![Page 1: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/1.jpg)
1
EE 616 Computer Aided Analysis of Electronic Networks
Lecture 12
Instructor: Dr. J. A. Starzyk, ProfessorSchool of EECSOhio UniversityAthens, OH, 45701
Note: materials in this lecture are from the notes of EE219A UC-berkeleyhttp://www- cad.eecs.berkeley.edu/~nardi/EE219A/contents.html
![Page 2: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/2.jpg)
2
Outline Transient Analysis of dynamical circuits
– i.e., circuits containing C and/or L Examples Solution of Ordinary Differential Equations (Initial Value
Problems – IVP)– Forward Euler (FE), Backward Euler (BE) and
Trapezoidal Rule (TR)– Multistep methods– Convergence
Methods for Ordinary Differential Equations
By Prof. Alessandra Nardi
![Page 3: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/3.jpg)
3
Ground Plane
Signal Wire
LogicGate
LogicGate
• Metal Wires carry signals from gate to gate.• How long is the signal delayed?
Wire and ground plane form a capacitor
Wire has resistance
Application Problems
Signal Transmission in an Integrated Circuit
![Page 4: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/4.jpg)
4
capacitor
resistor
• Model wire resistance with resistors.• Model wire-plane capacitance with capacitors.
Constructing the Model• Cut the wire into sections.
Application Problems
Signal Transmission in an IC – Circuit Model
![Page 5: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/5.jpg)
5
Nodal Equations Yields 2x2 System
C1
R2
R1 R3 C2
Constitutive Equations
cc
dvi C
dt
1R Ri v
R
Conservation Laws
1 1 20C R Ri i i
2 3 20C R Ri i i
1
1 2 21 1
2 22
2 3 2
1 1 1
0
0 1 1 1
dvR R RC vdt
C vdv
R R Rdt
1Ri
1Ci
2Ri
2Ci
3Ri
1v 2v
Application Problems
Signal Transmission in an IC – 2x2 example
![Page 6: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/6.jpg)
6eigenvectors
1
1 2 21 1
2 22
2 3 2
1 1 1
0
0 1 1 1
dvR R RC vdt
C vdv
R R Rdt
1 2 1 3 2Let 1, 10, 1C C R R R 1.1 1.0
1.0 1.1
A
dxx
dt
11 1 0.1 0 1 1
1 1 0 2.1 1 1A
Eigenvalues and Eigenvectors
Eigenvalues
Application Problems
Signal Transmission in an IC – 2x2 example
![Page 7: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/7.jpg)
1Change of variab (le ) ( ) ( ) (s ): Ey t x t y t E x t
1
1 2 1 2
10 0
0 0
0 0n n
n
A E E E E E
E
E
Eigendecomposition:
0
( )Substituting: ( ), (0)
dEy tAEy t Ey x
dt
11Multiply by ( ) : ( )dy t
E AE tdt
E y 0 0
10 0
0 0
( )n
y t
0Consider an ODE( )
( ), (0): dx t
Ax t x xdt
An Aside on Eigenanalysis
![Page 8: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/8.jpg)
8
( )Decoupling: ( ) ( ) (0)iti
i i i
dy ty t y t e y
dt
1 0 0
0 0
0 0
From last slide:( ) ( )
n
dy tdt
y t
DecoupledEquations!
1) Determine , E 1 0 0
3) Compute ( ) 0 0 (0)
0 0 n
t
t
e
y t y
e
102) Compute (0) y E x
4) ( ) ( ) x t Ey t
0Steps for solvi( )
( ), (0)ng dx t
Ax t x xdt
An Aside on Eigenanalysis
![Page 9: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/9.jpg)
9
1(0) 1v
2 (0) 0v
Notice two time scale behavior
• v1 and v2 come together quickly (fast eigenmode).• v1 and v2 decay to zero slowly (slow eigenmode).
