Circuit Theory

Prof. S.C. Dutta Roy

Department of Electrical Engineering
IIT Delhi
Lecture 37

Positive Real Functions


this is the thirty seventh lecture and we talk about positive real functions today
(Refer Slide Time: 00:00:30 min)

positive real function if we denote it by F of s is defined as a function F of s is PRF a function F
of s is positive real if and only if the definition very simple definition number one F of s is real
for s real it must be a real function and number two the real part of F of s real part of F of s
should be non-negative for real part of s non-negative okay
F of s is real for s real and real part of F of s non-negative for real part of s non-negative
real part of s non-negative means the right half plane isnt that right including the j omega axis
that is it is the closed right half plane
therefore what it means is that in the closed right half plane F of s should also be in the closed
right half of the F of s plane All right
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 1 of 32

in other words we can view this in terms of mapping of one complex plane that is s plane to F of
s plane
do not be scared with the word mapping what it simply mean is the following
(Refer Slide Time: 00:02:08 min)

that if we have an s plane sigma j omega and a F of s plane j real i am sorry j imaginary part of F
of s and this is the real part of F of s then it what what it means is the is that um when when s is
real this is the s plane and this is the F of s plane when s is real that is when s is on the real axis
the pink line F of s should also be real therefore F of s must lie on this line this is the mapping
that is the s plane this line maps into this line in the F of s plane thats the first part of the
definition
and the second part says that if if real part of s is non-negative which means that we take the
closed right half s plane when the real part of s is non-negative then real part of F of s should
also be non-negative
(Refer Slide Time: 00:03:02 min)
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 2 of 32


in other words the shaded area here maps on to the shaded area in the F of s plane this is what is
meant by mapping All right
it is an s to F of s mapping in such a manner that the real axis real axis of the s s plane maps on to
the real axis of the F of s plane and the right half of the s plane maps on to the right half of the F
of s plane thats the definition of a positive real function
(Refer Slide Time: 00:03:55 min)

ah a a question suppose i have a point lets say here lets say this point is sigma zero what would
be the possible location of on the F of s plane [Conversation between student and professor Not
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 3 of 32

audible ((00:04:05))] real axis positive half positive half so it would be somewhere here lets say
F of sigma All right
(Refer Slide Time: 00:04:20 min)

another question if you have understood the definition suppose i have a point here lets say
minus sigma one what would the corresponding F of minus sigma on b [Conversation between
Student and Professor - any point on the real axis((00:04:29 min))] [Laughter-any point] on the
real axis that is correct
so it is a one to one mapping of the positive real axis of the s plane on to the positive real axis of
the F of s plane it does not mean that the negative real axis maps on to the negative real axis no is
the point clear [Conversation between Student and Professor Not audible ((00: 04:52 min))]
[Laughter] why not because the definition only says that if real part of s is non-negative then real
part of F of s is non-negative it doesnt say that if real part of s is negative what will happen it
doesnt say [Noise] it only puts a constraint on the real part of s non-negative if it is negative
than F of sigma zero it has to be real isnt that right
if s is on this point then s is a real quantity and F of s has to be real
and therefore it must be somewhere on this line but it doesnt specify where it could be in the
right half of the uh real axis or in the left half of the real axis it could even be at the origin it
does not matter is the point clear all that it says is that real part of s is non negative [Noise]
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 4 of 32

that is the positive half of the real axis in the s plan maps on to the positive half of the real axis in
the F of s plane okay
this constraints must be kept in mind
a simple mathematics but the concepts are important and one if one loses sight of one small
concept one is able to make a Himalaya [Noise] okay a mistake which is of the dimension of
Himalayas
(Refer Slide Time: 00:06:15 min)

now very simple example
suppose F of s is a constant lets say K where K is real and non negative K is real and non
negative
obviously this function is real why because when s is real well its always real is always real so s
is real its also real when real part of s is non negative K it is a constant thats always non
negative so it is a therefore a resistance for example F of s equal to a resistance [Noise] or a
conductance is a positive real function a resistance is a positive real function a conductance is
also a positive real function
the impedance of an inductor F of s is equal to s times L this is a impedance of an inductor this
is also positive real because when s is real it is real and when real part of s is non negative the
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 5 of 32

