Difference between revisions of "Aufgaben:Exercise 3.8Z: Convolution of Two Rectangles"

From LNTwww
 
(3 intermediate revisions by the same user not shown)
Line 23: Line 23:
 
''Hints:''  
 
''Hints:''  
 
*This exercise belongs to the chapter  [[Signal_Representation/The_Convolution_Theorem_and_Operation|"The Convolution Theorem and Operation"]].
 
*This exercise belongs to the chapter  [[Signal_Representation/The_Convolution_Theorem_and_Operation|"The Convolution Theorem and Operation"]].
*It mainly refers to the page  [[Signal_Representation/The_Convolution_Theorem_and_Operation#Graphical_Convolution|"Graphical Convolution"]].
+
*It mainly refers to the page  [[Signal_Representation/The_Convolution_Theorem_and_Operation#Graphical_convolution|"Graphical convolution"]].
 
*The topic of this section is also illustrated in the interactive applet  [[Applets:Graphical_Convolution| "Graphical Convolution"]].
 
*The topic of this section is also illustrated in the interactive applet  [[Applets:Graphical_Convolution| "Graphical Convolution"]].
  
Line 48: Line 48:
 
+ The output signal  ${y(t)}$  has a trapezoidal shape.
 
+ The output signal  ${y(t)}$  has a trapezoidal shape.
 
- The spectrum is   ${Y(f)} = Y_0 \cdot \text{si}^{2}(\pi f T)$.
 
- The spectrum is   ${Y(f)} = Y_0 \cdot \text{si}^{2}(\pi f T)$.
+ With  $T = 2 \,\text{ms}$ , a triangular shape would result.
+
+ With  $T = 2 \,\text{ms}$,  a triangular shape would result.
  
  
Line 56: Line 56:
 
===Solution===
 
===Solution===
 
{{ML-Kopf}}
 
{{ML-Kopf}}
[[File:P_ID536__Sig_Z_3_8_a_neu.png|right|frame|To illustrate the graphical convolution $x(t) \star h(t)$]]
+
[[File:P_ID536__Sig_Z_3_8_a_neu.png|right|frame|To illustrate the convolution&nbsp; $x(t) \star h(t)$;<br>the abscissas  have been renamed:&nbsp; $\tau$]]
 
'''(1)'''&nbsp;  In general, the following applies to the convolution integral:
 
'''(1)'''&nbsp;  In general, the following applies to the convolution integral:
 
:$$y(t) = \int_{ - \infty }^{ + \infty } {x( \tau  ) \cdot h( {t - \tau } )}\hspace{0.1cm} {\rm d}\tau.$$
 
:$$y(t) = \int_{ - \infty }^{ + \infty } {x( \tau  ) \cdot h( {t - \tau } )}\hspace{0.1cm} {\rm d}\tau.$$
''Hints:''&nbsp; The abscissas in the graph opposite have been renamed&nbsp; $\tau$&nbsp;.
 
  
The signal value at time&nbsp; $t = 1 \,\text{ms}$&nbsp; ms can be calculated as follows:
+
The signal value at time&nbsp; $t = 1 \,\text{ms}$&nbsp; can be calculated as follows:
*Reflection of the impulse response&nbsp; ${h(\tau)}$,  
+
*Mirroring of the impulse response&nbsp; ${h(\tau)}$,  
*shift by&nbsp; $t = 1 \text{ ms}$&nbsp; to the right (violet curve in the sketch),
+
*shifting by&nbsp; $t = 1 \text{ ms}$&nbsp; to the right (violet curve in the sketch),
 
*multiplication of the two functions and integration.  
 
*multiplication of the two functions and integration.  
  