Application Problems
Signal Transmission in an IC – 2x2 example
![Page 10: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/10.jpg)
10
Circuit Equation Formulation
For dynamical circuits the Sparse Tableau equations can be written compactly:
For sake of simplicity, we shall discuss first order ODEs in the form:
riablescircuit va of vector theis where
)0(
0),,)(
(
0
x
xx
txdt
tdxF
),()(
txfdt
tdx
![Page 11: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/11.jpg)
11
Ordinary Differential Equations
Initial Value Problems (IVP)
Typically analytic solutions are not available
solve it numerically
.condition initial given the intervalan in
)(
),()(
:(IVP) Problem Value Initial Solve
00
00
x,T][t
xtx
txfdt
tdx
![Page 12: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/12.jpg)
12
Not necessarily a solution exists and is unique for:
It turns out that, under rather mild conditions on the continuity and differentiability of F, it can be proven that there exists a unique solution.
Also, for sake of simplicity only consider
linear case:
0),,( tydt
dyF
We shall assume that has a unique solution 0),,( tydt
dyF
00 )(
)()(
xtx
tAxdt
tdx
Ordinary Differential Equations Assumptions and Simplifications
![Page 13: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/13.jpg)
13
First - Discretize Time
Second - Represent x(t) using values at ti
ˆ ( )llx x t
Approx. sol’n
Exact sol’n
Third - Approximate using the discrete ( )ldx t
dtˆ 'slx
1 1
1
ˆ ˆ ˆ ˆExample: ( )
l l l l
ll l
d x x x xx t or
dt t t
Lt T1t 2t 1Lt 0
t t t
1t 2t 3t Lt0
3x̂ 4x̂1x̂
2x̂
Finite Difference Methods
Basic Concepts
![Page 14: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/14.jpg)
14
lt 1lt t
x
1( ) ( )slope l lx t x t
t
slope ( )ldx t
dt
1( ) ( ) ( )l l lx t x t t A x t
1
1
( ) ( )( ) ( )
or
( ) ( ) ( )
l ll l
l l l
x t x tdx t A x t
dt t
x t x t t A x t
Finite Difference Methods
Forward Euler Approximation
![Page 15: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/15.jpg)
15
1t 2t t
x
(0)tAx
11 ˆ( ) (0) 0x t x x tAx
3t
2 1 12 ˆ ˆ ˆ( )x t x x tAx
1ˆtAx
1 1ˆ ˆ ˆ( ) L L LLx t x x tAx
Finite Difference Methods
Forward Euler Algorithm
![Page 16: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/16.jpg)
16
lt 1lt t
x 1slope ( )l
dx t
dt
1( ) ( )slope l lx t x t
t
1 1( ) ( ) ( )l l lx t x t t A x t
11 1
1 1
( ) ( )( ) ( )
or
( ) ( ) ( )
l ll l
l l l
x t x tdx t A x t
dt t
x t x t t A x t
Finite Difference Methods
Backward Euler Approximation
![Page 17: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/17.jpg)
17
1t 2t t
x
1ˆtAx
2ˆtAx
1 11 ˆ ˆ( ) (0)x t x x tAx
Solve with Gaussian Elimination
1ˆ[ ] (0)I tA x x
1 1ˆ ˆ( ) [ ]L LLx t x I tA x
2 1 12 ˆ ˆ( ) [ ]x t x I tA x
Finite Difference Methods
Backward Euler Algorithm
![Page 18: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/18.jpg)
18
1
1
1
1 1
