new_s_spect_07_09

Upload: bhieestudents

Post on 07-Apr-2018

216 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/3/2019 New_S_Spect_07_09

    1/89

    1

    Spectral Analysis

    Lecture #7-9

    0 2000 4000 6000 8000 10000 1 2000 1 4000-4

    -3

    -2

    -1

    0

    1

    2

    3

    4

    Time m sec

    M a g n

    i t u

    d e

    i n

    V o

    l t s

  • 8/3/2019 New_S_Spect_07_09

    2/89

    2

    Table 1 Summary of FT PropertiesIntegration in time domain

    )f (G)t(g

    Diff. in freq. domain

    Diff. in time domain ) 2 ( )( j f d

    g t f d

    Gt

    2 ( ) ( )d

    G f d

    j g tf

    t

  • 8/3/2019 New_S_Spect_07_09

    3/89

    3

    Proof

    ( )( )g t G f

    Diff. in time domain) 2 ( )( j f

    d

    g t f d Gt

    Duality ( )d G f df

    2 ( ) ( ) j t g t

    Diff. in freq. domain

    ( )2 ( ) j t g t d G f df

  • 8/3/2019 New_S_Spect_07_09

    4/89

    4

    Examples

    Find the Fourier transform (FT) of -atg(t) = t e u(t)

    Solution

    -ate u(t) 1a + j 2 f

    Diff. in freq. domain

    ( )2 ( ) j t g t d G f df

    2 ( )at j t e u t ( ) 12d a j f df

    +

    ***

  • 8/3/2019 New_S_Spect_07_09

    5/89

    5

    ( )att e u t ( )212a j f +

    ***

  • 8/3/2019 New_S_Spect_07_09

    6/89

    6

    Example

    Find the Fourier transform (FT) of 2 -atg(t) = t e u(t)

    ( )att e u t ( )2

    1

    1 2 j f +Diff. in freq. domain

    ( )2 ( ) j t g t d G f df

    Solution

  • 8/3/2019 New_S_Spect_07_09

    7/89

    7

    2 ( )att e u t

    ( )3

    2

    1 2 j f +

    ***

  • 8/3/2019 New_S_Spect_07_09

    8/89

    8

    Table 1 Summary of FT Properties

    )f (G)t(g

    Area under G(f)

    Area under g(t) ( ) (0)g t d Gt

    =

    ( ) (0)G f d gf

    =

  • 8/3/2019 New_S_Spect_07_09

    9/89

    9

    Area under g(t)

