=
Try other values of modulation signal amplitude and measure A1 and A2 and thus
calculate m. Compare the values with the ratios of the modulation signal peak value to the
dc offset. Launch an Excel spreadsheet to tabulate your results
Note that the voltmeter reads peak to peak values.
Open the spectrum analyser and observe the spectrum of the modulated signal (monitor
point 4). Adjust the modulation amplitude using the Signal Level Control and observe the
Chapter 4
Modulation and Coding Principles Amplitude Modulation
53230
spectrum. Use the cursors to measure the relative levels of the two sidebands to the
carrier at m=1, 0.5 and 0.
Move the spectrum analyser probe (orange) to the modulation source (monitor point 3).
Measure modulating frequency using the cursor.
Return to the modulated output (monitor point 4) and measure the frequencies of the two
sidebands. Calculate the frequency difference between the carrier and the upper
sideband, and the carrier and the lower sideband.
Now measure the modulating frequency on the second channel. Compare the values.
Set the modulation index to 1 using the oscilloscope display.
Open the phasescope. Move the reference probe (yellow) to the carrier source (monitor
point 1) and the input probe (blue) to the modulated signal (monitor point 4). Note that the
display shows a signal with constant phase changing in amplitude between a radial point
to zero.
Change the modulation amplitude and note that the phase does not change but the
variation in amplitude does. What happens when the amplitude is zero?
Change the modulation source from the 62.5kHz Locked Sine Source to the Function
Generator. Set the Function Generator to Fast and the output to a sine wave. Adjust the
Frequency control and observe the spacing of the sidebands from the carrier on the
spectrum analyser.
Chapter 4
Modulation and Coding Principles Amplitude Modulation
53230
Practical 2: Demodulation with an Envelope Detector
Objectives and Background
Demodulation
Demodulation is the reverse process to modulation. In this case it takes the modulated
signal of a carrier and two sidebands and extracts the modulating signal from it. In this
instance this can be done very simply. If the modulated signal is passed through either a
full or half wave rectifier followed by a filter that passes only the modulation, then the
output follows the amplitude of the carrier. The resulting signal is the modulation plus the
dc offset, which can be removed. This type of demodulator is called an envelope
detector.
The diagram shows the output of the rectifier,
Note that the filter is usually simply a resistorcapacitor network. The time constant is
important, as it determines the magnitude of the residual carrier, or twicecarrier frequency
components. Mathematically, this demodulator can also be thought of as a multiplier that
takes the signal and multiplies it by its own carrier (this is because the carrier is switching
the diodes in the rectifier). The result of multiplying the two sidebands is the demodulated
signal and multiplying the carrier produces the dc offset in the output. In other words, the
envelope detector uses the carrier component to demodulate the signal.
In this Practical you will use the same setup as in Practical 1 to generate an AM double
sideband (dsb) signal and use an envelope detector to demodulate it. The envelope
detector uses a full wave rectifier, so that both positive and negative peaks of the carrier
contribute.
You will also see how the addition of a filter removes much of the twice carrier component.
Chapter 4
Modulation and Coding Principles Amplitude Modulation
53230
Block Diagram
Make Connections Diagram
Chapter 4
Modulation and Coding Principles Amplitude Modulation
53230
Chapter 4
Modulation and Coding Principles Amplitude Modulation
53230
Practical 2: Demodulation with an Envelope Detector
Perform Practical
Use the Make Connections diagram to make the connections on the hardware that are
required.
This practical uses the same AM generator circuit as Practical 1.
Use the voltmeter to set the dc Carrier offset to 0.25 volts. Use the oscilloscope and the
Signal Level Control to set the modulation index to approximately 0.5.
In the IQ Modulator block, set all of the controls to half scale.
Observe the output of the Envelope Detector (monitor point 5). Note that it reproduces
the modulating signal.
Chapter 4
Modulation and Coding Principles Amplitude Modulation
53230
Practical 3: Demodulation with a Product Detector
Objectives and Background
An alternative type of demodulator is called a product detector. As you saw in the
envelope detector, the demodulating process requires the modulated signal to be
multiplied by a frequency equal to, and in phase with, that of the carrier. In the envelope
detector the signal is simply multiplied by itself to achieve this. In the product detector, the
source of this signal is an external oscillator. This results in better demodulation because
the multiplying signal is not varying in amplitude and does not contain so much noise. The
action of such a demodulator is achieved by using a balanced modulator fed from an
oscillator. This oscillator is often referred to as a local oscillator as in a link it is, as viewed
from the receiver point of view, local as opposed to remote at the transmitting end. The
process of having a reference signal at the receiver at the same frequency and phase as
the original carrier is used in many of the demodulator processes.
Synchronising
Although it has been stated that the local oscillator needs to be on frequency and in
phase, so far it has not been explained how this is achieved. It is, in fact not easy. Other
assignments show how this can be achieved for other types of demodulators but in this
assignment, for simplicity, a sample of the original carrier is used to synchronise the local
oscillator. In practice, AM double sideband with full carrier is usually demodulated with
derivatives of the envelope detector.
As you will see later, the product detector is used for suppressed carrier single sideband
transmission. However, it is important that you understand that a product detector can be
used to detect full carrier AM.
The diagram shows the output of the multiplier,
You will also see how the addition of a filter can remove much of the twice carrier
component. A filter in this position is often referred to as a post detection filter.
Chapter 4
Modulation and Coding Principles Amplitude Modulation
53230
Block Diagram
Make Connections Diagram
Chapter 4
Modulation and Coding Principles Amplitude Modulation
53230
Chapter 4
Modulation and Coding Principles Amplitude Modulation
53230
Practical 3: Demodulation with a Product Detector
Perform Practical
Use the Make Connections diagram to make the connections on the hardware that are
required.
Again, you are using the same generator configuration to provide the AM waveform.
Use the voltmeter to make sure that the dc Carrier offset is set to 0.25 volts.
Use the Signal Level Control to set the modulation index to approximately 0.5.
In the IQ Modulator block, set all of the controls to half scale.
Use the oscilloscope to observe the output of the product detector (monitor point 5) and
the spectrum analyser to note the twice carrier frequency component. Observe the output
after the post detection filter (monitor point 6).
Chapter 5
Modulation and Coding Principles AM with Suppressed Carrier
53230
Amplitude Modulation with Suppressed Carrier
Objectives
To understand the concept of carrier suppression and its advantages
To investigate the spectrum of an amplitude modulated signal with suppressed carrier
To investigate demodulation of an amplitude modulated signal with suppressed carrier
To appreciate the advantages of single sideband suppressed carrier amplitude modulation
and to investigate its generation and demodulation
Chapter 5
Modulation and Coding Principles AM with Suppressed
Carrier
53230
Concepts of Modulation
A carrier is simply a single frequency of constant amplitude, phase and frequency. More
properly, this is called an unmodulated or plain carrier. In itself, it does not carry any
information. However, when referred to as an unmodulated carrier the implication is that
some information will be carried on it at some time. The carrier transports the information
to be carried, hence the name. As it is an oscillation it is sometimes also referred to as a
wave.
How is information to be carried? This information can be of many forms and can, by the
time it reaches the carrier, be either analogue or digital in form. Even if the information is
digital the process of transmission is analogue, because the real world is analogue. So, in
general, there is no difference between the processes involved in carrying analogue or
digital information. Information to be carried is often referred to as baseband. The reason
for this name will be come clearer later on.
In order to be decoded at the receiving end of a communications channel, some
characteristic of the carrier has to varied to represent differences in the baseband signal.
There are only three carrier characteristics that can be varied: its amplitude, its frequency
or its phase. Some schemes vary more than one of these characteristics and also, as you
will see, in some cases varying one will inadvertently vary another. So it is important not to
think of each in isolation.
The term modulation arises from the implication that some part of the carrier characteristic
is changing. When carrying information, the carrier is said to be modulated, and the sub
system responsible for doing this is called a modulator. The baseband information is
sometimes referred to as the modulation.
The opposite process to modulation is demodulation, in which the baseband signal is
recovered. The trick is to try and recover the baseband signal so that it is as near as
possible to the original, even when it has been severely weakened and distorted during
transmission.
Another consideration is to use as little transmission bandwidth as possible, so that as
many signals as possible can be sent down a cable or via a radio link as possible.
Transmission power is also important; usually the minimum that can be used to achieve a
usable output is desirable.
The concept of signaltonoise ratio will also be introduced and how it is a measure of the
quality of both the modulated and baseband signals.
The assignments will introduce all the modulation and demodulation concepts vital to an
understanding of information transmission.
Chapter 5
Modulation and Coding Principles AM with Suppressed Carrier
53230
Equations of Amplitude Modulation
The equation of a sinusoidal voltage waveform is given by :
v = V
max
.sin(t + )
where:
v is the instantaneous voltage
V
max
is the maximum voltage amplitude
is the angular frequency
is the phase
A steady voltage corresponding to the above equation conveys little information. To
convey information the waveform must be made to vary so that the variations represent
the information. This process is called modulation.
From the above equation, the basic parameters of such a waveform are:
its amplitude, V
max
its frequency, (or f)
its phase,
Any of these may be varied to convey information.
Amplitude Modulation
Amplitude modulation uses variations in amplitude (V
max
) to convey information. The wave
whose amplitude is being varied is called the carrier wave. The signal doing the variation
is called the modulation.
For simplicity, suppose both carrier wave and modulation signal are sinusoidal.
i.e.:
v
c
= V
c
sin
c
t (c denotes carrier) and
v
m
= V
m
sin
m
t (m denotes modulation)
We want the modulating signal to vary the carrier amplitude, V
c
, so that:
v
c
= (V
c
+ V
m
sin
m
t).sin
c
t
where (V
c
+ V
m
sin
m
t) is the new, varying carrier amplitude.
Expanding this equation gives:
v
c
= V
c
sin
c
t + V
m
sin
c
t. sin
m
t
which may be rewritten as
v
c
= V
c
[sin
c
t + m sin
c
t. sin
m
t]
Chapter 5
Modulation and Coding Principles AM with Suppressed
Carrier
53230
where m = V
m
/V
c
and is called the Modulation Index.
Now sin
c
t.sin
m
t = (1/2) [cos(
c
m
) t cos(
c
+
m
) t]
so, from the previous equation:
v
c
= V
c
[sin
c
t + m sin
c
t. sin
m
t]
we can express v
c
as:
v
c
= V
c
sin
c
t + (mV
c
/2) [cos(
c

m
) t] (mV
c
/2) [cos(
c
+
m
) t]
This expression for v
c
has three terms:
The original carrier waveform, at frequency
c
, containing no variations and thus carrying
no information
.A component at frequency (
c

m
), whose amplitude is proportional to the modulation
index. This is called the LOWER SIDE FREQUENCY.
A component at frequency (
c
+
m
), whose amplitude is proportional to the modulation
index. This is called the UPPER SIDE FREQUENCY.
It is the upper and lower side frequencies that carry the information. This is shown by the
fact that only their terms include the modulation index m. Because of this, the amplitudes
of the side frequencies vary in proportion to that of the modulation signal; the amplitude of
the carrier does not.
Sidebands
If the modulating signal is a more complex waveform, for instance an audio voltage from a
speech amplifier, there will be many side frequencies present in the total waveform.
This gives rise to components 2 and 3 in the last equation being bands of frequencies,
known as sidebands.
Hence we have the upper sideband and the lower sideband, together with the carrier.
Chapter 5
Modulation and Coding Principles AM with Suppressed Carrier
53230
Theory on Frequency Translation and Negative Frequencies
Translating from zero frequency
The modulation process can be thought of as that of frequency translation. The baseband
modulation is moved up in frequency by an amount equal to the carrier frequency.
Therefore zero Hz (i.e. dc) becomes the carrier frequency and the baseband becomes the
upper sideband.
From our observations that a dc offset in the modulation controls the amplitude of the
carrier, the upper sideband is spaced from the carrier by the modulation frequency and the
modulation amplitude controls the sideband amplitude this concept seems to have some
validity.
One major problem is: where does the lower sideband come from? It would appear to be
the result of a component on the other side of 0Hz (i.e. negative) and equal in amplitude to
the modulation. One interesting observation is that the power in the original modulation is
split equally between the upper and lower sideband, so the process cannot simply be the
result of turning the modulation into the upper sideband.
What is a negative frequency? Such a concept is being used as a tool to help model the
mathematics that explain how signal processing works. Modeling is becoming very
important now digital signal processing (DSP) is taking the place of analogue circuits in
many applications.
If you imagine a frequency to be generated by a rotating vector, then looking at it side on,
you can see the familiar sine wave. You would still see the same sine wave irrespective of
the direction the vector was rotating. However, it is the direction that actually determines if
the frequency is positive or negative. Now you can see why you cannot tell the difference,
looking simply at the sine wave from one side.
Suppose now you could look at the same rotating vector, still edge on, but now from
underneath. The result would be a cosine wave, but the relative signs of the two signals
(the sine and the cosine) would tell you which way the vector is rotating and hence if the
frequency is positive or negative.
Chapter 5
Modulation and Coding Principles AM with Suppressed
Carrier
53230
So if you have a conventional frequency described by a sine wave, you could view it as
actually containing both positive and negative frequencies in equal proportion. When you
translate this up, by multiplying it with a carrier, then the power is split equally between the
upper and lower sidebands exactly as would be expected.
If you wanted to generate only one sideband then you would have to make sure that the
original frequency contained only positive or negative frequencies. You do that by having
two sets of original signals, set at 90 degrees to each other, the relative signs of which tell
which direction the vectors are rotating in. If the outputs are then summed, only one
sideband will be generated. This is exactly what is done in Practical 2.
A similar problem occurs when a signal is demodulated or translated down in frequency. If
only one multiplier is used then both the upper and lower sidebands become a mixture of
positive and negative frequencies and simply appear a mixture at baseband. The way to
solve this is to translate with both sine and cosine versions of the local oscillator and
therefore have two baseband signals at 90 degrees. By suitable DSP processing any
signal can be demodulated. This is the basis of the zero IF receiver which you may learn
about later, and why IQ modulators and demodulators are so useful.
Chapter 5
Modulation and Coding Principles AM with Suppressed Carrier
53230
Theory on the Experimental Determination of the Modulation Index
This is most easily done by measuring the maximum and minimum values which the
instantaneous amplitude of the carrier reaches. Let us call these x and y.
Taking the equation for a sinusoidal carrier modulated by a sinusoidal waveform (see the
Modulation Maths Concept):
v
c
= V
c
[sin
c
t + m sin
c
t. sin
m
t]
and rearranging it, v
c
can be expressed as
v
c
= V
c
sin
c
t [1 + m sin
m
t]
so that the instantaneous amplitude of the carrier is
V
c
[1 + m sin
m
t]
Since sin w
m
t can vary between +1 and 1,
x = V
c
(1 + m) and y = V
c
(1 m).
To get the value of modulation index m from x and y, V
c
can be eliminated between these
equations by division, giving:
y/x = (1 m)/(1 + m).
Solving for m gives:
m = (x y)/(x + y)
Chapter 5
Modulation and Coding Principles AM with Suppressed
Carrier
53230
Practical 1: Double Sideband with Suppressed Carrier
Objectives and Background
Suppressed Carrier Transmission
It is fairly obvious that the sidebands carry all the modulation information and the carrier is
of constant amplitude and represents the dc offset. Of course, the envelope detector uses
the carrier to recover the modulating signal. The advantage of transmitting the carrier is
the simplicity of the system. However, the disadvantage is that a significant proportion of
the power in the modulated signal is in the carrier component that contains no information.
If the signal contained simply the sidebands then there would be significant power saving
which, for example in a large transmitter, might be tens of kilowatts.
How can the carrier be removed? As you may have guessed, by simply removing the dc
offset! When the offset is removed, the signals in the time domains look like this:
And in the frequency domain look like this:
Important things to note are:
In the time domain the carrier phase reversal with the sign of the modulating signal
The absence of the carrier in the frequency domain.
Chapter 5
Modulation and Coding Principles AM with Suppressed Carrier
53230
This type of modulator is often referred to as a balanced modulator. The signal it
generates is called double sideband suppressed carrier (DSBSC).
Chapter 5
Modulation and Coding Principles AM with Suppressed
Carrier
53230
Block Diagram
Make Connections Diagram
Chapter 5
Modulation and Coding Principles AM with Suppressed Carrier
53230
Chapter 5
Modulation and Coding Principles AM with Suppressed
Carrier
53230
Practical 1: Double Sideband with Suppressed Carrier
Perform Practical
Use the Make Connections diagram to make the required connections on the
hardware.
Open the oscilloscope and the spectrum analyser.
Set the modulation Signal Level Control to maximum.
Note that the signal on the spectrum analyser shows the two sidebands with the carrier
either missing or significantly reduced.
In the IQ Modulator block, adjust the lower balance control on the I modulator and see
that, when perfectly balanced, the carrier is removed completely.
Try adjusting the upper balance control of the I Modulator and see that it affects the
magnitude of the modulating frequency (62.5kHz) present in the output of the balanced
modulator. Leave the modulator adjusted for best carrier and modulation balance
(minimum amplitudes of each).
Use the cursor on the spectrum analyser to measure the spacing of the two sidebands
and confirm that it is twice the modulating frequency.
Examine the oscilloscope upper trace and observe the phase change between positive
and negative modulation peaks. You may need to expand the oscilloscope to do this. Do
not be concerned that this may be difficult to see as the phasescope will show it better.
Open the phasescope. Move the reference probe (green) to the carrier source. Note that
the phasescope now shows the phase reversal clearly. The magnitude of the modulated
signal varies between a radial value, down to zero amplitude and then to the same radial
value with opposite phase. Try adjusting the carrier balance (the lower control on the I
Modulator) and see the effect on the phasescope.