real part of F of s after all real part of F of s is equal to L times the real part of s and therefore
then this is non negative obviously this will also be non negative
therefore an inductor the impedance of an inductor is also a positive real function okay
(Refer Slide Time: 00:07:35 min)

lets take the impedance of a capacitor lets say F of s is equal to one over s C C is a real positive
quantity All right
and you see that the first the first part of the definition is obviously met that one s is real this
quantity is real
now lets look at the real part of F of S this is equal to one over C then real part of one over
sigma plus j omega All right and this is equal to one over C real part of this is sigma square plus
omega square and the numerator you get sigma
and obviously if sigma is greater than equal to zero what is sigma this is the real part of s if
sigma is greater than equal to zero then obviously real part of F of s is greater than equal to zero
the sign depends on sigma right
and therefore if the sigma is non negative then this the whole quantity is non negative in other
words one by s C impedance of a capacitor is also a positive real function
(Refer Slide Time: 00:08:42 min)
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 6 of 32


and in general one can conclude one can conclude that any network can [Noise] any network can
which consist of passive elements like resistance inductance and capacitance shall be shall have
an input impedance and driving point impedance which is positive real and this is why our
interest in positive real functions the driving point functions of passive networks are positive real
now um we have not yet proved it we have only said that of an inductance is positive real
impedance of a capacitor is positive real impedance of a resistor {capa}((00:09:22)) is positive
real and after it transform and contains three inductances and therefore the impedance should
also be positive real a combination of them which is contained in N it appears to reason it
appears to {intrusion}((00:09:40)) that it should also be positive real but lets prove it lets make
a rigorous proof and look at the proof very interesting proof
(Refer Slide Time: 00:09:49 min)
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 7 of 32


what we do is let the input impedance be Z of s okay let the input impedance is Z of s and let the
current going in be capital I agreed then for sinusoidal excitation that is s equal to j omega for
sinusoidal excitation the power that is absorbed by the network the power that goes into the
network shall be magnitude I square divided by two capital I is the peak value peak vector
magnitude I square divide by two this is the root mean square current multiplied by
[Conversation between Student and Professor ((00: 10: 32 min ))] no the real part it is the
resisted part which absorbs power the reactive part is the reactive part so this will be the real
part of Z of j omega and since the network is passive this would be greater than equal to zero
which means that the real part of Z of j omega is greater than equal to zero agreed this is the first
part of the [noise] first part of the proof
the second part well we have not proved anything we have simple said that the real part on the j
omega axis should be non negative real part of the j omega axis should be non negative that
doesnt mean that it should be non negative ah over the whole right half plane it doesnt mean
that
(Refer Slide Time: 00:11:29 min)
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 8 of 32


now to prove that part that part this is only a partial proof to prove that part let us let us modify
capital N let us modify capital N to N prime such that such that what was an inductance in N is
now what was a pure inductance N [Noise] let us add in N prime an inductance L in sinusoidal
resistance R okay that is N prime well make it into a more losing network All right
in sinusoidal every inductance we put a resistance such that the impedance of this element
impedance of this element in N was s L here it is s L plus R and i can write this is as L s plus
[Noise] R by L agreed i can write this is as L s plus alpha lets say
and let let the value of the alpha be the same for each inductor in other words what I am saying is
each inductor Li where i can go from zero to where i can go from one to n lets say each inductor
Li is augmented by resistance Ri in such a manner that Ri by Li is a constant equal to alpha
we can do that every inductor that i find i add a resistance alpha times Li thats all that it means
in series i add a resistance All right
similarly every capacitor C in N is augmented every capacitor lets say Cj is augmented in N
prime by a shunt conductance Gj such that that is i make each capacitor losing each capacitor in
the N and a lose is introduced in such a manner that Gj is equal to alpha times Cj All right
this alpha and this alpha is the same we can do that okay we are taking we take a box of resistor
and just go on adding resistor across each capacitor and serializing each inductor we can do that
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 9 of 32

now then the capacitance the admittance of this capacitor which was S Cj now becomes now
becomes Cj times S plus alpha is that clear it becomes S Cj plus Gj S Cj plus Gj i take Cj
common therefore I get Cj times S plus alpha is that okay
resistors in N will deep in-tagged that is resistor R K in N goes as resistor R K in N prime
(Refer Slide Time: 00:14:50 min)

