Part II Problems

Problem 1: [Step and delta responses]

(a) Find the unit impulse response w for the LTI operator 2D
2
+ 4D + 4I.
(b) Find the unit step response v for the same operator.
. .
(c) Verify that v = w (as it should be, since u = ).
(d) For each of the following functions, nd the LTI differential operator p(D) having it as
unit impulse response.
(i) 2u(t).
(ii) u(t)t.
(iii) u(t)t
2
.

PartIIProblemsandSolutions
Problem1: [Step and delta responses]
(a)Find the unit impulse response wfor the LTI operator 2D
2
+4D+4I.
(b)Find the unit step response vfor the same operator.
. .
(c)Verify that v=w(as it should be, since u=).
(d)For each of the following functions, nd the LTI differential operator p(D)having it as
unit impulse response.
(i) 2u(t).
(ii) u(t)t.
(iii) u(t)t
2
.
Solution: (a)Therootsofthecharacteristicpolynomialare1i,sothegeneralsolution
to the homogeneous equation is e
t
(acost+bsint). The unit impulse response for this
secondorderoperatorhasw(0) =0andw
.
(0+)=
1
2
.Therstforcesa=0andthesecond
givesb=
2
1
:w(t) =
1
2
u(t)e
t
sint.
(b)Fort>0,theunitstepresponseisasolutiontop(D)x=1.Inourcase,x
p
=
4
1
issuch
asolution, andthegeneralsolutionisthen x =
4
1
+e
t
(acost+bsint). Werequirerest
initialconditions:0=x(0) =
1
4
+aora=
1
4
. x
.
=e
t
((a+b)cost+ (ab)sint),so
0=x
.
(0) =a+bandb=
4
1
aswell:v=
4
1
u(t)(1e
t
(cost+sint)).
(c)v
.
=
4
1
e
t
((1+1)cost+ (11)sint) =
1
2
e
t
sint.
(d) (i) This function has a jump in value, so the operator must be of rst order. (aD+
bI)(2u) =2a(t) +2bu(t),sob=0anda=
2
1
: p(D) =
1
2
D.
(ii)Thisfunctionhasnojumpbutitsderivativedoes, sotheoperatormustbeofsecond
.. .
order. For t > 0, w(t) = t is the solution to a
2
= 0 with x(0) = 0 and x+a
1
x+a
0
x
x
.
(0)
1 1
t=0
a
2
dt
d
t . Plugin: a
1
+a
0
t = 0implies a
1
= a
0
= 0, and 1= impliesthat = =
a
2
. ..
= 1. So p(D) = D
2
. Oryoucanarguethatw(t) = u(t)and u(t)t,w(t) w(t) = (t), a
2
=
soa
2
(t) =(t)anda
2
=1.
(iii)Thisfunctionw(t)hasnojumpinvalueorderivative,butitssecondderivativedoes
jump: w
..
(t) = 2u(t). So w
(3)
(t) = 2(t). This means that we are looking for a third
order operator, a
3
D
3
+a
2
D
2
+a
1
D+a
0
I. t
2
is a solution to the homogeneous equation,
soa
2
2+a
1
2t+a
0
t
2
= 0,whichimpliesthata
0
= a
1
= a
2
= 0. w
..
(0) = 2impliesthat
a
3
=
2
1
andp(D) =
1
2
D
3
.Oryoucanarguethatw(t) =u(t)t
2
,w
.
(t) =u(t)2t,w
..
(t) =2u(t),
w
(3)
(t) =2(t),soa
3
w
(3)
(t) =(t)impliesthata
3
=
2
1
.
Part I Problems
In the next two problems, nd the unit impulse and the unit step response for
Problem 1: D + kI
Problem 2: D
2
+
0
2
I

PartIProblemsandSolutions
In the next two problems, nd the unit impulse and the unit step response for
Problem1: D+kI
Solution: (t) =u(t)e
kt
v(t) =u(t)
1
1 e
kt
k
Problem2: D
2
+
0
2
I
Solution: (t) =u(t)

1
0
sin (
0
t)
v(t) =u(t)
1
(1 cos(
0
t))

2
0
18.03SC Practice Problems 25
Step and delta responses
1. Find the unit step and unit impulse responses for the operator 2D + I, and graph
them.
2. Find the unit impulse response for the operator D
2
+ 2D, and graph it.
3. From your answer to 2., nd the solution to x + 2x = 3(t 1) with rest initial
conditions.
18.03SC Practice Problems 25
Step and delta responses
Solution suggestions
1. Findtheunitstepandunitimpulseresponsesfortheoperator2D+ I,andgraphthem.
The unit step response h = h(t) is the continuous solution that is zero for t < 0,
and is a solution of
2x + x= 1 for t> 0. (1)
This equation has particular solution x
p
= 1. The homogeneous system 2 x + x= 0
has general solution ce
t/2
, so the general solution of (1) is x= 1 + ce
t/2
.
Because there is no impulse at t = 0 the pre and post-initial conditions are the
same, i.e. x(0

