HW 06: Difference between revisions

From Class Wiki
Jump to navigation Jump to search
Fonggr (talk | contribs)
No edit summary
Fonggr (talk | contribs)
 
(5 intermediate revisions by the same user not shown)
Line 2: Line 2:
Figure out why <math>\int_{0}^{\infty} \cos(2\pi\,f\,u)\,du</math> seems to equal an imaginary odd function of frequency, but there is no j.
Figure out why <math>\int_{0}^{\infty} \cos(2\pi\,f\,u)\,du</math> seems to equal an imaginary odd function of frequency, but there is no j.


==Solution==
==Background==
This is the incorrect solution derived in class. Cosine is incorrect, because a real odd function of time, <math>\sgn(t)\,\!</math>,should map to an imaginary odd function of frequency.
 
===Proof===
{| border="0" cellpadding="0" cellspacing="0"
|-
|<math>F[o(t)]\,\!</math>
|<math>=\int_{-\infty}^{\infty}\,o(t)\,e^{-j\,2\,\pi\,f\,t}\,dt</math>
|-
|
|<math>=\int_{-\infty}^{\infty}\,o(t)\,\left[\cos(2\,\pi\,f\,t)+j\,\sin(2\,\pi\,f\,t)\right]\,dt</math>
|Euler's identity
|-
|
|<math>=\int_{-\infty}^{\infty}\,o(t)\,j\,\sin(2\,\pi\,f\,t)\,dt</math>
|Even function integrates out over symmetric limits
|-
|
|<math>=\int_{-\infty}^{\infty}\,\left[\mbox{Im }e(t) \mbox{ and an Im }o(f)\right]\,dt</math>
|-
|
|<math>=\mbox{Im }o(f)\,\!</math>
|Time integrates out
|}
*The odd function of time has no component (ie. 0) of frequency. Thus it is an even function in frequency.
 
===Functions===
*Even*Even=Even
*Odd*Odd=Even
*Odd*Even=Odd
 
==Incorrect Solution derived in class==
 
{| border="0" cellpadding="0" cellspacing="0"
|-
|<math>F\left[\frac{\sgn (t)}{2}\right]</math>
|<math>=\int_{-\infty}^{\infty} \frac{\sgn (t)}{2} e^{-j\,2\,\pi\,f\,t}\,dt</math>
|-
|
|<math>=\frac{1}{2}\left[\int_{-\infty}^{0}  -1\cdot e^{-j\,2\,\pi\,f\,t}\,dt+\int_{0}^{\infty} 1\cdot e^{-j\,2\,\pi\,f\,t}\,dt\right]</math>
|-
|
|<math>=\underbrace{\frac{1}{2}\int_{0}^{-\infty} e^{j\,2\,\pi\,f\,u}\,du}_{\begin{matrix}u=-t \\ du=-dt\end{matrix}}+\underbrace{\frac{1}{2}\int_{0}^{\infty} e^{-j\,2\,\pi\,f\,u}\,du}_{\begin{matrix}u=t \\ du=dt\end{matrix}}</math>
|-
|
|<math>=\int_{0}^{-\infty} \frac{e^{j\,2\,\pi\,f\,u} + e^{-j\,2\,\pi\,f\,u}}{2}\,du</math>
|-
|
|<math>=\int_{0}^{-\infty} \cos(2\,\pi\,f\,u)\,du</math>
|}
 
==Correct Solution==
{| border="0" cellpadding="0" cellspacing="0"
{| border="0" cellpadding="0" cellspacing="0"
|-
|-

Latest revision as of 16:38, 3 December 2008

Problem

Figure out why ∫0∞cos⁡(2πfu)du seems to equal an imaginary odd function of frequency, but there is no j.

Background

This is the incorrect solution derived in class. Cosine is incorrect, because a real odd function of time, sgn⁡(t),should map to an imaginary odd function of frequency.

Proof

F[o(t)] =∫−∞∞o(t)e−j2πftdt
=∫−∞∞o(t)[cos⁡(2πft)+jsin⁡(2πft)]dt Euler's identity
=∫−∞∞o(t)jsin⁡(2πft)dt Even function integrates out over symmetric limits
=∫−∞∞[Im e(t) and an Im o(f)]dt
=Im o(f) Time integrates out
  • The odd function of time has no component (ie. 0) of frequency. Thus it is an even function in frequency.

Functions

  • Even*Even=Even
  • Odd*Odd=Even
  • Odd*Even=Odd

Incorrect Solution derived in class

F[sgn⁡(t)2] =∫−∞∞sgn⁡(t)2e−j2πftdt
=12[∫−∞0−1⋅e−j2πftdt+∫0∞1⋅e−j2πftdt]
=12∫0−∞ej2πfudu⏟u=−tdu=−dt+12∫0∞e−j2πfudu⏟u=tdu=dt
=∫0−∞ej2πfu+e−j2πfu2du
=∫0−∞cos⁡(2πfu)du

Correct Solution

F[sgn⁡(t)2] =∫−∞∞sgn⁡(t)2e−j2πftdt
=12[∫−∞0−1⋅e−j2πftdt+∫0∞1⋅e−j2πftdt]
=12[∫0−∞1⋅e−j2πftdt+∫0∞1⋅e−j2πftdt]
=12∫0∞−ej2πfudu⏟u=−tdu=−dt+12∫0∞e−j2πfudu⏟u=tdu=dt
=∫0∞−ej2πfu+e−j2πfu2du
=∫0∞−jej2πfu−e−j2πfu2jdu
=∫0∞−jsin⁡(2πfu)du
≠∫0∞cos⁡(2πfu)du