Sie sind auf Seite 1von 29

# The Islamia University of Bahawalpur

## Department of Electronic Engineering

LAB MANUAL SIGNAL AND SYSTEMS EE-311 5thSemester

LAB EXPERIMENT # 06
Fourier Series in MATLAB

## Lab Instructor Signatures: Date:

OBJECTIVE:

Fourier Series

E.g.Compute and Plot the trigonometric & exponential Fourier spectra for the periodic signal

Code:

x = exp(-t/2);

## theta_n(1) = angle(D_n(1)); theta_n(2:M+1) = angle(D_n(2:M+1));

1
subplot (2, 2, 3); stem(n,C_n,'k');

Output:

Code:

## t=linspace (-pi/4,2*pi+pi/4, 1000); % Time vector exceeds one

period.

x = inline
(['mod(t,2*pi)/A.*(mod(t,2*pi)<A)+','((mod(t,2*pi)>=A)&(mod(t,2*pi)<
pi))'],'t','A');

## for n = 1:100, % Compute N remaining terms

2
D_n=1/(2*pi*n)*((exp(-j*n*A)-1)/(n*A) + j*exp(-j*n*pi));

end

## x_N = cumsum (sumterms);

A = pi/2;

Z=x(t,A);

plot(t,x_N(21,:),'k',t,Z,'k:');

## xlabel ('t'); ylabel ('Amplitude'); legend ('x 20(t)','x(t)',0);

Output:

Increasing N to 100,improves the approximation but does not reduce the overshoot.

## >>xlabel('t'); ylabel('Amplitude'); legend('x_100(t)','x(t)',0);

Output

3

1.

(a) Plot magnitude and phase graph of the following given sequence.

(b) Plot waveform by its magnitude and phase spectrum (Synthesis equation)

2.

(a) Plot magnitude and phase graph of the following given sequence.

(b) Plot waveform by its magnitude and phase spectrum (Synthesis equation)

4
Student Name: shaher yar 15ES10

Instructions
Print this grading sheet, write your name and roll number at the top, and give it to the
Instructor/lab Engineer during your lab check off. Include this as the cover page to

Check Off
During your in-lab check off, be prepared to show the Instructor/Lab Engineer the following:
Fourier Synthesis And Fourier Series Implementation on sine, cose, square and guassian
waves signals

Worksheet Score

## _______ (of 5) Discussion Topics / Q & A Instructor/Lab Engineer

5
LAB EXPERIMENT # 07
Implementation of Fourier Series And Fourier Synthesis in MATLAB

## Student Name : Shaher Yar Roll No:15ES10

OBJECTIVE:

Fourier Synthesis And Fourier Series Implementation on sine, cosine, square and Gaussian
waves signals

## Evaluating Fourier Transforms with MATLAB

In class we study the analytic approach for determining the Fourier transform of a continuous

time signal. In this tutorial numerical methods are used for finding the Fourier transform of

## has a Fourier transform:

X( jf ) = 4sinc(4π f )

This can be found using the Table of Fourier Transforms. We can use MATLAB to plot this

transform. MATLAB has a built-in sinc function. However, the definition of the MATLAB
sinc

function is slightly different than the one used in class and on the Fourier transform table. In

6
MATLAB:

sinc(x) =sin(π x)

Thus, in MATLAB we write the transform, X, using sinc(4f), since the π factor is built in to
the

function. The following MATLAB commands will plot this Fourier Transform:

>> f=-5:.01:5;

>> X=4*sinc(4*f);

>> plot(f,X)

In this case, the Fourier transform is a purely real function. Thus, we can plot it as shown
above.

In general, Fourier transforms are complex functions and we need to plot the amplitude and

phase spectrum separately. This can be done using the following commands:

>> plot(f,abs(X))

7
>> plot(f,angle(X))

Note that the angle is either zero or π. This reflects the positive and negative values of the

transform function.

## Performing the Fourier Integral Numerically

For the pulse presented above, the Fourier transform can be found easily using the table.