) = x(0
+
) = 0. We need to choose the constant c to t the post-
initial condition: x(0
+
) = 1 + c = 0. Thus, c = 1 and the unit step response h
is

h(t) = 1 e
t/2
u(t).
The unit impulse response w= w(t) is the solution that is zero for x< 0, a solution
of 2 x + x= 0 for x> 0, and satises x(0
+
) = 1/a
1
= 1/2, where a
1
is the coefcient
of x . Or, alternatively, but equivalently, the unit impulse response is the derivative
of the unit step response, so, using the product rule
1

1
w(t) = h

(t) = e
t/2
u(t) + 1 e
t/2
(t) = e
t/2
u(t).
2 2
The term (1 e
t/2
)(t) = 0 because at t= 0 the coefcient of (t) is 0. The graphs
of both are given below.
Figure 1: The unit step response h(t) and unit impulse response w(t) for 2D+ I.
2. FindtheunitimpulseresponsefortheoperatorD
2
+ 2D,andgraphit.
The unit impulse response for this operator is the function w(t) that is zero for t< 0
and satises the equation
x + 2x = 0
18.03SC Practice Problems 25 OCW 18.03SC
for t > 0 with post initial conditions x(0
+
) = 0 and x (0
+
) = 1/1 = 1, since the
operator is of order 2 and has leading coefcient 1.
By examining the characteristic polynomial, we see that homogeneous solutions
have the form c
1
e
2t
+ c
2
.
Now we use the post initial conditions to nd the right constants c
1
and c
2
. From
the condition on the function itself, c
1
+ c
2
= 0, and, from the condition on the rst
derivative, 2c
1
= 1. Thus, c
1
= 1/2, c
2
= c
1
= 1/2, and the unit impulse
response is
w(t) =
1

1 e
2t

u(t).
2
The graph of w(t) is given below.
Figure 2: The unit impulse response w(t) for D
2
+ 2D.
3. From your answer to 2., nd the solution to x + 2x = 3(t 1) with rest initial
conditions.
Using time invariance, we nd that a solution to x + 2x = 3(t 1) is
x= 3w(t 1) =
3

1 e
22t

u(t 1).
2
2
Unit Step and Unit Impulse Response: Introduction
In real life, we often do not know the parameters of a system (e.g. the
spring constant, the mass, and the damping constant, in a spring- mass-
dashpot system). We may not even know the order of the system. For
example, there may be many interconnected springs or diodes. Instead, we
often learn about a system by watching how it responds to various input
signals.
In this session we will study the response of a linear time invariant (LTI)
system from rest initial conditions to two standard and very simple signals:
the unit impulse (t) and the unit step function u(t). Reasonably enough
we will call these responses the unit impulse response and the unit step re-
sponse.
The theory of the convolution integral studied in the next session will
give us a method of dertemining the response of a system to any input once
we know its unit impulse response.
Because both (t) and u(t) are discontinuous at t = 0 we will have
to be careful with our denition of initial conditions. The most sensible
mathematical and physical way to do this is to dene our initial conditions
at 0

. As input an impulse causes a jump when it is applied. This means

that the conditions at 0
+
will be different than those at 0

. To distinguish
these two cases we will use the terms pre-initial conditions (at 0

) and post-
initial conditions (at 0
+
). We will be able to state precisely the effect of a unit
impulse on these conditions.

InitialConditions
1. Introduction
Before we try to solve higher order equations with discontinuous or
impulsiveinputweneedtothinkcarefullyaboutwhathappenstotheso-
lutionatthepointofdiscontinuity.
Recallthatwehavetheleftandrightlimitsofafunctionast 0:
x(0

) =limx(t) and x(0

+
)=limx(t).
t 0 t 0
(Notethatwecandenetheselimitsastgoestoanyvalue a.) Foracon-
tinuousfunctionthesetwolimitsarethesame,andtheyarebothequalto
x(0).
Fortheunitstepfunctionwehave
u(0

) =0, u(0
+
)=1, u(0)isundened.
Fortheunitimpulsefunction(t)wehave
(0

) =0, (0
+
)=0, (0) =.
Inthisunitourdifferentialequationswillalwayshaveinitialconditions
att = 0. Theaboveexamplesshowthatwhenthereisadiscontinuitywe
mightneedtodistinguishbetween0

and0
+
. Assuming xistheoutput,
.
wewilldothisbycalling x(0

), x(0

), . . . thepre-initialconditionsand
.
x(0
+
), x(0
+
), . . .thepost-initialcondition.
Important: Hereafterwhenwejustsayinitialconditionswewillmeanthe
pre-initialconditions. Incaseswherex(t)issmooththepreandpost-initial
conditionsarethesameandtheirisnoneedtodistinguishbetweenthem.
2. SimpleExamples
Example1. Considertheinitialvalueproblem
.
x=u(t), x(0

) =0.
0 fort<0
x(t) =
t fort>0.
Thisisasimplecalculusproblemandhassolution
t
x(t)

..