then what happens to the input impedance of N prime what happens to input impedance to N
prime
you see the only change that has happened is wherever s occurred in N s should be replaced s
plus alpha and therefore is this point clear whenever s occurs the impedance of inductor it is L
times s plus alpha in admittance of a capacitor it [Noise] is C time s plus alpha therefore
[Conversation between Student and Professor Not audible ((00: 15:20 min))] [Laughter-our
objective will be clear] in a moment
we are trying to derive another network N prime whose impedance now instead of s it shall
be[Conversation between Student and Professor Not audible ((00: 15:34 min))] s plus alpha
this is what we wanted to derive that is from the given network N whose {impi}((00:15:41))
input impedance [Noise] is Z of s we want to derive a network whose input impedance is Z of s
plus alpha
(Refer Slide Time: 00:15:49 min)
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 10 of 32


lets call this as Z one of s All right then by the same token by the same token as was applied to
capital N we must have real part of Z one of j omega as greater than equal to zero is that right for
N the real part on the j omega axis was greater than equal to zero there is no reason why N prime
should not also have the same property all we have done is added a few resistors we have not
changed the character of the network right
therefore the real part of Z one of j omega is greater than equal to zero which means that the real
part that is real part Z of j omega plus alpha is greater than equal to zero clear because what is Z
one of s it is Z of s plus alpha therefore this means that the real part of Z of j omega plus alpha is
the is greater than equal to zero
i argue now that alpha was a purely arbitrary quantity what was the constraint on alpha that alpha
should be non-negative i can not add a negative resistor
{alphaf}((00:17:07)) the resistor that I add with parallel with a capacitor or in series with an
inductor the only constraint is that alpha should be greater than equal to zero
if alpha is equal to zero we get the original network if alpha is non-zero we get a derived network
and therefore alpha is not a separate notation we could as well replace this by sigma where the
only qualification of the sigma is that sigma should be real and non-negative
but what is sigma plus j omega isnt that precisely [Conversation between Student and Professor
Not audible ((00: 17:43 min))] s therefore what we get from this derivation is that real part of Z
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 11 of 32

of s should be greater than equal to zero is that clear is a very beautiful uh derivation originally
due to cover we will have cover a (jarment jentlepen) ((00:17:59 min))
now what does it mean it means that Z of s is real for s real and real part of Z of s is greater than
equal to zero and therefore Z of s which is [Noise] the driving point impedance of an arbitrary
passive network we have specified nothing we only said it contains resistors capacitor and
inductor nothing else
it could also contains transformers because transformers have combinations of three inductors
and therefore we have proved that the driven point impedance of a passive network is positive
real and this is why we are interested in positive real functions the connection between the circuit
theory and positive real function is through driving [Noise] point impedance [Noise]
now it takes a moment to justify that this property of positive realness also [Noise] applies to a
driving point admittance function that is[Conversation between Student and Professor Not
audible ((00: 18:56 min))] yes [Conversation between Student and Professor Not audible ((00:
18:58 min))]would you please explain the proof [Laughter-okay]
what we did was lets go back uh what we did was we took an arbitrary network N which
contains resistors conductors and capacitors and we argued that the input impedance Z of s its
real part on the j omega axis must be positive why because the power absorbed by the network
must be positive must be non-negative okay
so the real part of Z of j omega [Noise] is greater than equal to zero from N when we derive a
network from N we derive a network okay by adding resistors in serial with inductor and
resistor in parallel capacitor in such a manner that instead of Z of s we get Z of s plus alpha okay
thats what we did here in N prime we added conductances in parallel capacitor and resistor in
series inductors in such a manner that each s L now becomes L times s plus alpha and each s C
becomes C times s plus alpha and therefore all that changes in N prime is that s is replaced by s
plus alpha
[Conversation between Student and Professor Not audible ((00: 20:26 min))] pardon me
[Conversation between Student and Professor Not audible ((00: 20:28 min))] if the network is
not passive than of course ((we cannot)) (00: 20:32 min) this argument shall not be valid if
capital p is non is can be negative then this argument is not valid yes
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 12 of 32

