Amplitude Modulation

1 Introduction
We are going to study Amplitude Modulation, a widely used family of continuous-wave modula-
tion system. The fundamental idea of AM is that a baseband or low-pass signal m(t) is used to
modulate a carrier signal which has a frequency much higher than the bandwidth of the signal
m(t). The message is the modulating waveform and the carrier is the modulated waveform.
This frequency-band shifting is a mandatory operation for various reasons. First, it allows us
to make the best use of the channel characteristics (bandwidth, noise level...). Technical reasons
also require the shift in frequency: there is a direct relationship between the frequency of a
signal and the size of the antenna that is needed to send or receive it. The higher the frequency,
the smaller the antenna. As you have probably noticed, the antenna of your FM stereo (87 Mhz
- 107 Mhz) is much bigger than the one of your cell phone (900 Mhz - 2400 Mhz).

2 A little bit of theory about AM

2.1 Definition of an AM signal
2.1.1 AM Large Carrier signal
An AM Large Carrier signal (AM-LC) S AM −LC (t) is defined as:

SAM −LC (t) = Ac (1 + Ka m(t))cos(2πfc t) (1)

| {z }| {z }
e(t) c(t)

where

• Ka is a constant.

• m(t) is the baseband message. For an audio signal, the spectrum of m(t) would typically
be in the range of 300 Hz - 3 kHz.

• fc is the carrier frequency (fc is usually set to 100 kHz with the TIMS module).

Notice that the signal SAM −LC (t) is expressed as the product of a low-frequency term e(t) and
a high-frequency term c(t). The low-frequency term e(t) contains a DC and an AC component.
The DC term will give rise to the carrier term in the AM-LC signal. The spectrum of an AM-LC
signal is shown on Figure 1.

1
 |M(f)| |S AM−LC (f)|
Carrier
Component

−W +W f −fc fc − W fc fc + W f

(a) Original message m(t) (b) AM-LC signal SAM −LC (t)

Figure 1: Spectrum of m(t) and SAM −LC (t)

2.1.2 Double Side Band Suppressed Carrier signal

A Double Side Band Suppressed Carrier signal (DSB-SC) S DSB−SC (t) is defined as:

SDSB−SC (t) = Ac m(t) cos(2πfc t) (2)

You can notice that the low-frequency term is the message signal m(t). There is no DC compo-
nent and therefore no carrier signal in S DSB−SC (t). This is why SDSB−SC (t) is referred to as
suppressed carrier. The spectrum of a DSB-SC signal is shown on Figure 2.

|M(f)| |S    (f)|
Upper
Lower Side Band
Side Band (USB)
(LSB)

 
+W f c c fc fc + W f

(a) Original message m(t) (b) DSB-SC signal SDSB−SC (t)

Figure 2: Spectrum of m(t) and SDSB−SC (t)

2.2 A useful example: the single tone message

Let us assume that m(t) = Am cos(2πfm t) where fc >> fm . The AM-LC signal s(t) has the
following form:

s(t) = Ac (1 + µ cos(2πfm t)) cos(2πfc t) (3)

where µ = Ka Am is called the modulation index or modulation factor.

2
 The modulation factor µ is generally expressed in percent and its magnitude can be measured
directly from the AM display itself using the following formula:

P −Q
µ= (4)
P +Q
where P and Q are defined on Figure 3.
AM−LC signal with µ = 50%
2.5

P
2

1.5

Q
1

0.5
s(t)

−0.5

−1

−1.5

0.1 0.15 0.2 0.25 0.3 0.35 0.4

Time t

Figure 3: Definition of P and Q for an AM-LC signal

2.3 AM generation
The most common way to generate AM signals is to multiply the carrier signal by the low-
frequency term. If an AM-LC signal is generated, the low-frequency term is 1 + K a m(t). If a
DSB-SC signal is generated, the low-frequency term is m(t). Figure 4 and Figure 5 show the
block-diagrams corresponding to both cases assuming a sine wave message.