InitialConditions OCW18.03SC
Itiseasytoseethatx(0
+
)=0,sothepost-initialconditionisthesame
as the pre-initial condition. This should not surprise us. Although the
rate of input jumps from 0 to 1, it is still only inputting an innitesimal
amount at a time. So, the response x(t) should be continuous. But, note
. .
thatx(0

) =0=x(0
+
)=1.
Example2. Considertheinitialvalueproblem
.
x=(t), x(0

) =0.
Weknowhowtointegrate(t)togetx(t)=u(t).
t
1
x = u(t)
Herethepre-initialconditionx(0

) =0doesnotmatchthepost-initial
conditionx(0
+
)=1.Theimpulsecausesajumpinthevalueofx.
Example3. ConsiderasecondorderIVP
.. .
x=u(t), x(0

) =0, x(0

) =0.
0 fort<0
Integratingtwiceweget x(t) =
t
2
/2 fort>0.
t
1
x(t)
. .
Again,itseasytocheckthatx(0

) = x(0
+
)andx(0

) = x(0
+
). That
..
is, the pre and post initial conditions are the same. (But, x(0

) = 0 =
x(0
+
)=1.)
Example4. Considertheinitialvalueproblem
.. .
x=(t), x(0

) =0, x(0

) =0.
.
Integratingoncegives
0 fort<0
x(t) =
t fort>0.
x(t)=u(t).Integratingasecondtimegives
t
x(t)
Checkingthepreandpostinitialconditionsgives
x(0

) = 0 = x(0
+
)
. .
x(0

) = 0 = x(0
+
)=1
2
InitialConditions OCW18.03SC
Inotherwords,x(t)itselfiscontinuous,butforthesecondorderequation
theinput(t)causedajumpintherstderivative.
Ifwecontinuedtheseexampleswedndthatforann
th
-orderequation
aninputof(t)causesajumpinthederivativeofordern1.
3. RestInitialConditions
Thecasewherex(t) =0fort<0iscalledrestinitialconditions.Ifwe
haveaDEofordernthistranslatesintopre-initialconditions
x(0

) =0, x
.
(0

) =0, . . . , x
(n1)
(0

) =0.
4. Conclusion
Aunitstepinputu(t)causesasmoothresponsewithmatchingpreand
post-initialconditions.Foraunitimpulseinput(t)thepreandpostinitial
conditions match except for the derivative one less than the order of the
equation.
3



First order Unit Step Response
1. Unit Step Response
Considertheinitialvalueproblem
.
x + kx = ru(t), x(0

) = 0, k,r constants.
Thiswouldmodel,forexample,theamountofuraniuminanuclearreactor
where we add uranium at the constant rate of r kg/year starting at time
t = 0andwherek isthedecayrateoftheuranium.
Asinthepreviousnote,addinganinnitesimalamount(r dt)atatime
leadstoacontinuousresponse. Wehavex(t) = 0fort < 0;andfort > 0
wemustsolve
.
x + kx = r, x(0) = 0.
Thegeneralsolutionisx(t) = (r/k) + ce
kt
.Tondc,weusex(0) = 0:
r r
0= x(0) =
k
+ c c =
k
.
Thus,inbothcasesandu-format
x(t) =
0
k
r
(1 e
kt
)
fort < 0
=
r
(1 e
kt
)u(t). (1)
fort > 0 k
Withr = 1,thisistheunit step response,sometimeswrittenv(t). To
bemoreprecise,wecouldwritev(t) = u(t)(1/k)(1 e
kt
).
Theclaimthatwegetacontinuousresponseistrue,butmayfeelabit
unjustied. Lets redo the above example very carefully without making
thisassumption.Naturally,wewillgetthesameanswer.
Theequationis
. 0 fort < 0
x + kx =
r fort > 0,
x(0

) = 0. (2)
Solvingthetwopiecesweget
c
1
e
kt
fort < 0
x(t) =
k
r
+ c
2
e
kt
fort > 0.
Thisgivesx(0

) = c
1
andx(0
+
) = r/k + c
2
.Ifthesetwoaredifferentthere
isajumpatt = 0ofmagnitude
x(0
+
) x(0

) = r/k + c
2
c
1
.