1( ( ) ( ))
21
( ( ) ( ))2
( ) ( )
1( ) ( ) ( ( ) ( ))
2
l l
l l
l l
l l l l
d dx t x t
dt dt
Ax t Ax t
x t x t
t
x t x t tA x t x t
t
x
1( ) ( )slope l lx t x t
t
slope ( )ldx t
dt
1slope ( )l
dx t
dt
1 1
1 1( ( ) ( )) ( ( ) ( ))
2 2l l l lx t tAx t x t tAx t
Finite Difference Methods
Trapezoidal Rule Approximation
![Page 19: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/19.jpg)
19
1t 2t t
x1ˆ (0)
2 2
t tI A x I A x
Solve with Gaussian Elimination
1 11 ˆ ˆ( ) (0) (0)
2
tx t x x Ax Ax
12 1
2
11
ˆ ˆ( )2 2
ˆ ˆ( )2 2
L LL
t tx t x I A I A x
t tx t x I A I A x
Finite Difference Methods
Trapezoidal Rule Algorithm
![Page 20: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/20.jpg)
20
1
1( ) (( ) )) )( (l
l
t
l l t
dx t x t x t A dx t xA
dt
lt 1lt
1
( )l
l
t
tAx d
1( )ltAx t BE
( )ltAx t FE
( ) ( )2 l l
tAx t Ax t
Trap
Finite Difference Methods
Numerical Integration View
![Page 21: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/21.jpg)
21
Finite Difference Methods - Sources of Error
![Page 22: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/22.jpg)
22
Trap Rule, Forward-Euler, Backward-Euler Are all one-step methods
Forward-Euler is simplest No equation solution explicit method. Box approximation to integral
Backward-Euler is more expensive Equation solution each step implicit method
Trapezoidal Rule might be more accurate Equation solution each step implicit method Trapezoidal approximation to integral
1 2 3ˆ ˆ ˆ ˆ is computed using only , not , , etc.l l l lx x x x
Finite Difference Methods
Summary of Basic Concepts
![Page 23: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/23.jpg)
23
( ) ( ( ), ( ))dx t f x t u t
dtNonlinear Differential Equation:
k-Step Multistep Approach: 0 0
ˆ ˆ ,k k
l j l jj j l j
j j
x t f x u t
Solution at discrete points
Time discretization
Multistep coefficients
2ˆ lx
lt1lt 2lt 3lt l kt
ˆ lx1ˆ lx
ˆ l kx
Multistep Methods
Basic Equations
![Page 24: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/24.jpg)
24
Multistep Equation:
1 1 1,l l l lx t x t t f x t u t
FE Discrete Equation: 1 11ˆ ˆ ˆ ,l l l
lx x t f x u t
0 1 0 11, 1, 1, 0, 1k
Forward-Euler Approximation:
Multistep Coefficients:
Multistep Coefficients:
BE Discrete Equation:
0 1 0 11, 1, 1, 1, 0k 1ˆ ˆ ˆ ,l l l
lx x t f x u t
Trap Discrete Equation: 1 11ˆ ˆ ˆ ˆ, ,
2l l l l
l l
tx x f x u t f x u t
0 1 0 1
1 11, 1, 1, ,
2 2k Multistep Coefficients:
0 0
ˆ ˆ ,k k
l j l jj j l j
j j
x t f x u t
Multistep Methods – Common Algorithms
TR, BE, FE are one-step methods
![Page 25: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/25.jpg)
25
Multistep Equation:
01) If 0 the multistep method is implicit 2) A step multistep method uses previous ' and 'k k x s f s
03) A normalization is needed, 1 is common 4) A -step method has 2 1 free coefficientsk k
How does one pick good coefficients?