[Conversation between Student and Professor Not audible ((00: 20:40 min))] thats right
[Conversation between Student and Professor Not audible ((00: 20:46 min))] why cannot oh
because the power in a sinusoidal excitation the power is absorbed normally in the resistive
components the reactive power is of no concern the power you see the reactive power can be
positive as well as negative and there is an exchange between inductor and capacitors overall the
average [Noise] reactivity [ ]((00:21:11)) but the average power absorbed in the resistor will be
non-negative okay so the it is on the passivity is on the basis of power absorbed by the network
All right and ((theref)) (00: 21:27 min) thats why it was it was on the j omega axis
[Conversation between Student and Professor Not audible ((00: 21:30 min))] um-huh a non
sinusoidal input can always be broken up into [Conversation between Student and Professor
Not audible ((00: 21:36 min))] sinusoidal input and therefore it would be a sum of such terms
okay
any other questions
so what we did was we argued that the real part of Z of j omega plus alpha is non-negative and
we argued that the alpha was purely our choice [Noise] alpha could be anything any non-
negative value is admissible and therefore we could replace it by sigma where sigma is greater
than equal to zero when sigma is greater than equal to zero then sigma plus j omega is [Noise] [
]((00: 22:08))
(Refer Slide Time: 00:22:15 min)


I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 13 of 32

and therefore real part of it is s it is s with sigma greater than equal to zero is the right half plane
omega could be positive or negative it doesnt matter
and therefore when s is in the right half plane then the sigma greater than equal to zero obviously
the real part of Z of s will greater than equal to zero which shows
[Conversation between Student and Professor Not audible ((00: 22:31 min))] no it didnt say
[Conversation between Student and Professor Not audible ((00: 21:37 min))] only in the right
half plane not all values
you see a sigma here alpha was positive quantity we could not add a negative resistor okay so we
prove that Z of s the driving point impedance of a passive network is always positive real
now we want to extend it to driving point admittance
(Refer Slide Time: 00:23:04 min)


and all that require all that required is to prove a theorem that if F of s is P R [Conversation
between Student and Professor Not audible ((00: 23:09 min))] yes [Conversation between
Student and Professor Not audible ((00: 23:11 min))] two things [Conversation between
Student and Professor Not audible ((00: 23:14 min))] okay [Conversation between Student and
Professor Not audible ((00: 23:16 min))] oh we dont have to prove it because Z of s after all is
made up of a sum one by s C and R each of them are real for s real and itll only be a series-
parallel and multiple series-parallel combination so all of them all of these quantity would be
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 14 of 32

real for s real okay that is trivial but this is not trivial that real part of Z of s is non-negative is
not a trivial proof
any other question
[Conversation between Student and Professor Not audible ((00: 23:49 min))] why did you say
that real part of Z one j omega is {great}((00: 23:56 min)) than [Conversation between Student
and Professor Not audible ((00: 23:57 min))] is the first case because the power has to be non-
negative okay
lets go back this has to be non-negative and this is the positive quantity magnitude square
divided by two and therefore this must be non-negative [Noise]
anything else [Noise]
i appreciate these questions because first time one {make}((00:24:20)) synthesis synthesis does
appear to be tough but its a very simple once you get over this fear of complex variables and
this {func}((00:24:28)) the functions are extremely simple form very nicely behaved well
behaved functions nothing to be scared of okay
All right as i said Z of s the driving point impedance function is positive real we want to extend it
to driving point admittance function all that is needed is to prove this theorem
(Refer Slide Time: 00:24:51 min)

that the F of s is P R then so is one by F of s that is if impedance is P R then the admittance also
be PR and it is extremely simple to prove
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 15 of 32

(Refer Slide Time: 00:25:03 min)


you see if F of s is PR then obviously one by F of s is real for s real agreed because F of s is real
for s real
(Refer Slide Time: 00:25:16 min)

and the the other thing that we have to prove is real part of one by F of s okay
suppose [Noise] suppose F of s is U plus j X U is its real part and X is its imaginary part
since its a function of s U is a function of sigma and omega its a two variable function
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 16 of 32

similarly X is a function of sigma and omega you see F of s s is sigma plus j omega I break this
up this complex function into a real part and an imaginary part
the real part shall be a sigma and omega imaginary part shall also be a function of sigma and
omega
(Refer Slide Time: 00:26:06 min)