FirstorderUnitStepResponse OCW18.03SC
The initial condition x(0

) = 0 implies c
1
= 0, so our solution looks
like

0 fort < 0
x(t) =
k
r
+ c
2
e
kt
fort > 0.
Tondc
2
wesubstitutethisintoourdifferentialequation(2).(Wemustuse
thegeneralizedderivativeifthereisajumpatt = 0.)Aftersubstitutionthe
leftsideof(2)becomes
. 0 fort < 0
x + kx = (r/k + c
2
)(t) +
kc
2
e
kt
+ r + kc
2
e
kt
fort > 0
0 fort < 0
= (r/k + c
2
)(t) +
r fort > 0.
Comparingthiswiththerightsideof(2)weseethatr/k + c
2
= 0,orc
2
=
r/k.Thisgivesexactlythesamesolution(1)wehadbefore.
Figure1showsthegraphoftheunitstepresponse(r = 1). Noticethat
itstartsat0andgoesasymptoticallyupto1/k.
t
1/k
v(t)
.
Figure1.Unitstepistheresponseofthesystemx + kx = f (t) when f (t) = u(t).
The Meaning of the Phrase Unit Step Response
Inthisnotelookedatthesystemwithequation
.
x + kx = f (t)
and we considered f (t) to be the input. As we have noted previously, it
sometimes makes more sense to consider something else to be the input.
Forexample,inNewtonslawofcooling
.
T + kT = kT
e
itmakesphysicalsensetocallT
e
,thetemperatureoftheenvironment,the
input.Inthiscasetheunitstepresponseofthesystemmeanstheresponse
totheinput T
e
(t) = u(t),i.e.thesolutionto
.
T + kT = ku(t).
2
UnitStepResponse:Post-initialConditions
Quiz: Considertheequation
.
v+kv=u(t)
withrestinitialconditions,v(0

) =0.
.
Forthesolutionv(t)whatisv(0
+
)?
Choices:
.
a) v(0
+
)=0
.
b) v(0
+
)=1/k
.
c) v(0
+
)=1
.
d) v(0
+
)=k
e) Noneofthese.
Answer: (c)
v(t)iscontinuoussov(0

) = v(0
+
) = v(0) = 0ThereforetheDEshows
.
v(0
+
)=u(0
+
)=1.
UnitStepResponse:Post-initialConditions
Quiz: Considertheequation
.
v+kv=u(t)
withrestinitialconditions,v(0

) =0.
.
Forthesolutionv(t)whatisv(0
+
)?
Choices:
.
a) v(0
+
)=0
.
b) v(0
+
)=1/k
.
c) v(0
+
)=1
.
d) v(0
+
)=k
e) Noneofthese.
Pickwhatyouthinkisthecorrectchoiceandthenlookattheanswer.
UnitStepResponse:Post-initialConditions
Quiz: Considertheequation
.
v+kv=u(t)
withrestinitialconditions,v(0

) =0.
.
Forthesolutionv(t)whatisv(0
+
)?
Thinkaboutyouranswerandthenlookatthechoices.

FirstorderUnitImpulseResponse
1. UnitImpulseResponse
Considertheinitialvalueproblem
.
x+kx=(t), x(0

) =0, k,rconstants.
Thiswouldmodel,forexample,theamountofuraniuminanuclearreactor
whereattimet = 0weadd1kilogramofuraniumallatonceandkisthe
decayrateoftheuranium.
Becauseoftherestinitialconditionswehave x(t) = 0for t < 0. The
effectoftheinputistocausetheamountx(t)tojumpfrom0to1att =0.
Thatis,x(0
+
)=1.Fort>0theinput(t) =0and,therefore,fort>0we
shouldsolve
.
x+kx=0, x(0) =1.
Thegeneralsolutionisx(t) =ce
kt
.Tondc,weusex(0) =1,whichgives
c=1.Thus,inbothcasesandu-format
x(t) =
0
e
kt
fort<0
= e
kt
u(t). (1)
fort>0
This is called the unit impulse response, which we denote w(t). In
somesenseitisthesimplestnontrivialsolution;youjustgivethesystema
unitkickatt = 0,standback,andwatchtheresult. Fort > 0itisjustthe
homogeneoussolutionwithinitialconditionx(0) =1.
2. GraphoftheUnitImpulseResponsew(t)
Figure1showsthegraphoftheunitimpulseresponse. Noticethatat
t=0itjumpstox=1andthendecaysexponentiallyto0.
t
1
w(t)
.
Figure1.Theunitimpulseresponseofthesystemx+kx.
3. (t)asalimitofboxfunctions
Originallywefound(t)asalimitofboxfunctionsofarea1.Inthissec-
tionwewillcomputetheunitimpulseresponseasthelimitoftheresponses