    2( )) ( j f tg tG f e dt

    =

    ( )G f df

    Substituting f=0

    (0)G = ( )g t dt

    Area under G(f)2( )( ) j f tG f dt e tg

    = Substituting t=0

    (0)g =

  • 8/3/2019 New_S_Spect_07_09

    10/89

    10

    Example

    Find

    tA tri

    ( )2sinA c f Solution

    ( )2sin 4c f df

    ( )2

    sin 4c f 14 4

    ttri

    ( )( 0)G f d f g

    =

    ( )2sin 4c f df

    = 14

    =

    ***

  • 8/3/2019 New_S_Spect_07_09

    11/89

    11

    Example

    Find

    ( )2222

    a t aa

    ef

    +

    Solution2

    11

    dtt

    +

    a=1

    ( )22

    1 2 t+f e

    ***

  • 8/3/2019 New_S_Spect_07_09

    12/89

    12

    ( )22

    1 2 t +f

    e

    ( ) 1 f Ga a

    g at

    4- Time scaling

    12

    a

    =

    ( )22

    1 t

    +

    ( )21

    1 t

    +

    ( ) (0)g t d Gt

    = 211 dtt

    =+ (0)G =

    ***

  • 8/3/2019 New_S_Spect_07_09

    13/89

    13

    Table 1 Summary of FT Properties

    )f (G)t(g

    Conjugate function ( ) ** ( )g G f t =

    Proof 2( )) ( j f tg tG f e dt

    = Taking conjugate

    ** 2( )) ( j f tG f g t e dt

    = 2* *( ) ( ) j f tG g t dtf e

    = ( ) ** ( )g G f t =

  • 8/3/2019 New_S_Spect_07_09

    14/89

    14

    Table 1 Summary of FT Properties

    )f (G)t(g

    Real function ( ) * ( )G f G f =

    Real & magnitude parts of G(f) have evensymmetry

    Imajinary & phase parts of G(f) have oddsymmetry

    ( ) ** ( )g G f t =

  • 8/3/2019 New_S_Spect_07_09

    15/89

    15

    Table 1 Summary of FT Properties

    )( ()g t G f

    Even function ( ) ( )G f G f =

    Proof

    Scaling in time domain

    ( ) 1 f Ga a

    g at

    a = -1

    ( ) 11 1

    f Gg t

    ( ) ( )g Gt f

    )( ()g t G f

  • 8/3/2019 New_S_Spect_07_09

    16/89

    16

    Table 1 Summary of FT Properties

    is real & eveng(t) is real & even ( )G f

    )( ()g t G f

    ( ) ( )G f G f = ( )( )g t g t= Even functionReal function ( ) * ( )G f G f = ( )* ( )G f G f =

    Real & even function

    ( )*

    ( )G f G f =

    G(f) is Real & even

    Proof

  • 8/3/2019 New_S_Spect_07_09

    17/89

    17

    Table 1 Summary of FT Properties

    is pure imaginary &

    odd

    g(t) is real & odd ( )G f

    )( ()g t G f

    ( ) ( )G f G f = ( )( )g t g t= odd functionReal function ( ) * ( )G f G f =

    Real & odd function ( ) * ( )G f G f =

    G(f) is pure imaginary & odd

    Proof

    ( )* ( )G f G f =

  • 8/3/2019 New_S_Spect_07_09

    18/89

    18

    Table 1 Summary of FT Properties

    Refer to Table 2 (FT pairs) and identify :

    a) Real & even signals: Verify that G(f) isreal & even.

    b) Real & odd signals: Verify that G(f) ispure imaginary & odd.

  • 8/3/2019 New_S_Spect_07_09

    19/89

    19

    Table 2 Summary of FT Pairs

    G (f)g (t)

    ( )sinA c f t

    A rect t

    A tri

    ( )2sinA c f

    ( )ate u t12a j f +

    ( )ate u t 12a j f

  • 8/3/2019 New_S_Spect_07_09

    20/89

    20

    Table 2 Summary of FT Pairs

    1

    1

    G (f)g (t)

    a t

    e

    ( )22

    2

    2

    a

    a f +( )t

    ( )f

    ( )0cos 2 f t ( ) ( )0 012

    f f f f + +

    ( )0sin 2 f t ( ) ( )0 012

    f f f f j

    +

  • 8/3/2019 New_S_Spect_07_09

    21/89

    21

    Table 2 Summary of FT Pairs

    G (f)g (t)1

    j f ( )sgn t

    ( )u t ( )1 12 2

    f j f

    +

    2te

    1t ( )sgn f

    2f e

  • 8/3/2019 New_S_Spect_07_09

    22/89

    22

    Table 2 Summary of FT Pairs

    Ex) Find the FT of the Gaussian pulse

    ( )2tg t e

    =

    Solution

    ( )2tg t e

    = ( )G f Property of Diff. in time domain

    ( ) 2 ( )d j f dt

    f t Gg

    22 ( )t

    dd

    f e jt

    f G

  • 8/3/2019 New_S_Spect_07_09

    23/89

    23

    Sol.

    Property of Diff. in freq. domain

    ( ) ( )2 j t g t d G f df

    22 ( )td

    df e j

    tf G

    ( )2

    2 ( )2t

    e t j f G f

    2)2 (t j t

    dG f e

    df

    2

    (2 2 )t j t e f G f

    * -j

  • 8/3/2019 New_S_Spect_07_09

    24/89

    24

    Sol.