Adjust the modulation Signal Level Control and note the effects of changing the
modulation amplitude. Think about how the three instrument traces are related. Set the
modulation back to maximum amplitude.
Remove connection 5 and add connection 4 (refer to the Make Connections diagram). Set
the output of the Function Generator to Fast and select a sine wave.
Now adjust the Frequency of the modulation and observe the effect of the change in
modulation frequency (you will see an effect more noticeably as you approach the
maximum frequency). Confirm that the sideband spacing is twice the modulation
frequency.
Chapter 5
Modulation and Coding Principles AM with Suppressed Carrier
53230
Chapter 5
Modulation and Coding Principles AM with Suppressed
Carrier
53230
Practical 2: Demodulation of Double Sideband Suppressed Carrier
Objectives and Background
You should already appreciate that the simple envelope detector only works because the
carrier is present to act as the multiplying signal. Clearly this sort of detector will not work
on a suppressed carrier signal. However, you have already seen that a product detector
provides a locally produced carrier, the local oscillator.
You have also seen that a product detector can demodulate AM with full carrier. In this
assignment you will see that it can also demodulate suppressed carrier signals. It has to
be at the same frequency as the original carrier and, in this instance, also in the same
phase. Later, you will see a system that offers even greater power saving and only
requires the correct frequency.
In this practical you will investigate what happens when the local oscillator is not at the
correct frequency and then use a carrier reference to lock the local oscillator on frequency
and in phase.
Multiplying the modulated signal with the constant amplitude local oscillator results in a
demodulated output. It also requires a filter to remove the twice carrier frequency
component. This type of detector is often referred to as a product detector as it multiplies
the signal with a local oscillator.
You will also try use an envelope detector and confirm that without the carrier it will not
operate correctly.
Note that with DSBSC there is no dc offset in the output, but that a product detector works
equally well when the carrier is present (resulting in a dc offset).
Chapter 5
Modulation and Coding Principles AM with Suppressed Carrier
53230
Block Diagram
Make Connections Diagram
Chapter 5
Modulation and Coding Principles AM with Suppressed
Carrier
53230
Chapter 5
Modulation and Coding Principles AM with Suppressed Carrier
53230
Practical 2: Demodulation of Double Sideband Suppressed Carrier
Perform Practical
Use the Make Connections diagram to make the required connections on the hardware.
In this practical you use the same dsb generator to produce a double sideband
suppressed carrier signal as in Practical 1.
Open the oscilloscope and spectrum analyser and confirm that the output is that of a
double sideband suppressed carrier signal. You should make sure that the balanced
modulator (in the IQ Modulator block) balance controls are adjusted such that the carrier
and the modulation frequency are suppressed as much as possible. Also, in the IQ
Demodulator block, set all the controls to mid scale.
You have connected up two detectors to the output: a product detector and an envelope
detector.
Firstly, you will investigate the product detector.
Move the oscilloscope Channel 1 probe (blue) to the output of the low pass filter (monitor
point 5). Compare modulation and the demodulated output on the two traces.
It is possible that the two waveforms may be the same, but it is more likely that the output
waveform will be slightly off frequency and will vary in amplitude.
Open the voltmeter. Adjust the frequency of the local oscillator with the dc Source voltage
control and see that if you get the frequency just right, such that the output is of constant
amplitude (no beat frequency). You will find that it is very difficult to do and clearly this is
unsatisfactory. You will probably get nearest to constant amplitude when the control
voltage is approximately zero.
Now add connection 13. This provides a synchronising signal to the local oscillator
Carefully adjust the frequency of the local oscillator using the dc Source control. You
should find that you can achieve synchronisation, with the required adjustment being
much easier.
Now to investigate the envelope detector.
Move the oscilloscope Channel 1 probe (blue) to the output of the envelope detector
(monitor point 6). Note that the output does not contain the correct frequency at all. You
should be able to work out what frequency it is and why it is there.
This confirms that the envelope detector is unsuitable for suppressed carrier systems.
Chapter 5
Modulation and Coding Principles AM with Suppressed
Carrier
53230
Practical 3: Generating Single Sideband with Suppressed Carrier
Objectives and Background
You have already seen that the sidebands in the modulated signal carry all the information
and that removing the carrier still enables the original baseband signal to be recovered.
You have also seen that there are two sidebands, one either side of the carrier. In fact,
they carry the same information and thus only one needs to be present to successfully
demodulate the signal.
Such a signal, with only one sideband present, is called single sideband, or SSB. Using
SSB further reduces the power in the transmitted signal and halves the required
bandwidth.
The simplest way to generate an SSB signal is to use a balanced modulator to produce
double sideband suppressed carrier and then remove one sideband using a bandpass
filter, centred on the other sideband. This is the method you will use in the practical. This
method of SSB generation is known as the filter method.
There are other methods of SSB generation, but the filter method is the simplest to
understand and is in very common usage in communication systems. It may be necessary
for the bandpass filter to have a very good shape factor because, at normal carrier and
audio frequencies, the upper and lower sidebands are quite close in frequency.
Another consideration is that the bandpass filter should offer significant attenuation to the
carrier, so that the balanced modulator need not be so accurately balanced. In practice
the balanced modulator might provide 30 db of carrier suppression and the filter a further
10db. The other sideband would normally be about 30 to 40 db down on the wanted one.
In order to achieve this, the SSB filter has several poles and is, in most cases, a ceramic
filter or crystal filter. Various filters are commercially available, with different
specifications depending on the application.
In the practical you will use a high modulating frequency so you can see clearly the
relationship between the various frequency components. This means that the filter
specification can be relaxed and thus a filter made from tuned (LC) circuits is used.
Separate filters are provided for upper and lower sidebands and the means is provided to
monitor the output of both.
You might be surprised that the output from the SSB filters is simply a sinusoidal signal
but, since we use sinusoidal carrier and modulating frequencies, the sum or difference of
the two must be a single frequency.
Chapter 5
Modulation and Coding Principles AM with Suppressed Carrier
53230
Upper or Lower Sideband?
An obvious question is: which sideband should be transmitted? The answer owes more to
convention than theory!
There is no reason why one sideband gives better results than the other, but general
practice seems to favour the upper sideband.
One convention is that with carrier frequencies below 10 MHz the lower sideband should
be used, but this is not always the case. The result of this is that many pieces of
communication equipment have to be able to deal with both upper sideband and lower
sideband SSB signals.
Chapter 5
Modulation and Coding Principles AM with Suppressed
Carrier
53230
Block Diagram
Make Connections Diagram
Chapter 5
Modulation and Coding Principles AM with Suppressed Carrier
53230
Chapter 5
Modulation and Coding Principles AM with Suppressed
Carrier
53230
Practical 3: Generating Single Sideband with Suppressed Carrier
Perform Practical
Use the Make Connections diagram to make the required connections on the hardware.
This practical uses a balanced modulator to generate a suppressed carrier signal. It is
followed by a pair of bandpass filters. One has its passband set to pass the lower
sideband and the other to pass the upper sideband.
Open the oscilloscope and the spectrum analyser.
Set the modulation Signal Level Control for maximum modulation. Note that the output of
the modulator (monitor point 3) has both sidebands present as expected.
Make sure that the modulator is properly balanced.
Move the spectrum analyser probe (orange) to the upper sideband filter output (monitor
point 4). Note that lower sideband has been significantly attenuated.
Move the spectrum analyser probe to the lower sideband filter (LSB) output (monitor point
5) and note that the upper sideband is attenuated.
The ratio between the wanted and unwanted sideband is called the sideband suppression.
How good this suppression is depends entirely on the quality of the filter. Measure the
sideband suppression in this case.
Notice that the amplitude of the wanted sideband at the output of the filter is also reduced
from that at the filter input. This is caused by the passband loss of the filter. All filters
introduce an undesirable loss, how much depends on the filter complexity and quality.
Move the oscilloscope Channel 2 probe (yellow) to the upper sideband filter (USB) output.
Note that it consists mainly of one frequency. This should be no surprise given the
spectrum analyser display. The amplitude ripple is caused by the residual carrier and
other sideband.
Refer back to the Make Connections diagram and remove connection 5 and add
connection 4. This will allow you to change the modulation frequency.
Set the Function Generator to Fast and select a sine wave.
By adjusting the Frequency you can see the relationship between the frequency of the
modulation and the frequencies of the sidebands. You will need to be close to maximum
frequency to see the effect.
If you enable the spectrum analyser second channel you can display both filter outputs at
the same time.
Chapter 5
Modulation and Coding Principles AM with Suppressed Carrier
53230
Chapter 5
Modulation and Coding Principles AM with Suppressed
Carrier
53230
Practical 4: Demodulating Single Sideband with Suppressed Carrier
Objectives and Background
You should now be familiar with the idea of a product detector and appreciate its ability to
demodulate a suppressed carrier signal.
It can also demodulate a signal with only one sideband. One important difference is that
the phase of the local oscillator no longer needs to be locked to that of the original carrier.
This is because any phase error has opposite effects on the lower and upper sidebands
which means that when both are present they do not combine to produce the baseband
signal.
If only one sideband is present then a local oscillator phase error only results in a phase
shift of the base band which is usually unimportant. Of equal importance is the fact that a
local oscillator frequency error causes similar problems when both sidebands are present.
With only one present, the effect of a local oscillator frequency error is an equal error in
the baseband signal frequency. Providing this is small it is not usually important.
This means that generating a local oscillator signal is much easier.
In this practical you will generate a SSB signal and then demodulate it with a product
detector. You will also see the effect of a local oscillator frequency error.
Chapter 5
Modulation and Coding Principles AM with Suppressed Carrier
53230
Block Diagram
Make Connections Diagram
Chapter 5
Modulation and Coding Principles AM with Suppressed
Carrier
53230
Chapter 5
Modulation and Coding Principles AM with Suppressed Carrier
53230
Practical 4: Demodulating Single Sideband with Suppressed Carrier
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
You are using the SSB generator configuration that you used in Practical 3, but adding to
it a demodulator.
The Practical uses a pair of demodulators: one for each sideband. Each demodulator
consists simply of a multiplier fed with a signal from a local oscillator. If the local oscillator
is at the frequency of the original carrier then both sidebands are mixed down to the same
frequency: the original modulation. Note that, in practice, only one of the sidebands is
transmitted and demodulated. Both are shown in this Practical so you can see that it does
not matter which one is used.
Ensure that the controls associated with the IQ Modulator and the two Multiplier circuits
being used are set to their mid positions.
Open the oscilloscope, the spectrum analyser and the voltmeter.
Set the Function Generator to Fast and select a sine wave. Open the frequency counter
and set the modulation frequency to approximately 65kHz, using the Frequency control of
the Function Generator.
Use the Signal Level Control to set the modulation to maximum. Note that the outputs
from the two sideband filters are essentially single frequency sine waves (monitor points 3
and 4).
Use the dc Source control and the voltmeter to set the Local Oscillator frequency dc
control voltage to zero, which should set the local oscillator to approximately that of the
carrier.
Use the IQ Modulator balance controls and the spectrum analyser to ensure that the
modulator is balanced correctly.
Move the oscilloscope Channel 1 probe (blue) to the outputs of the post detection filters
(monitor points 5 and 6), in turn, and note that they are approximately equal and of the
same frequency as the modulation. Try changing the modulation frequency and see what
happens. Note that, if you move outside the passband of the two sideband filters the
signal will disappear.
Put the oscilloscope Channel 1 probe (blue) on the upper sideband output (monitor point
5) and the Channel 2 probe (yellow) on the lower sideband (monitor point 6). Now change
the local oscillator control voltage (using the dc Source control) and note that the
frequency changes but in opposite directions. Note, therefore, that an error in local
oscillator frequency results in simply an equal error in output frequency.
Chapter 5
Modulation and Coding Principles AM with Suppressed
Carrier
53230
You can see this effect clearly by moving the spectrum analyser Channel 1 probe (orange)
to one output and the second channel probe (green) to the other. Enable the spectrum
analyser Ch2 (Ch2 Show) and also Alias Hi. Change the local oscillator frequency and
observe the result. You will need to lower the frequency of the spectrum analyser to see
the effect better.
Chapter 6
Modulation and Coding Principles SSB Generation using an IQ Modulator
53230
SSB Generation using an IQ Modulator
Objectives
To appreciate that a single sideband suppressed carrier signal may be produced using
phasing, rather than filtering, methods
To investigate the concept of the generation of a single sideband suppressed carrier
signal using an IQ modulator
Chapter 6
Modulation and Coding Principles SSB Generation using
an IQ Modulator
53230
Equations of Amplitude Modulation
The equation of a sinusoidal voltage waveform is given by :
v = V
max
.sin(t + )
where:
v is the instantaneous voltage
V
max
is the maximum voltage amplitude
is the angular frequency
is the phase
A steady voltage corresponding to the above equation conveys little information. To
convey information the waveform must be made to vary so that the variations represent
the information. This process is called modulation.
From the above equation, the basic parameters of such a waveform are:
its amplitude, V
max
its frequency, (or f)
its phase,
Any of these may be varied to convey information.
Amplitude Modulation
Amplitude modulation uses variations in amplitude (V
max
) to convey information. The wave
whose amplitude is being varied is called the carrier wave. The signal doing the variation
is called the modulation.
For simplicity, suppose both carrier wave and modulation signal are sinusoidal.
i.e.:
v
c
= V
c
sin
c
t (c denotes carrier) and
v
m
= V
m
sin
m
t (m denotes modulation)
We want the modulating signal to vary the carrier amplitude, V
c
, so that:
v
c
= (V
c
+ V
m
sin
m
t).sin
c
t
where (V
c
+ V
m
sin
m
t) is the new, varying carrier amplitude.
Expanding this equation gives:
v
c
= V
c
sin
c
t + V
m
sin
c
t. sin
m
t
which may be rewritten as
Chapter 6
Modulation and Coding Principles SSB Generation using an IQ Modulator
53230
v
c
= V
c
[sin
c
t + m sin
c
t. sin
m
t]
where m = V
m
/V
c
and is called the Modulation Index.
Now sin
c
t.sin
m
t = (1/2) [cos(
c
m
) t cos(
c
+
m
) t]
so, from the previous equation:
v
c
= V
c
[sin
c
t + m sin
c
t. sin
m
t]
we can express v
c
as:
v
c
= V
c
sin
c
t + (mV
c
/2) [cos(
c

m
) t] (mV
c
/2) [cos(
c
+
m
) t]
This expression for v
c
has three terms:
The original carrier waveform, at frequency
c
, containing no variations and thus carrying
no information
.A component at frequency (
c

m
), whose amplitude is proportional to the modulation
index. This is called the LOWER SIDE FREQUENCY.
A component at frequency (
c
+
m
), whose amplitude is proportional to the modulation
index. This is called the UPPER SIDE FREQUENCY.
It is the upper and lower side frequencies that carry the information. This is shown by the
fact that only their terms include the modulation index m. Because of this, the amplitudes
of the side frequencies vary in proportion to that of the modulation signal; the amplitude of
the carrier does not.
Sidebands
If the modulating signal is a more complex waveform, for instance an audio voltage from a
speech amplifier, there will be many side frequencies present in the total waveform.
This gives rise to components 2 and 3 in the last equation being bands of frequencies,
known as sidebands.
Hence we have the upper sideband and the lower sideband, together with the carrier.
Chapter 6
Modulation and Coding Principles SSB Generation using
an IQ Modulator
53230
Theory on Frequency Translation and Negative Frequencies
Translating from zero frequency
The modulation process can be thought of as that of frequency translation. The baseband
modulation is moved up in frequency by an amount equal to the carrier frequency.
Therefore zero Hz (i.e. dc) becomes the carrier frequency and the baseband becomes the
upper sideband.
From our observations that a dc offset in the modulation controls the amplitude of the
carrier, the upper sideband is spaced from the carrier by the modulation frequency and the
modulation amplitude controls the sideband amplitude this concept seems to have some
validity.
One major problem is: where does the lower sideband come from? It would appear to be
the result of a component on the other side of 0Hz (i.e. negative) and equal in amplitude to
the modulation. One interesting observation is that the power in the original modulation is
split equally between the upper and lower sideband, so the process cannot simply be the
result of turning the modulation into the upper sideband.
What is a negative frequency? Such a concept is being used as a tool to help model the
mathematics that explain how signal processing works. Modeling is becoming very
important now digital signal processing (DSP) is taking the place of analogue circuits in
many applications.
If you imagine a frequency to be generated by a rotating vector, then looking at it side on,
you can see the familiar sine wave. You would still see the same sine wave irrespective of
the direction the vector was rotating. However, it is the direction that actually determines if
the frequency is positive or negative. Now you can see why you cannot tell the difference,
looking simply at the sine wave from one side.
Suppose now you could look at the same rotating vector, still edge on, but now from
underneath. The result would be a cosine wave, but the relative signs of the two signals
(the sine and the cosine) would tell you which way the vector is rotating and hence if the
frequency is positive or negative.
Chapter 6
Modulation and Coding Principles SSB Generation using an IQ Modulator
53230
So if you have a conventional frequency described by a sine wave, you could view it as
actually containing both positive and negative frequencies in equal proportion. When you
translate this up, by multiplying it with a carrier, then the power is split equally between the
upper and lower sidebands exactly as would be expected.
If you wanted to generate only one sideband then you would have to make sure that the
original frequency contained only positive or negative frequencies. You do that by having
two sets of original signals, set at 90 degrees to each other, the relative signs of which tell
which direction the vectors are rotating in. If the outputs are then summed, only one
sideband will be generated. This is exactly what is done in Practical 2.