FirstorderUnitImpulseResponse OCW18.03SC
totheseboxfunctions. Themaintwopointsindoingthisare: rst,togain
morecomfortandfacilitywiththiscircleofideasandsecond,toconvince
youthatthedeltafunctionismuchnicertoworkwiththanboxfunctions.
Weinviteyoutocomparetheamountofworkrequiredforsolvingtheunit
impulsewiththeamountofworkneededintheunitstepcase.
Aquickreview:Denetheboxfunctionasu
h
(t) =

0 fort<0
1/h for0<t<h
0 forh<t.
Ithastotalarea1forall h > 0andthegraphofu
h
(t)becomesaspikeas
h 0,i.e.
limu
h
(t) =(t).
h 0
Fortheequation
.
x+kx=u
h
(t), x(0

) =0
thethreepiecesofthesolutionareeasilyfoundtobe
x(t) =

c
1
e
kt
fort<0
1
hk
+c
2
e
kt
for0<t<h
c
3
e
kt
forh<t.
Using the initial condition x(0

) = 0 and matching the value of x at the

endpoints of each piece we nd c
1
= 0, c
2
= 1/hk, c
3
= (e
kh
1)/hk.
Thisgivesthesolution
x(t) =

0 fort<0
1 kh
1)e
kt
kh
(e forh<t.
kh
e 1
hk
1
hk
(1e
kt
) for0<t<h
Lettingh 0thisbecomes(sincelim
h 0
=1)

0 fort<0
x(t) =
e
kt
for0<t.
Thislimitisexactlytheunitimpulseresponsew(t)wefoundinaprevious
note.
Figure 2 shows this graphically by plotting the input and output for
severalvaluesofh.
2
FirstorderUnitImpulseResponse OCW18.03SC
t
11
1/1
t
2
1
1/2
t
3
1
1/3
t
1
Figure2.Responsesforh=1,h=.5,h=.333,andh 0.
Theinputisplottedinblackandtheoutputinred. Noticehowtheoutput
risesfasterandgetscloserto1ash 0. Finally,inthelimitofsmallh,it
jumpsdirectlyto1.
TheMeaningofthePhraseUnitImpulseResponse
Exactlyasinthecaseoftheunitstepresponse,theunitimpulseresponse
meanstheresponseofthesystemwhentheinputisaunitimpulse. Inthis
notewelookedatthethesystem
.
x+kx= f(t)
and we considered f(t) to be the input. Suppose, instead, we have the
system
.
T+kT =kT
e
,
whereweconsiderT
e
tobetheinput.Thentheunitimpulseresponseisthe
responseofthesystemtoinputT
e
(t) =(t),i.e.thesolutionto
.
T+kT =k(t).
3
Unit Impulse Response: Post-initial Conditions
Quiz: Consider the equation
.
w + kw = (t)
with rest initial conditions, w(0

) = 0.
.
For the solution w(t) what is w(0
+
)?
Choices:
.
a) w(0
+
) = 0
.
b) w(0
+
) = 1/k
.
c) w(0
+
) = 1
.
d) w(0
+
) = k
e) None of these.
Answer: (d).
.
Using the DE we get w(0
+
) + kw(0
+
) = (0
+
). We know w(0
+
) = 1 and
.
(0
+
) = 0. Therefore w(0
+
) = k.
We could also look at the solution w(t) = e
kt
for t > 0. Thus w
.
(t) =
ke
kt
for t > 0. This implies w
.
(0
+
) = k.
Using the solution to the DE probably seems easier than the rst method,
but it is important to be able to draw conclusions without knowing the so-
lution.
Unit Impulse Response: Post-initial Conditions
Quiz: Consider the equation
.
w + kw = (t)
with rest initial conditions, w(0

) = 0.
.
For the solution w(t) what is w(0
+
)?
Choices:
.
a) w(0
+
) = 0
.
b) w(0
+
) = 1/k
.
c) w(0
+
) = 1
.
d) w(0
+
) = k
e) None of these.
Pick what you think is the correct choice and then look at the answer.
UnitImpulseResponse:Post-initialConditions
Quiz: Considertheequation
.
w+kw=(t)
withrestinitialconditions,w(0

) =0.
.
Forthesolutionw(t)whatisw(0
+
)?
Thinkaboutyouranswerandthenlookatthechoices.
Second order Unit Step Response
1. Unit Step Response
Wewillusetheexampleofanundampedharmonicoscillatorwithin-
put f (t) modeledby
..
mx + kx = f (t).
Theunit step response isthesolutiontothisequationwithinputu(t) and
rest initial conditions x(t) = 0 for t < 0. That is, it is the solution to the
initialvalueproblem(IVP)
.. .
mx + kx = u(t), x(0