However, for some functions, an integration will need to be performed to find the transform

using:

## or, for this example:

X( jf ) = 1e− j 2π ftdt

8
This integration can be performed using the trapz command in MATLAB. This command has

the form: trapz(x,y) where the integral of the function y is found with respect to the variable
of

## integration, x. Or, using MATLAB:

>> clear

>> f=0;

>> t=-2:.01:2;

>> trapz(t,exp(-j*2*pi*f*t))

ans =

4.0000

This is consistent with our earlier results. The value of the transform at f = 0 was found to be
4

using the transform from the table. In order to find the complete transform over a range of

>> clear

>> t=-2:.01:2;

>> k=0;

## >> for f=-5:.01:5

k=k+1;

X(k)=trapz(t,exp(-j*2*pi*f*t));

end

>> f=-5:.01:5;

>> plot(f,X)

9
This should match the transform plotted above. This technique can also be used to
approximate

the Fourier transform of an infinite duration signal. Say we want to find the amplitude
spectrum

## x(t ) = cos(2π100t ) + cos(2π 500t )

We begin by creating a vector, x, with sampled values of the continuous time function. If we

want to sample the signal every 0.0002 seconds and create a sequence of length 250, this will

cover a time interval of length 250*0.0002 = 0.05 seconds. A plot of this signal is generated

## using the following MATLAB code:

>> clear

>> N=250;

>> ts=.0002;

>> t=[0:N-1]*ts;

>> x=cos(2*pi*100*t)+cos(2*pi*500*t);

>> plot(t,x)

10
We can find approximate the Fourier transform integral for 0 ≤ f ≤ 800 Hz using:

>> k=0;

## >> for f=0:1:800

k=k+1;

X(k)=trapz(t,x.*exp(-j*2*pi*f*t));

end

>> f=0:800;

## >> plot(f, abs(X))

As expected the peaks in the spectrum are at 100 and 500 Hz the two frequencies contained
in

the signal. Theoretically, we expect to see impulse functions at these two frequencies and
zero at

11
every other frequency. This is not what we observe. This is because the MATLAB code only

approximates the transform. There are three elements that make the results approximate. The

sequence used to compute the transform is a sampled version of a continuous signal. The
value

of the sampling interval (Ts) and number of samples taken (N) affect the approximation.

Additionally, the trapz command is using a summation to approximate the integral. We can

observe the effects of the sampling interval and number of samples by changing their values.

For example, if we increase the sampling interval by a factor of 10 (to .002) and decrease the

number of samples to 25 (so that we still cover the same range total time interval), and repeat
the

## MATLAB code above we get:

In this case the approximation is less accurate. If we return the sampling interval to the
original

value of .0002 and increase the number of samples to 500, the approximation to the transform
isfound to be:

12
Note that the transform is more accurate than the original. This is expected because we are

included more cycles of the waveform in the approximation (increasing the limits of
integration).

## The Discrete Fourier Transform (DFT)

An alternative to using the approximation to the Fourier transform is to use the Discrete
Fourier

Transform (DFT). The DFT takes a discrete signal in the time domain and transforms that
signal

into its discrete frequency domain representation. This transform is generally the one used in

DSP systems. The theory behind the DFT is covered in ECE 351. In that course you will find

that the DFT of a signal can be used to approximate the continuous time Fourier transform.

## The Fast Fourier Transform (FFT)

Depending on the length of the sequence being transformed with the DFT the computation of

this transform can be time consuming. The Fast Fourier Transform (FFT) is an algorithm for

computing the DFT of a sequence in a more efficient manner. MATLAB provides a built in

command for computing the FFT of a sequence. In this section we will discuss the use of the

FFT to approximate the Fourier transform of signals. Recall that the DFT and FFT are
discrete

sequences

## will be obtained by sampling continuous time signals.

13
In general, if a continuous time function, x(t), is sampled every Ts seconds until N samples
are

collected, the DFT/FFT of this sequence of length N is also of length N. The components of
the

resulting transform correspond to frequencies spaced every 1/(N*Ts) Hz. For example, using
the

same two-frequency signal x(t) used above we can produce a sequence of samples of length
N=

## 250 spaced every Ts = .0002 seconds as shown previously.

However, this time we will find the amplitude spectrum of this signal using the fft command.

The resulting transform will contain N = 250 values. Since the frequency components are
spaced

## every 1/(N*Ts) Hz these correspond to frequency values from 0 to (N-1)/(N*Ts) Hz as shown

below.