Message

 
Sine wave (f m )

DC voltage Carrier sine wave

(f c )

Figure 4: AM-LC modulation

3
 Message


  
Sine wave (f m )

Carrier sine wave

(f c )

Figure 5: DSB-SC modulation

2.4 AM demodulation
In this Lab, we investigate two ways to demodulate AM signals: envelope detection and product
demodulation. Let us start with the envelope detection method.

2.4.1 Envelope detection

The envelope n(t) of the AM-LC signal S AM −LC (t) is defined as:

n(t) = |e(t)| = |Ac (1 + Ka m(t))| (5)

The envelope contains the information bearing message m(t). Strictly speaking, the message
should include the DC component but it is often ignored. At the receiver, electronic circuits
extract the envelope of the modulated signal, an operation called Envelope Detection (ED). An
envelope detector consists of a diode followed by a RC low-pass filter circuit as shown on Figure
6.

Diode

Received C R Demodulated
Signal Message

Figure 6: An envelope detector

The envelope detector output is then |e(t)|. However, we require that |e(t)| = e(t) so that
we can have direct access to the message signal m(t).

Therefore, the condition for correct demodulation is the following:

1 + Ka m(t) > 0 or |Ka m(t)| < 1 (6)

4
 If this condition is violated, s(t) is said to be overmodulated and it is no longer possible to
recover the original message using Envelope Detection. You will investigate the effect of overmod-
ulation on a speech signal and see how it affects its intellegibility. It is an interesting experiment.

We have seen that the ED works well when the envelope of the AM signal is proportional
to the message m(t) (for example when we want to demodulate an AM-LC signal with no
overmodulation). However, when the envelope is no longer proportional to m(t), which is the
case for DSB-SC signals, ED is irrelevant to demodulate. Therefore, we need another way to
demodulate: product demodulation.

2.4.2 Product demodulation

The main idea of product demodulation relies on the operation of frequency translation. The
frequency translation principle is the following: when a narrow-band signal r(t) is multiplied
with a cosine wave c(t) = cos(2πf1 t), the resulting signal y(t) = r(t)c(t) contains two signals
that correspond to the sum fc + f1 and the difference fc − f1 of the signal frequencies (see Figure
7 and Figure 8).

|R(f)|

c
fc f

Figure 7: Spectrum of the narrow-band signal r(t)

|Y(f)|

 f   f  f fc fc + f1 f

c 1 c c 1 c 1

Figure 8: Spectrum of the frequency translated signal y(t)

A band-pass filter can be used to select one component (the sum or the difference). The
output of the filter is the same signal as the original one but translated to another freqency.

A synchronous demodulator performs first a frequency translation operation on the received

signal r(t) by multiplying it with a sine wave c(t) = cos(2πf c t) tuned to the carrier frequency
fc . This is why the demodulator is said to be synchronous: the AM generator and the AM

5
demodulator use the same carrier frequency. The resulting signal y(t) = cos(2πf c t)r(t) has
therefore two components: the original baseband message and a high-frequency term centered
on the frequency 2fc . The signal y(t) is then passed to a low-pass filter that eliminates the
high-frequency term, allowing us to have access to message signal. The block diagram of a
synchronous demodulator is shown on Figure 9.

AM signal Low−pass filter Demodulated

Message

α Phase shifter

Carrier sine wave

(f c )

Figure 9: Synchronous product demodulator

As you can see on Figure 9, the synchronous demodulator contains a phase shifter. Phase
issues are also important when performing synchronous product demodulation. You will show
in the preliminary questions that optimal demodulation is achieved when the received signal
and the local carrier of the demodulator are in phase. In this case, the demodulator is said
to be coherent. From a spectral point of view, it means that the contribution to the output
of the Upper Sibe Band (USB) and Lower Side Band (LSB) are added coherently or in phase.
Therefore they reinforce each other. You will experiment that the original message can still be
recovered with non-coherent demodulation but there are some values of α that must be avoided
because they correspond to a zero output of the demodulator.