) = 0, x(0

) = 0.
Thiscouldbeanundampedspring-masssystemwithmassm andspring
constantk.Themassisatrestatequilibriumuntiltimet = 0whenasteady
forcestartstoactonit.
Forcerepresentsachangeinmomentumovertime. Aniteforce F(t)
canonlycauseanininitesimalchangeinmomentum(i.e. F(t) dt)atatime.
Therefore,themassdoesnotchangepositionabruptly,nordoesitchange
velocityinstantaneously.Becauseofthisweshouldexpectasolutionwhich
is continuous with continuous derivative. Only the acceleration experi-
encesadiscontinuity.
Fort < 0wealreadyknowthatx(t) = 0.Fort > 0theDEis
..
mx + kx = 1.
Thishasaconstantparticularsolutionx(t) = 1/k,andageneralhomoge-
neoussolution
x
h
(t) = c
1
cos(
n
t) + c
2
sin(
n
t), where
n
=

k/m.
Puttingthetwotogethergivesthegeneralsolution
x(t) = 1/k + c
1
cos(
n
t) + c
2
sin(
n
t) fort > 0.
. . .
Thecontinuityofx andx impliesx(0) = x(0

) = 0andx(0) = x(0

) = 0.
Thisallowsustondc
1
andc
2
.
0= x
.
(0) = 1/k + c
1
c
1
= 1/k
0= x(0) = c
2

n
c
2
= 0.






SecondorderUnitStepResponse OCW18.03SC
Theunit step response forthissystemis(inbothcasesandu-format)
0 fort < 0 1
x(t) =
1
k
(1 cos(
n
t)) fort > 0.
=
k
(1 cos(
n
t))u(t).
Asintherstordercase,wewillsometimesdenotethisv(t).
Theclaimthatwegetacontinuousresponseistrue,butmayfeelabit
unjustied. Lets redo the above example very carefully without making
thisassumption.Itwilltakemorework,butwewillgetthesameanswer.
IncasesformattheequationfortheIVPis
.. 0 fort < 0 .
mx + kx =
1 fort > 0,
x(0

) = 0, x(0

) = 0. (1)
Solvingthetwopiecesweget
c
1
cos(
n
t) + c
2
sin(
n
t) fort < 0
x(t) =
1/k + c
3
cos(
n
t) + c
4
sin(
n
t) fort > 0.
.
Thepre-initialconditionsx(0

) = x(0

) = 0easilyimplyc
1
= c
2
= 0. So
oursolutionlookslike
0 fort < 0
x(t) =
1/k + c
3
cos(
n
t) + c
4
sin(
n
t) fort > 0.
Tondc
3
andc
4
wesubstitutex(t) intoequation(1).
.
Tomeasurethejumpswecomputex(0
+
) = 1/k + c
3
andx(0
+
) = c
4

n
.
Weusethisaswecomputederivativesofx.
. 0 fort < 0
x(t) = (1/k + c
3
)(t) +
c
3

n
sin(
n
t) + c
4

n
cos(
n
t) fort > 0.
.. 0 fort < 0
x(t) = (1/k + c
3
)

(t) + c
4

n
(t) +
c
3

2
cos(
n
t) c
4

2
sin(
n
t) fort > 0.
n n
Sincetheright-handsideofequation(1)doesnothaveanydeltafunctions
..
orany

(t) thecoefcientsinfrontofthesetermsintheformulaforx must

be0:
1/k + c
3
= 0 c
3
= 1/k
c
4

n
= 0 c
4
= 0.
2
SecondorderUnitStepResponse OCW18.03SC
In the end, we have exactly the same solution as above for the unit step
response.
To summarize: the continuity assumptions follow because any jumps
.
in x(t) or x(t) would result in delta functions when x is substituted into
equation(1).
Thegeneralizedderivative

(t) isnotsomethingweveseenbefore. It
isoftencalledadoublet. Thereisanentiretheoryoftheseandothergener-
alizedfunctions,butwewillonlyuse(t) inthiscourse.
Figure 1 shows the graph of the unit step response (with k = 1 and
m = 0.5.
t
2/k
..
Figure1.Theunitstepresponseforthesystemmx + kx = u(t).
Ifweaddedsomedampingthehomogeneouspartofthesolutionwould
goto0andtheunitstepresponsewouldgoasymptoticallyto1/k.
The Meaning of the Phrase Unit Step Response
Aswenotedintherstordercase,theunitstepresponseistheresponseof
thesystemtoaunitstepinput.Forexample,ifoursystemis
.. . .
mx + bx + kx = by
andweconsidery tobetheinput,thentheunitstepresponseisthesolution
to
.. . . .. .
mx + bx + kx = bu(t) equivalently mx + bx + kx = b(t).
3
SecondorderUnitImpulseResponse
1. EffectofaUnitImpulseonaSecondorderSystem
Weconsiderasecondordersystem
.. .
mx+bx+kx= f(t). (1)
Our rst task is to derive the following. If the input f(t) is an impulse
c(ta),thenthesystemsresponseto f(t)hasthefollowingproperties.
.
1.Themomentummx(t)jumpsbycunitsatt=a.Thatis,
mx
.
(a
+
)mx
.
(a