    2)2 (t j t d G f e

    df

    2 (2 2 )t j t e f G f

    ( )

    d

    G f df =

    2 ( )f G f

    ( )G f =2f e

    2 2f te e

  • 8/3/2019 New_S_Spect_07_09

    25/89

    25

    Table 1 Summary of FT Properties

    11( () )g Gt f

    Multiplication intime domain

    ( ) ( ) ( ) ( )1 2g t g t g Gt f =

    22 ( () )g Gt f

    1 2( )* ( )G G f d

    =

    ( ) 1 2( ) ( )G f G f G f =

  • 8/3/2019 New_S_Spect_07_09

    26/89

    26

    Table 1 Summary of FT Properties

    11( () )g Gt f

    Convolution intime domain

    ( ) ( ) ( ) ( )1 2g t g t Gg f t=

    22 ( () )g Gt f

    ( ) ( )1 2 1 2( ) ( )g t g t g g t d

    =

    ( ) 1 2( ) ( )G f G f G f =

  • 8/3/2019 New_S_Spect_07_09

    27/89

    27

    Linear Time Invariant (LTI) Systems

    An LTI system is characterized by its impulseresponse (system output when the input

    h(t) is the impulse response of the system .

    LTI Systemx(t) y(t)

    x ).(t)= (t)

    LTI Systemx(t)= (t) y(t)=h(t)

  • 8/3/2019 New_S_Spect_07_09

    28/89

    28

    Linear Time Invariant (LTI) Systems

    Given the impulse response of an LTI system, thesystem response (output y(t) ) to an arbitraryinput x(t) can be calculated by two methods :

    1) Time domain method.2) Frequency domain method.

    LTI Systemx(t) y(t)=?

    LTI Systemx(t)= (t) y(t)=h(t)

  • 8/3/2019 New_S_Spect_07_09

    29/89

    29

    Time Domain method

    LTI Systemx(t)= (t) y(t)=h(t)

    (t) h (t)

    (t- ) h (t- )TI

    ( )x (t- ) ( )x h (t- )LTI

    ( ) ( )-

    x t = x (t- ) d

    ( )y t

    ( ) ( )-

    y t = x h(t- ) d

    ( )=x t h(t)

  • 8/3/2019 New_S_Spect_07_09

    30/89

    30

    Time Domain method

    LTI Systemx(t) y(t)

    ( ) ( )x t )y t = h(t

    ( ) ( ) ( )-

    y t = =x x h(t-t )h d(t)

    ( )=h t x(t)

    ( ) ( ) ( )-

    y t = =h h x(t-t )x d(t)

    Convolution processis cumulative

  • 8/3/2019 New_S_Spect_07_09

    31/89

    31

    Time Domain method

    LTI Systemx(t) y(t)

    ( ) ( ) ( )-

    y t = =x x h(t-t )h d(t)

    1) Plot x( )2) Plot h(- ) inversion of h( )

    3) Plot h(t-) shift h(-

    ) by t

    4) y(t) = area under the product ( )x h(t- )

  • 8/3/2019 New_S_Spect_07_09

    32/89

    32

    Frequency Domain method

    LTI Systemx(t) y(t)

    ( ) ( ) ( )-y t = =x x h(t-t )h d(t)

    H(f) Transfer function of the system.

    ( )Y(f)=X f H(f )

    X (f)

    H (f)

    y(t) 4) Y(f)IFT

    ( ) ( )x3 t) )y t = h(t FT2) h (t) FT

    1) x (t) FT

  • 8/3/2019 New_S_Spect_07_09

    33/89

    33

    Proof

    ( ) ( ) ( )-

    y t =x t h(t)= x h(t- ) d

    ( ) ( )2 j f t

    - y t=f e dY t

    ( ) 2

    -

    j f t

    -

    = ex h(t- ) d dt

    =

    ( )-

    x d =

    2( ) f H f e

    ***

  • 8/3/2019 New_S_Spect_07_09

    34/89

    34

    Proof

    ( ) ( ) 2( )-

    f x dY f = H ef

    ( )= H f

    ( )( ) = H f X f

    ( ) 2-

    j f ex d

  • 8/3/2019 New_S_Spect_07_09

    35/89

    35

    Example

    An LTI system has an impulse response

    Is excited by an input . Find the system

    output y(t) using:a) Multiplication in frequency domain

    b) Convolution in time domain.

