z 2 Z
0V    1 1
VIN Z2 Z2
Buffer Amplifier

VIN + V  VOUT
VOUT V  VIN
-
V  V VOUT  VIN

 Output voltage = Input voltage


 Input impedance is infinite
 Output impedance is zero
Single-Pole Passive Filter
R vout ZC 1 / sC
 
vin R  Z C R  1 / sC
vin C vout
1

sCR  1

 First order low pass filter


 Cut-off frequency = 1/CR rad/s
 Problem : Any load (or source) impedance
will change frequency response.
Single-Pole Active Filter

R
vin C
vout

 Same frequency response as passive filter.


 Buffer amplifier does not load RC network.
 Output impedance is now zero.
Low-Pass and High-Pass Designs

High Pass Low Pass

vout 1 vout 1
 
vin 1
1 vin sCR  1
sCR
Higher Order Filters
 You might think we could make higher order filters by
simply cascading N first order filters
 This doesn’t work
 The single pole of a first order filter must be purely
real (no imaginary part)
 The poles of a higher order filter usually need to be
complex
 Solution: Use second order sections, each one
synthesising a conjugate pair of complex poles
Summary
 Active filter designs aim to replace the
inductors in passive filters.
 Design techniques :
 Synthesis by sections
 Component simulation
 Operational simulation
 All based on op-amps – understanding of
basic op-amp circuits is essential.

You might also like