) =c.
2.Thepositionx(t)isunchangedatt=a.Thatis,
x(a
+
)=x(a

).
Recalltheargumentthatweusedbefore: If x(t)hadajumpat a then
. ..
x(t)wouldcontainamultipleof(ta).So,mx(t)wouldcontainamulti-
pleofthedoublet

(ta).Thisisimpossiblesincetheinput(ta)does
notcontainadoublet.Thisshowspoint(2)above.
.
Toshowpoint(1), wenotethatif mx(t)hasajumpof c unitsat ta
..
thenmx(t)containsthetermc(ta).Thisisneededtomaketheleft-hand
sideofequation(1)matchtherighthandsidewhen f(t) =c(ta).
Anotherwaytoshowpoints(1)and(2)isaphysicalargument.Aforce
actingonthemassovertimechangesitsmomentum. Infact,thebestway
tostateNewtonssecondlawisthat
dp
= f(t),
dt
wherep(t)isthemomentumofasystemand f(t)isanexternalforceacting
on the system. If a force f(t) acts over the time interval [t
1
,t
2
] the total
changeofmomentumduetotheforceis

t
2
f(t)dt.
t
1
Physicistscallthistheimpulseoftheforce f(t)overtheinterval[t
1
,t
2
]. If
a very large forceis applied over a very shorttime interval and has total
impulseof1theresultwillbeasuddenunitjumpinthemomentumofthe
system.
SecondorderUnitImpulseResponse OCW18.03SC
Forasecondordersystemtheunitimpulsefunctioncanbethought
ofasanidealizationofthisforce. Itisaforcewithtotalimpulse1applied
allatonce.
Athirdargumentthatwewillskipwouldbetosolveequation(1)with
a box function for input and take the limit as the box gets narrower and
talleralwayswitharea1.
2. UnitImpulseResponse
Weconsideronceagainthedampedharmonicoscillatorequation
.. .
mx+bx+kx= f(t).
Theunitimpulseresponseisthesolutiontothisequationwithinput f(t) =
(t)andrestinitialconditions: x(t) =0fort <0. Thatis,itisthesolution
totheinitialvalueproblem(IVP)
.. . .
mx+bx+kx=(t), x(0

) =0, x(0

) =0.
Thiscouldbeandampedspring-masssystemwithmassm,dampingcon-
stantbandspringconstantk. Themassisatrestatequilibriumuntiltime
t=0whenitishitbyasuddenverybriefveryintenseforce,ratherlikeget-
tinghitontheheadbyahammer. Theeffectistoincreasethemomentum
instantaneously,withoutchangingthepositionofthemass.
Letw(t)denotethesolutionweseek. Therestinitialconditionstellus
thatw(t) =0fort<0. Weknowfromsection1thattheeffectoftheinput
istocauseaunitjumpinthemomentumatt=0andnochangeinposition.
Wealsoknowthat,fort > 0,theinput(t) = 0. Puttingthistogether,for
t>0thew(t)satisestheequation
.. . .
mw+bw+kw=0, w(0) =1/m, w(0) =0.
Thisisahomogeneousconstantcoefcientlineardifferentialequationwhich
wehavelotsofpracticeinsolving.
Example1. Findtheunitimpulseresponseforthesystem
.. .
2x+8x+26x= f(t). (2)
Solution. Wewillusethestandardnotationw(t)fortheunitimpulsere-
sponse. Wearelookingfortheresponsefromrestto f(t) =(t). Weknow
2

SecondorderUnitImpulseResponse OCW18.03SC
w(t) = 0fort < 0. Att = 0theinputcausesaunitjumpinmomentum,
.
i.e.,2w(0
+
)=1.So,fort>0wehavetosolve
2w
..
+8w
.
+26w=0, w
.
(0
+
)=1/2, w(0
+
)=0.
Therootsofthecharacteristicpolynomialare23i.Whichimplies
w(t) =c
1
e
2t
cos(3t) +c
2
e
2t
sin(3t), fort>0.
Theinitialconditionsgive
0=w(0
+
)=c
1
,
.
1/2=w(0
+
)=2c
1
+3c
2
c
2
=1/6.
Thus,theunitimpulseresponse(inbothcasesandu-format)is
0 fort<0 1
w(t) =
1
6
e
2t
sin(3t) fort>0
=
6
e
2t
sin(3t)u(t). (3)
Figure1thegraphoftheunitimpulseresponse.Noticethatatt=0the
.
graphhasacorner.Thiscorrespondstotheslopewjumpingfrom0to1/2.
Fort>0thegraphdecaysto0whileoscillating.
t
Figure1.Theunitimpulseresponseforthesystem2
.. .
x+8x+26x.
3. CheckingExample1bySubstitution
Withanydifferentialequationyoucanverifyasolutionbypluggingit
intotheequation.Wewilldothatwithexample1togainsomemoreinsight
intowhywegetthesolution.
First,wetakethederivativesofthesolutioninequation(3)fort>0
w
.
(t) =
1
6
e
2t
(2sin(3t) +3cos(3t))fort>0
w
..
(t) =
1
6
e
2t
(5sin(3t)12cos(3t))fort>0
Nextwelookatthejumpsatt=0
w(0