A similar problem occurs when a signal is demodulated or translated down in frequency. If
only one multiplier is used then both the upper and lower sidebands become a mixture of
positive and negative frequencies and simply appear a mixture at baseband. The way to
solve this is to translate with both sine and cosine versions of the local oscillator and
therefore have two baseband signals at 90 degrees. By suitable DSP processing any
signal can be demodulated. This is the basis of the zero IF receiver which you may learn
about later, and why IQ modulators and demodulators are so useful.
Chapter 6
Modulation and Coding Principles SSB Generation using
an IQ Modulator
53230
Theory on the Experimental Determination of the Modulation Index
This is most easily done by measuring the maximum and minimum values which the
instantaneous amplitude of the carrier reaches. Let us call these x and y.
Taking the equation for a sinusoidal carrier modulated by a sinusoidal waveform (see the
Modulation Maths Concept):
v
c
= V
c
[sin
c
t + m sin
c
t. sin
m
t]
and rearranging it, v
c
can be expressed as
v
c
= V
c
sin
c
t [1 + m sin
m
t]
so that the instantaneous amplitude of the carrier is
V
c
[1 + m sin
m
t]
Since sin w
m
t can vary between +1 and 1,
x = V
c
(1 + m) and y = V
c
(1 m).
To get the value of modulation index m from x and y, V
c
can be eliminated between these
equations by division, giving:
y/x = (1 m)/(1 + m).
Solving for m gives:
m = (x y)/(x + y)
Chapter 6
Modulation and Coding Principles SSB Generation using an IQ Modulator
53230
Practical 1: Generating SSB with an IQ Modulator
Objectives and Background
The generation of an SSB signal can be achieved by a number of methods. They fall into
two categories: filtering out the unwanted sideband with a bandpass filter, or by using
phase to cancel it out.
The filter method has been used extensively because it is quite easy to do, reliable and
controllable, even though the filter performance requirement is quite exacting.
There are several variations on the phase method, but all use a 90 degree shift to achieve
cancellation. These methods have no requirement for a filter and therefore sound quite
attractive. However, producing signals of equal amplitude but with 90 degree phase shifts
is quite difficult.
The simplest of the phasing methods requires two carriers at 90 degrees, fed to two
balanced modulators. The two modulation signals at 90 degrees feed the two modulators
and the outputs are combined. This generates SSB. Which sideband is cancelled
depends on the polarities of the phase shifts.
The generation of two single frequency carriers at 90 degrees is quite simple.
In general, modulation comprises a band of frequencies. Maintaining a constant amplitude
and a 90 degree phase difference over this band is extremely difficult with conventional
analogue circuits. For this reason, the various phasing methods have not been widely
used. Now that digital signal processing (DSP) is available, with its ability to generate such
signals, phasing methods are more practicable.
In this practical you will be using analogue methods to demonstrate the principle. The two
modulators, fed with two 90 degree carriers, are used in many modulation schemes and
are referred to as IQ modulators. This is from the terms In phase and Quadrature. The
word quadrature means at 90 degrees, from the mathematical term quadrant.
The modulation is generated in a special way with a circuit that is also used in other
assignments. The circuit comprises an integrator circuit and a sine/cosine angle generator.
In this assignment it is not important for you to understand how this works. The modulation
frequency control is by a dc voltage and the outputs of the sine/cosine generator are two
sine waves of equal amplitude and differing in phase by 90 degrees. If the dc voltage is
exactly zero the frequency output is zero. With a positive dc voltage applied the cosine
output leads the sine, and for negative voltages the sine leads the cosine.
You will notice that the two sine waves are not perfect, particularly at higher frequencies.
This illustrates well the difficulties of implementing this type of process using analogue
components.
In the practical you will see that generating an SSB signal can be achieved without the use
of filters.
Chapter 6
Modulation and Coding Principles SSB Generation using
an IQ Modulator
53230
Chapter 6
Modulation and Coding Principles SSB Generation using an IQ Modulator
53230
Block Diagram
Make Connections Diagram
Chapter 6
Modulation and Coding Principles SSB Generation using
an IQ Modulator
53230
Chapter 6
Modulation and Coding Principles SSB Generation using an IQ Modulator
53230
Practical 1: Generating SSB with an IQ Modulator
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the Integrator to Fast. Open the voltmeter and set the dc Source voltage to
maximum. This voltage controls the frequency of the modulation.
Open the oscilloscope and note the two signals at 90 degrees. Adjust the dc Source
voltage see the frequency change. Confirm the phase difference changes sign with
negative dc control voltage input.
Open the phasescope and confirm the phase difference on the modulating signal (it may
not be exactly 90 degrees, due to circuit component tolerances).
Now to examine the phase relationship between the two carrier inputs.
Move the oscilloscope Channel 2 probe (yellow) to the Carrier Source output (monitor
point 1). Notice that this is also the reference channel for the phasescope. Place the
oscilloscope Channel 1 probe (blue) on the 45 degree Phase Shift output (monitor point
2). You will have to change the timebase and use the X expand button on the oscilloscope
to see the phase difference clearly. Adjust the Variable Phase Shift control on the Carrier
Source variable phase block to obtain 45 degrees on the phasescope.
Now, move the oscilloscope Channel 1 probe to the +45 degree output (monitor point 3).
A phase of +45 degrees should be seen.
Just to check, move the yellow probe to monitor point 2. The total 90 degrees difference
between the two carrier inputs should now be seen.
Move the two probes back to the modulation blue on monitor point 5 and yellow on
monitor point 6.
Open the spectrum analyser and note that the output is one frequency, as you would
expect for SSB with single tone modulation.
Move the green probe to the carrier source and switch on the spectrum analyser channel
2 (Ch2 Show). This now acts as a reference frequency to show where the carrier would
be. Note that the lower sideband has been removed. You will probably have to select
Alias Hi and lower the frequency on the spectrum analyser to see this clearly.
Change the modulation frequency by adjusting the dc Source voltage. Note that the SSB
output frequency changes.
Momentarily, move the green probe back to monitor point 7 and adjust the dc Source to
zero volts. Move it back to monitor point 1 and see how the SSB output frequency
compares with the carrier frequency.
Chapter 6
Modulation and Coding Principles SSB Generation using
an IQ Modulator
53230
Now adjust the dc Source to give you a negative voltage. See that the upper sideband is
now removed. Note what happens to the phase on the phasescope.
Move the blue and yellow probes back to monitor points 2 and 3, respectively, to display
the phase between the two carriers on the phasescope.
Deselect the spectrum analyser second channel and try adjusting the carrier phase (with
the Variable Phase Shift control associated with the Carrier Source block). Note that only
at 90 degrees is the other sideband cancelled completely. Also try adjusting the carrier
balance controls on the modulators and note that this affects the cancellation of the
carrier.
Now, refer back to the Make Connections diagram and try removing connection 9 or 10 so
that only one of the balanced modulators is working. You should see that the signal
reverts to double sideband suppressed carrier. Notice that the wanted sideband increases
in amplitude when the other disappears. Measure the difference and try to account for it.
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
Amplitude Shift Keying
Objectives
To appreciate the principle of amplitude shift keying and its relationship to analogue
amplitude modulation
To understand the terms bit rate and symbol rate associated with digitally modulated
signals
To generate a twolevel (binary) amplitude shift keyed signal and investigate the spectrum
and bandwidth associated with it
To investigate multilevel ASK
To investigate the demodulation of an ASK signal
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
Equations of Amplitude Modulation
The equation of a sinusoidal voltage waveform is given by :
v = V
max
.sin(t + )
where:
v is the instantaneous voltage
V
max
is the maximum voltage amplitude
is the angular frequency
is the phase
A steady voltage corresponding to the above equation conveys little information. To
convey information the waveform must be made to vary so that the variations represent
the information. This process is called modulation.
From the above equation, the basic parameters of such a waveform are:
its amplitude, V
max
its frequency, (or f)
its phase,
Any of these may be varied to convey information.
Amplitude Modulation
Amplitude modulation uses variations in amplitude (V
max
) to convey information. The wave
whose amplitude is being varied is called the carrier wave. The signal doing the variation
is called the modulation.
For simplicity, suppose both carrier wave and modulation signal are sinusoidal.
i.e.:
v
c
= V
c
sin
c
t (c denotes carrier) and
v
m
= V
m
sin
m
t (m denotes modulation)
We want the modulating signal to vary the carrier amplitude, V
c
, so that:
v
c
= (V
c
+ V
m
sin
m
t).sin
c
t
where (V
c
+ V
m
sin
m
t) is the new, varying carrier amplitude.
Expanding this equation gives:
v
c
= V
c
sin
c
t + V
m
sin
c
t. sin
m
t
which may be rewritten as
v
c
= V
c
[sin
c
t + m sin
c
t. sin
m
t]
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
where m = V
m
/V
c
and is called the Modulation Index.
Now sin
c
t.sin
m
t = (1/2) [cos(
c
m
) t cos(
c
+
m
) t]
so, from the previous equation:
v
c
= V
c
[sin
c
t + m sin
c
t. sin
m
t]
we can express v
c
as:
v
c
= V
c
sin
c
t + (mV
c
/2) [cos(
c

m
) t] (mV
c
/2) [cos(
c
+
m
) t]
This expression for v
c
has three terms:
The original carrier waveform, at frequency
c
, containing no variations and thus carrying
no information
.A component at frequency (
c

m
), whose amplitude is proportional to the modulation
index. This is called the LOWER SIDE FREQUENCY.
A component at frequency (
c
+
m
), whose amplitude is proportional to the modulation
index. This is called the UPPER SIDE FREQUENCY.
It is the upper and lower side frequencies that carry the information. This is shown by the
fact that only their terms include the modulation index m. Because of this, the amplitudes
of the side frequencies vary in proportion to that of the modulation signal; the amplitude of
the carrier does not.
Sidebands
If the modulating signal is a more complex waveform, for instance an audio voltage from a
speech amplifier, there will be many side frequencies present in the total waveform.
This gives rise to components 2 and 3 in the last equation being bands of frequencies,
known as sidebands.
Hence we have the upper sideband and the lower sideband, together with the carrier.
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
Intersymbol Interference
Intersymbol interference is a particular type of distortion applicable to digital signals. It
simply refers to the fact that the present symbol may be distorted by the values of the
symbols on either side of it.
For example, if a post detection filter had insufficient bandwidth and the signal did not
have time to reach its maximum output during a 1 symbol, if the previous symbol was
zero, then this would be regarded as intersymbol interference.
More subtle problems may occur if there are reflections in a cable, or on radio signals,
causing energy from other symbol periods to arrive at the same time.
All communication systems use filtering to maximize the signaltonoise ratio or prevent
other signals causing interference. Any filtering will cause some intersymbol interference
and it is necessary to find the right compromise between too little filtering and too much
distortion. Some systems, such as GMSK (Gaussian Minimum Shift Keying), are designed
to tolerate significant distortion, in order to reduce their occupied bandwidth.
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
Symbol Rate and Bit Rate
The concepts of symbols, bits, symbol rate and bit rate are important terms in digital
communications.
The concept of a bit (a binary digit) should be familiar as a one or zero in a binary data
stream. The bit rate is simply the rate at which the bits change. For example, imagine a
system that digitized an audio signal at 32k samples per second, each sample being
digitized at 256 possible levels. This means each sample is an 8 bit word. In order to send
this stream over a simple link it would have to be turned into serial data. This means the
serial data stream would run at 32k x 8 = 256k bits per second. This is the bit rate. In this
example we are assuming that there is no extra data for synchronization or for error
correction.
These bits are then modulated onto the carrier in some form. In order to be modulated
they have to be converted to change some parameter of the carrier: its amplitude,
frequency or phase. In a simple system there would be only two states: off or on, one
frequency or the other, one of two phases etc. These states are called symbols.
In the simplest binary system there are only two symbols and each bit has two possible
states so the bits are directly mapped to symbols. This means that the symbol rate is
equal to the bit rate.
There is no reason why there have to be only two possible carrier states. In an amplitude
shift keying (ASK) system there could be more than two possible amplitude states, or in
phase shift keying (PSK) system there could be other possible phases than zero and 180
degrees. If there you had a PSK system with four possible states then each transmitted
data symbol can be decoded as being one of four states. Therefore, not one but two bits
can be carried per symbol. Now, if the bit rate remains the same, we only need to transmit
symbols at half the rate. In such a system the symbol rate is half the bit rate. If there were
16 symbols available then 4 bits per symbol could be carried and the symbol rate would
be one quarter the bit rate. Such systems are called Mary , where M is the number of
possible symbols, sometimes referred to as the order of the modulation scheme.
In such a system the bit rate (B) is:
M S B
2
log =
where S is the symbol rate and M the number of possible symbols.
To avoid confusion this bit rate is sometimes called the gross bit rate
It is important to remember that it is the symbol rate that is the rate at which the carrier
changes state. Therefore, it determines the occupied bandwidth.
It is clear that for a given bandwidth, the higher the order of the modulation scheme the
less bandwidth is used. However there is a penalty to be paid. When demodulated, the
higher the order of the scheme the more likely there are to be errors. This is obvious
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
because, for example, it is clearly easier to detect the difference between 0 and 180
degrees than zero, 90, 180, and 270.
There is another compromise to be made if error correcting data is added in that, although
adding extra data reduces the number of errors, the bit rate has to rise, with a
consequential increase in occupied bandwidth and received noise.
In order to calculate the amount of useful data that can be transmitted through a digital
system, first find the symbol rate. Then calculate the bit rate by using the number of bits
per symbol. The useful data, sometimes referred to as the payload, can then be
calculated by subtracting the extra data added for error correction, data identification and
synchronisation.
In a multiplexed system more than one data stream may be present and you may have to
find out what proportion of the data stream is allocated to a particular set of data. In very
complex systems this proportion may not even be constant!
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
Sampling
Signals in the real the world are analogue. In a digital communications system the first
process is to turn these analogue signals into digital format.
The signals could be anything: speech, television or representing the pH of a liquid, for
example. However, the common factor linking analogue signals is that they are time
continuous. This means that they are varying in time in a smooth manner. The diagram
shows a typical time continuous varying signal.
A digital signal is a series of discrete numbers that describes the signal, where each
number represents the signal at a particular point in time. This means that analogue signal
has to be sampled at various points in time and each value converted to a digital
number. This concept of sampling is very important to understand.
In order for the digital signal to be useful, three further factors have to be considered:
the sampling has to be regular;
the time interval between samples has to be short enough to follow the fastest changes in
the analogue signal;
in a digital signal not only is the time domain in discrete steps but so is the signal itself.
For example a signal may be represented by zero to fifteen amplitude states, which might
mean that some of the finer detail may be lost. The number of steps to which the signal is
digitised is an important consideration.
The terms used to describe these digitising parameters are:
the rate at which the signal is sampled regularly is called the sampling rate;
Time
Signal
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
the number of levels in the digital signal is called the resolution;
the resolution is often a power of two as this represents steps in the number of bits in a
binary system.
For example 16 levels requires 4 bits and 256 levels requires 8 bits.
The following diagram shows the same signal but sampled and digitised to 8 levels
Note that the output steps between the available levels and is timed at the sampling
points. Note also that some of the detail of the signal has been lost due to both the lack of
resolution and the low sampling rate. In a digital system the choice of resolution and
sampling rate must be made very carefully.
If the sampling rate is far too low, then the wrong waveshape can be produced from time
repetitive signals. This effect is called aliasing and is described in another Theory section.
There are several methods of implementing both the analogue to digital process and the
digital to analogue process and these are described in another Theory section.
Time
2 Si
Sampling
points
Available
levels
Digitised
output
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
Practical 1: Generating Amplitude Shift Keying
Objectives and Background
In this assignment you will generate an amplitude shift keyed (ASK) signal.
Amplitude shift keying is simply an amplitude modulation where the modulation is not a
continuous analogue signal, where all levels are present, but a digital one where only a
few levels are present. The simplest from of digital modulation comprises only two levels
and is called binary keying.
The name keying as referred to digital modulation originates from the oldest form of digital
modulation: Morse code. Characters are represented by sequences of dots and dashes in
the Morse code. Morse was sent in the very early days of communications along cables by
simply turning a voltage off and on. When radio was developed the same code simply
turned the carrier off and on to represent the dots and dashes. The operator used a hand
operated switch to form the code and this switch was referred to as a key. Hence the
carrier was keyed and this name remains with us today.
Morse code sent in this way was binary amplitude shift keying (ASK), because it changed
the amplitude of the carrier between two levels: off and on. ASK can exist as an amplitude
shift between any two levels but onoff keying is used because it is easier to tell the
difference between on and off than between on and slightly on.
In this Practical, a balanced modulator with a dc offset is used (exactly as was used to
produce AM double sideband) and the modulation, sometimes referred to as the data, is
represented by a square wave signal. You can think of this as simply a stream of ones and
zeros. In a real system the sequence of ones and zeros would be data, but not necessarily
its raw form. Various encoding methods are used to help with the synchronisation of both
carrier and bit rate recovery. For the purposes of understanding the concepts, how the
data is encoded is unimportant.
It is also important that you understand the terms bit rate and symbol rate, as it is the
symbol rate that determines the minimum bandwidth that the signal occupies and the ratio
of symbol rate to bit rate gives a measure of the efficiency of the system. If you do not
understand these terms refer to the Concept resources.
You should also be aware that very sudden changes in amplitude in a signal mean that
high order harmonics are present which, of course, means more occupied bandwidth.
There is no purpose in having very sharp transitions, providing that the transitions are
sharp enough not to take so long to reach one state from another that it is impossible to
decode. This problem is called intersymbol interference Use the Concept resources for
more information on this.