so what is the real part [noise] real part of one over F of s it is real part of one over U plus j X
which is equal to real part of U square plus X square and in the numerator [Noise] it is U
[Conversation between Student and Professor Not audible ((00: 26:24 min))] pardon me
[Conversation between Student and Professor Not audible ((00: 26:26 min))] okay U minus j X
this is equal to U divided by U square plus X square All right
and obviously this is non-negative if U is non-negative
when is U non-negative that is real part of s non-negative is that clear because by hypothesis F of
s is P R so its real part is non-negative for real part of s non-negative okay
therefore this quantity is non-negative if U is non-negative but U is non-negative in turn when
sigma is non negative [Noise] therefore we prove that the real part of this reciprocal function is
non-negative for the right half plane and therefore one by F of s is P R
(Refer Slide Time: 00:27:31 min)
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 17 of 32


and therefore our theorem now states that all DPIs and this I is now an immittance i m m i t t a n
c e it is impedance as well as admittance all DPIs of passive networks are positive real okay
now the question of testing that before i test lets look at some useful properties
(Refer Slide Time: 00:28:07 min)

properties of positive real functions some useful properties
one of them that i have seen is if F is P R then so is one over F this we have already proved thats
how we we stated the theorem that all DPIs are all DPIs of passive networks are P R suppose F
one and F two are P r then what can you say about F one plus F two [Conversation between
Student and Professor Not audible ((00:28:46 min))] that is also P r the proof is trivial because
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 18 of 32

real part of F is equal to real part F one plus real part F two and both of them are individually
[Noise] non-negative so the sum is non-negative
what can you say about F one minus F two [Conversation between Student and Professor Not
audible ((00: 29:03 min))] not necessarily P r
because the real part due to subtraction may become negative not necessarily P r
(Refer Slide Time: 00:29:19 min)

All right next poles and zeros of positive real functions
where are they okay cannot lie in the in the right half plane why not [Noise] right half s plane
why not poles and zeros of positive real function cannot lie in the right half of the s plane what is
the reason
[Conversation between Student and Professor Not audible ((00: 30:00 min))] pardon me
[Conversation between Student and Professor Not audible ((00: 30:03 min))] which polynomial
it has something to do with the real part it is a passive network you see all positive real functions
this is interchangeable impedance admittance or positive real functions all driving per
{limittences}((00:30:20)) are PRFs it can be shown that all PRFs can be realized by driving point
{limittences}((00:30:26)) of passive network
now if there is pole in the right half plane what does it mean it means that it means instability
that its a real part real part is no longer non negative in the right half plane
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 19 of 32

what does instability mean that the output can grow indefinitely without an input that is the
unstable condition All right
that is possible only when the when the network is able to generate power where as a passive
network the power has always to be non-negative so they cannot lie in the right half right half of
the s plane
[Conversation between Student and Professor Not audible ((00: 31:07 min))] no that is not
correct because for a transfer function which is not a P r not necessarily a P r that can be zero in
the right half plane for a transfer function there can be
(Refer Slide Time: 00:31:27 min)

for example i can have a function s minus one [Noise] by s plus one it is an all pass function as
you know the poles and zero are in mirror image symmetry did you do this in the class
[Conversation between Student and Professor Not audible ((00: 31:37 min))] we have okay
(Refer Slide Time: 00:31:40 min)
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 20 of 32


but it is not P r if it is a positive real function then the function as well as its reciprocal have to be
P r and therefore we cannot have poles or zeros in the right half plane okay
(Refer Slide Time: 00:31:53 min)

number four poles on the j omega axis this is important and we shall we shall go slow on this
poles on the j omega axis must be simple why simple because if they are multiple then it leads to
instability instability okay poles on the j omega axis must be simple now im adding something
to this for a positive real function with [Noise] real positive residual with real positive residual
poles on the j omega axis must be simple with real positive residual okay
now why is it so let us look at this
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 21 of 32