>> clear

>> N=250;

>> ts=.0002;

>> deltaf=1/(N*ts);

>> t=[0:N-1]*ts;

>> x=cos(2*pi*100*t)+cos(2*pi*500*t);

>> Xf=fft(x);

>> f=[0:N-1]*deltaf;

>> plot(f,abs(Xf))

14
Note that the spectrum shows four components. Two are at the expected frequencies of 100
and

500 Hz. The other two are at 4500 and 4900 Hz, frequencies that do not appear in the signal.

This is due to the periodic nature of the DFT. Only frequencies up to 0.5/Ts correspond to the

actual frequencies in the Fourier transform. We can produce the spectrum plot that only
shows

these frequencies and shows the negative frequency components by applying the fftshift
function

as shown.

>> Xf_shift=fftshift(Xf);

## >> plot([-N/2:N/2-1]*deltaf, abs(Xf_shift))

15
The values of the sequence length, N, and the time sampling interval, Ts, will have an effect
on

the accuracy of the spectrum that is calculated. First, we will increase the spacing between

samples by a factor of two, or Ts = 0.0004. From the sampling theorem we know that the
slower

we sample, the lower the frequency that we can accurately represent. Repeating the FFT:

>> clear

>> N=250;

>> ts=.0004;

>> deltaf=1/(N*ts);

>> t=[0:N-1]*ts;

>> x=cos(2*pi*100*t)+cos(2*pi*500*t);

>> Xf=fft(x);

>> Xf_shift=fftshift(Xf);

## >> plot([-N/2:N/2-1]*deltaf, abs(Xf_shift))

We observe that the maximum frequency is now 1250 Hz, instead of 2500 Hz as in the
original

computation and the frequency components of 100 and 500 Hz are still represented correctly.
In

general, the maximum frequency represented is given by 0.5/Ts. Say we increase the
sampling

interval to Ts = 0.00125

>> clear

>> N=250;

>> ts=.00125;

>> deltaf=1/(N*ts);

>> t=[0:N-1]*ts;

>> x=cos(2*pi*100*t)+cos(2*pi*500*t);

>> Xf=fft(x);

16
>> Xf_shift=fftshift(Xf);

## >> plot([-N/2:N/2-1]*deltaf, abs(Xf_shift))

It appears as if the signal, x, has frequency content at 100 Hz (correct) and 300 Hz
(incorrect).

This incorrect component is due to the aliasing effect and the fact that the signal has been

sampled at too low of a frequency. The sampling frequency (1/Ts) always needs to be at least
two

times the highest frequency component in the signal being transformed, or in our example at

## least 2*500 Hz = 1000Hz or Ts < 1/1000 = .001.

Next we will observe the effect of N, the number of samples taken. With Ts = .0004 seconds
we

repeat the computation of the FFT with N = 100, 200, and 400 samples. The results are
shown

below.

Note that the effect of larger N is to increase the resolution of graph. This is due to the fact
that

the frequency spacing is given by 1/NTs, or in these three cases 25 Hz, 12.5 Hz, and 6.25 Hz,

respectively. For a given sampling interval, Ts, as N is increased, the length of time that the

continuous time signal is sampled increased (NTs). Thus, we would expect that the resulting

## Guidelines for Using the fft Command

In general, when the fft command is used to produce the amplitude and/or phase spectrum of
a

continuous time signal values for N and Ts must be selected. The following guidelines should
be

1. Select Ts as large as possible but so that the highest frequency component in your signal

## is less than 1/2Ts.

2. After determining the value of Ts, select N so that 1/NTs, the frequency resolution is

## small enough to accurately display your frequency components

17

Plot sine wave, its phase spectrum and its phase angle.
Code
Fs = 150;
t = 0:1/Fs:1;
f = 5;
x = sin(2*pi*t*f);
nfft = 1024;
X = fft(x,nfft);
X = X(1:nfft/2);
mx = abs(X);
f = (0:nfft/2-1)*Fs/nfft;
figure(1);
plot(t,x);
theta_X = angle(X);
figure(2);
plot(t,angle(x));
title('Sine Wave phase');
xlabel('Time (s)');
ylabel('phase');
figure(3);
plot(f,mx);
title('power spectrum of a Sine Wave');
xlabel('Frequency (Hz)');
ylabel('power');