Since ASK is amplitude modulation with full carrier, then is it possible to have ASK with
suppressed carrier? The answer is yes but, because the phase reverses and the
amplitude stays the same to represent the two symbols, it is actually regarded as phase
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
modulation. This particular form of modulation is binary phase shift keying with a phase
shift of 180 degrees and is explored fully in the assignments on phase modulation.
In this Practical you will see what a binary ASK signal looks like and how a premodulation
filter controls unnecessary occupied bandwidth.
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
Block Diagram
Make Connections Diagram
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
Practical 1: Generating Amplitude Shift Keying
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Open the oscilloscope and the spectrum analyser.
Set the modulation Signal Level Control and the IQ Modulator block controls to
approximately half scale.
Set the Function Generator to Fast and the output to a square wave.
Open the voltmeter and set the Carrier offset voltage to +0.25 volts using the dc Source
control. Close the voltmeter.
Use the oscilloscope cursor to adjust the Frequency on the function generator so that the
bit period is about 20S. Adjust the modulation Signal Level Control so that, for a bit
zero, the amplitude of the carrier is almost zero. The oscilloscope should now show
amplitude shift keying. Note that the sidebands on the spectrum analyser show that the
occupied bandwidth is extremely large.
Change the Frequency of the function generator and note the effect.
Open the phasescope.
Move the reference probe (yellow) for the phase scope to the carrier (monitor point 1) and
note the constellation, showing constant phase with amplitude from a value (your one
state amplitude) to zero.
Return the probe to the data signal (monitor point 2).
Refer to the Make Connections diagram and remove connection 1 and add connections
10 and 11. This places a premodulation filter in circuit. Adjust the function generator back
to 20S bit period. Notice that the bandwidth has been significantly reduced and the rapid
amplitude changes on the oscilloscope have been smoothed.
If you increase the frequency of the function generator you will see that if the bit rate is too
near the filter cutoff then significant intersymbol interference takes place.
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
Practical 2: Generating Multi Level Amplitude Shift Keying
Objectives and Background
In Practical 1 you generated simple binary ASK. It is possible to have ASK that contains
more than one level. In this practical you will investigate 4level ASK.
The method of generating it is similar to that used for binary ASK, but the data source is
the microprocessor which, with its digital to analogue converter, has the ability to generate
a voltage containing four levels representing a stream of random 2bit numbers. In
practice, these 2bit numbers would be mapped from data containing a wider data format.
The important point is that the symbol rate is half the bit rate. So, for a given bandwidth,
twice the bit rate can be transmitted. Of course, the signal could have any number of
levels but demodulation becomes more and more difficult and the advantages over
analogue AM diminish. In fact, with an infinite number of levels it becomes analogue AM!
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
Block Diagram
Make Connections Diagram
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
Practical 2: Generating Multilevel Amplitude Shift Keying
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Open the voltmeter and set the Carrier offset voltage to +0.25 volts using the dc Source
control.
Set the modulation Signal Level Control and the IQ Modulator block controls to
approximately half scale.
Open the oscilloscope and look at the signals. You should see that the modulating data
(blue trace) contains a finite number of different levels within the waveform. Decrease the
oscilloscope timebase, if necessary, to see this more clearly. The modulated signal (yellow
trace) contains the same number of carrier amplitudes. This is 16level ASK.
Think about the relationship between symbol rate and bit rate. Work out how much higher
the bit rate is for the same symbol rate as binary ASK.
Change to 4level and 8level and observe the effects. Set the data to 4 levels.
Close the oscilloscope and open the phasescope. Move the reference probe (blue) to the
carrier source (monitor point 1). You should be able to see the four carrier amplitudes as
constellation points. It is easier to see the points if you enable the Persistence function on
the phasescope. Increase the size of the phasescope and change to 8 level data. You
should be able to see the 8 constellation points.
Using 16 level data you may have difficulty see the points separately. With the dc offset
producing ASK with full carrier, the phase is constant. If you reduce the dc Source, this will
no longer be the case.
When any multilevel signal is demodulated the closer the constellation points are
together, the more difficult it is to determine which symbol is actually being sent.
Note that it is easier to identify the type of modulation using the phasescope.
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
Practical 3: Demodulating Amplitude Shift Keying
Objectives and Background
The demodulation of ASK is achieved in exactly the same way as for analogue AM. The
output from the demodulator would then be decoded in some way to regenerate whatever
data was being sent. To achieve this may need bit synchronisation.
In this Practical you will use both an envelope detector and a product detector and you will
see that the results are similar. The product detector offers some advantages when
operating on a noisy signal but requires that an onfrequency and inphase local oscillator
be generated. In general, because ASK has rather poor performance in the presence of
noise, it is only used in simple systems with simple demodulators.
An interesting aside is that Morse code is still used employing ASK to turn a carrier off and
on and works extremely well at very low signaltonoise ratios. The reason for this is that
the demodulator output is an audio tone, which is then fed to one of the best decoders in
the world the human ears and brain!
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
Block Diagram
Make Connections Diagram
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
Chapter 7
Modulation and Coding Principles Amplitude Shift Keying
53230
Practical 3: Demodulating Amplitude Shift Keying
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Ensure that the balance controls associated with the IQ Modulator and IQ Demodulator
blocks are set to their mid positions.
Open the oscilloscope and the spectrum analyser.
Open the voltmeter and set the dc Carrier offset voltage to +0.25 volts using the lefthand
dc Source control.
Set the Function Generator to Fast and select a square wave output.
Set the modulation Signal Level Control to about half scale.
Adjust the Function Generator Frequency so you can see at least one cycle of data on
the screen. Use the Signal Level Control to adjust the modulation level to 100%. This
would correspond with the greatest probability of receiving data with no errors. Note we
are not using a premodulation filter, so the spectrum is very wide.
Move the oscilloscope Channel 1 probe (blue) to the envelope detector output. Note that
the output signal follows the modulation, although the edges are not so fast.
Now move the oscilloscope Channel 2 probe (yellow) to the product detector output and
adjust the righthand dc Source control voltage to the local oscillator for a steady output
trace. Note that the product detector output also follows the modulation.
Adjust the modulation level using the Signal Level Control and compare the output
waveforms. Note that the output from the envelope detector is at maximum when the
modulation is 100%. Increasing the signal level further does not affect the amplitude of the
output from the envelope detector.
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
Frequency Modulation
Objectives
To appreciate the concepts of frequency modulation and to understand the term
deviation
To generate a frequency modulated signal by direct oscillator frequency shift
To investigate the spectrum and bandwidth of a frequency modulated signal and to
appreciate the use of Bessel functions to determine the spectrum
To appreciate and use Carsons Rule for the determination of bandwidth
To understand the operation of a phase locked loop
To investigate the demodulation of a frequency modulated signal using a phase locked
loop
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
Bessel Function and FM
Modulation
The equation of a sinusoidal voltage waveform is given by:
v = V
max
.sin(t+)
where:
v is the instantaneous voltage,
V
max
is the maximum voltage amplitude,
is the angular frequency,
is the phase.
A steady voltage corresponding to the above equation conveys little information.
To convey information the waveform must be made to vary so that the variations
represent the information. This process is called modulation.
Any of these may be varied to convey information.
Frequency Modulation
Frequency modulation uses variations in frequency to convey information.
The wave whose frequency is being varied is called the carrier wave. The signal doing the
variation is called the modulating signal.
For simplicity, suppose both carrier wave and modulating signal are sinusoidal; ie:
v
c
= V
c
sin
c
t
(c denotes carrier) and
v
m
= V
m
cos
m
t
(m denotes modulation)
What is Frequency?
If the frequency is varying, how can it be defined?
You can no longer count the number of cycles over a longish interval to determine the
cycles per second. Instead, frequency is defined as the rate of change of phase.
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
This is consistent with the simple definition because, at a constant (angular) frequency
radians/second, the phase is changing at radians per second, which is /2 cycles per
second.
Since the instantaneous frequency can only be defined by reference to the phase, the
phase must be examined in order to arrive at an expression for the frequencymodulated
signal.
Phase of the FM Signal
For the unmodulated carrier v
c
= V
c
sin
c
t, the phase is:
=
c
t
The modulating signal varies the carrier frequency,
c
, so that its frequency takes the
form:
=
c
+ D cos
m
t
(where D denotes the peak value of the deviation).
It is related to the amplitude of the modulating signal v
m
by the 'frequency slope' of the
frequency modulator (VCO), say k radians/s per V.
The peak value of v
m
produces deviation D, so:
D = k V
m
The total phase change undergone at time t is found by integrating the angular frequency.
It is
= (
c
+ D cos
m
t) dt
=
c
t + (D/
m
) sin
m
t
(If you are not familiar with integration you will have to take this result on trust).
So the FM signal can be expressed as:
V
c
sin [
c
t + (D/
m
) sin
m
t]
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
Modulation Index
In the expression for the FM signal:
V
c
sin [
c
t + (D/
m
) sin
m
t]
the coefficient D/
m
turns out to be quite important and is given the name modulation
index.
It is often represented by the Greek letter beta, .
So we may write the FM signal as:
v
c
= V
c
sin (
c
t + sin
m
) t
where is the modulation index D/
m
.
In this expression, the factor sin (
c
t + sin
m
)t (let us call it F) is of the form sin(a + b),
which can be expanded to sin a cos b + cos a sin b.
Applying this expansion to F, we get:
F = sin
c
t cos(sin
m
) t + cos
c
t sin (sin
m
) t
FM Sidebands
These complicated functions can be expanded, using mathematics too elaborate to
explain here, into a series of terms like this:
F = J
0
( ) sin
c
t+ J
1
( ) [ sin (
c
+
m
)t  sin (
c

m
)t ]
+ J
2
( ) [ sin (
c
+ 2
m
)t  sin (
c
 2
m
)t ]
+ J
3
( ) [ sin (
c
+ 3
m
)t  sin (
c
 3
m
)t ]
+ J
4
( ) [ sin (
c
+ 4
m
)t  sin (
c
 4
m
)t ]
+ ...
where J
0
( ), J
1
( ), J
2
( ) etc are constants whose values depend only on . They are
called Bessel Functions.
There is an infinite series of these functions, and so an infinite number of FM sidebands.
But, in practice the values of the Bessel functions become very small as the series goes
on. For example, when = 2
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
J
0
(2) = 0.224
J
1
(2) = 0.577
J
2
(2) = 0.353
J
3
(2) = 0.129
J
4
(2) = 0.034
J
5
(2) = 0.007
A Practical Approximate Rule
Because the higherorder sidebands become very small, in practice the bandwidth of the
FM signal may be restricted to a finite bandwidth.
The practical rule that is used, often called Carsons Rule, is to take the bandwidth
required as:
B = 2 ( F
d
+ F
m
)
where B is the bandwidth, F
d
the deviation and F
m
is the bandwidth of the modulation,
all in the same units.
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
The Phase Locked Loop
A phase locked loop (PLL) is a sub system that enables an oscillator to be synchronized
in frequency and phase to an incoming signal. The block diagram shows the building
blocks that make up a phase locked loop.
Imagine that the voltage controlled oscillator (VCO) is oscillating near to the incoming
signal frequency. The output of the phase/frequency comparator is a signal that
represents the frequency error between the VCO and the incoming signal. This signal is
applied to the frequency control input of the VCO, which then changes its frequency to be
equal to the incoming signal. The output of the comparator then compares the phases of
the two signals and uses the VCO frequency control to match the two phases. The system
is now in lock. If either the signal or the VCO moves in phase with respect to each other
the comparator output moves the VCO so that the two are always locked together.
In fact most PLLs only use a phase comparator (detector). This is because phase
detectors, when presented with two different frequencies, produce an ac signal equal in
frequency to the difference between them. This has the effect of swinging the VCO up and
down in frequency and, as it passes the signal frequency, the loop locks.
Loop Stability
One of the problems that will almost certainly arise, unless steps are taken to stop it, is
instability. The loop relies on the system operating with negative feedback, i.e. if the VCO
moves, the polarity of the control signal brings it back. This is easily done when the
system is operating at, or near to, dc. However, a problem arises if you consider the loop
moving in response to a fast changing frequency. The control signal will contain an ac
component. All systems are subject to delays and phase shifts, which become more
significant at higher frequencies. Remembering that 180 degrees phase shift is equivalent
to inverting a signal, inevitably there is going to be a frequency at which the phase shift
round the loop is enough to cause the polarity to reverse and positive feedback will be
applied. This results in the system oscillating back and forth at the frequency which
produces the positive feedback.
There is another subtle problem that results in the design of a PLL system being more
difficult then you might imagine. Remember that we are using the frequency of an
oscillator to control its phase. Now phase is the integral of frequency and so there is
already a 90 degree shift caused by the VCO. This means that there only has to be
another 90 degrees of phase shift before instability, not 180 as you might have thought.
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
The same problem occurs, for example, in a mechanical position control system that uses
speed control, because position is the integral of speed. The phase lock loop is a control
system and exactly the same mathematics can be used to describe a PLL as is used to
describe a position control servo.
Of course, instability can only occur if there is enough gain in the loop at the problem
frequency. This is where the loop filter can solve the problem, as it reduces gain at higher
frequencies while maintaining control over phase. There will be a frequency that the loop
filter produces 90 degrees of phase shift, but the gain will be low and so instability will not
arise. The critical frequency is when the overall loop gain is 1 and the overall added phase
shift must be less than 90 degrees at this point. The amount by which it is less is called
phase margin and, in practice, should be about 45 degrees for good stable performance.
The design of the loop filter is not simple and has to be done knowing all the gains and
phase shifts in the system.
Many phase comparators produce a control signal that contains a significant amount of
high frequency energy but this is not normally a problem as it is removed by the loop filter.
In PLLs that use phase only comparators, the bandwidth of the loop filter also determines
the range over which the loop will lock on, or capture, a signal as the comparator only
generates an ac signal off lock. The range over which the loop will capture a signal is
called its capture range. The range over which the loop will remain locked, once lock is
achieved, is called the lock range. The time to achieve lock can be important and is
referred to as lock time.
Phase Comparator
A number of circuits will operate as phase comparators or detectors. A multiplier is often
used. If the two inputs of a multiplier are fed with two signals that are at the same
frequency, but with different phases, the output will comprise a twicefrequency
component and a dc component that represents the phase error. There are some
important restrictions to this, in that it will only operate over 180 degrees and has zero
output at 90 degrees, not zero. The graph shows the output voltage for such a detector
plotted against input phase difference.
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
This shows that the output repeats for 180 to 360, albeit in the other polarity.
This phase range problem is not significant as, in a properly operating loop, the gain is
such that only a small error has to occur before the VCO is corrected.
There are other types of comparator. The logic function exclusive OR is exactly the same
in action as the multiplier and is often used in digital circuits. More complex digital circuits
have advantages, such as: acting as frequency comparators as well as phase
comparators; operating over 360 degrees; and having less ac signal component in the
output. Most of these are based on circuits using D type flipflops. These D type
comparators have the disadvantage of making the loop much less tolerant of noise in the
signal. They are used in applications such as frequency synthesizers, while the multipliers
and OR gates are used in applications such as demodulators or carrier reference
recovery.
The design of PLLs is a complex compromise of performance parameters, the relative
importance of each performance parameter depends, to a large extent, on the application.
Applications
Many phase lock loops are used to recover some sort of constant frequency component to
provide a reference for a demodulator.
Another very common application is in frequency synthesizers, where an oscillator is
frequency divided to some low frequency and a PLL locks it to an external reference. By
changing the divider ratio, different frequencies that are all multiples of the reference
frequency can be generated or synthesised.
The PLL can also be used to demodulate FM as, when locked to the FM signal, the VCO
tracks the frequency modulation. Therefore the control signal to the VCO contains the
modulation, plus a dc component. This dc component can easily be removed using a high
pass filter. In this case the loop filter bandwidth must be high enough to pass the
modulation or the VCO will not be able to follow and the loop will come out of lock. In most
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
cases the output is passed through a post detection filter, which will remove any remaining
high frequency components but, because it is outside the loop, will not affect loop stability.
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
Practical 1: Concepts of FM
Objectives and Background
In this Practical you will investigate frequency modulation (FM).
In frequency modulation, only the frequency of the carrier is changed, the amplitude
remains constant.
With no modulation the carrier frequency is constant and is at its nominal frequency.
When modulation is applied the frequency is moved above the nominal carrier frequency
when the modulation signal is positive and below the nominal frequency when the
modulation is negative.
The nominal frequency is sometimes referred to as the centre frequency. Normally the
frequency changes equally above and below the centre frequency. The amount by which
the carrier is varied above and below is called the deviation. Note that the total frequency
change during a full modulation cycle is therefore twice the deviation.
The modulation frequency can, of course, be of any frequency or band of frequencies and,
like amplitude modulation, sidebands are produced. However, the mathematics of this
process for FM is more complex than for AM.
In FM an important parameter that determines the characteristics of the signal is the ratio
of the deviation to the maximum modulating frequency. If this value is quite small (less
than 1 for example) most of the energy is contained in the first few sidebands. This is
similar to AM, where the bandwidth is determined by the maximum modulating frequency.
When the ratio is larger the overall bandwidth is determined mainly by the deviation, rather
than the maximum modulating frequency. These two conditions are grouped together by
the rather vague terms narrow band FM and wideband FM.
The ratio of deviation to maximum modulating frequency is often referred to by the Greek
letter beta and is defined mathematically by:
m
f
f
=
Where f is the frequency deviation and f
m
is the maximum modulating frequency.
As the exact determination of the bandwidth is quite difficult to calculate, an approximation
is often used to obtain a figure for the bandwidth where most of the energy is confined.
This approximation is known as Carsons Rule.