    Solution

    234

    th(t) = rect 22

    4t

    x(t) = rect

  • 8/3/2019 New_S_Spect_07_09

    36/89

    36

    b) Frequency Domain method

    ( )Y(f) = X f 3) H(f)

    X (f) =

    t-4y(t)=24 tri 4

    ( ) 82 j f Y(f)= 96 sinc 4f e4)

    4t-2

    1) x (t) = 2 rect

    FT

    4t-2

    2) h (t) = 3 rect

    H (f) =

    ( ) 82 j f = 96 sinc 4f e

    IFT

    FT

    ***

  • 8/3/2019 New_S_Spect_07_09

    37/89

    37

    a) Frequency domain method

    4t-2

    x (t) = 2 rect

    4t-2

    h (t) = 3 rect

    x(t)

    4t

    2

    t

    h(t)3

    4

    4t-4y (t) = 24 tri

    8

    y(t)

    t

    24

    4

  • 8/3/2019 New_S_Spect_07_09

    38/89

    38

    b) Time Domain method

    4t-2

    x (t) = 2 rect

    4t-2

    h (t) = 3 rect

    ( ) ( ) ( )-

    y t = = xx t h h(t- ) d(t)

    x(t)

    4t

    2

    t

    h(t)3

    4

  • 8/3/2019 New_S_Spect_07_09

    39/89

    39

    b) Time Domain method

    h( )3

    4

    ( ) ( )-

    y t = x h(t- ) d

    x( )

    4

    2

    h(- )3

    -4

    tt-4

    h(t- )3

  • 8/3/2019 New_S_Spect_07_09

    40/89

    40

    Case 1 : No overlapping

    ( ) ( )-y t = x h(t- ) d

    0t

    =

    x( )

    4

    2

    tt-4

    h(t- )3

    ***

  • 8/3/2019 New_S_Spect_07_09

    41/89

    41

    Case 2 : Partial overlapping

    ( ) ( )-

    y t = x h(t- ) d

    0t 4 0t

    40 t

    t

    x( )h(t- )

    6

    =

    x( )

    4

    2

    h(t- )

    tt-4

    3

    ***

  • 8/3/2019 New_S_Spect_07_09

    42/89

    42

    Case 3 : Partial overlapping

    ( ) ( )-

    y t = x h(t- ) d

    4 4t 4t

    =

    84 t

    x( )h(t- )

    t-4 4

    6

    =

    x( )

    4

    2

    tt-4

    3

    ***

  • 8/3/2019 New_S_Spect_07_09

    43/89

    43

    Case 4 : No overlapping

    ( ) ( )-

    y t = x h(t- ) d

    h(t- )

    tt-4

    3

    4 4t

    8t

    =

    x( )

    4

    2

    ***

  • 8/3/2019 New_S_Spect_07_09

    44/89

    44

    b) Time domain method

    ( )

    0 0t

    y t =

    y(t)

    4

    t

    24

    6 0 4t t

    48 6 4 8t t 0 8t

    424 4

    ttri

    =

    8

  • 8/3/2019 New_S_Spect_07_09

    45/89

    45

    Notes on convolution process

    Start of y =start of x

    + start of h

    =0+0 = 0

    End of y = end of x

    + end of h

    =4+4 = 8

    x()

    4

    2

    t

    h(t)3

    4y(t)

    84t

    24

  • 8/3/2019 New_S_Spect_07_09

    46/89

    4646

    impulse response

    of an LTI system

    LTI Systemx(t)= (t) y (t)=h(t)

    h (t) impulse response

    LTI Systemx(t) y(t)=x(t) h(t)

    = (t) h(t)

    (t) h(t)=h(t) h(t) for all

    A h l i

  • 8/3/2019 New_S_Spect_07_09

    47/89

    4747

    Another solution

    x(t)

    4t

    2

    t

    h(t)3

    4

    dx(t)=dt

    t

    dx(t)dt2

    04

    -2

    dx(t) h(t)=

    dt

    t6

    4

    dx(t) h(t)

    dt

    -6

    8

    ***

    A th l ti

  • 8/3/2019 New_S_Spect_07_09

    48/89

    4848

    Another solution

    y(t) x(t) h(t)=

    t

    -

    d= x( ) h( ) ddt

    t

    6

    4

    dx(t) h(t)dt

    -6

    8

    y(t)

    4

    t

    24

    8

    ***

    l

  • 8/3/2019 New_S_Spect_07_09

    49/89

    49

    Example

    An LTI system has an impulse response

    Is excited by an input . Find the system

    output y(t) using:a) Convolution in time domain.

