6.

200 Supplementary Notes: Math of Complex Num-

bers and Sinusoidal Functions
Prof. Karl K. Berggren, Dept. of EECS
April 6, 2023

The algebra of working with sinusoidal functions is made easier by

having a facility with trigonometric identities, but easier still by being
completely comfortable working with complex numbers in Cartesian
and polar coordinates. These notes provide some tips and tricks to
make acquiring these skills easier. Of course, in the end skills are
honed through practice, which notes can’t provide. A few exercises
are available at the end of the notes.

Math of Sinusoidal Functions and the Complex Plane

Sinusoidal functions are intrinsically related to circles in general and

to rotary motion in particular. If you imagine looking at an object
moving along a circular orbit or path from the side, so that you only
see the object moving up and down, the plot of its displacement from
its mean position vs. time will be a sinusoidal function.1 1
Luckily, YouTube exists so you don’t
The basic parameters that define a sinusoidal signal are its fre- need to visualize anything ever again:
have a look at this video https://
quency, phase, and amplitude. We see them here plotted out and youtu.be/Ohp6Okk_tww to clarify the
expressed analytically. concept described.

v, i Figure 1: Sinusoidal functions have

three key parameters that define them,
ϕT/2π x (t) = X cos (ωt − ϕ) their amplitude, phase, and radial
frequency. These are typically con-
ceptualized relative to a cosine wave
for reasons discussed later in the text.
The temporal period and temporal
phase shift can be calculated from these
parameters using basic formulas.
t

T = 2π/ω
 6.200 supplementary notes: math of complex numbers and sinusoidal functions 2

Relating Outputs to Inputs

One key task of understanding how signals propagate through elec-

trical systems is relating the output to the input. Linear systems will
always increase the output proportionately when the input increases,
and shift the output signal phase according to the input signal phase.
Additionally, the signal output frequency will always equal the signal
input frequency. Thus the absolute choice of frequency, amplitude,
and phase is typically uninteresting to us. What interests us is how
the amplitude is increased or decreased and how the phase is shifted
by the system. Thus we need to develop a mathematics for mapping
between sinusoids.
We can quantify the mapping between two sinusoids with the
same frequency as shown in figure ??—as a shift in time and a scal-
ing of the y axis, thus we only need two parameters to specify any
sinusoid relative to any other sinusoid. See figure ??.

Figure 2: Any two sinusoids with

frequency ω are related by a time
shift (i.e. a relative phase) and a scale
change.

B
A
t

∆ϕ 2π
T

The mathematical description of this mapping can be expressed as

follows:
t→−ϕ/ω
A cos(ωt) =====⇒ B cos(ωt + ϕ) (1)
× B/A

where B/A provides the amplitude scaling and the time axis is
shifted by −ϕ/ω.
Such a notation is profoundly awkward—can you imagine having
to draw arrows all over the place with superscripts and subscripts
just to define a function? One might as well simply write it out as a
sentence... 2 2
If you ever feel bad about doing too
much algebra, you can console yourself
with the knowledge that originally,
algebra was described textually... as in:
“We find that our unknown variable
plus two will have the same value
as four of that unknown variable
squared.” Symbols weren’t used until
centuries later. Yikes!
 6.200 supplementary notes: math of complex numbers and sinusoidal functions 3

Polar Notation

Instead, we will be using polar notation for complex numbers exten-

sively. It will turn out that in polar notation, this type of mapping
is much easier than using the awkward notation in (??). Although
you may have used polar notation before, we will review it briefly to Im
ensure it is fresh in our minds for later discussion.
|e j ϕ |
In polar notation, complex numbers are described in terms of their
amplitude and phase, thus we write: ejϕ
ℑ[e j ϕ ] ∠e j ϕ
Re
Ae j ϕ ≡ A cos ϕ + j A sin ϕ (2)
ℜ[e j ϕ ]
√
where j = −1.
Figure 3: Polar notation provides a
Note that here ℑ{e j ϕ } = sin ϕ and ℜ{e j ϕ } = cos ϕ where the compact representation for complex
ℑ symbol represents the imaginary part, and the ℜ symbol the real numbers that is well suited to systems
part. Of course the |e j ϕ | term equals 1, but in general for an arbitrary undergoing rotary motion in the com-
plex plane. The complex coordinate is
complex number a + b j = Xe j ϕ , where X and ϕ are real numbers, called a “phasor” in analogy to a vector.
√
this magnitude would be X = a2 + b2 . In particular, the x (real) component of
the phasor will trace out a cosine wave
Combining the two concepts, we can describe the cosines as being if the angle ϕ increases linearly in time.
the real part of complex numbers, i.e. using a similar notation as
before in (??):

