Theoretical Physics

Prof. Ruiz, UNC Asheville, doctorphys on YouTube

Chapter V Notes. Transfer Functions

V1. Driven Oscillations and an LP Filter. We return to our low-pass (LP) filter
with R = 1 , C = 1 , and f (t ) = 1 for 1 second from t = 0 to t = 1 . The initial
charge on the capacitor is q (0) = 0 . We have already solved this problem.

The charge on the capacitor is

given by the following integral.

t
q (t ) = ∫ f (u ) g (t − u )du
0

This solution is the convolution of

our input voltage function f (t )
and the capacitor-decay response function g (t ) = e−t . The g (t ) is our unshifted

Green's function G (t , 0) . We can also write our general answer with the following
Green's function notation.

t
q (t ) = ∫ G (t , u ) f (u )du with G (t , u ) = e− (t −u )
0

We would like to find the solution for a sinusoidal driving force.

We have f (t ) = sin ω t and G (t , u ) = e− (t −u ) .

The solution is

t
q (t ) = ∫ e− (t −u ) sin(ω u ) du
0

It is easier to integrate an exponential so we use the real-imaginary trick.

t
q (t ) = Im ∫ e− (t −u ) eiω u du
0

Michael J. Ruiz, Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License

 t t
q (t ) = e−t Im ∫ eu eiω u du = e−t Im ∫ e(1+ iω ) u du
0 0

 e(1+ iω )u t

q (t ) = e Im 
−t

 1 + iω 0


−t  e(1+iω )t − 1
q (t ) = e Im  
 1 + iω 

−t  (et eiω t − 1) (1 − iω ) 
q (t ) = e Im  
 1 + iω (1 − iω ) 

 (et ei ω t − 1)
−t 
q (t ) = e Im 2
(1 − iω ) 
 1 + ω 

 (e i ω t − e − t ) 
q (t ) = Im 2
(1 − iω ) 
 1 + ω 

 ei ω t   e−t 
q (t ) = Im  2
(1 − iω )  − Im  2
(1 − iω ) 
 1 + ω   1 + ω 
Taking the imaginary part of the second term we have the following.

 ei ω t  ω e−t
q (t ) = Im  2
(1 − iω )  + 2
1 + ω  1 + ω
Now we proceed to work on the first term with help from the Euler formula.

Michael J. Ruiz, Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License

 −t
 (cos ω t + i sin ω t )  ωe
q (t ) = Im  2
(1 − iω )  + 2
 1 + ω  1+ ω

(sin ω t − ω cos ω t ) ω e−t

q (t ) = 2
+ Note that q(0) = 0 .
1+ ω 1+ ω2

Check out the second part. It is the transient component. It goes to zero as t goes to
infinity. We are interested in the steady-state solution so we throw this last term away.
Then,
sin ω t − ω cos ω t ω e− t
q (t ) = for the steady state and
q (t ) transient =
1+ ω2 1+ ω2
The dimensions do not look correct in these equations because we have let R =1
and C = 1 . We can make things look dimensionally correct by a nice trick. We know
RC has dimensions of time since we saw its appearance earlier in the exponent as
−
t 1
RC
e . Therefore we can define c ω = = 1 and use it as needed to get the right
RC
dimensions for frequency. The "c" in the subscript stands for cut-off. To make

sin ω t − ω cos ω t
q (t ) =
1+ ω2
1
look dimensionally correct, we need
ωc = = 1 in front of the sine and a ωc2 in
RC
the denominator.
ωc sin ω t − ω cos ω t
q (t ) =
ωc2 + ω 2
Now it looks more friendly dimensionally by inspection. Now watch this next trick to
show that we have a sinusoidal oscillation that is out of phase with respect to our
voltage driving function f (t ) = sin ω t .

Michael J. Ruiz, Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License

Do you remember the following trig identity we derived in the first chapter?

sin(α − β ) = sin α cos β − cos α sin β

With α =ωt we have sin(ω t − β ) = sin ω t cos β − cos ω t sin β

Compare the following two equations.

ωc sin ω t − ω cos ω t
q (t ) =
ωc2 + ω 2
sin(ω t − β ) = sin ω t cos β − cos ω t sin β
Now check out the magic triangle below.

ωc = ω 2 + ωc2 cos β

ω = ω 2 + ωc2 sin β

ω
Then we have
tan β = and the following.
ωc

ω 2 + ωc2 cos β sin ω t − ω 2 + ωc2 sin β cos ω t

q (t ) =
ω 2 + ωc2

sin ω t cos β − cos ω t sin β

q (t ) =
ω 2 + ωc2

sin(ω t − β )
q (t ) =
ω 2 + ωc2

Michael J. Ruiz, Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License

 With Vin (t ) = f (t ) = sin ω t = V0 sin ω t

sin(ω t − β )
Vout (t ) = ωc V0 ωc
. Why the out in front?
ω 2 + ωc2
Hint: Dimensional Analysis!

Comparing the amplitudes, we obtain

Vout (t ) ωc
T≡ =
Vin (t ) .
ω 2 + ωc2

where the output has phase given by

ω
tan β = .
ωc
Now you can see below why this is a low-pass filter.