In real-life, the local oscillators used in AM transmitters and AM receivers are implemented
using crystals. Even if a manufacturer can guarantee that all produced crystals have a frequency
that lie within a certain range, they are not exactly the same. It means that the crystals used in
the modulator and in the demodulator will not have exactly the same frequency. The channel
also introduces a drift in frequency (dispersion, attenuation, Doppler effect...)
As a result of these different phenomena, the modulated signal that goes to the receiver
has a carrier frequency close but different of the one of the local oscillator used in the demod-
ulator. Therefore, we need to track these phase and frequency drifts: this operation is called
synchronization.

2.5 The Costas receiver

The Costas receiver is a practical synchronous and coherent receiver. The first step when doing
synchronization is to generate at the receiver a signal that has exactly the same frequency as the
received signal. There is no point in trying to cancel the phase difference between the received
AM signal and the signal that will be used to demodulate it if these signals do not even have
the same frequency.
Deriving the local carrier from the received signal itself is called carrier acquisition. Carrier

6
acquisition is performed using a Phase Locked Loop (PLL). A PLL includes a non-linear feedback
loop and its analysis is very complex. You can refer to your textbook for some additional
explanations. A simplified block-diagram of the PLL is shown on Figure 10. When the PLL is
”locked”, the output of the VCO has exactly the same frequency as the input signal.

Input signal Low−pass filter Output

(Loop filter)

VCO
Control voltage

Figure 10: A Phase Locked Loop

A lot of techniques exist to derive the carrier information from the received signal. However,
they all assume the existence of a component, however small, at carrier frequency. How can we
then demodulate DSB-SC signals? An analysis can show that, as a non-linear arrangement, it
is not so much the largest frequency component as the central frequency component to which
the PLL will lock. The amplitude of the central component need not be large (under some
conditions, it can even be zero). Non linearities will generate energy at the carrier frequency.
We will not spend more time on the PLL as you will use it to demodulate FM signals in the
“FM” Lab.

The Costas receiver is shown on Figure 11 and it consists of two coherent detectors supplied
with the same input signal. The two detectors have the structure of a PLL but their individual
local oscillators are in phase quadrature with each other. The upper path is referred to as the
In-phase coherent detector. The lower path is referred to as the Quadrature-phase coherent
detector.
The Costas receiver has not only the property to derive a carrier from the received signal
(even if there is no component at the carrier frequency in the received signal) but it is also
a coherent demodulator. It is typically used in Quadrature Carrier Multiplexing where the
received signal r(t) has the following form:

r(t) = Ac m1 (t) cos(2πfc t) + Ac m2 (t) sin(2πfc t) (7)

If the Costas receiver is fed with r(t), the I-channel output will be (1/2)A c m1 (t) and the
Q-channel output will be (1/2)Ac m2 (t).

3 Objectives
• Generate AM-LC signals and demodulate them using envelope detection.

• Generate DSB-SC signals and demodulate them using product demodulation. Investigate
the phase and frequency issues that can occur at the receiver.
• Implement the Costas receiver (a synchronous and coherent receiver)

7
 I branch

Low−pass filter
Output



   Loop filter

VCO

−π/2

 
Low−pass filter
Output
Q branch

Figure 11: The Costas receiver

4 Preliminary questions
1. Derive the Fourier transform of the AM-LC signal S AM −LC (t) defined as:

SAM −LC (t) = Ac (1 + Ka m(t))cos(2πfc t) (8)

| {z }| {z }
e(t) c(t)

2. Assume that we want to demodulate a DSB-SC signal r(t) whose equation is given by:

r(t) = Ac m(t) cos(2πfc t) (9)

where

• m(t) is the message signal

• fc is the carrier frequency
• Ac is the carrier amplitude

Demodulation of r(t) is performed with the synchronous product demodulator shown on

Figure 9. Assuming that the signal at the output of the phase shifter is cos(2πf c t + α),
derive the equation of the demodulated message.
Show that there is a value of α for which the output of the demodulator is zero. What is
the optimal value of α?

3. When performing envelope detection, the capacitor alternately charges during the rising
of the modulated waveform and discharges between consecutive peaks of the modulated
waveform. Using the notations of Figure 6, explain in a few lines the constraints to which
the following relation corresponds:

8
 1 1
<< RC << (10)
fc W
where W is the bandwidth of the message signal.