(Refer Slide Time: 00:32:47 min)

first we say why the residue should be real
but first let us see if we have a function F of s with lets say two poles one at j omega i and the
other at minus j omega i while complex poles must occur in conjugate pairs why [Conversation
between Student and Professor Not audible ((00: 33:01 min))] because its function is real for s
real okay
so if j omega i is a pole minus j omega it should also be a pole and in F of s if i expand in partial
fraction we shall have Ki divided by S minus j omega i that would be the partial fraction
corresponding to j omega i
and for corresponding to minus j omega i it should be s plus j omega i then a residue which has
to be complex conjugate of this [Noise] residue why is it so that is this residue must be Ki star
why
is the point clear[Conversation between Student and Professor Not audible ((00: 33:48 min))]
okay i could lets lets write this as Ki prime
lets write this as Ki prime now you add this two terms [Conversation between Student and
Professor Not audible ((00: 33:59 min))] pardon me [Conversation between Student and
Professor Not audible ((00: 34:02 min))] when you add them up the function must be real
function and it cannot be real unless Ki prime is equal to Ki star is this point clear [noise]
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 22 of 32

therefore therefore and in addition it says in addition it says that Ki must be real if it is real i will
tell you why it why it cannot be complex in a minute [Conversation between Student and
Professor Not audible ((00: 34:29 min))] uh-huh if Ki is real obviously both the residue should
be same identical lets assume for a moment that there we are
if they are real then obviously if Ki is equal to Ki star equal to real then obviously the sum of this
two terms would be twice Ki s divided by s square plus omega i square
and this shows that Ki the residual must be real otherwise it cannot give rise to a real function
okay is that clear [Noise]
suppose for argument sake let we do have the residue which is complex lets say we have K one
plus j K two divided by divided by s plus lets say [noise] s minus j omega i
if i combine this with K one minus j K two will be sum be here
[Conversation between Student and Professor Not audible ((00: 35:31 min))] pardon me
[Conversation between Student and Professor Not audible ((00: 35:33min))] it will be real
[Conversation between Student and Professor Not audible ((00: 35:35min))] pardon me
[Conversation between Student and Professor Not audible ((00: 35:42min))] okay in other
words i can have a residue like this K one plus j K two divided by s plus j omega i plus K one
minus j K two divided by s plus j omega i [Conversation between Student and Professor Not
audible ((00: 36:03min))] no some shall be negative [Conversation between Student and
Professor Not audible ((00: 36:07min))] it cannot be negative why not [Conversation between
Student and Professor Not audible ((00: 36:15min))] okay
(Refer Slide Time: 00:36:28 min)
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 23 of 32



same as the last [Noise] now the sum of these two term whether you take it in this form or in this
form the sum of these two term shall be this
this Ki can this be complex [Conversation between Student and Professor Not audible ((00:
36:38 min))] no it cannot be complex okay for a pair of poles on the j omega axis Ki is the
residue okay Ki is the residue that is Ki is equal to the function let me write in this next page
[Conversation between Student and Professor Not audible ((00: 36:58min))] uh this point is not
clear okay
(Refer Slide Time: 00:37:02 min)

I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 24 of 32

if you add this two you can get a real quantity and the quantity must be of these form two Ki s
divided by s square plus omega i square isnt that right
where this Ki cannot be complex it must be real if it is be complex then the whole quantity will
be complex it cannot be a real function
and Ki as you see is given by F of s multiplied by s square plus omega i square divided by s
twice s at s square equal to minus omega i square now this point not clear [Conversation between
Student and Professor Not audible ((00: 37:48 min))] no [Conversation between Student and
Professor Not audible ((00: 37:51 min))] yes [Conversation between Student and Professor
Not audible ((00: 37:52 min))] okay [Conversation between Student and Professor Not audible
((00: 37:59 min))]
(Refer Slide Time: 00:38:05 min)

fine lets lets clear this s square plus omega i square i get K one plus j K two s plus j omega i
plus K one minus j K two s minus j omega i okay
this is equal to K one s [Noise] K one s twice K one s [Conversation between Student and
Professor Not audible ((00: 38:27 min))]
minus twice K two [Conversation between Student and Professor Not audible ((00: 38:29
min))] omega i square [Conversation between Student and Professor Not audible ((00: 38:34
min))] no only omega i divided by s square plus omega i square
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 25 of 32