    b) Multiplication in frequency domain

    Solution

    234

    th(t) = rect 32

    6t

    x(t) = rect

    ) Ti D i h d

  • 8/3/2019 New_S_Spect_07_09

    50/89

    5050

    x(t)

    6t

    2d x(t)=dt

    2

    06 t

    dx(t)dt

    -2

    dx(t) h(t)=

    dt

    t

    h(t)3

    4

    t6

    6

    dx(t) h(t)

    dt

    -6

    104

    a) Time Domain method ***

  • 8/3/2019 New_S_Spect_07_09

    51/89

    ) Ti D i th d ( th l ti )

  • 8/3/2019 New_S_Spect_07_09

    52/89

    52

    a) Time Domain method (another solution)

    6t-3

    x (t) = 2 rect

    4

    t-2h (t) = 3 rect

    t

    x(t)2

    6

    ( ) ( ) ( )-

    y t = = xx t h h(t- ) d(t)

    4t

    h(t)3

    b) Time Domain method

  • 8/3/2019 New_S_Spect_07_09

    53/89

    5353

    b) Time Domain method

    ( ) ( )-

    y t = x h(t- ) d

    h( )

    4

    3 h(- )

    -4

    3

    tt-4

    h(t- )3

    x( )2

    6

  • 8/3/2019 New_S_Spect_07_09

    54/89

    5454

    Case 1 : No overlapping

    ( ) ( )-

    y t = x h(t- ) d

    0t

    =

    x( )2

    6

    tt-4

    h(t- )3

    ***

  • 8/3/2019 New_S_Spect_07_09

    55/89

    5555

    Case 2 : Partial overlapping

    ( ) ( )-

    y t = x h(t- ) d

    0t 4 0t

    40 t

    x( )h(t- )

    t

    6

    x( )2

    6

    tt-4

    h(t- )3

    ***

    =

  • 8/3/2019 New_S_Spect_07_09

    56/89

    5656

    Case 3 : complete overlapping

    ( ) ( )-y t = x h(t- ) d

    4 0t 6t

    64 t

    x( )h(t- )

    t

    6

    t-4

    x( )2

    6

    tt-4

    h(t- )3

    ***

    =

    =

  • 8/3/2019 New_S_Spect_07_09

    57/89

    5757

    Case 4 : Partial overlapping

    ( ) ( )-

    y t = x h(t- ) d

    4 6t 6t

    106 t

    6

    x( )h(t- )

    6

    t-4

    x( )2

    6

    tt-4

    h(t- )3

    ***

    =

    =

    Case 5 : No overlapping

  • 8/3/2019 New_S_Spect_07_09

    58/89

    5858

    Case 5 : No overlapping

    ( ) ( )-

    y t = x h(t- ) d

    4 6t

    10t

    x( )2

    6

    t

    x( )h(t- )

    6

    t-4

    ***

    =

    a) Time domain method

  • 8/3/2019 New_S_Spect_07_09

    59/89

    59

    a) Time domain method

    ( )

    0 0t

    y t =

    6 0 4t t

    60 6 6 10t t 0 10t

    y(t)

    10t

    24

    24 4 6t

    64

    No Overlapping

    No Overlapping

    Partial Overlapping

    Partial Overlapping

    Complete Overlapping

    a) Time domain method

  • 8/3/2019 New_S_Spect_07_09

    60/89

    60

    b) Show that

    5( ) 30

    5t

    y t tri =

    56

    1t

    tri

    a) Time domain method

    ( ) ( )2 2 10( ) 150 sin 5 6 sin j f Y f c f c f e =

    y(t)

    10t

    24

    64

    30

    Example

  • 8/3/2019 New_S_Spect_07_09

    61/89

    6161

    Example

    An LTI system has an impulse response h(t) shown Isexcited by an input x(t) shown. Find the system output y(t) .