By Carsons Rule, the bandwidth B is given by
) ( 2
m
f f B + =
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
This is a most useful approximation. Remember that some signal energy will be outside
this bandwidth but, for practical purposes, the bandwidth estimated from Carsons rule is a
good approximation.
The exact energy distribution in an FM signal is determined by mathematical functions
called Bessel functions. They are quite complex and more information is given in the
Concept section. An interesting characteristic of FM is that, unlike AM, where the
amplitude of the carrier is unaffected by the modulation, energy moves from the carrier to
sidebands. This is a fairly obvious consequence of the fact that the amplitude remains
constant. So, the overall power in the signal remains constant and therefore the power in
the sidebands has to come from the carrier.
At certain values of beta, the carrier disappears. This can be produced by modulating with
a single tone. It also gives us a way of setting deviation by choosing a test modulating
frequency that gives a carrier null with the wanted value of deviation.
The lowest value of beta that gives zero carrier is called the 1
st
Bessel Null. The value is
in fact equal to 2.405, irrespective of anything else.
In this Practical you will use the voltage controlled oscillator to generate frequency
modulation and examine the signal in the time and frequency domains.
You may notice that the VCO is slightly nonlinear and, for equal voltage swing, it moves
further down in frequency than up. This is often the case with VCOs, but is made worse in
this case by the need to have rather wide deviation to provide clear displays.
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
Block Diagram
Make Connections Diagram
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
Practical 1: Concepts of FM
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the modulation Signal Level Control to minimum.
Open the oscilloscope and the spectrum analyser. Note the unmodulated carrier (blue
trace).
Increase the modulation to maximum, using the Signal Level Control, and note the
sidebands on the spectrum analyser.
Adjust the oscilloscope timebase so you can see only one cycle of the modulation and the
individual cycles of the carrier. Toggle the Neg Trig function on and off several times
whilst observing the traces. If you look very carefully you can see the frequency changes
in time with the modulation. Positive and negative peaks in the modulation result in higher
and lower carrier frequencies respectively. You may see this more clearly if you increase
the size of the oscilloscope.
Note that there is no change in carrier amplitude. Look at the spectrum analyser and note
that, even though the modulation is a single frequency sine wave, there is more than one
set of sidebands.
Open the frequency counter and measure the frequency of the modulation. Measure the
spacings of the sidebands on the spectrum analyser and compare the values with the
modulation frequency.
Refer to the Make Connections diagram and remove connection 3 and add connection 5.
Set the Function Generator to Fast and select a sine wave.
Use the Signal Level Control to set the amplitude of the modulation to maximum, thus
giving maximum deviation, and then vary the frequency of the modulation using the
Function Generator Frequency control. Note the effect on the spectrum analyser.
Even at low frequencies the bandwidth cannot be less than the deviation. In general, when
you can see the individual sidebands the value of beta (i.e. the modulation index) is small,
whereas if the value of beta is large then the individual sidebands are masked by the
frequency deviation.
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
Practical 2: Carsons Rule and Bessel Function Null
Objectives and Background
In this Practical you will measure the deviation of a frequency modulated signal. You will
use a dc voltage of the same magnitude as the sinewave modulating signal to drive the
VCO and will measure the frequency using the counter.
You will then use the Bessel null method to measure the deviation with an ac signal, then
estimate the bandwidth using Carsons rule. In the process, you will see the carrier reduce
in amplitude and disappear at certain values of beta.
Remember that in frequency modulation the carrier power remains essentially constant at
all times. However, in practice, it does vary in amplitude by a very small amount caused by
deficiencies in the modulator and the amplifiers that follow it. The parameter that
describes this is called residual AM (residual amplitude modulation) and is often
measured in percentage AM modulation for maximum deviation. In a practical system it
should be fractions of a percent.
AM can also be produced in the receiver if any filter that the signal passes through has
passband ripple. In practice, any AM that is produced is usually removed by passing the
signal through a limiting amplifier prior to demodulation.
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
Block Diagram
Make Connections Diagram
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
Practical 2: Carsons Rule and Bessel Function Null
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
In this Practical you are going to measure the deviation using a dc voltage to drive the
modulator and then use the Bessel null method to confirm the value using ac modulation.
The first measurement is the amplitude of the ac modulation source.
Open the oscilloscope, spectrum analyser, frequency counter and voltmeter.
Set the Function Generator to Fast and select a sine wave.
Adjust the Frequency control on the Function Generator block and use the frequency
counter to set the modulation frequency to approximately 10kHz. Set the voltmeter to ac
pp and measure the peak to peak amplitude of the sine wave. The peak value is half of
this value.
Refer to the Make Connections diagram and remove connection 2 and add connection 5,
so that the modulation source is now the dc Source.
Set the voltmeter to dc and adjust the dc Source voltage to minus half the ac peak to peak
value that you measured. This should set the carrier to the frequency corresponding to
when the modulation is at its most negative. Move the frequency counter probe (yellow) to
the carrier output (monitor point 1), measure the frequency and note the value.
Now set the dc Source voltage to plus half the ac peak to peak value and measure the
carrier frequency. Note the value. You can also see the frequency change on the
oscilloscope and the spectrum analyser.
The difference between these two carrier frequencies is the peak to peak frequency
change, i.e. twice the deviation. Calculate the deviation and note it down.
Remove connection 5 and replace connection 2 so the modulation source is again the
Function Generator.
Move the frequency counter probe (yellow) back to the modulation (monitor point 2).
Slowly increase the modulation frequency, while observing the spectrum analyser display.
You should be able to see the individual sidebands and the carrier. Note that the carrier
amplitude starts to decrease. There should be a frequency between 15kHz and 20kHz
where the carrier disappears. Adjust the modulation frequency carefully so that the carrier
is as near to zero amplitude as you can. Note the frequency of the modulation.
Remembering that the first Bessel null occurs when beta is 2.405, the deviation can be
calculated using
m
f f 405 . 2 =
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
Where f is the deviation and f
m
is the 1
st
Bessel null frequency.
Compare the value you measured from the dc measurement and that from using the
Bessel null method. Due to the design of the modulator, the deviation at high frequency
will be slightly different than at low frequency, such that the results will not be exactly the
same for both methods.
Now use Carsons rule to calculate the bandwidth for this combination of modulation
frequency and deviation.
Use the spectrum analyser and cursors to see how this compares with the occupied
bandwidth. Remember, bandwidth is measured at the points each side of the peak and 3
db below it.
You can also calculate the control sensitivity for the VCO thus:
p p
V
f
s
=
2
Where S is the sensitivity, f is the deviation and V
pp
is the peak to peak magnitude of the
ac modulation.
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
Practical 3: Demodulation of FM using a Phase Locked Loop
Objectives and Background
There are a number of methods available for demodulating a frequency modulated signal
(FM). One method widely used is the phase locked loop (PLL). You will find a full
explanation of the how a PLL works in the Concepts section.
PLLs are used in many applications to track and generate frequencies, as well as to
demodulate FM or to provide local oscillator reference signals. The mathematics is
complex, but it is important that you understand the principle of how they work.
The FM generator that you will use in this Practical is the VCO that you have already used
modulated by the function generator. The PLL is made from the local oscillator, which is
also a VCO, and multipliers used as a phase detector. A loop filter and a post detection
filter complete the demodulator.
In the Practical you will see how the PLL operates as a demodulator and how the loop
bandwidth limits the maximum modulating frequency.
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
Block Diagram
Make Connections Diagram
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
Practical 3: Demodulation of FM using a Phase Locked Loop
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the IQ Demodulator controls to mid scale.
Open the oscilloscope and the spectrum analyser.
Set the Function Generator to Fast and select a sine wave output. This provides the
modulation source for the FM generator.
Open the frequency counter and use the Frequency control on the Function Generator to
set the frequency of the modulation to 8kHz.
Open the voltmeter and use the Signal Level Control to set the ac pp modulation
amplitude to 0.2 volts. Set the dc Source control to about half scale to provide a dc offset
to the Loop Filter of approximately zero volts.
Set the Loop Filter Compensation to Fast.
From the spectrum analyser, note that the output is FM.
You can calculate the deviation that is in use by using the value you calculated for VCO
voltage sensitivity and the current modulation amplitude of 0.2 volts peak to peak.
Move the oscilloscope Channel 1 probe (blue) to the output of the multiplier (monitor point
3) and the Channel 2 probe (yellow) to the VCO output (monitor point 1). You should be
able to see the demodulator working and reproducing the modulation. If the display is
unstable then adjust the dc Source control to lock the loop. Increase the timebase of the
oscilloscope and you should be able to see the twice carrier frequency component.
Move the oscilloscope Channel 1 probe (blue) to the loop filter output (monitor point 4)
and note that the high frequency component has been removed. Move the spectrum
analyser probe (orange) to see that, even after the loop filter (monitor point 4), there is
some high frequency noise present. Now look at the post detection filter output (monitor
point 5) and see that it has been significantly reduced.
While monitoring the loop filter output with the oscilloscope, increase the modulation
(using the Signal Level Control) and note that, if the deviation is too wide, the loop unlocks
at peaks of deviation. Return the modulation voltage to 0.2 volts peak to peak and adjust
the modulation frequency (using the Function Generator Frequency control). Note that
above a certain value the demodulator output decreases. This is due to the loop filter
bandwidth.
Note that you may have to readjust the dc Source voltage to make sure the loop stays
locked.
Chapter 8
Modulation and Coding Principles Frequency Modulation
53230
You can also try changing the modulation signal to a square wave. You will have to reduce
the deviation and the modulation frequency.
Chapter 9
Modulation and Coding Principles Frequency Modulation using an IQ Modulator
53230
Frequency Modulation using an IQ Modulator
Objectives
To appreciate that a frequency modulated signal can be produced using an IQ modulator
and the advantages of this method
To understand that two carrier signals in quadrature (90 degrees apart) are required for
this form of modulation
To investigate the generation of a frequency modulated signal using the IQ modulator
method
Chapter 9
Modulation and Coding Principles Frequency Modulation
using an IQ Modulator
53230
Bessel Function and FM
Modulation
The equation of a sinusoidal voltage waveform is given by:
v = V
max
.sin(t+)
where:
v is the instantaneous voltage,
V
max
is the maximum voltage amplitude,
is the angular frequency,
is the phase.
A steady voltage corresponding to the above equation conveys little information.
To convey information the waveform must be made to vary so that the variations
represent the information. This process is called modulation.
Any of these may be varied to convey information.
Frequency Modulation
Frequency modulation uses variations in frequency to convey information.
The wave whose frequency is being varied is called the carrier wave. The signal doing the
variation is called the modulating signal.
For simplicity, suppose both carrier wave and modulating signal are sinusoidal; ie:
v
c
= V
c
sin
c
t
(c denotes carrier) and
v
m
= V
m
cos
m
t
(m denotes modulation)
What is Frequency?
If the frequency is varying, how can it be defined?
You can no longer count the number of cycles over a longish interval to determine the
cycles per second. Instead, frequency is defined as the rate of change of phase.
Chapter 9
Modulation and Coding Principles Frequency Modulation using an IQ Modulator
53230
This is consistent with the simple definition because, at a constant (angular) frequency
radians/second, the phase is changing at radians per second, which is /2 cycles per
second.
Since the instantaneous frequency can only be defined by reference to the phase, the
phase must be examined in order to arrive at an expression for the frequencymodulated
signal.
Phase of the FM Signal
For the unmodulated carrier v
c
= V
c
sin
c
t, the phase is:
=
c
t
The modulating signal varies the carrier frequency,
c
, so that its frequency takes the
form:
=
c
+ D cos
m
t
(where D denotes the peak value of the deviation).
It is related to the amplitude of the modulating signal v
m
by the 'frequency slope' of the
frequency modulator (VCO), say k radians/s per V.
The peak value of v
m
produces deviation D, so:
D = k V
m
The total phase change undergone at time t is found by integrating the angular frequency.
It is
= (
c
+ D cos
m
t) dt
=
c
t + (D/
m
) sin
m
t
(If you are not familiar with integration you will have to take this result on trust).
So the FM signal can be expressed as:
V
c
sin [
c
t + (D/
m
) sin
m
t]
Chapter 9
Modulation and Coding Principles Frequency Modulation
using an IQ Modulator
53230
Modulation Index
In the expression for the FM signal:
V
c
sin [
c
t + (D/
m
) sin
m
t]
the coefficient D/
m
turns out to be quite important and is given the name modulation
index.
It is often represented by the Greek letter beta, .
So we may write the FM signal as:
v
c
= V
c
sin (
c
t + sin
m
) t
where is the modulation index D/
m
.
In this expression, the factor sin (
c
t + sin
m
)t (let us call it F) is of the form sin(a + b),
which can be expanded to sin a cos b + cos a sin b.
Applying this expansion to F, we get:
F = sin
c
t cos(sin
m
) t + cos
c
t sin (sin
m
) t
FM Sidebands
These complicated functions can be expanded, using mathematics too elaborate to
explain here, into a series of terms like this:
F = J
0
( ) sin
c
t+ J
1
( ) [ sin (
c
+
m
)t  sin (
c

m
)t ]
+ J
2
( ) [ sin (
c
+ 2
m
)t  sin (
c
 2
m
)t ]
+ J
3
( ) [ sin (
c
+ 3
m
)t  sin (
c
 3
m
)t ]
+ J
4
( ) [ sin (
c
+ 4
m
)t  sin (
c
 4
m
)t ]
+ ...
where J
0
( ), J
1
( ), J
2
( ) etc are constants whose values depend only on . They are
called Bessel Functions.
There is an infinite series of these functions, and so an infinite number of FM sidebands.
But, in practice the values of the Bessel functions become very small as the series goes
on. For example, when = 2
Chapter 9
Modulation and Coding Principles Frequency Modulation using an IQ Modulator
53230
J
0
(2) = 0.224
J
1
(2) = 0.577
J
2
(2) = 0.353
J
3
(2) = 0.129
J
4
(2) = 0.034
J
5
(2) = 0.007
A Practical Approximate Rule
Because the higherorder sidebands become very small, in practice the bandwidth of the
FM signal may be restricted to a finite bandwidth.
The practical rule that is used, often called Carsons Rule, is to take the bandwidth
required as:
B = 2 ( F
d
+ F
m
)
where B is the bandwidth, F
d
the deviation and F
m
is the bandwidth of the modulation,
all in the same units.
Chapter 9
Modulation and Coding Principles Frequency Modulation
using an IQ Modulator
53230
The Phase Locked Loop
A phase locked loop (PLL) is a sub system that enables an oscillator to be synchronized
in frequency and phase to an incoming signal. The block diagram shows the building
blocks that make up a phase locked loop.
Imagine that the voltage controlled oscillator (VCO) is oscillating near to the incoming
signal frequency. The output of the phase/frequency comparator is a signal that
represents the frequency error between the VCO and the incoming signal. This signal is
applied to the frequency control input of the VCO, which then changes its frequency to be
equal to the incoming signal. The output of the comparator then compares the phases of
the two signals and uses the VCO frequency control to match the two phases. The system
is now in lock. If either the signal or the VCO moves in phase with respect to each other
the comparator output moves the VCO so that the two are always locked together.
In fact most PLLs only use a phase comparator (detector). This is because phase
detectors, when presented with two different frequencies, produce an ac signal equal in
frequency to the difference between them. This has the effect of swinging the VCO up and
down in frequency and, as it passes the signal frequency, the loop locks.
Loop Stability
One of the problems that will almost certainly arise, unless steps are taken to stop it, is
instability. The loop relies on the system operating with negative feedback, i.e. if the VCO
moves, the polarity of the control signal brings it back. This is easily done when the
system is operating at, or near to, dc. However, a problem arises if you consider the loop
moving in response to a fast changing frequency. The control signal will contain an ac
component. All systems are subject to delays and phase shifts, which become more
significant at higher frequencies. Remembering that 180 degrees phase shift is equivalent
to inverting a signal, inevitably there is going to be a frequency at which the phase shift
round the loop is enough to cause the polarity to reverse and positive feedback will be
applied. This results in the system oscillating back and forth at the frequency which
produces the positive feedback.
There is another subtle problem that results in the design of a PLL system being more
difficult then you might imagine. Remember that we are using the frequency of an
oscillator to control its phase. Now phase is the integral of frequency and so there is
already a 90 degree shift caused by the VCO. This means that there only has to be
Chapter 9
Modulation and Coding Principles Frequency Modulation using an IQ Modulator
53230
another 90 degrees of phase shift before instability, not 180 as you might have thought.
The same problem occurs, for example, in a mechanical position control system that uses
speed control, because position is the integral of speed. The phase lock loop is a control
system and exactly the same mathematics can be used to describe a PLL as is used to
describe a position control servo.
Of course, instability can only occur if there is enough gain in the loop at the problem
frequency. This is where the loop filter can solve the problem, as it reduces gain at higher
frequencies while maintaining control over phase. There will be a frequency that the loop
filter produces 90 degrees of phase shift, but the gain will be low and so instability will not
arise. The critical frequency is when the overall loop gain is 1 and the overall added phase
shift must be less than 90 degrees at this point. The amount by which it is less is called
phase margin and, in practice, should be about 45 degrees for good stable performance.
The design of the loop filter is not simple and has to be done knowing all the gains and
phase shifts in the system.
Many phase comparators produce a control signal that contains a significant amount of
high frequency energy but this is not normally a problem as it is removed by the loop filter.
In PLLs that use phase only comparators, the bandwidth of the loop filter also determines
the range over which the loop will lock on, or capture, a signal as the comparator only
generates an ac signal off lock. The range over which the loop will capture a signal is
called its capture range. The range over which the loop will remain locked, once lock is
achieved, is called the lock range. The time to achieve lock can be important and is
referred to as lock time.