  
The product is also rectangular with the height&nbsp; $2 \text{ V} \cdot 300 \; \text{1/s}$&nbsp; and width&nbsp; $1 \,\text{ms}$. This results in for the area:
+
The product is also rectangular with the height&nbsp; $2 \text{ V} \cdot 300 \; \text{1/s}$&nbsp; and width&nbsp; $1 \,\text{ms}$. This results for the area:
 
:$$y( {t = 1\;{\rm{ms}}} ) \hspace{0.15 cm}\underline{= 0.6\;{\rm{V}}}{\rm{.}}$$
 
:$$y( {t = 1\;{\rm{ms}}} ) \hspace{0.15 cm}\underline{= 0.6\;{\rm{V}}}{\rm{.}}$$
The green rectangle illustrates the calculation of the second signal value. Now the resulting rectangle is twice as wide after the multiplication and we get:
+
The green rectangle illustrates the calculation of the second signal value.&nbsp; Now the resulting rectangle is twice as wide after the multiplication and we get:
 
:$$y( {t = 2\;{\rm{ms}}} ) = 2\;{\rm{V}} \cdot {\rm{300}}\;{1}/{{\rm{s}}} \cdot 2\;{\rm{ms}}\hspace{0.15 cm}\underline{={\rm{1.2}}\;{\rm{V}}}{\rm{.}}$$
 
:$$y( {t = 2\;{\rm{ms}}} ) = 2\;{\rm{V}} \cdot {\rm{300}}\;{1}/{{\rm{s}}} \cdot 2\;{\rm{ms}}\hspace{0.15 cm}\underline{={\rm{1.2}}\;{\rm{V}}}{\rm{.}}$$
  
  
'''(2)'''&nbsp; Because of the symmetry of ${y(t)}$ with respect to the time&nbsp; $t = 2.5\, \text {ms}$&nbsp; holds:
+
'''(2)'''&nbsp; Because of the symmetry of&nbsp; ${y(t)}$&nbsp; with respect to the time&nbsp; $t = 2.5\, \text {ms}$&nbsp; holds:
 
:$$y( {t = 3\;{\rm{ms}}} ) = y( {t = 2\;{\rm{ms}}} ) \hspace{0.15 cm}\underline{= {\rm{1}}{\rm{.2}}\;{\rm{V}}}{\rm{,}}$$
 
:$$y( {t = 3\;{\rm{ms}}} ) = y( {t = 2\;{\rm{ms}}} ) \hspace{0.15 cm}\underline{= {\rm{1}}{\rm{.2}}\;{\rm{V}}}{\rm{,}}$$
 
:$$y( {t = 4\;{\rm{ms}}} ) = y( {t = 1\;{\rm{ms}}} )\hspace{0.15 cm}\underline{ = 0.6\;{\rm{V}}}{\rm{.}}$$
 
:$$y( {t = 4\;{\rm{ms}}} ) = y( {t = 1\;{\rm{ms}}} )\hspace{0.15 cm}\underline{ = 0.6\;{\rm{V}}}{\rm{.}}$$
  
  
[[File:P_ID537__Sig_Z_3_8_c.png|right|frame|Convolution result&nbsp; $y(t)$]]
+
[[File:P_ID537__Sig_Z_3_8_c.png|right|frame|Overall result&nbsp; $y(t)$]]
'''(3)'''&nbsp;  In subtasks&nbsp; '''(1)'''&nbsp; and&nbsp; '''(2)'''&nbsp; the signal values were calculated at discrete time points.
+
'''(3)'''&nbsp;  Proposed solutions <u>1 and 3</u> are correct:
 +
*In subtasks&nbsp; '''(1)'''&nbsp; and&nbsp; '''(2)'''&nbsp; the signal values were calculated at discrete time points.
 
*All points are to be connected by straight line segments, since the integration over rectangular functions of increasing width results in a linear course.  
 
*All points are to be connected by straight line segments, since the integration over rectangular functions of increasing width results in a linear course.  
 
*This means:&nbsp; The output signal&nbsp; ${y(t)}$&nbsp; is trapezoidal.
 