    Solution21

    t

    h(t)

    1

    1

    t

    x(t)1

    -1

    2

    -1

    Solution x(t)***

  • 8/3/2019 New_S_Spect_07_09

    62/89

    6262

    Solution

    t

    dx(t)

    dt1

    -1

    2

    1

    d x(t)=dtd

    x(t) h(t)=dt

    1t

    ( )1

    -12

    -1

    1

    -2

    21t

    h(t)1

    dx(t) h(t)

    dt

    t1

    12

    3

    -2

    4

    Solutiond

    ***

  • 8/3/2019 New_S_Spect_07_09

    63/89

    6363

    y(t) x(t) h(t)=

    dx(t) h(t)

    dt

    21

    t1

    3 4

    -2

    t

    -

    d= x( ) h( ) d

    dt

    y(t)

    21t

    1

    3 4

    -1

    Example

  • 8/3/2019 New_S_Spect_07_09

    64/89

    6464

    p

    An LTI system has an impulse response h(t) shown. ItIs excited by an input x(t) shown. Find an expression for

    the system output in the interval .

    Solution

    2 t

    h(t)

    3

    2 t

    x(t)4

    4

    2 t 4

    ( )

  • 8/3/2019 New_S_Spect_07_09

    65/89

    6565

    ( ) ( )-

    y t = x h(t- ) d

    2

    x( )4

    4

    -2

    h(- )

    32

    h( )

    3

    ***

  • 8/3/2019 New_S_Spect_07_09

    66/89

    6666

    Complete overlapping

    ( ) ( )-

    y t = x h(t- ) d

    42 t

    2

    x( )4

    4

    2

    2

    2 * 3t

    = d

    t

    h(t- )3

    t-2( )

    2

    8 2 * 3t

    + d

    =

    ***

  • 8/3/2019 New_S_Spect_07_09

    67/89

    6767

    26= t +36 t-36

    ( )y t =

    =

    =

    Example***

  • 8/3/2019 New_S_Spect_07_09

    68/89

    68

    An LTI system has an impulse responseIs excited by an input . Find:a) The system transfer function (H(f))

    b) The system output y(t).

    Solution

    ( )- t

    h(t) = e tu( )- tx(t) = e tu

    ) ( ) ( )F

    tT

    a h t e u t

    = ( ) =H f

    ) ( ) ( )F

    tT

    b x t e u t

    = ( ) =X f

    Solution***

  • 8/3/2019 New_S_Spect_07_09

    69/89

    69

    ( ) =H f ( ) =X f

    ( ) ( )( )Y f X f H f =

    ( ) ( )1

    2 2 j f j f =

    + +

    ( ) ( )2 2C D

    j f j f = +

    + +

    ( )=y t

    Partial fraction

    Solution1

    ***

  • 8/3/2019 New_S_Spect_07_09

    70/89

    70

    ( ) ( )1

    ( ) 2 2Y f j f j f = + + ( ) ( )2 2C D

    j f j f = ++ +

    C =( )

    1

    =

    D =

    ( )

    1

    =

    ( ) ( )1 t ty t e e u t

    =

    Example

  • 8/3/2019 New_S_Spect_07_09

    71/89

    71

    An LTI system has an impulse responseIs excited by an input . Find: the systemoutput y(t) using time domain analysis.

    Solution

    ( )- t

    h(t) = e tu( )- tx(t) = e tu

    x( )

  • 8/3/2019 New_S_Spect_07_09

    72/89

    72

    ( ) ( )-

    y t = x h(t- ) d

    x( )1 - e

    h( )

    1 - e

    h(- )1

    0

    h(t- )1

    t

    ( )- t e

    Case 1 : No overlapping

  • 8/3/2019 New_S_Spect_07_09

    73/89

    73

    ( ) ( )-y t = x h(t- ) d

    x()1 - e

    Case 1 : No overlapping

    =

    0t

    h(t- )1

    t

    ( )- t e

    Case 1 : Partial overlapping

    ***

  • 8/3/2019 New_S_Spect_07_09

    74/89

    74

    ( ) ( )-

    y t = x dh(t- )

    x()1 - e

    pp g

    t

    h(t- )1( )- t e

    0t

    ( )

    0

    tt

    = ee d

    =

    Case 1 : Partial overlapping

    ***

  • 8/3/2019 New_S_Spect_07_09

    75/89

    75

    x()1 - e

    pp g

    t

    h(t- )1( )- t e

    0t

    ( )( )

    0

    tt

    y t = e e d

    ( )y t =

    ( )1 t t= e e

  • 8/3/2019 New_S_Spect_07_09

    76/89

    76

    ( )( )

    0

    10t t

    0 t

    y t =e e t

    ( ) ( )1 t ty t e e u t

    =

    System PropertiesMemory & memoryless systems

  • 8/3/2019 New_S_Spect_07_09

    77/89

    77

    Memory & memoryless systems

    A system is memoryless if its output at agiven time depends only at the input at thistime.