Phase Comparator
A number of circuits will operate as phase comparators or detectors. A multiplier is often
used. If the two inputs of a multiplier are fed with two signals that are at the same
frequency, but with different phases, the output will comprise a twicefrequency
component and a dc component that represents the phase error. There are some
important restrictions to this, in that it will only operate over 180 degrees and has zero
output at 90 degrees, not zero. The graph shows the output voltage for such a detector
plotted against input phase difference.
Chapter 9
Modulation and Coding Principles Frequency Modulation
using an IQ Modulator
53230
This shows that the output repeats for 180 to 360, albeit in the other polarity.
This phase range problem is not significant as, in a properly operating loop, the gain is
such that only a small error has to occur before the VCO is corrected.
There are other types of comparator. The logic function exclusive OR is exactly the same
in action as the multiplier and is often used in digital circuits. More complex digital circuits
have advantages, such as: acting as frequency comparators as well as phase
comparators; operating over 360 degrees; and having less ac signal component in the
output. Most of these are based on circuits using D type flipflops. These D type
comparators have the disadvantage of making the loop much less tolerant of noise in the
signal. They are used in applications such as frequency synthesizers, while the multipliers
and OR gates are used in applications such as demodulators or carrier reference
recovery.
The design of PLLs is a complex compromise of performance parameters, the relative
importance of each performance parameter depends, to a large extent, on the application.
Applications
Many phase lock loops are used to recover some sort of constant frequency component to
provide a reference for a demodulator.
Another very common application is in frequency synthesizers, where an oscillator is
frequency divided to some low frequency and a PLL locks it to an external reference. By
changing the divider ratio, different frequencies that are all multiples of the reference
frequency can be generated or synthesised.
The PLL can also be used to demodulate FM as, when locked to the FM signal, the VCO
tracks the frequency modulation. Therefore the control signal to the VCO contains the
modulation, plus a dc component. This dc component can easily be removed using a high
pass filter. In this case the loop filter bandwidth must be high enough to pass the
Chapter 9
Modulation and Coding Principles Frequency Modulation using an IQ Modulator
53230
modulation or the VCO will not be able to follow and the loop will come out of lock. In most
cases the output is passed through a post detection filter, which will remove any remaining
high frequency components but, because it is outside the loop, will not affect loop stability.
Chapter 9
Modulation and Coding Principles Frequency Modulation
using an IQ Modulator
53230
The IQ Modulator
The IQ modulator is a most useful building block in communications systems. It is
available as an integrated circuit with different models operating over a wide range of
frequencies.
It comprises two balanced modulators with their carrier inputs fed from the same source
but one shifted by 90 degrees. The two modulation inputs are available for the user. The
outputs of the two modulators are then summed.
The name IQ modulator comes from In phase and Quadrature. The term quadrature
simply means at 90 degrees.
The diagram shows the basic IQ modulator.
Provided the phase shift is 90 degrees at the carrier frequency then, in vector terms, the
output with respect to the input is shown below:
This means that the output is a signal at the carrier frequency and its phase will depend
on the values of the I and Q modulation inputs. Notice that the amplitude will also vary,

+
+I
I
output
Chapter 9
Modulation and Coding Principles Frequency Modulation using an IQ Modulator
53230
because the output is the result of summing two equal values at right angles. It is
therefore 1.414 times the value of when only one input is present.
In mathematical terms the output is the a+jb (complex) sum of the I and Q modulation
inputs. The two modulation inputs can be two quite separate signals. This is how QAM is
generated.
If the output is required to be a phase vector with constant amplitude, the angle of which is
determined by a single input, that input signal has to be processed to generate suitable I
an Q signals.
Since an output is required that is a vector represented by r, where r is the required
constant radius and is the variable angle, and what we have is a+jb, this is done using
the equivalent of changing the mapping in the normal way.
i.e. for an input representing an angle of
sin
mod
M I =
cos
mod
M Q =
Where M is the magnitude of the required signal to drive the modulators.
By generating both the I and Q modulation inputs by processing a single input
representing angle, the output vector is of constant length and is driven round a circle
rather than a square.
Q
+Q
+I
I
output
Chapter 9
Modulation and Coding Principles Frequency Modulation
using an IQ Modulator
53230
This is, of course, a phase modulator.
In practical terms, the accuracy of the processing of all these signals depends on the
accuracy with which the 90 degree phase shifts can be maintained.
Chapter 9
Modulation and Coding Principles Frequency Modulation using an IQ Modulator
53230
Practical 1: Generating Frequency Modulation using an IQ Phase
Modulator
Objectives and Background
In this practical you will generate frequency modulation (FM) by using an IQ modulator.
Since there is a good method of generating FM by using direct modulation of a voltage
controlled oscillator, why do you need another method? The answer lies in the use of
available building blocks in both analogue electronics and, more especially, digital signal
processing (DSP).
A voltage controlled oscillator with a linear control characteristic is not easy to design,
especially at high frequency, and very difficult to implement with DSP. On the other hand,
the IQ modulator is commonly available, both as an integrated circuit and as DSP code.
For these two reasons alone the ability to generate FM using an IQ modulator is rather
attractive.
In order to understand how such a system works you have to understand the relationship
between frequency and phase. Indeed, this relationship is important if you are to have a
thorough understanding of modulators in general.
One of the definitions of frequency is: rate of change of phase. Imagine two sine waves of
equal frequency and in phase. Consider one zero crossing point where, at the moment,
the two signals are coincident. If one signal increases in frequency by a small fixed
amount, the zero crossing point on the higher frequency sine wave will start to advance
with respect to the other. This means that the phase of the higher frequency signal is
advancing at a constant rate, equal in terms of complete cycles to the difference in
frequency. Consequently, for a constant frequency difference, the phase is increasing
linearly. Mathematically, phase is the integral of frequency.
If you consider the opposite process, i.e. making a change in phase between two signals,
the following happens. In order to advance the phase you have to increase the frequency
for an instant until the phase moves to where you want it and then return it so the phase
does not change further. This means that frequency is the differential of phase. One
important point to remember is that you cannot change the phase without changing the
frequency and you cannot change the frequency without changing the phase.
The two diagrams below show the two processes.
Chapter 9
Modulation and Coding Principles Frequency Modulation
using an IQ Modulator
53230
As you can see, after a frequency change the phase difference increases for ever (or as
long as the two frequencies are different). Since phase is measured as an angle, one
might suppose that the angle increases to an infinite value. This is true but, because
angular measure represents a circle, the angle repeats every 360 degrees (or 2 radians).
In fact, if you as an observer were not there, at the moment the frequency changed you
would have no idea what the original phase was.
This relationship between frequency and phase provides the clue as to how frequency
modulation can be produced by what is a phase modulator.
The IQ modulator is a block that can produce any phase output from a phase reference
input in response to a control signal. How they work is covered in more detail in the
Concepts section.
By processing the I and Q modulation inputs with sine and cosine functions the output can
be any phase at constant amplitude. If the signal represents frequency, how is that signal
Chapter 9
Modulation and Coding Principles Frequency Modulation using an IQ Modulator
53230
processed to drive the phase modulator? Suppose that the input were a constant, non
zero, positive level. The output of a frequency modulator would be a constant frequency,
higher than the nominal carrier frequency. This would mean that the phase is increasing
linearly with time. If the input were applied to an integrator then the output would be an
increasing signal which would represent phase. This signal could then be applied to the
sine/cosine processor that generates the I and Q signals. The only problem occurs in that
the integrator output would soon be very large and limit. This is solved by resetting the
integrator at a value equal to plus or minus 360 degrees.
This Practical shows such a system operating. An added advantage is that, unlike with a
VCO, there is a carrier reference. This means that, by using the phasescope, you can
actually see the frequencies generated by the modulator as a rotating vector with respect
to the nominal carrier frequency. An ac signal can then be applied and the spectrum
inspected to confirm that FM has been produced.
In the course of the practical you will notice that some of the signals are not perfect. This
is caused by the fact that the sine/cosine processor and the integrator are implemented in
analogue circuitry so that you can see the individual processing blocks working. Due to
bandwidth limitations and delays the outputs contain some small irregularities. In modern
systems these imperfections can be overcome by implementing the whole process by
DSP.
Chapter 9
Modulation and Coding Principles Frequency Modulation
using an IQ Modulator
53230
Block Diagram
Make Connections Diagram
Chapter 9
Modulation and Coding Principles Frequency Modulation using an IQ Modulator
53230
Chapter 9
Modulation and Coding Principles Frequency Modulation
using an IQ Modulator
53230
Practical 1: Generating Frequency Modulation Using an IQ Phase
Modulator
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the switch in the Integrator block to Slow.
Set the Signal Level Control and the IQ Modulator controls to half scale.
Open the voltmeter and use it to set the dc Source to give a voltage to the integrator of
zero volts (monitor point 4).
Open the phasescope and use it to adjust the Variable Phase Shift control associated
with the Carrier Source so that the two carrier signals to the modulators are exactly 90
degrees apart. Open the oscilloscope and check the phase difference. You will need to
use the X expand to see the two sinewaves properly.
Move the Channel 1 oscilloscope probe (blue) to the output of the summing block (monitor
point 8). You should now be able to see on the oscilloscope the output waveform on the
upper trace moving with respect to the lower trace. This shows a difference in frequency.
Use the dc Source control to reduce this frequency difference to as small as possible. You
can also use the modulation Signal Level Control to reduce the input to the integrator in
order to make the adjustment finer. Note that the difference between the frequencies
reverses at around zero volts.
When the rate is slow enough you should be able to see the vector on the phasescope
rotating. You may need to close the oscilloscope temporarily to be able to see this rotation
clearly, as closing unwanted test instruments has the effect oFMaximising the refresh rate.
Use the dc Source control to adjust the dc voltage around zero and observe that a
difference in frequency is a continuing change of phase. Note that the direction of rotation
is determined by the sign of the frequency difference.
Close the phasescope. Open the spectrum analyser and the frequency counter.
Set the switch in the Integrator block to Fast.
Refer to the Make Connections diagram. Remove connection 2 and add connection 18.
This has the same effect as increasing the deviation.
Set the Signal Level Control to maximum. Adjust the dc Source control and note the
frequency moving on the spectrum analyser and on the frequency counter. Some
unwanted sidebands will be seen on the analyser due to glitches (irregularities) in the
sine/cosine processor at higher frequencies.
Chapter 9
Modulation and Coding Principles Frequency Modulation using an IQ Modulator
53230
Refer to the Make Connections diagram. Remove connection 1 and add connection 19.
This changes the modulation source. Move the Channel 2 oscilloscope probe (yellow) to
the integrator input (monitor point 4). Set the Function Generator switch to Fast, set the
modulation amplitude (Signal Level Control) to maximum and select a sine wave.
As you adjust the modulation frequency you should be able to see the carrier reducing in
amplitude on the spectrum analyser as you would expect in an FM signal. You may need
to maximise the size of the spectrum analyser and select Alias Hi to see this effect
clearly.
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Frequency Shift Keying
Objectives
To appreciate the principle of frequency shift keying and its relationship to analogue
frequency modulation
To generate a twolevel (binary) frequency shift keyed signal and investigate the spectrum
and bandwidth associated with it
To investigate the demodulation of an FSK signal
To understand the concept oFMinimum shift keying and its use to limit the bandwidth of
an FSK signal
To generate and subsequently demodulate a minimum shift keyed signal
To appreciate the concept of multilevel FSK (MFSK) and to generate 4, 8 and 16 level
MFSK signals
To investigate the spectrum occupied by an MFSK signal and its relationship to symbol
rate
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Bessel Function and FM
Modulation
The equation of a sinusoidal voltage waveform is given by:
v = V
max
.sin(t+)
where:
v is the instantaneous voltage,
V
max
is the maximum voltage amplitude,
is the angular frequency,
is the phase.
A steady voltage corresponding to the above equation conveys little information.
To convey information the waveform must be made to vary so that the variations
represent the information. This process is called modulation.
Any of these may be varied to convey information.
Frequency Modulation
Frequency modulation uses variations in frequency to convey information.
The wave whose frequency is being varied is called the carrier wave. The signal doing the
variation is called the modulating signal.
For simplicity, suppose both carrier wave and modulating signal are sinusoidal; ie:
v
c
= V
c
sin
c
t
(c denotes carrier) and
v
m
= V
m
cos
m
t
(m denotes modulation)
What is Frequency?
If the frequency is varying, how can it be defined?
You can no longer count the number of cycles over a longish interval to determine the
cycles per second. Instead, frequency is defined as the rate of change of phase.
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
This is consistent with the simple definition because, at a constant (angular) frequency
radians/second, the phase is changing at radians per second, which is /2 cycles per
second.
Since the instantaneous frequency can only be defined by reference to the phase, the
phase must be examined in order to arrive at an expression for the frequencymodulated
signal.
Phase of the FM Signal
For the unmodulated carrier v
c
= V
c
sin
c
t, the phase is:
=
c
t
The modulating signal varies the carrier frequency,
c
, so that its frequency takes the
form:
=
c
+ D cos
m
t
(where D denotes the peak value of the deviation).
It is related to the amplitude of the modulating signal v
m
by the 'frequency slope' of the
frequency modulator (VCO), say k radians/s per V.
The peak value of v
m
produces deviation D, so:
D = k V
m
The total phase change undergone at time t is found by integrating the angular frequency.
It is
= (
c
+ D cos
m
t) dt
=
c
t + (D/
m
) sin
m
t
(If you are not familiar with integration you will have to take this result on trust).
So the FM signal can be expressed as:
V
c
sin [
c
t + (D/
m
) sin
m
t]
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Modulation Index
In the expression for the FM signal:
V
c
sin [
c
t + (D/
m
) sin
m
t]
the coefficient D/
m
turns out to be quite important and is given the name modulation
index.
It is often represented by the Greek letter beta, .
So we may write the FM signal as:
v
c
= V
c
sin (
c
t + sin
m
) t
where is the modulation index D/
m
.
In this expression, the factor sin (
c
t + sin
m
)t (let us call it F) is of the form sin(a + b),
which can be expanded to sin a cos b + cos a sin b.
Applying this expansion to F, we get:
F = sin
c
t cos(sin
m
) t + cos
c
t sin (sin
m
) t
FM Sidebands
These complicated functions can be expanded, using mathematics too elaborate to
explain here, into a series of terms like this:
F = J
0
( ) sin
c
t+ J
1
( ) [ sin (
c
+
m
)t  sin (
c

m
)t ]
+ J
2
( ) [ sin (
c
+ 2
m
)t  sin (
c
 2
m
)t ]
+ J
3
( ) [ sin (
c
+ 3
m
)t  sin (
c
 3
m
)t ]
+ J
4
( ) [ sin (
c
+ 4
m
)t  sin (
c
 4
m
)t ]
+ ...
where J
0
( ), J
1
( ), J
2
( ) etc are constants whose values depend only on . They are
called Bessel Functions.
There is an infinite series of these functions, and so an infinite number of FM sidebands.
But, in practice the values of the Bessel functions become very small as the series goes
on. For example, when = 2
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
J
0
(2) = 0.224
J
1
(2) = 0.577
J
2
(2) = 0.353
J
3
(2) = 0.129
J
4
(2) = 0.034
J
5
(2) = 0.007
A Practical Approximate Rule
Because the higherorder sidebands become very small, in practice the bandwidth of the
FM signal may be restricted to a finite bandwidth.
The practical rule that is used, often called Carsons Rule, is to take the bandwidth
required as:
B = 2 ( F
d
+ F
m
)
where B is the bandwidth, F
d
the deviation and F
m
is the bandwidth of the modulation,
all in the same units.
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
The Phase Locked Loop
A phase locked loop (PLL) is a sub system that enables an oscillator to be synchronized
in frequency and phase to an incoming signal. The block diagram shows the building
blocks that make up a phase locked loop.
Imagine that the voltage controlled oscillator (VCO) is oscillating near to the incoming
signal frequency. The output of the phase/frequency comparator is a signal that
represents the frequency error between the VCO and the incoming signal. This signal is
applied to the frequency control input of the VCO, which then changes its frequency to be
equal to the incoming signal. The output of the comparator then compares the phases of
the two signals and uses the VCO frequency control to match the two phases. The system
is now in lock. If either the signal or the VCO moves in phase with respect to each other
the comparator output moves the VCO so that the two are always locked together.
In fact most PLLs only use a phase comparator (detector). This is because phase
detectors, when presented with two different frequencies, produce an ac signal equal in
frequency to the difference between them. This has the effect of swinging the VCO up and
down in frequency and, as it passes the signal frequency, the loop locks.
Loop Stability
One of the problems that will almost certainly arise, unless steps are taken to stop it, is
instability. The loop relies on the system operating with negative feedback, i.e. if the VCO
moves, the polarity of the control signal brings it back. This is easily done when the
system is operating at, or near to, dc. However, a problem arises if you consider the loop
moving in response to a fast changing frequency. The control signal will contain an ac
component. All systems are subject to delays and phase shifts, which become more
significant at higher frequencies. Remembering that 180 degrees phase shift is equivalent
to inverting a signal, inevitably there is going to be a frequency at which the phase shift
round the loop is enough to cause the polarity to reverse and positive feedback will be
applied. This results in the system oscillating back and forth at the frequency which
produces the positive feedback.
There is another subtle problem that results in the design of a PLL system being more
difficult then you might imagine. Remember that we are using the frequency of an
oscillator to control its phase. Now phase is the integral of frequency and so there is
already a 90 degree shift caused by the VCO. This means that there only has to be
another 90 degrees of phase shift before instability, not 180 as you might have thought.
