Advanced Measurements and Digital circuits EE-705

Real filters

Approximation to Ideal filters

Presented by Supervisor
Eng. Marwan Magdy Prof. Alaa Eldin Khalil
Reg.No# 22127307
slide no#1
Content
 Introduction to filters
 Types of Filters
 Design with Normalized Analog Filters
 Approximation to Ideal filters
 Butterworth Filters
 Chebyshev Filters
 Elliptic Filters
 Bessel Filters
 Comparison between four filter types
 References

slide no#2
 Introduction to filters
 Detection of a wanted signal may be impossible if unwanted
 signals and noise are not removed sufficiently by filtering
 Electronic filters allow some signals to pass, but stop others
 Filters allow some signal frequencies applied at their input terminals to pass Ideal Filter Response
through to their output terminals with little or no reduction in signal level
 Passband: Range of signal frequencies that are allowed to pass through a
filter, with little or no change to signal level
 Passband cutoff frequency: Passband edge where there is a 3 dB reduction
in signal amplitude (half-power point)
 Stopband: Range of signal frequencies that are reduced in amplitude by
an amount specified in the design, and effectively prevented from passing,
is called the stopband
 Transition Zone : Range of frequencies in between the passband and the
stopband where the reduction in signal amplitude (also known as the
attenuation) changes rapidly

Practical Filter Response

slide no#3
 Types of Filters
Bandpass Bandstop
Lowpass Highpass

Smooth Pass band

Pass band Ripple
Passband
and Butterworth or Bessel Chebyshev

Stopband Pass band

Stop band
Response Ripple
&
Stop
band
Ripple
Inverse Chebyshev Elliptic slide no#4
 Design with Normalized Analog Filters

 A normalized filter is one in which the passband Design :

cutoff point is at = 1 rad/s
 Select the type of response required
 Passive filters are normalized for a 1  load
 Determine the filter order using the frequency
impedance
response graphs
 The reason for normalization is to make the
 Use normalized analog filter tables to obtain a set
calculation of values simple
of normalized component values
 Passive analog filters can be designed using the
 Scale the obtained normalized component
tables of normalized component values
values for the frequency, impedance, and
 One set of normalized component values can be frequency response (lowpass, highpass….) as
used to design passive lowpass, highpass, required
bandpass, and bandstop filters with any load
impedance

slide no#5
 Approximation to Ideal filters
 Many types of filters, they bear the names of the men who developed them. The common
ones are Butterworth, Chebychev, Elliptic, Bessel

Wilhelm cauer
German British Engineer Russian German Mathematician
Mathematician 1930 Mathematician &
Otto Brune
American Engineer
20th century slide no#6
 Bessel Response

 Bessel response is smooth in the passband, and attenuation rises

smoothly in the stopband
 Far away from the cutoff point the attenuation rises at 6n dB/octave, where
n is the filter order and an octave is the doubling of frequency.
 For example, a third-order filter will give an 18 dB/octave rise in attenuation
 Natural cutoff frequency for the Bessel response is that which gives
a 1 s delay
 This is not a constant value, but depends on filter order
 To simplify design, Bessel response can be scaled to give a 3 dB at =1 Normalization
for all filter orders Factors

slide no#7
  Bessel LC Values Rs =  or Rs = 0
Bessel Response

Bessel Normalized Lowpass

Filter Component Values
 Attenuation vs. frequency for different order

 Bessel LC Values Rs = 1
 Bessel Response Filter Examples

 Fifth-Order Lowpass Rs >1  Sixth-Order Lowpass Rs 1

 Fifth-Order Lowpass Rs =1  Sixth-Order Lowpass Rs 1

slide no#9
 Bessel filter

 Advantages:
• Best step response-very little overshoot
• Smooth roll off.
• No ripples.
• Shortest time delay
 Disadvantages
• Very slow roll off speed.
 Application : includes audio systems, communication systems, and other
applications where minimal distortion is desired.

slide no#10
 Butterworth Filter AMP
maximally flat response
 Butterworth filter response has a smooth passband and a
smooth increase in stopband attenuation. −20𝑑𝐵/𝑑𝑒𝑐𝑎𝑑

➢The Butterworth characteristic provides a very flat amplitude response in the

pass band and a roll-off rate of −20𝑑𝐵/𝑑𝑒𝑐𝑎𝑑𝑒/𝑝𝑜𝑙𝑒.

➢It differs from Bessel response in that the attenuation in the stopband rises by
6n dB/octave almost immediately outside the passband

➢Filters with the Butterworth response are normally used when all frequencies
in the passband must have the same gain.

➢The Butterworth response is often referred to as a

maximally flat response because the passband response offers the
steepest roll-off without inducing a passband ripple.