) =0, w(0
+
)=0
. .
w(0

) =0, w(0
+
)=1/2
3
SecondorderUnitImpulseResponse OCW18.03SC
Nowwecancomputethefullgeneralizedderivatives
w
.
(t) =
1
6
e
2t
(2sin(3t) +3cos(3t))u(t)
w
..
(t) =
1
2
(t) +
1
6
e
2t
(5sin(3t)12cos(3t))u(t)
Finallywesubstitutewforxinequation(2)
2w
..
(t) = (t)

3
5
e
2t
sin(3t)4e
2t
cos(3t)
8w
.
(t) =

3
8
e
2t
sin(3t) +4e
2t
cos(3t)
13
26w(t) =
3
e
2t
sin(3t)
.. .
2w+8w+26w = (t).
TheMeaningofthePhraseUnitImpulseResponse
As weve noted several times already, the response to a given input de-
pendsonwhatweinourequationweconsidertobetheinput. Forexam-
ple,ifoursystemis
.. . .
mx+bx+kx=by
and we consider y to be the input, then the unit impulse response is the
solutionto
.. .
.
.. .
.
mx+bx+kx=b(t) equivalently mx+bx+kx=b(t).
.
(Here,iswhatwevecalledadoublet.)
4
UnitImpulseResponse:Post-initialConditions
..
Quiz: Let w(t)bethesolutionto mx+kx = (t)withrestinitialcondi-
.
tions.Whatisw(0
+
)?
Choices:
.
a) w(0
+
)=0
.
b) w(0
+
)=
m
.
c) w(0
+
)=k
.
d) w(0
+
)=k/m
.
e) w(0
+
)=1/m
f) Noneofthese.
Answer: (e).
The unit impulse input causes a unit jump in momentum. Starting from
. .
restthismeansmw(0
+
)=1orw(0
+
)=1/m.
Unit Impulse Response: Post-initial Conditions
..
Quiz: Let w(t) be the solution to mx + kx = (t) with rest initial condi-
.
tions. What is w(0
+
)?
Choices:
.
a) w(0
+
) = 0
.
b) w(0
+
) =
m
.
c) w(0
+
) = k
.
d) w(0
+
) = k/m
.
e) w(0
+
) = 1/m
f) None of these.
Pick what you think is the correct choice and then look at the answer.
Unit Impulse Response: Post-initial Conditions
..
Quiz: Let w(t) be the solution to mx + kx = (t) with rest initial condi-
.
tions. What is w(0
+
)?
Think about your answer and then look at the choices.
Higher Order Unit Impulse Response
We can extend our reasoning in the rst and second order cases to any
order. Consider an n
th
order system with DE
a
n
x
(n)
+ a
n1
x
(n1)
+ . . . + a
1
x

+ a
0
x = f (t) , (1)
where we take f (t) to be the input. The equation for the unit impulse re-
sponse of this system is
a
n
x
(n)
+ a
n1
x
(n1)
+ . . . + a
1
x

+ a
0
x = (t), with rest IC. (2)
The effect of the function input is to cause a jump in the n 1
st
derivative
at time t = 0, while the lower order derivatives do not jump. That is, the
system is put in the state
x(0
+
) = 0, x

(0
+
) = 0, . . . , x
(n2)
(0
+
) = 0, x
(n1)
(0
+
) = 1/a
n
.
To show this we use the same reasoning as in the second order case. Sup-
pose there was a jump in a lower derivative. For example, suppose
x
(n3)
(0
+
) = b = 0.
Then the expression for x
(n2)
(t) contains b(t), which implies that x
(n1)
(t)
contains b

(t) and x
(n)
(t) contains b

(t). This is impossible because the

right-hand side of (2) does not have any derivatives of the delta function.
Since x
n1
(t) has a jump of x
(n1)
(0
+
) = 1/a
n
at t = 0, its derivative
a
n
x
(n)
(t) has a unit impulse, (t), at t = 0.
We conclude that the solution to (2) is 0 for t < 0 and for t > 0 it is
exactly the same as the solution to
a
n
x
(n)
+ a
n1
x
(n1)
+ . . . + a
1
x

+ a
0
x = 0
with initial conditions
x(0) = 0, x

(0) = 0, . . . , x
n2
(0) = 0, x
n1
(0) = 1/a
n
.