4. Derive the formula of the modulation factor µ:

P −Q
µ= (11)
P +Q
This formula is also valid for µ > 1, provided the magnitude of P and Q are interpreted
correctly. Can you see how P and Q are defined for µ > 1?

5 List of modules required

• 1 AUDIO OSCILLATOR

• 1 ADDER

• 3 MULTIPLIERS

• 1 UTILITIES

• 1 VCO

• 1 PHASE SHIFTER

• 2 TUNEABLE LPF

6 Experiments
6.1 AM generation
6.1.1 AM-LC signal generation
1. Patch up according to the block diagram shown on Figure 12.

G
Message AM signal
Sine wave (f m )
g

DC voltage Carrier sine wave

(f c )

Figure 12: AM-LC generation

9
 Output of the ADDER ⇒ CH1-A
Sine wave message ⇒ CH1-B
Scope settings:
AM-LC signal ⇒ CH2-A
Sine wave message ⇒ Ext. Trigger

The DC term is provided by the VARIABLE DC built-in module. Use the AUDIO OSCIL-
LATOR module to generate the sine wave message and the MASTER SIGNALS module
to generate the 100 kHz carrier.

2. Setting up µ = 100%
Use the FREQUENCY COUNTER to set the AUDIO OSCILLATOR to about 1 kHz.
Turn both G and g fully counter-clockwise. This removes both the DC and the AC parts
of the low-frequency term from the output of the ADDER .

3. Switch the scope selector to CH1-A. This is the ADDER output. Make sure that the scope
is set up to respond to DC. Turn the front panel control on the VARIABLE DC module
almost fully counter-clockwise. This will provide a voltage of about -2 V.

4. Rotate the gain g of the ADDER clockwise to adjust the DC term at the output of the
ADDER to a convenient positive value, say 2 V.

5. Whilst watching the scope trace of CH1-A rotate the ADDER gain control G clockwise.
Superimposed on the DC output of the ADDER will appear the message sinewave. Adjust
the gain until the lower crests of the sinewave are exactly coincident with the origin. The
AC and DC term of the low-frequency term have now exactly the same amplitude

6. Generating the AM signal

Make sure that the MULTIPLIER is switched to accept DC when you connect it to the
output of the ADDER and the MASTER SIGNALS module. You can select either the sine
wave or the cosine wave from the MASTER SIGNALS module as a carrier. In order to
trigger on the original message, you need to select External Trigger in the Trigger Source
settings of scope VI and select Edge as Trigger Type.

7. Display the AM signal on the scope and visualize also the spectrum display. Note that the
envelope of the signal is exactly that of the message waveform.

8. What happens when you change the frequency of the AUDIO OSCILLATOR? What hap-
pens when you modify the gain controls of the ADDER? Which parameters are you mod-
ifying?

9. Set up and demonstrate to your Lab instructor a 50% modulated waveform. Print your
graph.
Note: to print a specific window, select it and pull down the File menu and select Print
Window and click then on Print. If the display is not stable, you need to stop the VI to
see exactly what you are about to print and do as above.

10. Adjust the message frequency to about 5 kHz. Measure the amplitude of each peak in the
frequency domain. What is the total power in the sidebands? What is the total power in

10
 the carrier? Compute the modulation efficiency η defined as:
power in the sidebands
η = (12)
total power
power in the sidebands
= (13)
power in the sidebands + power in the carrier

11. Set up and demonstrate to your Lab instructor a 130% modulated waveform.

6.1.2 DSB-SC signal generation

1. The block diagram for DSB-SC signal generation is shown on Figure 13.

Message


  
Sine wave (f m )

Carrier sine wave

(f c )

Figure 13: DSB-SC generation

The scope settings are the same as for AM-LC generation.

2. Switch the MULTIPLIER to the AC position. This will remove all DC component from
the input signal of the MULTIPLIER. You can now visualize a DSB-SC signal.

3. Does the envelope have the same shape as the original message? Compare the spectrum of
the DSB-SC signal to the one of an AM-LC signal. From a ”power” point of view, DSB-
SC is said to be more efficient that AM-LC. Can you give an explanation for this statement?