Vout ωc ω
lim = lim = c =1
Low Frequencies: ω →0 V
in
ω →0
ω 2 + ωc2 ωc

Vout ωc
lim = lim =0
High Frequencies: ω →∞ V ω →∞ 2 2
in ω + ωc
One final note, electrical engineers prefer the following conventions.

sin(ω t + φ )
Vin (t ) = V0 sin ω t Vout (t ) = ωc V0
and
ω 2 + ωc2
ω
φ = −β and
tan φ = tan( −β ) = − tan β = −
ωc

Michael J. Ruiz, Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License

 Vout
V2. Phasors. Engineers often just need Vin and the phase angle φ . The phasor
was developed to arrive at these quickly.

V = IR with input voltage V (t ) = V0 cos ω t

V0 cos ω t = I (t ) R
The current is in phase with the input voltage source.

V0 cos ω t
I (t ) =
R
q
V= with input voltage V (t ) = V0 cos ω t
C
q(t )
V0 cos ω t = q (t ) = CV0 cos ω t
C
dq(t )
i (t ) = = −ω CV0 sin ω t
dt
The current is not in phase with the input voltage source.

Applied Voltage: V (t ) = V0 cos ω t

Current: i (t ) = −ω CV0 sin ω t

See the figure. The current leads the
voltage by 90 degrees. The lines with
arrows in the figure are called phasors.

You use the real part of the phasor for

the applied voltage and current.

Michael J. Ruiz, Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License

Remember our trick to rotate by 90 degrees? You multiply a complex number z by i,
"making a left face." So in "phasor" language we have the following.

Phasor V (t ) = V0e jω t , where j = −1 because i is reserved for current.

It is understood you take the real part: V (t ) = V0 Re e jω t = V0 cos ω t

Phasor Current i (t ) = jω CV (t )

Check: i (t ) = Re [ jω CV (t ) ] = Re 
 jω CV0 e jω t
)  = −ω CV0 sin ω t
So in "phasor" language we write: i (t ) = jω CV (t )

Comparing this to Ohm's Law: i (t ) R = V (t ) , the effective resistance depends on

frequency. We call this the impedance.
1
V (t ) = i (t ) Z and i (t ) = V (t ) / Z , where
Z= .
jω C
V3. Transfer Function. Engineers are usually interested in the relative
Vout (t )
transmission amplitude
T= through the filter and the phase φ . Watch this
Vin (t )
powerful method using phasors. But first, recall the answer we obtained using the
Green's function:

Vout (t ) ωc
T= =
Vin (t ) ω 2 + ωc2

ω
tan φ = −
ωc
We write the regular voltage equation with impedances. Note that the impedance for the
resistor is simply R since there is no phase shift for the resistor.

Michael J. Ruiz, Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License

 Phasors: Vin (t ) = i (t ) Z R + i (t ) Z C and Vout (t ) = i(t ) Z C

 1  1
Vin (t ) = i(t )  R + Vout (t ) = i (t )
 jω C  and
jω C
Transfer Function

Vout (t ) 1/ ( jω C ) 1
H (ω ) = = H (ω ) =
Vin (t ) R + 1 1 + jω RC
jω C
The Magnitude of the Transfer Function is our Transmission

1
T = H (ω ) =
1 + jω RC
1 1
PV1 (Practice Problem). From z = x + iy = x + y 2 2
, show that
= .
z z
1
T = H (ω ) =
1 + ω 2 ( RC )2
1
ω
Remember our basic definition from before: c
=
RC (the cutoff frequency).
1
T = H (ω ) =
1 + ω 2 / (ωc ) 2

Vout (t ) ωc
T= =
Vin (t ) ω 2 + ωc2
This is our answer from before.

Michael J. Ruiz, Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License

 1 1 − jω RC
H (ω ) = =
From
1 + jω RC 1 + (ω RC ) 2 we get the phase.

Im H (ω ) −ω RC ω
tan φ = tan φ = tan φ = −
Re H (ω ) 1 ωc

1 1
H (ω ) = T = H (ω ) =
Summary:
1 + jω RC 1 + ω 2 / (ωc ) 2

sin(ω t + φ )
Vin (t ) = V0 sin ω t Vout (t ) = ωc V0
and
ω 2 + ωc2

Courtesy Inductiveload, Wikimedia Commons

ω
The phase of the voltage across the capacitor is behind since
tan φ = − <0.
ωc
π
The phase ranges as follows:
− ≤φ ≤ 0.
2
ω
PV2 (Practice Problem). Sketch T (ω ) = H (ω ) and
φ = − tan −1 for all
ωc
input frequencies ω . At which frequencies is the phase 0 and −π / 2 ?

Michael J. Ruiz, Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License

V4. Transfer Function and Transforms.

dq q q
vin (t ) = R+ vout (t ) =
dt C C

1 1
ℑ{Vin (t )} = ( jω ) RQ(ω ) + Q(ω ) ℑ{Vout (t )} = Q(ω )
C C

Vout (ω ) Q(ω ) / C 1
= =
Vin (ω ) jω RQ(ω ) + Q(ω ) / C 1 + jω RC

Vout (ω )
= H (ω )
Vin (ω )

In transform space, the ratio of the Fourier transform of the output to the Fourier
transform of the input is the transfer function!

PV3 (Practice Problem). Show that the above can be arrived at from the Laplace
Transform if you take q (0) =0 and make the substitution s = jω after you take
the Laplace Transform.

Michael J. Ruiz, Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License