EECE 301 Signals & Systems Prof.

Mark Fowler
Discussion #10
Laplace Transform Examples

1/19

Examples of Solving Differential Equations using LT

Notice how easy this is! -LT converts the differential equation into an algebraic equation -We can easily solve an algebraic equation for an output Y(s) -We can do partial fraction expansion -We can use the LT table But, this is where the hard part lies although it is easy for certain inputs.

2/19

Example 6.29: 2nd Order Differential Equation

d 2 y (t ) dy (t ) Given +6 + 8 y (t ) = 2 x(t ) 2 dt dt
Assume that the system has ICs given by:

& y (0 ) = 1 and y (0 ) = 2
Find the output y(t) for t 0 when the input is x(t) = u(t)

x(t )
1

Input starts at t=0

3/19

Applying the LT to this D.E. Gives:

d 2 y (t ) dy (t ) +6 + 8 y (t ) = 2 x(t ) 2 dt dt

& [s Y ( s) y (0 )s y(0 )]+ 6[sY ( s) y (0 )] + 8Y ( s)

2

= 2 X ( s)

Solving for Y(s) algebraically gives:

& y ( 0 ) s + y ( 0 ) + 6 y (0 ) 2 Y ( s) = + 2 X (s) 2 s + 6s + 8 s + 6s + 8
Using the specific ICs gives:

2 s +8 Y (s) = 2 + s 2 + 6s + 8 X ( s) s + 6s + 8
IC part H(s) Transfer Function (notice how this comes directly from the D.E.)
4/19

Notice how the IC part can be easily found using PFE and an LT table (For linear, constant coefficient differential equations it will always be like that!)

2 s +8 Y (s) = 2 + 2 s + 6s + 8 X ( s) s + 6s + 8
IC part H(s) Transfer Function (notice how this comes directly from the D.E.)

Notice that the input part may not be easy to do PFE/ILT if the input X(s) is complicated.

But in control systems and sometimes in electronics we are often interested in how the system responds to a step function Called the step response

Note : x (t ) = u(t ) X ( s ) =

1 s
5/19

 2 s+8 + 2 Y ( s) = 2 s + 6s + 8 s( s + 6 s + 8)
IC part First compare to this:
Ae nt sin n 1 2 t + u(t ) where : A=

Input part

[(

) ]
(
n

Is 0< || < 1?
2

n
2

+1

n 1 2 = tan n
1

s + 2 2 s + 2n s + n

= 8 =1
And identify:
2 n = 8 n = 2 2

2n = 6 = 6 / 2n = 6 / 4 2 = 1.06

No!!! So factor!!
6/19

Factored!!!

2 2 s+8 s+8 Y ( s) = 2 + s( s 2 + 6s + 8) = ( s + 4)( s + 2) + s( s + 4)( s + 2) s + 6s + 8

IC part Doing PFE:
>> [R,P,K]=residue([1 8],[1 6 8]) R= -2 3 P= -4 -2 K= [] >> [R,P,K]=residue(2,[1 6 8 0]) R= 0.2500 -0.5000 0.2500 P= -4 -2 0 K= []

Input part

IC part

Input part

2 3 0.25 0.5 0.25 Y ( s) = + s + 2 + s + 4 + s + 2 + s s + 4

IC part Transient part SS part
7/19

 2 3 0.25 0.5 0.25 Y ( s) = + s + 2 + s + 4 + s + 2 + s s + 4

IC part Using the LT table: Transient part SS part

e bt u(t ), b real or complex

1 , b real or complex s+b

1 / 4e 0 t = 1 / 4

y (t ) = 2e 4 t + 3e 2 t + 1 / 4e 4 t 1 / 2e 2 t + 1 / 4 , t 0
IC part Transient part SS part Specifying this means we can leave off the u(t)s
8/19

] [

y (t ) = 2e 4 t + 3e 2 t + 1 / 4e 4 t 1 / 2e 2 t + 1 / 4 , t 0
IC part Transient part SS part

] [

Note: IC part & Transient part have the same kind of decaying parts: These come from the systems characteristic polynomial!

e 4t & e 2t

Note: The book combined Y(s) into one big thing and found y(t) from that. Same answerbut we cannot see the impact of three parts!

