You are on page 1of 21

# DIGITAL SIGNAL PROCESSING

Butterworth Filter

Lecture by
R. Sivarajan, Assistant Professor

Magnitude Response
The magnitude function of the butterworth low pass filter is
given by

---------------(1)
where, N is the order of the filter, and
c is the cut off frequency.

## Lecture by R. Sivarajan, AP/ECE/APEC

Magnitude Response

## Lecture by R. Sivarajan, AP/ECE/APEC

Magnitude Response
As shown in the figure above, the function is monotonically
decreasing, where the maximum response is unity at =0.
The ideal response is shown by dashed line.
It can be seen that the magnitude response approaches the
ideal low pass characteristics as the order N increases.
For values <c, H(j) =1, For values >c, the values
of H(j) decreases.
At =c, the curve pass through 0.707 which corresponds
to -3 dB point.

## Normalized Butterworth Filter

From equation (1), we can get the magnitude square function
of a normalized butterworth filter as (

----------------------(1a)
Let us derive the transfer function of a stable filter. For this
purpose, substituting =s/j, we can write

## Normalized Butterworth Filter

The above equation tells us that this function has poles in the
LHS as well as in RHS because of the presence of two factor
H(s) and H(-s).
If H(s) has roots in the LHS, then the H(-s) has corresponding
roots in the RHS.
These roots we can get by equating denominator to zero, i.e.,
For N odd,




-------------------(2)
Lecture by R. Sivarajan, AP/ECE/APEC

## Normalized Butterworth Filter

For N odd,



------------------(3)
We know that for N odd, the roots can be obtained from
equation (2).
For N = 3,

## Normalized Butterworth Filter

All the poles are located in the s plane as shown in the figure
below.

## Normalized Butterworth Filter

It is found that the angular separation between the poles is
given by 360/2N, which in this case is equal to 60o and all the
poles lie on the circle.
To ensure stability, considering only the poles that lie in the
left half of the s plane, we can write the denominator of the
transfer function H(s) as
Therefore the transfer function of a third order butterworth
filter for cut off frequency c=1 rad/sec is
---------------(4)
Lecture by R. Sivarajan, AP/ECE/APEC

## Normalized Butterworth Filter

As we interested on a poles that lie in the left half of the s
plane, the same can be found by using
where
----------(5)
Now for N=3,

## Normalized Butterworth Filter

Hence the denominator of the transfer function is
So, the transfer function for normalized butterworth filter is

## Normalized Butterworth Filter

Hence the denominator of the transfer function is

## Similarly we can find the transfer function of a normalized

butterworth filter for any order of a filter as shown in table
below.
Lecture by R. Sivarajan, AP/ECE/APEC

## Normalized Butterworth Filter

N

Denominator of H(s)

1
2
3
4
5
6
7
Lecture by R. Sivarajan, AP/ECE/APEC

## Unnormalized Butterworth Filter

Equation 5 gives us the pole locations of butterworth filter for
c= 1 rad/ sec and are known as normalized poles.
In general, the unnormalized poles are given by
---------------------(6)
The transfer function of such butterworth filter can be obtained
by substituting s  s/
c in the transfer function of the
butterworth filter shown in table above.

## Order of the Filter

In equation (1a), the filter was restricted to -3 dB attenuation at
c .
Now let the maximum pass band attenuation in positive dB
is p (<3 dB) at pass band frequency p and s is the
minimum stop band attenuation in positive dB at the stop
band frequency s.
Now the magnitude function can be written as


Lecture by R. Sivarajan, AP/ECE/APEC

## Order of the Filter

From the figure above, we can find that at =p, the
attenuation is equal to p,






--------------(7)
Lecture by R. Sivarajan, AP/ECE/APEC

## Order of the Filter

Referring to the above figure, at =s,





Lecture by R. Sivarajan, AP/ECE/APEC







## Order of the Filter

Taking log on both side,


-------------------(8)

## Since this expression normally does not result in an integer

value, we therefore round off to the next highest value.
--------------------(9)
Lecture by R. Sivarajan, AP/ECE/APEC

where,
and