THE LINEAR SYSTEM

Definition, Characteristics, Analysis Methods

DEFINITION: A linear system is a system whose output is proportional to its input in the following
manner: 1. The total output is the sum of all the component effects. That is, if a cause c1 has an effect e1 and if another cause c2 has an effect e2, then when both causes act on the system, the effect will be e1 + e2. This is called additivity. If c1 e1 and c2 e2 then c1 + c2 e1 + e2 . 2. If a cause increases k-fold then the effect also increases k-fold. This is called homogeneity. If c e then for all k, kc ke . The combination of the additivity and homogeneity properties is called superposition and can be expressed as:

k1c1 + k2 c2 k1e1 + k2 e2

CHARACTERISTICS OF THE LINEAR EQUATION

The closer the input f(t) is to the characteristic mode, the stronger is the system response (output). This is the resonance phenomenon. A systems time constant is equal to the width of its impulse response h(t) and represents the time lag before a system fully responds to an input (rise time). A large time constant indicates a sluggish system.

Linear Equation:
Output Input

Q ( D) y (t ) = P (D) f (t )

Stability is determined by the systems roots. If the real part of all roots is negative, the system is said to be (asymptotically) stable. That is, in the absence of input the output will tend toward zero. If there are unrepeated roots with the real part equal to zero and no roots having positive real parts then the system is said to be marginally stable. That is, it tends toward some non-zero value or bounded oscillatory mode. If there are repeated roots with the real part equal to zero and/or any roots having positive real values then the system is unstable. That is, output grows to plus/minus infinity.

Polynomial from which the characteristic equation, modes and roots are derived.

Polynomials of the D (differential) operator.

Tom Penick

[email protected]

www.teicontrols.com/notes 05/09/98

TIME-DOMAIN METHOD OF ANALYSIS Zero-Input + Zero-State Total Response = c e

Zero-Input response The zero-input response is the component of the system output that results from the effect of the initial conditions only (zero input). Given the linear equation
ZERO-INPUT n jt j j =1 ZERO-STATE

+ f ( t ) h( t )

system output

result of initial conditions with the input held to zero, composed only of characteristic mode terms

zero-input zero-state Total Response = response + response

Q ( D ) y (t ) = P ( D ) f ( t )
if the input f (t ) equals zero, then the LHS Q ( D ) y (t ) must also equal zero. So to this to be true we now have Q ( D ) = 0 . Q ( D ) is a function involving the D-operator (differentiation).

result of system input with initial conditions at zero, composed of both characteristic and non-characteristic mode terms

calculate the zero-input response y 0 (t ) we have: Q ( D ) y 0 (t ) = 0 . Since Q ( D ) must equal 0 for Characteristic Equation: Q ( ) = 0 where: Q ( ) is a polynomial function formed by substituting for the D-operator in the function Q ( D ) , which is found in the linear equation.

By substituting for the D-operator in the polynomial function Q(D) and setting the result equal to zero, we obtain the characteristic equation Q() = 0. Solving for we get the characteristic roots of the linear equation. From the roots, the characteristic modes are derived in the form et. where is a root. In the case where there are repeated roots, that is more than one with the same value, unique characteristic modes are created in the form et, tet, t2et, etc. If we have a pair of complex characteristic roots, in the form j, the characteristic modes may be written in two ways. Complex Roots: + j , j
two possible forms:

The solution is formed by setting y0(t) with constants c and equal to the sum of the characteristic modes, each multiplied by a constant. The values of these constants may be determined if initial conditions are known.

c1e ( + j )t , c2 e ( j ) t t 2) ce cos( t + )
1)

with constants c1 and c2

Tom Penick

[email protected]

www.teicontrols.com/notes 05/09/98

Zero-State response The zero-state response is the component of the system output that results from the effect of the input only (zero initial conditions). It is found by convolving the input f (t ) with the unit impulse response h (t ) . (see Convolution.pdf)

Zero-State Response:

f ( t ) h (t ) =
where:

f ( ) h ( t ) d

f (t ) is the system input h (t ) is the unit impulse response, i.e. the output of the
system in response to a unit impulse at the input

If the system is causal, the integration may be taken from 0 to .

Unit Impulse Response: The unit impulse response h (t ) is the output of a system when the input is the unit impulse "function" (t ) . When the unit impulse response is known, the output can be found for any input function. (see ImpulseResponse.pdf for a sample problem)

h (t ) = bn (t ) + [ P( D ) y 0 (t )] u (t ) , m n (most situations) h (t ) = [ P( D ) y0 (t )] u (t ) , n < m (not practical)

where, from the system equation Q( D) y(t ) = P( D) f (t ) : m = the order of the highest order term in the function P(D) n = the order of the highest order term in the function Q(D) bn = the coefficient of the nth order term in the function P(D). If the term is absent then bn = 0. (t) = the unit impulse function y0(t) = the zero-input response 0, t < 0 (the unit function) u (t ) = 1, t 0

The unit impulse function (not a true function) represents a single high magnitude pulse occurring at t = 0. The width of the function approaches zero and the area under the curve is equal to one. It is defined only at t = 0.

Unit Impulse Function: (t ) = 0, t 0

(t ) dt = 1

Tom Penick

[email protected]

www.teicontrols.com/notes 05/09/98

THE CLASSICAL METHOD OF ANALYSIS Natural + Forced Response Q ( D )[ y n (t ) + y f (t )] = P( D ) f (t ) ,

with Q ( D ) y n (t ) = 0 (see Classical.pdf) Natural response yn(t) The natural response consists of all of the characteristic mode terms found in the total response. The natural response has the same characteristic modes as the zero-input response but the arbitrary constants are different. The natural response is the reaction of the circuit without excitation. Forced response yf(t) The forced response consists of all non-characteristic mode terms found in the total response. The forced response is the reaction of the circuit to external excitation.
system output sum of the characteristic modes

natural forced Total Response = response + response

composed of only those zero-state response terms which are not characteristic modes

LINEARITY IN CIRCUIT ELEMENTS

Mapping Inputs to Outputs

THE CAPACITOR

and similarly, THE INDUCTOR

i( t ) = C

dv ( t ) dt

(v to i)

The map that takes v to i is linear. f(v) = i, with

v (t )

i (t )

v( t ) = L

di( t ) dt di( t ) dt

(i to v)

i (t ) v (t )
+

f ( v )( t ) = C

dv( t ) dt
(i to v)

The map that takes i to v is linear. f(i) = v, with

f ( i )( t ) = L i( t ) = i( 0 ) +

The map of i to v is not linear unless v(0) = 0.

The map of v to i is not linear unless i(0) = 0.

1 t v ( t ) = v( 0 ) + i( )d C 0

1 t v ( )d L 0

(v to i)

Tom Penick

[email protected]

www.teicontrols.com/notes 05/09/98