9/19

Plots for ex. 6.29

Zero state system takes time to respond to step input Zero state transient

Eventually the ICs decay away, the ZS transient dies and the system reaches its steady-state response

10/19

Now Modify Example 6.29: 2nd Order Differential Equation Was

d 2 y (t ) dy (t ) +6 + 8 y (t ) = 2 x(t ) 2 dt dt
Now

2 s+8 Y ( s) = 2 + s( s 2 + 6 s + 8) s + 6s + 8
IC part Input part

d 2 y (t ) dy (t ) + + 8 y (t ) = 2 x (t ) 2 dt dt

2 s+8 Y ( s) = 2 + 2 s + s + 8 s( s + s + 8)
IC part Input part

= 8 =1
And identify:
2 n = 8 n = 2 2

2n = 1 = 1 / 2n = 1 / 4 2 = 0.18
Is 0< || < 1? Yes Complex Roots so use one of the 2nd order LT Pairs

11/19

 2 s+8 + 2 Y ( s) = 2 s + s + 8 s( s + s + 8)

Do PFE

s + 8 0.25 0.125 + j 0.0225 0.125 j 0.0225 = 2 + s + ( s + 0.5 j 2.78) + ( s + 0.5 + j 2.78) s + s + 8 s +1 s + 8 0.25 = 2 + s 0.25 s 2 + s + 8 s + s + 8

Recombine to get back quadratic

Ae nt sin n 1 2 t + u(t ) where : A=

[(

) ]
(
n

n
2

+1

n 1 2 = tan n
1

s + 2 2 s + 2n s + n

12/19

s +1 s + 8 0.25 + 0.25 2 Y ( s) = 2 s + s + 8 s + s + 8 s

LT Table

y (t ) = 2.87e 0.5t sin(2.78t + 0.36) + 0.25 0.25e 0.5t sin( 2.78t + 1.4)

Zero-Input Response

Zero-State Response

Total Response
13/19

Compare the Two Cases:

H ( s) =

2 2 = s 2 + 6s + 8 ( s + 4)( s + 2)
j

H ( s) =

2 2 = s 2 + s + 8 ( s + 0.5 j 2.78)( s + 0.5 + j 2.78)

14/19

Ex. 6.38 RLCLC Circuit In analyzing circuits we often -Have zero ICs -Have no specific input signal in mind We only want to find the transfer function and/or the frequency response

Suppose you must analyze the following circuit to find the current through C1 (You might need to do that because in the physical circuit there is a device being driven by that current and that device has an input impedance that is modeled as a capacitor)
Re-Drawn Version of Fig. 6.13 but equivalent!!

C1

x(t )
C2

device

y (t )

15/19

Now, replace device by its capacitor model and find the transfer function between: & input = voltage x(t) output = current y(t) Ls Use s-domain impedances 1/Cs And analyze the circuit as if X(s) is a DC voltage source. Can find that (see book for details):

This is the whole idea behind the LT approach!!

C1C2 L2 s 3 + C1s H ( s) = C1C2 L1L2 s 4 + RC1C2 L2 s 3 + [L1 (C1 + C2 ) + L2C2 ]s 2 + R(C1 + C2 ) s + 1

Note that there are 6 adjustable coefficients but only 5 adjustable component values Cant achieve all possible coefficient combinations
16/19

A zero at the origin forces low frequencies to be attenuated

A purely-imaginary-roots term puts conjugate zeros on the j axis nulls one specific frequency

(1 / L1 ) s (s 2 + 1 / C2 L2 ) H ( s) = 4 s + ( R / L1 ) s 3 + [L1 (C1 + C2 ) + L2C2 ] / C1C2 L1L2 s 2 + R (C1 + C2 ) / C1C2 L1L2 s + 1 / C1C2 L1L2

17/19

To see what this circuit can do we look at its frequency response. This is valid because all poles are in the region of convergence. (True for any passive RLC circuit) Get H() by replacing s j The following plot shows |H()| for two sets of component values similarities Note but differences!! How do we choose the component values to do what we want?!

STAY TUNED!!!

18/19

Due to a pole near the j axis

Due to zeros on the j axis

1570 Hz 157 Hz

157 kHz
19/19