problem of fourier series

Discussion in 'Homework Help' started by bhuvanesh, Mar 6, 2015.

  1. bhuvanesh

    Thread Starter Member

    Aug 10, 2013
    268
    2
    Choose the function f(t) ; −∞ < < +∞, for which a Fourier series cannot be defined.
    (a) 3 sin(25t) (b) 4 cos(20t + 3) + 2 sin (10t)
    (c) e-|t| sin(25t) (d) 1

    fourier series are available for all periodic signal .its obvious that all the four are periodic.But somewhere i hear that fourier series is not possible for constant .could someone give me hint for the answer ,so try to come up. Thank you
     
    Last edited: Mar 6, 2015
  2. MrAl

    Well-Known Member

    Jun 17, 2014
    2,418
    488
    Hi,

    How is '1' periodic?
    That's example (d).
     
  3. bhuvanesh

    Thread Starter Member

    Aug 10, 2013
    268
    2
    yes 1 is periodic with period infinity (or we can make period as we need and its extending from minus infinity to plus infinity )
     
  4. WBahn

    Moderator

    Mar 31, 2012
    17,715
    4,786
    What are the requirements for the Fourier series to exist?
     
    bhuvanesh likes this.
  5. bhuvanesh

    Thread Starter Member

    Aug 10, 2013
    268
    2
    it should be periodic ,
    it should finite discontinuous over one period
    it should absolutely integrable over any period
    it should not have no more that finite number of maxima and minima over single period
    thats all i think
     
  6. anhnha

    Active Member

    Apr 19, 2012
    773
    45
    bhuvanesh likes this.
  7. WBahn

    Moderator

    Mar 31, 2012
    17,715
    4,786
    Where is this somewhere that you heard this?

    Take the basic equation for x(t) in terms of it's Fourier series components and set all of them equal to zero except Ao (the coefficient of the cosine term at zero frequency) and what do you get?

    But if they are being nitpicky, then ask yourself if a constant is absolutely integrable over one period.
     
    bhuvanesh likes this.
  8. WBahn

    Moderator

    Mar 31, 2012
    17,715
    4,786
    Do you understand what each of these mean?
     
    bhuvanesh likes this.
  9. bhuvanesh

    Thread Starter Member

    Aug 10, 2013
    268
    2
    except this "it should not have no more that finite number of maxima and minima over single period"i know all
     
  10. WBahn

    Moderator

    Mar 31, 2012
    17,715
    4,786
    The simplest example would be a sine wave at an infinite frequency. In lay and imprecise terms, the problem is that such a waveform effectively becomes multivalued.
     
    bhuvanesh likes this.
  11. WBahn

    Moderator

    Mar 31, 2012
    17,715
    4,786
    Okay. So is a constant absolutely integrable over one period?
     
    bhuvanesh likes this.
  12. bhuvanesh

    Thread Starter Member

    Aug 10, 2013
    268
    2
    no, it is not. Integrating constant over one period leads to infintiy. At the same time look my waveform for e-|t| sin(25t). There at infinite time signal amplitude leading infinite so it would not be periodic ,would't be?
     
  13. WBahn

    Moderator

    Mar 31, 2012
    17,715
    4,786
    Sketch that waveform and consider the effect of the dying exponential on the integral. Notice that it is important that the exponent be |t| and not just t. Why?
     
    bhuvanesh likes this.
  14. bhuvanesh

    Thread Starter Member

    Aug 10, 2013
    268
    2
    e-|t| sin(25t).waveform attached
     
  15. WBahn

    Moderator

    Mar 31, 2012
    17,715
    4,786
    Not attached.

    But that's okay.

    Consider a simpler waveform.

    x(t) = e^-|t|

    Is this function absolutely integrable?

    If so, then any function y(t) for which |y(t)| <= |x(t)| must also be absolutely integrable. Why?
     
    bhuvanesh likes this.
  16. bhuvanesh

    Thread Starter Member

    Aug 10, 2013
    268
    2
    yes it integrable over -infinity to infinty .it gives zero

    but it not integrable for x(t) = e^|t|,right?

    if x(t) is integrable ,sure then y(t) is integrable since y(t) is less than or equal to x(t).

    x(t)= e-|t| sin(25t)

    here how would you say |y(t)| <= |x(t).why it may not like |y(t)| => |x(t)?
     
  17. WBahn

    Moderator

    Mar 31, 2012
    17,715
    4,786
     
  18. bhuvanesh

    Thread Starter Member

    Aug 10, 2013
    268
    2
    sorry i was doing wrong calculation
    x(t)= e-|t| sin(25t)
    Because of magnitude we are going to have only e^-x ,where x= positive value for (minus infinity to plus infinty )
    yes,its giving both positive and negative values,example e^ -1 it gives positive value but e^-10 gives negative value,but for any value its giving bounded value ,i mean its not giving infinity so its integrable,right?

    if |x(t)| gives bounded value(not infinity ),then |y(t)| gives less than the bounded values of |x(t)|,right?
     
  19. WBahn

    Moderator

    Mar 31, 2012
    17,715
    4,786
    What is the numerical value of e^-10 ?
     
  20. bhuvanesh

    Thread Starter Member

    Aug 10, 2013
    268
    2
    -7.28171817154
     
Loading...