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
The same problem occurs, for example, in a mechanical position control system that uses
speed control, because position is the integral of speed. The phase lock loop is a control
system and exactly the same mathematics can be used to describe a PLL as is used to
describe a position control servo.
Of course, instability can only occur if there is enough gain in the loop at the problem
frequency. This is where the loop filter can solve the problem, as it reduces gain at higher
frequencies while maintaining control over phase. There will be a frequency that the loop
filter produces 90 degrees of phase shift, but the gain will be low and so instability will not
arise. The critical frequency is when the overall loop gain is 1 and the overall added phase
shift must be less than 90 degrees at this point. The amount by which it is less is called
phase margin and, in practice, should be about 45 degrees for good stable performance.
The design of the loop filter is not simple and has to be done knowing all the gains and
phase shifts in the system.
Many phase comparators produce a control signal that contains a significant amount of
high frequency energy but this is not normally a problem as it is removed by the loop filter.
In PLLs that use phase only comparators, the bandwidth of the loop filter also determines
the range over which the loop will lock on, or capture, a signal as the comparator only
generates an ac signal off lock. The range over which the loop will capture a signal is
called its capture range. The range over which the loop will remain locked, once lock is
achieved, is called the lock range. The time to achieve lock can be important and is
referred to as lock time.
Phase Comparator
A number of circuits will operate as phase comparators or detectors. A multiplier is often
used. If the two inputs of a multiplier are fed with two signals that are at the same
frequency, but with different phases, the output will comprise a twicefrequency
component and a dc component that represents the phase error. There are some
important restrictions to this, in that it will only operate over 180 degrees and has zero
output at 90 degrees, not zero. The graph shows the output voltage for such a detector
plotted against input phase difference.
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
This shows that the output repeats for 180 to 360, albeit in the other polarity.
This phase range problem is not significant as, in a properly operating loop, the gain is
such that only a small error has to occur before the VCO is corrected.
There are other types of comparator. The logic function exclusive OR is exactly the same
in action as the multiplier and is often used in digital circuits. More complex digital circuits
have advantages, such as: acting as frequency comparators as well as phase
comparators; operating over 360 degrees; and having less ac signal component in the
output. Most of these are based on circuits using D type flipflops. These D type
comparators have the disadvantage of making the loop much less tolerant of noise in the
signal. They are used in applications such as frequency synthesizers, while the multipliers
and OR gates are used in applications such as demodulators or carrier reference
recovery.
The design of PLLs is a complex compromise of performance parameters, the relative
importance of each performance parameter depends, to a large extent, on the application.
Applications
Many phase lock loops are used to recover some sort of constant frequency component to
provide a reference for a demodulator.
Another very common application is in frequency synthesizers, where an oscillator is
frequency divided to some low frequency and a PLL locks it to an external reference. By
changing the divider ratio, different frequencies that are all multiples of the reference
frequency can be generated or synthesised.
The PLL can also be used to demodulate FM as, when locked to the FM signal, the VCO
tracks the frequency modulation. Therefore the control signal to the VCO contains the
modulation, plus a dc component. This dc component can easily be removed using a high
pass filter. In this case the loop filter bandwidth must be high enough to pass the
modulation or the VCO will not be able to follow and the loop will come out of lock. In most
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
cases the output is passed through a post detection filter, which will remove any remaining
high frequency components but, because it is outside the loop, will not affect loop stability.
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Intersymbol Interference
Intersymbol interference is a particular type of distortion applicable to digital signals. It
simply refers to the fact that the present symbol may be distorted by the values of the
symbols on either side of it.
For example, if a post detection filter had insufficient bandwidth and the signal did not
have time to reach its maximum output during a 1 symbol, if the previous symbol was
zero, then this would be regarded as intersymbol interference.
More subtle problems may occur if there are reflections in a cable, or on radio signals,
causing energy from other symbol periods to arrive at the same time.
All communication systems use filtering to maximize the signaltonoise ratio or prevent
other signals causing interference. Any filtering will cause some intersymbol interference
and it is necessary to find the right compromise between too little filtering and too much
distortion. Some systems, such as GMSK (Gaussian Minimum Shift Keying), are designed
to tolerate significant distortion, in order to reduce their occupied bandwidth.
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Symbol Rate and Bit Rate
The concepts of symbols, bits, symbol rate and bit rate are important terms in digital
communications.
The concept of a bit (a binary digit) should be familiar as a one or zero in a binary data
stream. The bit rate is simply the rate at which the bits change. For example, imagine a
system that digitized an audio signal at 32k samples per second, each sample being
digitized at 256 possible levels. This means each sample is an 8 bit word. In order to send
this stream over a simple link it would have to be turned into serial data. This means the
serial data stream would run at 32k x 8 = 256k bits per second. This is the bit rate. In this
example we are assuming that there is no extra data for synchronization or for error
correction.
These bits are then modulated onto the carrier in some form. In order to be modulated
they have to be converted to change some parameter of the carrier: its amplitude,
frequency or phase. In a simple system there would be only two states: off or on, one
frequency or the other, one of two phases etc. These states are called symbols.
In the simplest binary system there are only two symbols and each bit has two possible
states so the bits are directly mapped to symbols. This means that the symbol rate is
equal to the bit rate.
There is no reason why there have to be only two possible carrier states. In an amplitude
shift keying (ASK) system there could be more than two possible amplitude states, or in
phase shift keying (PSK) system there could be other possible phases than zero and 180
degrees. If there you had a PSK system with four possible states then each transmitted
data symbol can be decoded as being one of four states. Therefore, not one but two bits
can be carried per symbol. Now, if the bit rate remains the same, we only need to transmit
symbols at half the rate. In such a system the symbol rate is half the bit rate. If there were
16 symbols available then 4 bits per symbol could be carried and the symbol rate would
be one quarter the bit rate. Such systems are called Mary , where M is the number of
possible symbols, sometimes referred to as the order of the modulation scheme.
In such a system the bit rate (B) is:
M S B
2
log =
where S is the symbol rate and M the number of possible symbols.
To avoid confusion this bit rate is sometimes called the gross bit rate
It is important to remember that it is the symbol rate that is the rate at which the carrier
changes state. Therefore, it determines the occupied bandwidth.
It is clear that for a given bandwidth, the higher the order of the modulation scheme the
less bandwidth is used. However there is a penalty to be paid. When demodulated, the
higher the order of the scheme the more likely there are to be errors. This is obvious
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
because, for example, it is clearly easier to detect the difference between 0 and 180
degrees than zero, 90, 180, and 270.
There is another compromise to be made if error correcting data is added in that, although
adding extra data reduces the number of errors, the bit rate has to rise, with a
consequential increase in occupied bandwidth and received noise.
In order to calculate the amount of useful data that can be transmitted through a digital
system, first find the symbol rate. Then calculate the bit rate by using the number of bits
per symbol. The useful data, sometimes referred to as the payload, can then be
calculated by subtracting the extra data added for error correction, data identification and
synchronisation.
In a multiplexed system more than one data stream may be present and you may have to
find out what proportion of the data stream is allocated to a particular set of data. In very
complex systems this proportion may not even be constant!
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Practical 1: Generating and Demodulating Frequency Shift Keying
Objectives and Background
Frequency shift keying (FSK) is the transmission of digital data using frequency
modulation. The simplest form of digital data, a binary bit stream, only contains two levels
and these would be directly mapped to two frequencies. This means that the carrier simply
switches between two discrete frequencies. There are also systems that use multilevel
digital signals in order to conserve bandwidth by increasing the symbol rate.
One advantage of frequency shift keying is that there are no amplitude changes; it is the
frequency that moves, therefore advancing or retarding the phase linearly during a bit
period. A signal that has no changes in amplitude can be passed, without distortion,
through amplifiers that have non linear amplitude characteristics. All amplifiers are slightly
nonlinear and this can result in unwanted sidebands being produced and thus the
occupied bandwidth increases. This is often referred to as spectrum regrowth.
Frequency modulated signals, analogue or digital, can be passed without problems
through amplifiers that are highly nonlinear. The advantage of this is that amplifiers can
be made very power efficient at the expense of linearity and this is important where heat
dissipation or battery life is an issue.
A VCO can be used to generate FSK by simply feeding the control input with two voltage
levels, representing one and zero. The magnitude of the voltage change is the deviation,
referred to in FSK as the frequency shift.
The value of frequency shift can vary widely. It can be a few percent of the carrier
frequency or a factor of two to one, depending on the application. The occupied bandwidth
depends not only on the frequency shift but also, of course, on the rate at which the
frequency is switched. This is the bandwidth of the keying signal. Carsons rule still
applies, as FSK is simply FM with a square wave signal as the modulation.
There is usually some effort made to limit the maximum rate of frequency shift by using a
premodulation filter. This has to be wide enough not to introduce too much intersymbol
interference.
FSK has been used from the early days of digital communication systems, when the most
advanced technology was a radio link and a teleprinter. It is still widely used today in
applications as diverse as radio systems and telephone modems.
In this Practical you will generate FSK using a VCO modulated by a square wave signal
that represents a bit stream. You will demodulate it using a PLL. The techniques are very
similar to those used for analogue FM.
The next Practical will address the concepts oFMinimum shift keying and aggressive pre
modulation filtering, which together reduce bandwidth requirements.
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Block Diagram
Make Connections Diagram
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Practical 1: Generating and Demodulating Frequency Shift Keying
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Initially you have connected up the circuit such that the Premodulation filter is not in
circuit.
Open the oscilloscope and the frequency counter.
Set the Function Generator to Fast and select a square wave. Set the Signal Level
Control to full scale to give maximum modulation. Set the IQ Demodulator controls to half
scale.
Set the Frequency of the Function Generator to about 7kHz. This is the frequency of the
modulation.
Note the data signal (square wave) on the lower oscilloscope trace and the upper trace
(the carrier) showing no amplitude variation.
Increase the oscilloscope timebase to maximum speed and use the x expand to see the
individual cycles of the carrier. Change the trigger to Channel 1 by deselecting Y2 Trig.
You should be able to see the carrier changing between two frequencies.
Use the Defaults button to return the oscilloscope to the original settings.
Open the spectrum analyser and note that the two possible frequencies for the carrier are
clearly visible. Adjust the modulation amplitude using the Signal Level Control and note
how the frequency shift changes.
Use the Function Generator Frequency control to increase the frequency of the data
(modulating) signal to about 40kHz. Note that the spectrum now shows a number of
sidebands and the bandwidth is greater than the frequency shift.
Refer to the Make Connections diagram and remove connection 3 and add connections 2
and 4. This has now connected the Premodulation filter into the circuit.
Move the oscilloscope Channel 2 probe (yellow) to the output of the Premodulation filter
(monitor point 3). Note, using the spectrum analyser, that the bandwidth of the signal has
been reduced and, on Channel 1 of the oscilloscope, that the output of the premodulation
filter shows the edges of the data signal are less sharp.
Adjust the Frequency of the data signal and note the effect as the frequency nears the
cutoff of the filter.
Ensure that the Loop Filter Compensation switch is set to Fast.
Set the Frequency of the modulating data signal to 3kHz.
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Open the voltmeter and the use it to set the amplitude to 0.15 volts ac peak to peak. Move
the oscilloscope Channel 1 probe (blue) to the output of the post detection filter (monitor
point 4).
Set the oscilloscope timebase to 50S per division. You should be able to see the phase
lock loop demodulating the signal. You may need to adjust the dc offset into the loop filter
for the loop to lock (use the dc Source control).
Use the Signal Level Control to increase the amplitude of the modulation and thus
increase the frequency shift above the PLL loop filter bandwidth. You will see that the loop
cannot maintain lock.
Use the Function Generator Frequency control to increase the data (modulation)
frequency. Note that as the frequency is increased the demodulated output becomes more
sinusoidal as the frequency nears the loop filter cutoff.
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Practical 2: Minimum Shift Keying
Objectives and Background
The control of signal bandwidth is an important consideration in a communications
system.
Of course, there are other considerations, such as noise immunity, or how easy the
system is to implement. This last consideration might be very important in a system such
as a mobile phone where the size and power consumption of a handset is critical.
Frequency shift keying (FSK) has the advantage of there being no amplitude variation and
so is able to pass through high efficiency, nonlinear amplifiers without distortion.
In Practical 1 you investigated frequency shift keying and it may have seemed that the
choice of frequency shift is somewhat arbitrary. However, you saw that the smaller the
shift the narrower the signal bandwidth (remembering that it cannot be less than the
bandwidth of the data signal). However, a very small shift could be almost impossible to
detect. What might be the optimum value to keep the bandwidth low but still make
demodulation easy?
Such a system is called minimum shift keying (MSK) and can be shown to be when the
shift is made half the symbol rate. MSK has another important feature, the understanding
of which depends on the relationship between frequency and phase. In MSK, the phase
advances when the upper frequency is sent such that it reaches +90 degrees at the end of
the symbol. When the lower frequency is sent the phase retards and is 90 degrees at the
end of the symbol. This means that phase detection techniques can be used both to
modulate and demodulate MSK. This is quite attractive, as generating an accurate
frequency deviation is very difficult but, by using the I and Q techniques, the detection of
90 degree phase shifts is easier.
All these features make MSK very suitable for mobile phone systems. A derivative of MSK
called GMSK (Gaussian Minimum Shift Keying) is used in the GSM phone network.
In this Practical you will use an IQ modulator to generate an MSK signal. A simple phase
demodulator is then used to demodulate the signal. The output of the demodulator
represents the phase changes in signal, not the frequency changes. However, as we know
that the derivative of phase is frequency, the output can simply be passed through a
differentiator block to recover the original data. Unfortunately, such a block has a high
pass filter characteristic and therefore has the effect of making any noise in the system
more significant. In a real MSK system, the phase detector output would be used directly
and processed so to minimize noise.
Note also that in this practical only a stream of ones and zeros is used. In a real system
there will be situations when two ones or two zero follow each other. This means that the
total phase shift will be greater than plus or minus 90 degrees. This would be dealt with by
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
using an IQ demodulator, rather than the simple phase demodulator used in this Practical.
Note also that, when you examine the MSK signal with the phasescope, there is a
continuous frequency difference between the carrier and the modulated signal. This is not
significant, since it is the phase difference at the end of each symbol that is important.
However, it does make it difficult to see the 90 degree shift on the phasescope. In the
Practical, this problem is resolved by using the local oscillator of the demodulator as the
reference channel of the phasescope. The local oscillator is locked to the residual carrier
of the modulated signal by a phase lock loop and hence follows the carrier frequency.
As you have seen, the bandwidth of the modulated signal depends on the deviation, the
symbol rate and the rate of change at the symbol transitions. By adding a filter in the data
signal, such that its magnitude only just reaches the symbol value at the end of the
symbol, the occupied bandwidth is minimized. The filter has to have a characteristic such
that, while providing filtering, the phases of all the harmonic components of the signal are
preserved as much as possible. Of the various types of filter available the Gaussian filter
offers the best compromise. By adding such a filter the bandwidth is minimized and the
system is referred to as Gaussian Minimum Shift Keying, or GMSK.
In this Practical you will use a square wave to represent data. This is the equivalent of a
series of ones and zeros. Note that, since a symbol is a one or zero and each cycle of the
square wave is a one and a zero, the equivalent symbol rate is twice the square wave
frequency.
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Block Diagram
Make Connections Diagram
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Practical 2: Minimum Shift Keying
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the Integrator switch to Fast, the Function Generator switch to Fast and the Loop
Filter Compensation switch to Slow. Set the IQ Modulator and IQ Demodulator
controls to half scale.
Open the phasescope and use the Variable Phase Shift control associated with the
Carrier Source block to set the I and Q carrier phase difference to 90 degrees.
The first part of the Practical is to estimate the deviation sensitivity of the frequency
modulator block. You will do this by using dc voltages to set the modulation input and then
measuring the resulting frequencies.
Open the voltmeter and frequency counter. Use the dc Source control to set the voltage
to the Integrator to +0.4 volts and measure the output frequency. Note the value and set
the dc voltage to 0.4v. Measure the frequency again. Now calculate the frequency
difference divided by the voltage difference. This will be the modulation sensitivity in kHz
per volt.
Refer to the Make Connections diagram and remove connection 20 and add connection 2.
This changes the modulation source to the Function Generator output.
Open the oscilloscope. Select a square wave from the Function Generator. Move the
oscilloscope Channel 2 probe (yellow) to monitor point 2.
Move the frequency counter probe (orange) to the Signal Level Control output (monitor
point 3). Set the Function Generator Frequency to about 15kHz. Set the voltmeter to ac
pp and use the Signal Level Control to set the ac amplitude to about 0.2 volts pp. Move
the phasescope main channel probe (blue) to the MSK Generator output (monitor point 5)
and set the phasescope to Constellation.
Note that you can see the phase shift but the reference position is moving. In reality it is
moving round at a speed that the phasescope cannot track. Move the reference probe
(yellow) to the local oscillator VCO output (monitor point 6). You should now have a
stable display, with the phase varying over an arc of phases (probably 20 degrees total
variation, centred on somewhere less than 90 degrees). If you do not, adjust the dc
Source control to lock the VCO and thus give a stable display.
Use the Signal Level Control to increase the modulation amplitude until the arc over
which the phase shift varies is 90 degrees. This is not easy to estimate but do your best.
Now measure the modulation ac pp amplitude on the voltmeter. By using the value you
have calculated for the frequency sensitivity of the modulator, you can now calculate the
frequency deviation required for this 90 degrees shift per symbol. Remember here that the
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
symbol rate is represented by twice the function generator frequency. This should confirm
that the deviation is half the symbol rate for MSK.