Better phase response linearity than Chebychev but not Bessel.

slide no#11
 Butterworth Normalized Lowpass Component Values
Butterworth Filter
 Butterworth LC Values Rs =  or Rs = 0

 Butterworth passive lowpass filters have a ladder

network of series inductors with shunt capacitors
at their connection nodes
 First component in this ladder can be either a
series inductor or a shunt capacitor

First Component is Shunt C

 Butterworth LC Values Rs = 1

First Component is Series L

slide no#12
Butterworth Filter
➢Figure shows a sketch of the magnitude response of a Butterworth
filter.
➢The magnitude of function for an 𝑁𝑡ℎ-order Butterworth filter with
a passband edge 𝜔𝑝 is given by
 At𝜔=𝜔𝑝

 Thus, the parameter 𝜖determines the maximum variation in

passband transmission, 𝐴𝑚𝑎𝑥, according to
 Conversely , given 𝐴𝑚𝑎𝑥, the value of 𝜖 can be determined from
 At the edge of the stopband ,𝜔=𝜔𝑠, the attenuation of the
Butterworth filter can be obtained by substituting 𝜔=𝜔𝑠.

slide no#13
 Natural mode identification

 The natural modes of an Nth-order Butterworth filter can be determined

from the graphical construction shown Figure.
 ➢Observe that the natural modes lie on a circle of radius
and are spaced by equal angles of 𝝅/𝑵,with the first mode at an angle 𝝅/𝟐𝑵
from the +𝑗𝜔 axis.
 Once the 𝑁 natural modes 𝑝1,𝑝2,…,𝑝𝑁 have been found, the transfer
function can be written as

 where 𝑲is a constant equal to the required dc gain of the filter.

slide no#13
 Design Procedure

 To find a Butterworth transfer function that meets transmission specifications

of the low pass filter, we perform the following procedure:
1)Determine 𝜖.
2)Determine the required filter order as the lowest integer value of 𝑁that
results in 𝐴𝜔𝑠≥𝐴𝑚𝑖𝑛.
3)Determine the 𝑁 natural modes.
4)Determine 𝑇(𝑠).

slide no#15
 Butterworth Filter
 Advantages
• Reasonable roll-off rate for high order filter.
• Smooth & reasonably flat response in the passband and stopband.
• Easy to construct and tune.
• Reasonably smooth phase shift.
 Disadvantages
• Sharp cutoff filter need high order and therefore lots of components.
• Amplitude response is NOT actually flat in the passband, particularly near cutoff.
• Not possible to modify shape of amplitude response once the filter order is
chosen.
 Application : It is most often used in audio processing applications where flat
passband response is necessary.

slide no#16
 Chebyshev Filter
 Chebyshev response has ripples in the passband but a smooth increase in stopband
attenuation.
• By allowing the passband response to have ripples, the stopband attenuation
rises sharply just beyond the cutoff frequency.
 Graphs for different passband ripples

1dB Ripple 0.25dB Ripple 0.1dB Ripple 0.01dB Ripple

slide no#17
Chebyshev Tables: Equal Source and
Load Impedances

slide no#18
 Chebyshev Filter
 The magnitude of the transfer function of an 𝑁𝑡ℎ-order Chebyshev filter with a passband edge (ripple
bandwidth) 𝜔𝑝 is given by

 At the passband edge, 𝜔=𝜔𝑝, the magnitude function is given by

 Amax, the parameter 𝜖 determines the passband ripple according to
 Conversely, given 𝐴𝑚𝑎𝑥, the value of 𝜖 is determined from
 The attenuation achieved by the Chebyshev filter at the stopband edge (𝜔=𝜔𝑠)
is found as
 With the aid of a calculator, this equation can be used to determine the order 𝑁 required to obtain a specified 𝐴𝑚𝑖𝑛
by finding the lowest integer value of 𝑁 that yields 𝐴(𝜔𝑠)≥𝐴𝑚𝑖𝑛.
 As in the case of the Butterworth filter, increasing the order N of the Chebyshev filter causes its magnitude function to
approach the ideal brick-wall low-pass response. The poles of the Chebyshev filter are given by

 Finally, the transfer function of the Chebyshev filter can be written as

Where 𝐾 is the dc gain that the filter is required to have.
slide no#19
 Design Procedure

 To summarize, given low-pass transmission specifications, the transfer function of a Chebyshev filter
that meets these specifications can be found as follows:
 1)Determine 𝜖.
 2)Determine the order required.
 3)Determine the poles.
 4)Determine the transfer function.

slide no#20
 Chebyshev Filter
 Advantage :
• Amplitude Response can have flatter (but not smoother ripples) response in the
passband and stopband.
• Steeper roll off compared to Butterworth.
• Roll off can be traded for passband or stopband ripple. (ie if passband ripples is larger,
roll off increases.)
• Easier to tune and less expensive to build due to lower order required.
 Disadvantage:
• Passband amplitude response not smooth.
• Greater phase shift and group delay.
• Greater sensitivity to component values compare to Butterworth.

slide no#21
Elliptic filter
 Cutoff frequency Fc is the frequency where the response is down by the ripple amplitude ,
not 3dB.
 Stopband is defined as starting where the attenuation first reaches the same value as the
tops of the attenuation minima

slide no#22
 Elliptic filter
 The elliptic filters is characterized by ripple that exists in both the passband
and stopband. The passband ripple (can be controlled) of the elliptic filter is
similar to the Chebyshev filter, however the selectivity is greatly improved.
 Steepest roll-off compared to Butterworth and Chebychev.
 Elliptic filters are particularly sensitive to capacitance values and therefore
capacitors used (series/parallel) must be as close as possible to the design
value.
 Advantage
• Fastest roll off speed of all the filters among the four.
• Ideal for applications that want to effectively eliminate the
frequencies in the immediate neighborhood of passband.
 Disadvantage
• Ripples in both pass and stop bands.
• Requires more components than other filters.
• Largest time delay and non-linear phase, that leads to phase
distortion. High complexity.

slide no#23
Attenuation of sixth order approximation

slide no#24
slide no#25
 References

 Analog and Digital Filter Design, 2nd ed., by Steve Winder

slide no#26
Slide no# 27