× B ejϕ ωt+ϕ
ℜ[|Ae{zj ωt}] ====
A
⇒ ℜ[|Be j{z } ]. (3)
Ṽin Ṽout

This form has the advantage that the transformation can be mapped
onto a single step, a multiplication by a complex number! So now
instead of having to think about things like scaling and time trans-
lation, we can just multiply by a complex number and take the real
part of both sides later:

A jϕ
Ṽout = Ṽin e (4)
B

vout = ℜ[Ṽout ]
B jϕ
= ℜ[Ṽin e ]
A
B (5)
= ℜ[ Ae j ωt e j ϕ ]
A
= ℜ[ Be j (ωt+ϕ) ]
= B cos (ωt + ϕ).
When complex numbers are expressed in polar notation, they are
typically called phasors. Figure ?? shows an example phasor drawn
on the complex plane.3 3
As far as I’m concerned, this is the
most awesome name in all of science—
one only hopes that Star Trek realized
how cool they were being when they
created the “phaser”.
 6.200 supplementary notes: math of complex numbers and sinusoidal functions 4

Polar vs Cartesian Coordinate Systems

It may seem at first that the use of polar coordinates serves little pur-
pose other than to confuse. After all, rectilinear coordinate systems
seem more natural to those familiar with an urban landscape, where
roads criss-cross and we grow accustomed to the cardinal directions
on the compass.4 But in practice, polar coordinates serve a key alge- 4
With apologies to residents of Can-
braic function: they make it easy... indeed trivial... to multiply and berra, Australia, which is layed out in a
polar geometery.
divide complex numbers. While your instinct may be to work with
complex numbers in the cartesian notation it is essential that you
learn to work with polar notation when multiplying or dividing
complex numbers.
Let’s consider the problem of trying to calculate the sum, differ-
ence, product, and quotient of two complex numbers:

X = Ae jϕA = a + bj (6)
jϕB
Y = Be = c + dj. (7)

The results of these calculations are evidently best calculated in one

or the other notation:

X + Y = ( a + c) + (b + d) j (8)
X − Y = ( a − c) + (b − d) j (9)
j ( ϕ A + ϕB )
X · Y = ABe (10)
X A
= e j ( ϕ A − ϕB ) . (11)
Y B
From these examples, it is evident that polar coordinates are best
used when multiplying or dividing complex numbers, while Carte-
sian coordinates should be used when summing or subtracting com-
plex numbers. Converting between the two notations is not too hard.
Given a complex number in Cartesian coordinates, X = a + bj, one
can derive the polar form by looking at the graphic below.
Im

√
| a + bj| = a2 + b2

a+bj

ℑ[ Ae j ϕA ] = b ∠( a + b j ) = tan−1 (b/a)

Re

ℜ[ Ae j ϕA ] = a
 6.200 supplementary notes: math of complex numbers and sinusoidal functions 5

By inspecting the graphic below, and then by using geometric and

√
trigonometric theorems one can show that a + bj = a2 + b2 e atan(b/a) .5 5
The one significant subtlety to this
definition of the polar coordinate sys-
tem is that care must be taken to get the
Various Representations of Sinusoids sign of the arctangent function correct.
Arctangent is typically defined only
between −π/2 and π/2, but of course
A variety of representations of sinusoids can be used, some of which one can easily encounter a complex
use complex numbers, and some of which are entirely real. They number in one of the a < 0 quadrants.
In this case, care must be taken to note
are all, of course, equivalent, but we prefer the complex notations the sign of a and select the correct
because they make the algebra easier to work with and provide im- quadrant of the complex plane. To
proved intuition about what is going on. We will describe them all handle such scenarios, the ∠ operator is
convenient, where ∠( a + b j ) provides
here, and finally describe how to convert between them. the arctangent with the assumption that
the quadrant of the complex plane was
chosen appropriately.
Summing Two Sinusoids Representation
It is somewhat surprising, and distinctly non-obvious, that any sinu-
soidal function centered around zero (i.e. without a DC offset) can be
expressed as the simple sum of a sine and a cosine function with the
same period and phase, but different amplitudes.
To demonstrate this mathematically, take an arbitrary sinusoidal
function of the form A cos (ωt + ϕ). Using the trigonometric identity
for summing angles,6 we find immediately: 6
cos (α + β) = cos α cos β − sin α sin β.

A1 cos (ωt + ϕ) = A1 cos ωt cos ϕ − A1 sin ωt sin ϕ (12)

= A1 cos ϕ cos ωt − A1 sin ϕ sin ωt (13)
= A2 cos ωt + A3 sin ωt (14)

where A2 = A1 cos ϕ and A3 = − A1 sin ϕ. This approach to expres-

sion of a sinusoidal function has advantages in certain situations, but
has the distinct disadvantage of not lending itself to easy interpreta-
tion from the point of view of signal amplitude and phase.