[Conversation between Student and Professor Not audible ((00: 38:43 min))] [Noise] k when s
is real the function is real okay but [Conversation between Student and Professor Not audible
((00: 38:55 min))] the real part [Noise] real part real part of this function can be negative
when s equals to j omega for example look at this when s equals to j omega this contributes to an
imaginary part this contributes to a real part and depending on K two it can be positive or
negative [Conversation between Student and Professor Not audible ((00: 39:21 min))]
[laughter] pardon me [Conversation between Student and Professor Not audible ((00: 39:23
min))] then {property}((00:39:24)) is not satisfied
so the this cannot be allowed in other words the residue K one [Conversation between Student
and Professor Not audible ((00: 39:32 min))] should be positive and real
this is the proof we did it in a slightly round about fashion but [Conversation between Student
and Professor Not audible ((00: 39:40 min))] K one must be positive and real it cannot be
[Conversation between Student and Professor Not audible ((00: 39:44 min))] no it is not
sufficient i coming to the [Conversation between Student and Professor Not audible ((00: 39:49
min))] oh this is what i was going to {describe}((00:39:52)) this is what i was going to
(Refer Slide Time: 00:39:55 min)

you see if F of s is lets say of this form some numerator divided by s square plus omega not
square multiplied by lets say some polynomial D of s then obviously we have a pair of poles on
the j omega axis and plus minus j omega not then the residue K not at this pole at either of the
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 26 of 32

two poles is defined as s squared plus omega not squared multiplied by F of s divided by twice s
at s squared equal to minus omega not square
which simply means that what we have done is that we had expanded F of s in partial fraction we
have expanded this in K zero by s plus j omega not
plus K zero divided by s minus j omega not plus other terms that is what K zero K zero is called
the residual at the pole at plus minus j omega not and this residue must be real and positive why
positive
you are you are convinced about the realness of it okay why positive suppose K zero is negative
suppose K zero is negative
(Refer Slide Time: 00:41:27 min)

after all the term that comes is twice K zero s divided by s square plus omega not square
[Conversation between Student and Professor Not audible ((00: 41:34 min))] [Noise] I didnt
follow [Conversation between Student and Professor Not audible ((00: 41:46 min))] yeah
[Conversation between Student and Professor Not audible ((00: 41:49 min))] that is correct the
real part condition should be validated and therefore K zero must be real and positive All right
this point must {so keen}((00:42:07)) that if a positive real {functio}((00:42:09)) [Conversation
between Student and Professor Not audible ((00: 42:09 min))] pardon me [Conversation
between Student and Professor Not audible ((00: 42:10 min))] oh what is the real part of this
real part in the numerator ((00:42:18)) twice K zero multiplied by sigma and if K zero is negative
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 27 of 32

obviously sigma non-negative will make the real part negative okay and therefore K zero must
be real and positive [Noise]
the sum of this discussion sum and substance of this discussion is that if a positive real function
has a pole on the j omega axis this pole cannot take any liberty it must be constraint
the constraint is that the pole must be simple number one and the residual at this pole must be
real and positive okay if F of s if a positive real function has a pole on the j omega axis it must be
simple and the residual there it must be real and positive
(Refer Slide Time: 00:43:13 min)

next suppose F of s [Conversation between Student and Professor Not audible ((00: 43:19
min))] fifth property yes
suppose F of s now we we bring you a further restriction that our positive real function that we
shall deal with are all rational if you have a finite number of resistors capacitors and inductors
obviously the impedance function or admittance function or any function any network function
has to be rational what is a rational function [Conversation between Student and Professor Not
audible ((00: 43:45 min))] what is P s and what is Q s [Conversation between Student and
Professor Not audible ((00: 43:50 min))] both are what is a polynomial [Conversation between
Student and Professor Not audible ((00: 43:53 min))] polynomials [Conversation between
Student and Professor Not audible ((00: 43:58 min))] thats it a polynomial you are not saying
the key thing [Conversation between Student and Professor Not audible ((00: 44:05 min))]
what is the number of terms in the polynomial [Conversation between Student and Professor
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 28 of 32