*This means:&nbsp; The output signal&nbsp; ${y(t)}$&nbsp; is trapezoidal.
Line 85: Line 85:
 
*The associated spectrum is complex and reads:
 
*The associated spectrum is complex and reads:
 
:$$Y(f) = 6 \cdot 10^{ - 3} \;{{\rm{V}}}/{{{\rm{Hz}}}} \cdot {\mathop{\rm si}\nolimits} ( {2\;{\rm{ms}}\cdot{\rm{\pi }}f} ) \cdot {\mathop{\rm si}\nolimits} ( {3\;{\rm{ms}}\cdot{\rm{\pi }}f}) \cdot {\rm{e}}^{ - {\rm{j \hspace{0.05cm}\cdot \hspace{0.05cm} 2 \hspace{0.05cm}\cdot \hspace{0.05cm}2.5\;{\rm{ms}}\hspace{0.05cm}\cdot \hspace{0.05cm} \pi }}f} .$$
 
:$$Y(f) = 6 \cdot 10^{ - 3} \;{{\rm{V}}}/{{{\rm{Hz}}}} \cdot {\mathop{\rm si}\nolimits} ( {2\;{\rm{ms}}\cdot{\rm{\pi }}f} ) \cdot {\mathop{\rm si}\nolimits} ( {3\;{\rm{ms}}\cdot{\rm{\pi }}f}) \cdot {\rm{e}}^{ - {\rm{j \hspace{0.05cm}\cdot \hspace{0.05cm} 2 \hspace{0.05cm}\cdot \hspace{0.05cm}2.5\;{\rm{ms}}\hspace{0.05cm}\cdot \hspace{0.05cm} \pi }}f} .$$
*If the input pulse&nbsp; ${x(t)}$&nbsp; had the duration&nbsp; $T = 2\, \text {ms}$, had the duration&nbsp; ${y(t)}$&nbsp; would show a triangular waveform between&nbsp; ${t = 0}$&nbsp; and&nbsp; $t = 4  \text { ms}$&nbsp;.
+
*If the input pulse&nbsp; ${x(t)}$&nbsp; would have the duration&nbsp; $T = 2\, \text {ms}$, the duration&nbsp; ${y(t)}$&nbsp; would show a triangular waveform between&nbsp; ${t = 0}$&nbsp; and&nbsp; $t = 4  \text { ms}$.&nbsp; The maximum&nbsp; $1.2 \, \text {V}$&nbsp; would then only result at the time&nbsp; $t = 2 \, \text {ms}$.  
*The maximum&nbsp; $1.2 \, \text {V}$&nbsp; would then only result at the time&nbsp; $t = 2 \, \text {ms}$.  
+
 
  
  
Proposed solutions <u>1 and 3</u> are therefore correct.
 
 
{{ML-Fuß}}
 
{{ML-Fuß}}
  

Latest revision as of 13:56, 24 May 2021

The convolution of two rectangles  $x(t)$  and  $h(t)$

At the input of a causal LTI system (i.e. linear and time-invariant)

  • with a rectangular impulse response  ${h(t)}$  of duration  $2 \,\text{ms}$ ,
  • a rectangular pulse  ${x(t)}$  of duration  $T = 3 \,\text{ms}$  and amplitude  $A = 2\,\text{ V}$  is applied. 


The rectangular functions each start at the time  $t = 0$.

In this task you are to calculate the output signal  ${y(t)}$  with the help of the "Graphical Convolution". 

As you can easily check, the output signal  ${y(t)}$

  • differs from zero only in the range between  $0$  and  $5 \, \text{ms}$,  and
  • is symmetrical at the time  $t = 2.5 \, \text{ms}$.



Hints:



Questions

1

Calculate the signal values at the times   $t = 1 \,\text{ms}$  and  $t = 2 \,\text{ms}$.

$y(t = 1 \,\text{ms})\ = \ $

 $\text{V}$
$y(t = 2 \,\text{ms})\ = \ $

 $\text{V}$

2

Determine the signal values for the times  $t = 3 \,\text{ms}$  and  $t = 4 \,\text{ms}$  by exploiting the symmetry properties.

