CT Filters & Freq Resp

EECE 301
Signals & Systems

Prof. Mark Fowler
Note Set #15
C-T Systems: CT Filters & Frequency Response
1/14
Ideal Filters
Often we have a scenario where part of the input signals spectrum comprises
what we want and part comprises something we do not want. We can use a
filter to remove (or filter out) the bad part.
x (t )
y (t )
H()
Called a Filter in this case
X ( )
Case #1:
Undesired
Part
Spectrum of the
Input Signal
H ( )
In this case, we want

a filter like this:
Mathematically:
1,
H ( )
0, otherwise
A filter that
passes
low
frequencies
is called a
low-pass
filter
Passband
Stopband
2/14
Case #2:
Input Signal
Spectrum
Undesired
Part
X ( )
We then want:
H ( )
0,
H ( )
1, otherwise
A filter that
passes
high
frequencies
is called a
high-pass
filter
Stopband
Passband
3/14
X ( )
Case # 3:
Undesired Part
H ( )
A filter that stops middle

frequencies is called a
band-stop filter
X ( )
Case #4:
Desired Part
H ( )
A filter that passes middle

frequencies is called a
band-pass filter
4/14
What about the phase of an IDEAL filters H()?

Wellwe could tolerate a small delay in the output so
Put in the signal
we want passed
xg (t )
y ( t ) x g ( t td )
H()
Want to get out the

signal we want
passed but we can
accept a small delay
From the time-shift property of the FT then we need:
Y ( ) X g ( )e
jt d
Thus we should treat the exponential term here as H(), so we have:
H ( ) e jtd 1
H ( ) e
jtd
td
For in the
pass band
of the filter
Line of slope td
Linear Phase
5/14
So for an ideal low-pass filter (LPF) we have:

H ( )
p2()
1e jtd ,
H ( )
0, otherwise
H ( ) p2 ( )e jtd
H ( )
Phase is undefined in stop band:
Slope = t d
Summary of Ideal Filters
0 0e j
0 ?
1. Magnitude Response:
a. Constant in Passband
b. Zero in Stopband
i.e. phase is undefined

for frequencies outside
the ideal passband
2. Phase Response
a. Linear in Passband (negative slope = delay)
b. Undefined in Stopband
6/14
Example of the effect of a nonlinear phase but an ideal magnitude

0 Hz
1 Hz
2 Hz
3 Hz
x (t ) 9 5 cos(2t ) 3 cos(2 2t ) cos(2 3t )

H()
y (t ) 9 5 cos(2t ) 3 cos(2 2t ) cos(2 3t )

4
10
Filter has FLAT

Passband
Filter has
Non-Linear Phase
Point of this Example

A filter with an ideal magnitude response but nonideal phase response can still degrade a signal!!!
7/14
Are Ideal Filters Realizable? (i.e., can we actually MAKE one?)

Sadly No!!
So a big part of CT filter design focuses on how to get close to the ideal.
Cant Get an Ideal Filter Because they are Non-Causal!!!
For the ideal LPF we had H ( ) p2 ( )e jtd
Now consider applying a delta function as its input: x(t) = (t) X() = 1
Then the output has FT Y ( ) X ( ) H ( ) p2 ( )e
From the FT Table: 2 sinc 2 t / 2
jtd
2 p2 ( )
Linear Phase
Imparts Delay
So the response to a delta (applied at t = 0) is: y (t ) ( / ) sinc ( / )(t td )

x(t) = (t)
y(t)
Ideal
LPF
t
Starts before input starts
Thus, system is non-causal!
8/14
Plotting Frequency Response of Practical Filters

Although weve previously shown the plots of Freq. Resp. using the actual
numerical values of |H()| it is VERY common to plot its decibel values.
Decibel: a logarithmic unit of measure for a ratio between two powers

P1
10 log 10
P2
bel
decibel
Know
These!
P1/P2
(non-dB)
P1/P2
(dB)
Decibel Power Rules
1000 = 103
30 dB
30 dB is P ratio of 1000
100 = 102
20 dB
10 = 101
10 dB
1 = 100
0 dB
0.1 = 10-1
10 dB
-10 dB is P ratio of 0.1
0.01 = 10-2
20 dB
0.001 = 10-3
30 dB
Another Rule to Know!!