    Resistive circuits are memoryless circuits.

    RC circuits are memory circuits.

    Memory & memoryless systems

  • 8/3/2019 New_S_Spect_07_09

    78/89

    78

    Examples of memoryless systems:

    The simple identity system y(t)=x(t)The amplifier system y(t) = A x(t)

    Examples of memory systems:

    The delay system y(t)=x(t-t 0)

    The accumulator :

    ( )( )t

    y t x d

    =

    System Causality

  • 8/3/2019 New_S_Spect_07_09

    79/89

    79

    Causal system: Output at any time depends onlyon the input at present and last time (not futuretime).

    Called non-predictive system

    y(t) = x(t)+ x(t-1) is a causal system

    y(t) = x(t)+ x(t+1) is a non- causal system

    All memoryless system are causal systems.

    System Causality***

  • 8/3/2019 New_S_Spect_07_09

    80/89

    80

    Find whether the following systems are causal or not

    1) y(t) = x(-t)

    2) y(t) = x(t) cos (t+1)

    System Stability

  • 8/3/2019 New_S_Spect_07_09

    81/89

    81

    Stable system: Bounded input bounded output

    Find whether the following system is stable or not :

    ( ) ( )t

    y t x d

    = The integrator system

    System Stability***

  • 8/3/2019 New_S_Spect_07_09

    82/89

    82

    3) The integrator system

    is a non-stable system

    Let x(t)=u(t)

    ( ) ( )t

    y t x d

    = ( )x

    1

    t

    ( )y t( ) =y t

  • 8/3/2019 New_S_Spect_07_09

    83/89

    Memory & memoryless systems

  • 8/3/2019 New_S_Spect_07_09

    84/89

    8484

    A system is memoryless if its output at agiven time depends only at the input at this time.

    An LTI system is memoryless if h(t)= k (t)

    y(t) = k x(t)

    LTI Systemx(t)= (t) y (t)=h(t)

    ( ) ( ) ( ) ( ) ( )y(t)= x t h t t h t h t = =

    LTI Systemx(t) y (t)=x(t) ( ) ( )h t t=

    LTI Systemx(t) y (t)=k x(t) ( ) ( )h t k t=

    System Causality

  • 8/3/2019 New_S_Spect_07_09

    85/89

    8585

    A DT system is causal if its output at any timedepends only on the input at present and lasttime (not future time).

    For an LTI system y(t) = x(t) h(t)

    If h( ) exists at any ve value of , y(t) will dependon future input.

    An LTI system is causal if, h(t) =0 t < 0

    ( )-

    x(t= h - ) d

    System Stability

  • 8/3/2019 New_S_Spect_07_09

    86/89

    8686

    Stability Bounded input Bounded output

    It can be proved that, an LTI system is stable if

    h(t) is absolutely summable .

    Ex) The following are the impulse response of a CTLTI system. Determine whether each system is

    causal and/or stable or not. Justify your answer.

    ( )h t d t

    <

    4) ( ) ( 2)ta h t e u t= 4) ( ) ( 1 )tb h t e u t

    =

    Solution4 t h(t)

    ***

  • 8/3/2019 New_S_Spect_07_09

    87/89

    8787

    ) ( ) ( 2)= a h t e u tt

    2

    ( )

    = h t d t 42

    te d t

    4

    24

    =

    te 8

    4

    =

    e e8

    4

    = e =bounded

    Causality

    h(t)=0 for all t < 0Stability

    Solutionh(t)4t

    ***

  • 8/3/2019 New_S_Spect_07_09

    88/89

    8888

    ( )

    = h t d t

    Unstable

    Causality

    Stability

    t-1

    ) ( ) ( 1 )b h t e u t=

  • 8/3/2019 New_S_Spect_07_09

    89/89

    89

    Conclusions