Not audible ((00: 43:11 min))] so it is a finite series that is the essential [Laughter-part] of the
definition its a finite series containing positive integer [Noise] powers of the variable okay
now the question is what can be the difference in degree between P and Q we use the simple
[Noise] definition
degree of P is simply written as P and the temperature degree okay
what can be the difference between the degrees of P and Q [Conversation between Student and
Professor Not audible ((00: 44:40 min))]
the difference so i dont have to prove this side this is this is equal to what are the possible values
[Conversation between Student and Professor Not audible ((00: 44:47 min))] can it be zero can
this degree be equal [Conversation between Student and Professor Not audible ((00: 44:51
min))] yes if the degrees are equal then what is the behavior at infinity [Conversation between
Student and Professor Not audible ((00: 44:57 min))] a constant okay
now can the degree be differ by one [Conversation between Student and Professor Not audible
((00: 45:01 min))] yes then what is the behavior at infinity a pole or a zero now what is a pole of
F of s is a zero of one over F of s
and what is a zero of F of s is a pole of one over F of s
therefore a difference in degree of one is permitted what about two
no because then we shall have multiple pole or zeros at infinity so the degree difference between
P and Q can be at the most one zero or one All right
(Refer Slide Time: 00:45:45)
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 29 of 32


what does this mean this means that the highest powers if i write this as s to the n plus etcetera
and s to the n k then it means that the highest powers can differ only by zero or one All right they
cannot differ by more than one
now m can be greater than n or n can be greater than m it doesnt matter
what can you say about the lowest powers suppose the constant term here s zero and there is no
constant term here suppose the final term is b one s is this permitted is it permitted to have b zero
equal to zero [Conversation between Student and Professor Not audible ((00: 46:26 min))]
constant
suppose b zero is zero then we should have a pole at the origin if s zero is zero and b zero is non-
zero then there is a zero at the origin can we have in the denominator b zero and b [Conversation
between Student and Professor Not audible ((00: 46:44 min))] that is correct [Conversation
between Student and Professor Not audible ((00: 46:46 min))]
what does this mean that the lowest powers in numerator and denominator cannot differ by more
than one okay so this is the sixth property lowest powers in numerator and denominator cannot
differ by more than unity All right
so if i combine the two if i combine the two i can say that if the function is positive real these are
all necessary conditions they are not sufficient isnt that right even if these properties are true the
function may not be positive real because we have said the mapping of the real part yet therefore
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 30 of 32

these are all necessary condition not sufficient but we can say that the lowest power as well as
the highest power cannot differ by more than one for example [Conversation between the
Student and Professor-Not audible ((00: 47:43 min ))] pardon me [Conversation between the
Student and Professor-Not audible ((00: 47:45 min ))] why it is so okay lets take some example
(Refer Slide Time: 00:47:54)

suppose I have s square plus s plus one divided by lets say three s square s squared plus three s
then obviously at s equal to zero the function is a pole
on the other hand if three s is also absent then also it is a pole [Conversation between Student
and Professor Not audible ((00:48:15 min))] double pole and therefore the lowest powers
cannot differ by more than one
is the point clear
lets take some example [Conversation between Student and Professor Not audible ((00: 48:24
min))] oh that we have already [Conversation between Student and Professor Not audible ((00:
48:29 min))] elaborated if it is not
okay what does the pole on j omega axis cause in the time domain time domain it is either a sine
function or a cosine function if there a multiple poles for example pole of double order then this
is multiplied by t [Conversation between Student and Professor Not audible ((00: 48:47 min))]
from definition of PRF
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 31 of 32

well instability means that the real part condition shall be valid
instability means that the {rea}((00:48:57)) thats the power can be negative okay All right
now lets take a few examples [Noise] suppose i take s plus one divided by s square minus s plus
one can this be P r [Conversation between Student and Professor Not audible ((00: 49:17
min))] this can be P s [laughter]
then tricky question is it the denominator there is a negative sign what does it mean there must be
poles in the right half plane that means this this polynomial is not {high rates}((00:49:34)) okay
that is the point that i wanted to bring
(Refer Slide Time: 00:49:38 min)

if i write a network function as P of s by Q of s one of the necessary property is that both P and Q
must be {high rates}((00:49:47)) okay must be {high rates}((00:49:49)) All right
suppose i have s cubed plus s squared plus one divided by lets say five s cubed plus six s
squared plus twice s plus two [Conversation between Student and Professor Not audible ((00:
50:04 min))] [Noise] numerator is not {high rates}((00:50:06)) because there is a missing term
okay we will have more fun next time on examples and going into more details of positive real
functions
I
I
T
M
A
D
R
A
S
/
N
M
E
I
C
T
Circuit Theory Lecture 37
Prof. S. C. Dutta Roy Page 32 of 32