Want the highest accuracy
0 0
ˆ ˆ ,k k
l j l jj j l j
j j
x t f x u t
Multistep Methods
Definition and Observations
![Page 26: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/26.jpg)
26
Definition: A finite-difference method for solving initial value problems on [0,T] is said to be convergent if given any A and any initial condition
0,
ˆmax 0 as t 0lT
lt
x x l t
exactx
tˆ computed with
2lx
ˆ computed with tlx
Multistep Methods – Convergence Analysis
Convergence Definition
![Page 27: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/27.jpg)
27
Definition: A multi-step method for solving initial value problems on [0,T] is said to be order p convergent if given any A and any initial condition
0,
ˆmaxpl
Tl
t
x x l t C t
0for all less than a given t t
Forward- and Backward-Euler are order 1 convergent
Trapezoidal Rule is order 2 convergent
Multistep Methods – Convergence Analysis
Order-p Convergence
![Page 28: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/28.jpg)
28
Multistep Methods – Convergence Analysis
Two types of error
made.been haserror previous no assuming
,solution theof eexact valu theand ˆ value
computed ebetween th difference theis at method
n integratioan of (LTE)Error Truncation Local The
11
1
)x(tx
t
ll
l
exactly.known iscondition initial only the that assuming
,solution theof eexact valu theand ˆ value
computed ebetween th difference theis at method
n integratioan of (GTE)Error Truncation Global The
11
1
)x(tx
t
ll
l
![Page 29: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/29.jpg)
29
For convergence we need to look at max error over the whole time interval [0,T]– We look at GTE
Not enough to look at LTE, in fact:– As I take smaller and smaller time steps t,
I would like my solution to approach exact
solution better and better over the whole time interval, even though I have to add up LTE
from more time steps.
Multistep Methods – Convergence Analysis
Two conditions for Convergence
![Page 30: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/30.jpg)
30
1) Local Condition: One step errors are small (consistency)
2) Global Condition: The single step errors do not grow too quickly (stability)
Typically verified using Taylor Series
All one-step methods are stable in this sense.
Multistep Methods – Convergence Analysis
Two conditions for Convergence
![Page 31: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/31.jpg)
31
Definition: A one-step method for solving initial value problems on an interval [0,T] is said to be consistent if for any A and any initial condition
1ˆ0 as t 0
x x t
t
One-step Methods – Convergence Analysis
Consistency definition
![Page 32: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/32.jpg)
32
One-step Methods – Convergence Analysis
Consistency for Forward Euler
![Page 33: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/33.jpg)
33
One-step Methods – Convergence Analysis
Convergence Analysis for Forward Euler
Forward-Euler definition
Expanding in t about yields
where is the "one-step" error bounded by
![Page 34: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/34.jpg)
34
One-step Methods – Convergence Analysis
Convergence Analysis for Forward Euler
![Page 35: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/35.jpg)
35
One-step Methods – Convergence Analysis
A helpful bound on difference equations
![Page 36: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/36.jpg)
36
One-step Methods – Convergence Analysis
A helpful bound on difference equations
![Page 37: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/37.jpg)
37
One-step Methods – Convergence Analysis
Back to Convergence Analysis for Forward Euler
![Page 38: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/38.jpg)
38
• Forward-Euler is order 1 convergent• The bound grows exponentially with time interval• C is related to the solution second derivative• The bound grows exponentially fast with norm A.
One-step Methods – Convergence Analysis
Observations about Convergence Analysis for FE
![Page 39: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/39.jpg)
39
Summary
Transient Analysis of dynamical circuits– i.e., circuits containing C and/or L
Examples Solution of Ordinary Differential Equations
(Initial Value Problems – IVP)– Forward Euler (FE), Backward Euler (BE) and
Trapezoidal Rule (TR)– Multistep methods– Convergence
![Page 40: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/40.jpg)
40
Multistep Methods - Local Truncation Error
![Page 41: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/41.jpg)
41
Local Truncation Error (cont’d)
![Page 42: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/42.jpg)
42
Examples
![Page 43: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/43.jpg)
43
Examples (cont’d)
![Page 44: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/44.jpg)
44
Determination of Local Error
![Page 45: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/45.jpg)
45
Implicit Methods
![Page 46: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/46.jpg)
46
Convergence
![Page 47: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/47.jpg)
47
Convergence (cont’d)
![Page 48: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/48.jpg)
48
Convergence (cont’d)
![Page 49: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/49.jpg)
49
Other methods
![Page 50: EE 616 Computer Aided Analysis of Electronic Networks Lecture 12](https://reader036.vdocuments.site/reader036/viewer/2022081520/56814daf550346895dbb04f1/html5/thumbnails/50.jpg)
50
Summary