Note on triggering issues

You have noticed that even if the envelope of the AM signal is stable, it is not the case
of the AM signal itself. It is drifting a little bit which prevents you from looking at fine
details in the signal. This a triggering issue that you would also have with a real scope.
The reason is that the rate 1/fm at which the scope sweeps is not proportional to the
period 1/fc of the AM signal, which makes each sweep start at a different point of the AM
signal, giving thus the impression that the DSB-SC signal is moving.
To make consecutive sweeps identical and therefore to have a nice stable display, it is
necessary that fm be a sub-multiple of fc . This special condition can be arranged with
TIMS by choosing the 2kHz MESSAGE sinusoid from the MASTER SIGNALS module.
The frequency of this signal is actually 100/48 kHz (approximately 2.08 kHz) , an exact
sub-multiple of the carrier frequency. You can use the 2kHz MESSAGE instead of the
AUDIO OSCILLATOR and see how the modulated message looks like.

11
6.2 AM Demodulation
Let us now demodulate the AM signals that we have generated.

6.2.1 Envelope detection

1. Patch up according to the block diagram shown on Figure 14.

AM signal Diode + LPF Demodulated signal

Figure 14: Envelope detector

Original message ⇒ CH1-A

AM-LC signal ⇒ CH1-B
Scope settings:
Demodulated message ⇒ CH2-A
Original message ⇒ Ext. Trigger

The Diode + LPF module is provided by the UTILITIES module. Its specification are
available in the TIMS User Manual. Use the AM-LC signal (sine wave message and
µ < 100%) as an input to your demodulator.

2. Display both the AM modulated message and the demodulated message. Make sure you
can visualize the moments during which the capacitor discharges. Print the display and
explain the relation between the two curves. Print the Fourier spectrum of the demodulated
message in the range 0 - 300 kHz. How can you explain the presence of components at
higher frequencies?

3. By varying the bandwidth of the message, determine what bandwidth message can effec-
tively be demodulated with this value of RC.

4. Connect the output of the Diode + LPF stage to the RC LPF stage of the UTILITIES
module. Visualize the signal before and after filtering. What has changed? If you were
asked, as an engineer, to design the RC LPF, how would you define the constraints on its
cut-off frequency?
Note on the envelope of the AM signal
It is important to note that the envelope always contains a DC component. This is due to
the mathematical operation of taking the absolute value of the low-frequency term in the
AM signal (it is a good exercise to prove this statement).
Usually, we do not pay a lot of attention to this DC term. However, it is a direct measure
of the strength of the carrier and it can be a very useful indication at the receiver. You can
visualize it by changing the carrier amplitude using the g GAIN control of the ADDER.

5. A speech signal is available at the TRUNKS PANEL module (channel # 1). Use this
signal as your message? Look at the scope and spectrum display. You can listen to the

12
 demodulated signal using the HEADPHONE AMPLIFIER built-in module (select the
position IN of the LPF SELECT switch). How does the speech signal sound when there
is overmodulation?

6.2.2 Product demodulation

Recall that product demodulation is used to demodulate signals (like DSB-SC signals) whose
envelope is not proportional to the original message.

Synchronous demodulation

A synchronous demodulator requires a local carrier that has exactly the same frequency as
the carrier from which the modulated signal was derived. A ”laboratory” way to implement
it is to use the same carrier (a stolen carrier) to modulate and to demodulate. In real-life
applications, the carrier frequency has to be derived from the received signal itself.

1. Patch according to the diagram shown on Figure 15.

AM signal

  Demodulated
Message

α Phase shifter

Carrier sine wave

(f c )

Figure 15: Synchronous product demodulator

Original message ⇒ CH1-A

DSB-SC signal ⇒ CH1-B
Scope settings: Signal before LPF ⇒ CH2-A
Demodulated message ⇒ CH2-B
Original message ⇒ Ext. Trigger

The low-pass filter is provided by the TUNEABLE LPF module. Use a DSB-SC signal as
an input for the demodulator.

2. What is the effect of α on the demodulated message? What about the Fourier domain?
Confirm that there is a value of α for which the amplitude of the demodulated message is
maximum and a value of α for which the output of the demodulator is zero.