P1/P2 = 2 ~ 3 dB P1/P2 = 1/2 ~ -3 dB
9/14
But |H()| relates Voltages (or current) not POWER!!!
A cos(0t )
Input voltage amplitude
h(t)
H()
=A
| H (0 ) |
Convert to Powers
Input Power = A2/2
Output Power = A2|H(0)|2/2
A H (0 ) cos(0t H (0 ))
Output voltage amplitude = A|H(0)|
Output Voltage Amplitude

Input Voltage Amplitude
2
2
A H ( ) / 2
Pout
10 log10
10 log10
2
A /2
Pin
10 log10 H ( )
20 log10 H ( )
20 log10(|H()|)
Decibel value for |H(0)|

10/14
In addition to using decibels for the |H()| it is also common to use a

logarithmic scale for the frequency axis
Linear Axis:
Log Axis:
1000
2000
0 0
10
3000
4000
4
6
20
10
6000
10
5000
f (Hz)
80
60
8000
10
f (Hz)
40
7000
10
9000
10000
10
200
100
We may be just as interested in 0 1 kHz as we are in 1 10 kHz

But the linear axis plot has the 0 1 kHz region all scrunched up
However the log axis allows us to expand out the lower
frequencies to see them better!
11/14
Simplest Real-World Lowpass Filter: RC Circuit
Input Signal
v(t)
y(t) Output Signal
1. Convert capacitor into impedance: Z c ( )

2. Imagine input as phasor:
Ae
1
jC
Small impedance at high

Large impedance at low
3. Now analyze the circuit as if it were a DC circuit with a complex voltage

in (the phasor) and complex resistors (the impedances):
Ae
R
1/jC
Here
use
Voltage
Divider.
Z c ( )
1
x
x
R Z c ( )
1 j RC
Now find the output phasor

as a function of the input
phasor the thing that
multiplies the input phasor
is ALWAYS the Freq Resp !
1
H ( )
1
j
RC
12/14
Now we can plot this
|H(f)|
0.5
0
0
1000
2000
3000
4000
1000
2000
3000
4000
5000
f (Hz)
6000
7000
8000
9000 10000
5000
6000
7000
8000
9000
<H(f)
0
-0.5
-1
-1.5
f (Hz)
10000
RC=1.5915e-4;
f=0:10:10000;
H=1./(1 + j*2*pi*f*RC);
subplot(2,1,1)
plot(f,abs(H))
grid
xlabel('f (Hz)')
ylabel('|H(f)|')
subplot(2,1,2)
plot(f,angle(H))
grid
xlabel('f (Hz)')
ylabel('<H(f)')
Although these are correct plots we usually prefer to use:

dB for the magnitude axis (but not the angle axis!)
log axis (rather than linear) for the frequency axis
o But keep in mind that when using a log axis a linear phase
will NOT be a straight line!!!
13/14
Using log scale allows us to see

that this filter is quite flat up
to about 200 Hz!
Instead we can plot this using dB and log axes

0
|H(f)| (dB)
-5
-10
-15
-20
-25 0
10
10
10
f (Hz)
10
10
<H(f)
-0.5
-1
-1.5 0
10
10
10
f (Hz)
10
10
RC=1.5915e-4;
f=1:10:10000;
H=1./(1 + j*2*pi*f*RC);
subplot(2,1,1)
semilogx(f,20*log10(abs(H)))
grid
Use
xlabel('f (Hz)')
20
ylabel('|H(f)|')
here!
subplot(2,1,2)
semilogx(f,angle(H))
grid
xlabel('f (Hz)')
ylabel('<H(f)')
14/14

CT Filters & Freq Resp

Uploaded by

Copyright:

Available Formats

CT Filters & Freq Resp

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CT Filters & Freq Resp

Uploaded by

Copyright:

Available Formats

EECE 301

Signals & Systems

In this case, we want

A filter that stops middle

A filter that passes middle

What about the phase of an IDEAL filters H()?

Want to get out the

From the time-shift property of the FT then we need:

Thus we should treat the exponential term here as H(), so we have:

So for an ideal low-pass filter (LPF) we have:

Phase is undefined in stop band:

i.e. phase is undefined

Example of the effect of a nonlinear phase but an ideal magnitude

x (t ) 9 5 cos(2t ) 3 cos(2 2t ) cos(2 3t )

y (t ) 9 5 cos(2t ) 3 cos(2 2t ) cos(2 3t )

Filter has FLAT

Point of this Example

Are Ideal Filters Realizable? (i.e., can we actually MAKE one?)

So the response to a delta (applied at t = 0) is: y (t ) ( / ) sinc ( / )(t td )

Plotting Frequency Response of Practical Filters

Decibel: a logarithmic unit of measure for a ratio between two powers

Decibel Power Rules

-10 dB is P ratio of 0.1

-20 dB is P ratio of 0.01

-30 dB is P ratio of 0.001

Another Rule to Know!!

But |H()| relates Voltages (or current) not POWER!!!

Output Voltage Amplitude

Decibel value for |H(0)|

In addition to using decibels for the |H()| it is also common to use a

We may be just as interested in 0 1 kHz as we are in 1 10 kHz

Simplest Real-World Lowpass Filter: RC Circuit

y(t) Output Signal

1. Convert capacitor into impedance: Z c ( )

Small impedance at high

3. Now analyze the circuit as if it were a DC circuit with a complex voltage

Now find the output phasor

Now we can plot this

Although these are correct plots we usually prefer to use:

Using log scale allows us to see

Instead we can plot this using dB and log axes

You might also like