Output

18
19

Plot cosine wave, its phase spectrum and its phase angle.
Code
%%
Fs = 150;
t = 0:1/Fs:1;
f = 5;
x=cos(2*pi*t*f);
nfft = 1024;
X = fft(x,nfft);
X = X(1:nfft/2);
mx = abs(X);
f = (0:nfft/2-1)*Fs/nfft;
figure(1);
plot(t,x);
theta_X= angle (X)
title('cose Wave Signal');
xlabel('Time (s)');
ylabel('Amplitude');
figure(2);
plot(t,angle(x));
title('ANGLE Spectrum of a cose Wave');
xlabel('Frequency (Hz)');
ylabel('Power');
figure(3);
plot(f,mx);
title('power spectrum of a cose Wave');
xlabel('Frequency (Hz)');
ylabel('power');

Output

20
21

Plot cosine wave with shift, its phase spectrum and its phase angle.

Code
Fs = 150;
t = 0:1/Fs:1;
f = 5;
pha = 1/3*pi;
x = cos(2*pi*t*f + pha);
nfft = 1512;
X = fft(x,nfft);
X = X(1:nfft/2);
mx = abs(X);
f = (0:nfft/2-1)*Fs/nfft;
figure(1);
plot(t,x);
theta_X= angle (X)
title('shifted cose Wave Signal');
xlabel('Time (s)');
ylabel('Amplitude');
figure(2);
plot(t,angle(x));
title('ANGLE Spectrum of a cose Wave');
xlabel('Time (s)');
ylabel('phase');
figure(3);
plot(f,mx);
title('power spectrum of a cose Wave');
xlabel('Frequency (Hz)');
ylabel('power');

Output

22
23

Plot square wave, its phase spectrum and its phase angle.

Code
Fs = 150;
t = 0:1/Fs:1;
f = 5;
x = square(2*pi*t*f);
nfft = 1524;
X = fft(x,nfft);
X = X(1:nfft/2);
mx = abs(X);
f = (0:nfft/2-1)*Fs/nfft;
figure(1);
plot(t,x);
theta_X= angle (X)
title('SQUARE Wave Signal');
xlabel('Time (s)');
ylabel('Amplitude');
figure(2);
plot(t,angle(x));
title('ANGLE Spectrum of a SQUARE Wave');
xlabel('time (s)');
ylabel('phase');
figure(3);
plot(f,mx);
title('power spectrum of a SQUARE Wave');
xlabel('Frequency (Hz)');
ylabel('power');

Output

24
25

Plot square pulse, its phase spectrum and its phase angle.

Code
Fs = 150;
t = -0.5:1/Fs:0.5;
w = .2;
x = rectpuls(t,w);
nfft = 1024;
X = fft(x,nfft);
X = X(1:nfft/2);
mx = abs(X);
f = (0:nfft/2-1)*Fs/nfft;
figure(1);
plot(t,x);
theta_X = angle(X);
title('SQUARE PULSE Signal');
xlabel('Time (s)');
ylabel('Amplitude');
figure(2);
plot(t,angle(x));
title('PHASE OF square pulse signal');
xlabel('Time (s)');
ylabel('phase');
figure(3);
plot(f,mx);
title('power spectrum of a square pulse');
xlabel('Frequency (Hz)');
ylabel('power');
Output

26
27

Plot square pulse, its phase spectrum and its phase angle.

Code
Fs = 60;
t = -.5:1/Fs:.5;
x = 1/(sqrt(2*pi*0.01))*(exp(-t.^2/(2*0.01)));
nfft = 1512;
X = fft(x,nfft);
X = X(1:nfft/2);
mx = abs(X);
f = (0:nfft/2-1)*Fs/nfft;
figure(1);
plot(t,x);
theta_X = angle(X);
title('Gaussian PULSE Signal');
xlabel('Time (s)');
ylabel('Amplitude');
figure(2);
plot(t,angle(x));
title('Gaussian pulse signal phase');
xlabel('Time (s)');
ylabel('phase');
figure(3);
plot(f,mx);
title('power spectrum of a guassian pulse signal');
xlabel('Frequency (Hz)');
ylabel('power');

Output

28
29