Negative Frequency Representation

A formulation of sinusoidal signals can be developed by using so-
called negative frequencies. Of course in real space all frequencies are
positive, but in the complex plane, signals that rotate clockwise are
termed negative frequency. This notation can be derived as follows.

 
e j (ωt+ϕ) + e − j (ωt+ϕ)
A1 cos (ωt + ϕ) = A1 (15)
2
A e j ϕ A e − j ϕ − j ωt
= 1 e j ωt + 1 e (16)
2 2
= A4 e j ωt + c.c. (17)
 6.200 supplementary notes: math of complex numbers and sinusoidal functions 6

where c.c represents the complex conjugation operation, and where

A4 = A1 e j ϕ /2, and we have used the fact that cos x = (e j x +
e − j x )/2.
Notice that although the final expression appears to be complex,
the sum of a complex number with its conjugate is always real, so in
fact it is real (as it should be).

Analytic Signal Representation

An alternative approach to dealing with sinusoidal signals in lin-
ear systems is to use the so-called “analytic signal” approach. This
approach is extremely common and popular in the electrical engi-
neering community. In this approach, the cos (ωt + ϕ) expression is
viewed as the real part of a complex exponential, i.e.:

A1 cos (ωt + ϕ) = ℜ[ A1 cos (ωt + ϕ) + A1 j sin (ωt + ϕ)] (18)

j ωt+ϕ
= ℜ[ A1 e ] (19)
= ℜ[ A1 e j ϕ e j ωt ] (20)
j ωt
= ℜ[ A5 e ] (21)

where A5 = A1 e j ϕ .

Comparing Representations of Sinusoidal Functions

There are 4 equivalent mathematical languages that one can use
when representing a sinusoidal function. One can translate between
them.

f (t) = A1 cos (ωt + ϕ) (22)

= A2 cos (ωt) + A3 sin (ωt) (23)
j ωt
= A4 e + c.c. (24)
= ℜ[ A5 e j ωt ]. (25)

It is convenient to be able to switch between notations at will. For

example, the cos(ωt + ϕ) notation is convenient for plotting and
visualizing, while the complex notations are convenient for working
the math. So being able to start in one notation, switch to another,
then switch back, is an invaluable skill. This is readily done by using
algebra. I have summarized the conversions in table ??. The columns
represent what you know, while the rows are the coefficients that
you’re seeking.
 6.200 supplementary notes: math of complex numbers and sinusoidal functions 7

Polar Cartesian Complex conj. Table 1: Conversion formulas between

the various coefficients provided in
Coefficient A1 , ϕ A ,A A4 A5
q 2 3 (??) through (??). We have used some
A1 A1 A22 + A23 2| A4 | | A5 | standard identities, and where one has
to be careful, as always, when taking
ϕ ϕ atan(− A3 /A2 ) ∠ A4 ∠ A5 the arctangent to make sure you are
A2 A1 cos ϕ A2 2ℜ[ A4 ] ℜ[ A5 ] in the correct quadrant of the complex
plane. The columns represent what you
A3 − A1 sin ϕ A3 −2ℑ[ A4 ] −ℑ[ A5 ] know, while the rows represent values
A1 j ϕ
A4 2 e A2 /2 − j A3 /2 A4 A5 /2 you are seeking.
A5 A1 e j ϕ A2 − jA3 2A4 A5

Exercises

Here we provide some exercises for yourself to check if you are un-
derstanding this material. Try not to refer to notes when working
these exercises–try to answer these questions just with paper and
pencil in front of you.

1. Prove that the sum of any two sinusoidal functions with the same
frequency results in another sinusoidal function with the same
frequency.

2. Derive the expression given in the table above for A4 and A5 in

terms of A2 and A3 .
√
3. Given cos (ωt + ϕ) = (cos(ωt) + sin(ωt)) / 2 solve for ϕ. Now
√
do the same thing for (− cos(ωt) + sin(ωt)) / 2

4. Plot the the values 1, 1 + j , j , −1 + j , −1, −1 − j , − j , and 1 −

j on the complex plane. Now write these expressions in their
equivalent polar coordinates.

Glossary and Definitions

Phasor Complex number expressed as a vector from the origin in the

complex plain.

Acknowledgements

Thanks to the many students, TAs, and colleagues who have con-
tributed to the formulation of the ideas in these notes, and to Rinske
Wijtmans for help formatting the final document.