3. Optional
Use the speech signal as the original message and listen to the demodulated message whilst
varying α.

13
Asynchronous demodulation

1. Use a Voltage Controled Oscillator (VCO) as a local carrier for the demodulator in the
previous setup. The VCO has to be adjusted around 100 kHz. Make sure that the
switch is on the HI position and that the on-board switch SW2 is on the VCO position.
Disconnect any input and connect the output to the FREQUENCY COUNTER. Adjust
the frequency f0 until the signal delivered by the VCO has a frequency around 100 kHz.
For this experiment, there is no need for a PHASE-SHIFTER but you can keep it if you
want to. It will not affect your observations.
2. Visualize the demodulated signal when the message is a single tone sine wave. Is the trace
of the demodulated message stable? What about the speech signal? How can you explain
your observations.
3. Change the frequency f0 of the VCO and investigate how it affects the demodulated
message.

6.3 The Costas receiver

Let us now implement a receiver that is used in real-life applications: the Costas receiver.
1. Setting up the Costas receiver
Before patching according to the block-diagram shown on Figure 16, follow these instruc-
tions:
• Set the two TUNEABLE LPF to their WIDE range and tune them to their widest
bandwidth. The two TUNEABLE LPF will be used to implement the low-pass filters
in the I branch and the Q branch.
• Use the RC LPF in the UTILITIES module to filter the control signal going to the
VCO.
• Before patching in the PHASE SHIFTER, set the on-board toggle switch to the HI
range. Adjust the phase shift to approximately 90 ◦ .
• Before inserting the VCO, make sure that the on-board switch SW2 is on the VCO
position. Select the HI frequency range with the front panel toggle switch.
I-channel output ⇒ CH1-A
DSB-SC signal ⇒ CH1-B
Scope settings:
Q-channel output ⇒ CH2-A
I-channel output ⇒ Ext. Trigger
Use the DSB-SC signal available at the TRUNKS PANEL (channel # 2) as the input of
the Costas receiver.
2. Detecting the lock of the loop
To detect the lock of the loop, the natural way consists in displaying the reference carrier
and the VCO output on the two channels of the scope and to synchronize the scope on
the reference carrier. The lock will then be indicated by the two plots becoming stable. In
real life, the reference carrier is not available and other methods have to be used to detect
the lock of the loop.

14
 I branch

   
Output



   Loop filter

VCO

−π/2


    

Output
Q branch

Figure 16: The Costas receiver

3. Disable the feedback loop by turning the GAIN control of the VCO fully counter-clockwise.

4. Use the FREQUENCY COUNTER to tune the VCO to within a few hundred Hertz of
100 kHz.

5. Visualize the I-channel and the Q-channel outputs. Slowly increase the VCO GAIN control
until the VCO locks to the DSB-SC carrier, as indicated by the scope traces becoming
stationary with respect to each other and by observing that the FREQUENCY COUNTER
now reads 100.00 kHz.

6. Observe the demodulated output from the filter of the I arm. If it is not exactly sinusoidal,
fine tune the VCO to get a cleaner signal. Notice that the frequency of the VCO will not
change as it is locked to the carrier. What is the output of the Q arm? Does this make
sense?

7. If time permits, use the Quadrature Carrier Multiplexing signal which is available on
Channel # 3 of the TRUNKS PANEL as an input to the Costas receiver. Identify each
signal.

8. Optional
Note: for this experiment, you need an extra PHASE SHIFTER. Ask your Lab instructor
if there is one available.
A technique of interest is to replace the VCO signal with a stolen carrier connected, via
a PHASE SHIFTER, into the loop. This simulates the locked VCO and allows static
observations of all points of the loop for various values of the phase angle α.

9. Use the DSB-SC signal available at channel # 2 of the TRUNKS PANEL as the input of
the Costas receiver. Replace the ”locked” VCO signal by a 100 kHz signal coming from

15
the MASTER SIGNALS module whose phase can be adjusted with a PHASE SHIFTER.
You can notice by changing α that the message appears at the output of both the I and
Q low-pass filters. Is there an optimal value of α? To which value does it correspond?

16