$y(t = 3 \,\text{ms})\ = \ $

 $\text{V}$
$y(t = 4 \,\text{ms})\ = \ $

 $\text{V}$

3

Which of the following statements are true?

The output signal  ${y(t)}$  has a trapezoidal shape.
The spectrum is   ${Y(f)} = Y_0 \cdot \text{si}^{2}(\pi f T)$.
With  $T = 2 \,\text{ms}$,  a triangular shape would result.


Solution

To illustrate the convolution  $x(t) \star h(t)$;
the abscissas have been renamed:  $\tau$

(1)  In general, the following applies to the convolution integral:

$$y(t) = \int_{ - \infty }^{ + \infty } {x( \tau ) \cdot h( {t - \tau } )}\hspace{0.1cm} {\rm d}\tau.$$

The signal value at time  $t = 1 \,\text{ms}$  can be calculated as follows:

  • Mirroring of the impulse response  ${h(\tau)}$,
  • shifting by  $t = 1 \text{ ms}$  to the right (violet curve in the sketch),
  • multiplication of the two functions and integration.


The product is also rectangular with the height  $2 \text{ V} \cdot 300 \; \text{1/s}$  and width  $1 \,\text{ms}$. This results for the area:

$$y( {t = 1\;{\rm{ms}}} ) \hspace{0.15 cm}\underline{= 0.6\;{\rm{V}}}{\rm{.}}$$

The green rectangle illustrates the calculation of the second signal value.  Now the resulting rectangle is twice as wide after the multiplication and we get:

$$y( {t = 2\;{\rm{ms}}} ) = 2\;{\rm{V}} \cdot {\rm{300}}\;{1}/{{\rm{s}}} \cdot 2\;{\rm{ms}}\hspace{0.15 cm}\underline{={\rm{1.2}}\;{\rm{V}}}{\rm{.}}$$


(2)  Because of the symmetry of  ${y(t)}$  with respect to the time  $t = 2.5\, \text {ms}$  holds:

$$y( {t = 3\;{\rm{ms}}} ) = y( {t = 2\;{\rm{ms}}} ) \hspace{0.15 cm}\underline{= {\rm{1}}{\rm{.2}}\;{\rm{V}}}{\rm{,}}$$
$$y( {t = 4\;{\rm{ms}}} ) = y( {t = 1\;{\rm{ms}}} )\hspace{0.15 cm}\underline{ = 0.6\;{\rm{V}}}{\rm{.}}$$


Overall result  $y(t)$

(3)  Proposed solutions 1 and 3 are correct:

  • In subtasks  (1)  and  (2)  the signal values were calculated at discrete time points.
  • All points are to be connected by straight line segments, since the integration over rectangular functions of increasing width results in a linear course.
  • This means:  The output signal  ${y(t)}$  is trapezoidal.
  • The associated spectrum is complex and reads:
$$Y(f) = 6 \cdot 10^{ - 3} \;{{\rm{V}}}/{{{\rm{Hz}}}} \cdot {\mathop{\rm si}\nolimits} ( {2\;{\rm{ms}}\cdot{\rm{\pi }}f} ) \cdot {\mathop{\rm si}\nolimits} ( {3\;{\rm{ms}}\cdot{\rm{\pi }}f}) \cdot {\rm{e}}^{ - {\rm{j \hspace{0.05cm}\cdot \hspace{0.05cm} 2 \hspace{0.05cm}\cdot \hspace{0.05cm}2.5\;{\rm{ms}}\hspace{0.05cm}\cdot \hspace{0.05cm} \pi }}f} .$$
  • If the input pulse  ${x(t)}$  would have the duration  $T = 2\, \text {ms}$, the duration  ${y(t)}$  would show a triangular waveform between  ${t = 0}$  and  $t = 4 \text { ms}$.  The maximum  $1.2 \, \text {V}$  would then only result at the time  $t = 2 \, \text {ms}$.