On the oscilloscope you should be able to see the phase changing over the 90 degree
range. Move the oscilloscope channel 1 probe (blue) to the output of the integrator
(monitor point 4). The signal should be an integrated square wave (i.e. a triangle). You will
have to adjust the oscilloscope timebase to see this clearly. Compare this to the phase
detector output at the post detection filter (monitor point 7).
The differentiated output at monitor point 8 should be similar the original data (on monitor
point 3), although the differentiated output will probably be somewhat rounded.
Move the Y2 probe (blue) back to the phase detector output (monitor point 7) and increase
the Signal Level Control to increase the deviation, so the phase shift is greater than 90
degrees. Observe the fact that the output is no longer a triangle wave but has become
distorted.
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Practical 3: Multilevel Frequency Shift Keying
Objectives and Background
In Practical 1 binary FSK was generated, where two frequencies were transmitted. It is
possible to have a system where the data causes the carrier to switch between one of,
say, 4 or 8 frequencies. Such as system is called multi frequency shift keying or MFSK.
Many different systems exist, using anything from 4 to 50 frequencies. They can be very
efficient but, as the number of frequencies increases, so does the difficulty in
demodulating them.
Using more than two frequencies will increase the bit rate for a constant symbol rate.
MFSK can also be used to increase the noise immunity of a system by keeping the bit rate
the same but increasing the number of frequencies and hence reducing the symbol rate.
Note also that the bandwidth depends on the symbol rate BUT cannot be less than the
total range of the FSK frequencies. This means that, in a system that uses MFSK to
increase noise immunity, the bandwidth is often increased as a consequence.
Demodulation of higher order MFSK is usually performed by using DSP, because
implementing all the required analogue filtering would need so much electronic hardware.
In this Practical you will see 4, 8 and 16 frequencies being used.
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Block Diagram
Make Connections Diagram
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Practical 3: Multilevel Frequency Shift Keying
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the Signal Level Control to half scale.
Open the oscilloscope and note the lower trace (yellow) showing the signal being applied
to the VCO. Use the buttons on the block diagram to change to 8 level and 4 level data
and check that the number of levels corresponds with the changes in data format. You
may need to increase the size of the oscilloscope to verify this.
Increase the timebase speed on the oscilloscope and change the trigger to Channel 1
(deselect Y2 Trig). Note the carrier frequency (blue trace) changing.
Set the Compensation switch associated with the Loop Filter to Fast.
Move the Channel 1 probe (blue) to the post detection filter demodulated output (monitor
point 4). The PLL may not be locked so you may not see demodulated recognisable data.
Turn the modulation deviation down (using the Signal Level Control) and use the dc
Source control to lock the loop.
You should be able to see the multilevel data on the demodulated output. Again you
should appreciate that although more data is carried per symbol it is more likely that the
wrong symbol will be recovered.
Move the oscilloscope Channel 1 probe back to the VCO output (monitor point 1). Use the
button on the block diagram to select 4 level data.
Open the spectrum analyser. Use the default button on the oscilloscope to return the
settings to the original ones.
Use the Signal Level Control to turn the modulation deviation to maximum. You should
be able to see a typical FSK spectrum on the analyser.
Change the data to 8 level and 16 level. The spectrum becomes less like a binary FSK
signal.
In order to see the individual frequencies, use the buttons on the block diagram to reduce
the symbol rate to 5 symbols per second. You can now see the individual frequencies on
the spectrum analyser.
Increase the oscilloscope timebase speed and use the expand control so you can see
individual carrier cycles. Trigger off Channel 1 (deselect Y2 Trig) and now see how the
frequency is changing. Change the number of levels in the data and compare the number
of individual frequencies shown on the spectrum analyser to the number of levels in the
data.
Chapter 10
Modulation and Coding Principles Frequency Shift Keying
53230
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
Phase Modulation
Objectives
To appreciate the principle of phase modulation
To generate a phase modulated signal using an IQ modulator and investigate the
spectrum and bandwidth associated with it
To investigate the demodulation of a phase modulated signal using residual carrier and a
phase locked loop
To investigate the demodulation of a phase modulated signal using a frequency
demodulator
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
The Phase Locked Loop
A phase locked loop (PLL) is a sub system that enables an oscillator to be synchronized
in frequency and phase to an incoming signal. The block diagram shows the building
blocks that make up a phase locked loop.
Imagine that the voltage controlled oscillator (VCO) is oscillating near to the incoming
signal frequency. The output of the phase/frequency comparator is a signal that
represents the frequency error between the VCO and the incoming signal. This signal is
applied to the frequency control input of the VCO, which then changes its frequency to be
equal to the incoming signal. The output of the comparator then compares the phases of
the two signals and uses the VCO frequency control to match the two phases. The system
is now in lock. If either the signal or the VCO moves in phase with respect to each other
the comparator output moves the VCO so that the two are always locked together.
In fact most PLLs only use a phase comparator (detector). This is because phase
detectors, when presented with two different frequencies, produce an ac signal equal in
frequency to the difference between them. This has the effect of swinging the VCO up and
down in frequency and, as it passes the signal frequency, the loop locks.
Loop Stability
One of the problems that will almost certainly arise, unless steps are taken to stop it, is
instability. The loop relies on the system operating with negative feedback, i.e. if the VCO
moves, the polarity of the control signal brings it back. This is easily done when the
system is operating at, or near to, dc. However, a problem arises if you consider the loop
moving in response to a fast changing frequency. The control signal will contain an ac
component. All systems are subject to delays and phase shifts, which become more
significant at higher frequencies. Remembering that 180 degrees phase shift is equivalent
to inverting a signal, inevitably there is going to be a frequency at which the phase shift
round the loop is enough to cause the polarity to reverse and positive feedback will be
applied. This results in the system oscillating back and forth at the frequency which
produces the positive feedback.
There is another subtle problem that results in the design of a PLL system being more
difficult then you might imagine. Remember that we are using the frequency of an
oscillator to control its phase. Now phase is the integral of frequency and so there is
already a 90 degree shift caused by the VCO. This means that there only has to be
another 90 degrees of phase shift before instability, not 180 as you might have thought.
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
The same problem occurs, for example, in a mechanical position control system that uses
speed control, because position is the integral of speed. The phase lock loop is a control
system and exactly the same mathematics can be used to describe a PLL as is used to
describe a position control servo.
Of course, instability can only occur if there is enough gain in the loop at the problem
frequency. This is where the loop filter can solve the problem, as it reduces gain at higher
frequencies while maintaining control over phase. There will be a frequency that the loop
filter produces 90 degrees of phase shift, but the gain will be low and so instability will not
arise. The critical frequency is when the overall loop gain is 1 and the overall added phase
shift must be less than 90 degrees at this point. The amount by which it is less is called
phase margin and, in practice, should be about 45 degrees for good stable performance.
The design of the loop filter is not simple and has to be done knowing all the gains and
phase shifts in the system.
Many phase comparators produce a control signal that contains a significant amount of
high frequency energy but this is not normally a problem as it is removed by the loop filter.
In PLLs that use phase only comparators, the bandwidth of the loop filter also determines
the range over which the loop will lock on, or capture, a signal as the comparator only
generates an ac signal off lock. The range over which the loop will capture a signal is
called its capture range. The range over which the loop will remain locked, once lock is
achieved, is called the lock range. The time to achieve lock can be important and is
referred to as lock time.
Phase Comparator
A number of circuits will operate as phase comparators or detectors. A multiplier is often
used. If the two inputs of a multiplier are fed with two signals that are at the same
frequency, but with different phases, the output will comprise a twicefrequency
component and a dc component that represents the phase error. There are some
important restrictions to this, in that it will only operate over 180 degrees and has zero
output at 90 degrees, not zero. The graph shows the output voltage for such a detector
plotted against input phase difference.
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
This shows that the output repeats for 180 to 360, albeit in the other polarity.
This phase range problem is not significant as, in a properly operating loop, the gain is
such that only a small error has to occur before the VCO is corrected.
There are other types of comparator. The logic function exclusive OR is exactly the same
in action as the multiplier and is often used in digital circuits. More complex digital circuits
have advantages, such as: acting as frequency comparators as well as phase
comparators; operating over 360 degrees; and having less ac signal component in the
output. Most of these are based on circuits using D type flipflops. These D type
comparators have the disadvantage of making the loop much less tolerant of noise in the
signal. They are used in applications such as frequency synthesizers, while the multipliers
and OR gates are used in applications such as demodulators or carrier reference
recovery.
The design of PLLs is a complex compromise of performance parameters, the relative
importance of each performance parameter depends, to a large extent, on the application.
Applications
Many phase lock loops are used to recover some sort of constant frequency component to
provide a reference for a demodulator.
Another very common application is in frequency synthesizers, where an oscillator is
frequency divided to some low frequency and a PLL locks it to an external reference. By
changing the divider ratio, different frequencies that are all multiples of the reference
frequency can be generated or synthesised.
The PLL can also be used to demodulate FM as, when locked to the FM signal, the VCO
tracks the frequency modulation. Therefore the control signal to the VCO contains the
modulation, plus a dc component. This dc component can easily be removed using a high
pass filter. In this case the loop filter bandwidth must be high enough to pass the
modulation or the VCO will not be able to follow and the loop will come out of lock. In most
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
cases the output is passed through a post detection filter, which will remove any remaining
high frequency components but, because it is outside the loop, will not affect loop stability.
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
Intersymbol Interference
Intersymbol interference is a particular type of distortion applicable to digital signals. It
simply refers to the fact that the present symbol may be distorted by the values of the
symbols on either side of it.
For example, if a post detection filter had insufficient bandwidth and the signal did not
have time to reach its maximum output during a 1 symbol, if the previous symbol was
zero, then this would be regarded as intersymbol interference.
More subtle problems may occur if there are reflections in a cable, or on radio signals,
causing energy from other symbol periods to arrive at the same time.
All communication systems use filtering to maximize the signaltonoise ratio or prevent
other signals causing interference. Any filtering will cause some intersymbol interference
and it is necessary to find the right compromise between too little filtering and too much
distortion. Some systems, such as GMSK (Gaussian Minimum Shift Keying), are designed
to tolerate significant distortion, in order to reduce their occupied bandwidth.
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
The IQ Modulator
The IQ modulator is a most useful building block in communications systems. It is
available as an integrated circuit with different models operating over a wide range of
frequencies.
It comprises two balanced modulators with their carrier inputs fed from the same source
but one shifted by 90 degrees. The two modulation inputs are available for the user. The
outputs of the two modulators are then summed.
The name IQ modulator comes from In phase and Quadrature. The term quadrature
simply means at 90 degrees.
The diagram shows the basic IQ modulator.
Provided the phase shift is 90 degrees at the carrier frequency then, in vector terms, the
output with respect to the input is shown below:
This means that the output is a signal at the carrier frequency and its phase will depend
on the values of the I and Q modulation inputs. Notice that the amplitude will also vary,
because the output is the result of summing two equal values at right angles. It is
therefore 1.414 times the value of when only one input is present.

+
+I
I
output
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
In mathematical terms the output is the a+jb (complex) sum of the I and Q modulation
inputs. The two modulation inputs can be two quite separate signals. This is how QAM is
generated.
If the output is required to be a phase vector with constant amplitude, the angle of which is
determined by a single input, that input signal has to be processed to generate suitable I
an Q signals.
Since an output is required that is a vector represented by r, where r is the required
constant radius and is the variable angle, and what we have is a+jb, this is done using
the equivalent of changing the mapping in the normal way.
i.e. for an input representing an angle of
sin
mod
M I =
cos
mod
M Q =
Where M is the magnitude of the required signal to drive the modulators.
By generating both the I and Q modulation inputs by processing a single input
representing angle, the output vector is of constant length and is driven round a circle
rather than a square.
This is, of course, a phase modulator.
Q
+Q
+I
I
output
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
In practical terms, the accuracy of the processing of all these signals depends on the
accuracy with which the 90 degree phase shifts can be maintained.
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
Practical 1: Generating Phase Modulation using an IQ Modulator
Objectives and Background
In this Practical you will see how to generate an analogue phase modulated signal using
an IQ modulator.
There are other ways of producing phase modulation, some use some type of reactance
modulation of a tuned circuit in an amplifier. However, the IQ modulator is the most
accurate and repeatable method and, due to the availability of integrated circuit IQ
modulators, it is a common and inexpensive technique.
The IQ modulator on its own cannot produce an output vector from one input signal
representing phase. The input signal is first processed into its sine and cosine
components and these two signals are then applied to the I and Q modulation inputs. How
IQ modulators work is covered in the Resources section.
In this Practical you will first use a dc signal to drive the phase modulator and see the
results on both the oscilloscope and the phasescope. Then you will use a sinusoidal signal
to generate an analogue phase modulated signal.
The magnitude of the phase modulation is the phase index and is the magnitude of the
phase deviation either side of zero. Since it is an angle, it is measured in degrees or
radians.
This is similar to the definition of frequency deviation in an FM system. In frequency
modulation there is almost no limit to the magnitude of the deviation but in phase
modulation the limit is clearly 180 degrees. A phase deviation of say 270 degrees is
exactly the same as 90 degrees. If the phase deviation is more than 360 degrees it
becomes a form of frequency modulation.
Note that, if the output is frequency multiplied the phase index is multiplied. This is shown
in the Practical by adding a frequency multiplier. This effect is important because, for
example, a 180 degree shift becomes a 360 degree shift when multiplied by two. This can
be used in digital PSK to recover a carrier.
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
Block Diagram
Make Connections Diagram
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
Practical 1: Generating Phase Modulation using an IQ Modulator
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the Signal Level Control to maximum and the IQ Modulator controls to half scale.
Open the voltmeter and set the dc Source to give a control voltage of zero.
Open the phasescope and use the Variable Phase Shift control associated with the
Carrier Source to adjust the I and Q carriers to be 90 degrees apart in phase.
Move the phasescope main probe (blue) to the modulator output (monitor point 6). Note
that the phase is close to zero (taking into account the +45 shift at the reference monitor
point 2). Open the oscilloscope and confirm that the phase is around zero. The exact
value is unimportant.
Now, adjust the dc Source control voltage and you will see that the phase moves. This
should be easily seen on both the oscilloscope and the phasescope.
You can estimate the phase modulator sensitivity by measuring the voltage required to
move the phase a total of 180 degrees, i.e. 90 degrees in either direction.
Set the phase difference to be 90 degrees.
Move the oscilloscope Channel 1 probe (blue) to the output of the frequency multiplier
(monitor point 7). Open the frequency counter and note the reading. Move the frequency
counter probe (orange) to the frequency multiplier output (monitor point 7). You should be
able to see from the counter and the oscilloscope that the frequency has been doubled.
Note the relative phase of the peaks and troughs of the carrier reference and the doubled
modulator output.
Return the oscilloscope Channel 1 probe (blue) to the modulator output (monitor point 6)
and set the phase to be zero. Notice that the phase of the doubled output has moved by
180 degrees.
Move the oscilloscope Channel 1 probe to the modulation (monitor point 3). Set the
Function Generator to Fast and select a sine wave output. Refer to the Make
Connections diagram and remove connection 3 and add connection 2. This makes the
modulation source the Function Generator. Move the frequency counter probe to the
modulation (monitor point 3) and set the frequency to about 15kHz. Return the
oscilloscope Channel 1 probe (blue) to the modulator output (monitor point 6).
Set the Signal Level Control to minimum. Open the spectrum analyser. Increase the
modulation and you will see the phase modulation on the phasescope. Set the
phasescope to Constellation to see it more clearly. As you increase the modulation you
will see on the spectrum analyser that side bands appear when modulation index is about
plus and minus 60 degrees.
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
Note that the amplitude of the carrier reduces as the modulation index is increased (by
increasing the Signal Level Control) but does not disappear. In fact it disappears only
when the output switches between zero and 180 degrees. This is not the case here, as
the modulation contains all levels between zero and 180. This effect is examined more
closely in the Assignment on Phase Shift Keying.
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
Practical 2:Demodulation of Phase Modulation using Residual Carrier
Reference
Objectives and Background
In this Practical you will use the phase modulator from Practical 1 and demodulate the
output. There is sufficient residual carrier to enable a PLL to lock a local oscillator and
provide a phase reference. As the PLL already contains a multiplier, acting as a phase
detector, you simply need to take the output from it before the loop filter. A post detection
filter will remove the twice carrier frequency component.
Of course, the phase detector will only function up to plus and minus 90 degrees, but that
is the maximum phase shift available anyway.
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
Block Diagram
Make Connections Diagram
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
Practical 2: Demodulation of Phase Modulation using Residual Carrier
Reference
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the IQ Modulator and the IQ Demodulator controls to half scale. Set the Loop Filter
Compensation switch to Fast.
Open the oscilloscope, voltmeter and frequency counter. Set the Function Generator to
Fast and select a sine wave output at a Frequency of 15kHz.
Move the phasescope main channel probe (blue) to the I carrier signal (monitor point 1).
Open the phasescope and use the Variable Phase Shift control associated with the
Carrier Source to set the IQ carrier phase difference to 90 degrees.
Move the phasescope main channel probe (blue) to the modulator output (monitor point 4)
and adjust the modulation amplitude to confirm that phase modulation is being generated.
Set the Signal Level Control to give a phase modulation index of 90 degrees (it may be
easier to set up if you use the Constellation setting on the phasescope).
Move the oscilloscope Channel 1 probe (blue) to the post detection filter output (monitor
point 5). Move the oscilloscope second channel probe (yellow) to the modulation input
(monitor point 3).
Change the timebase so you can see the modulation. You should be able to see the
demodulated output on the upper trace. It may be necessary to adjust the dc offset into
the loop filter (using the dc Source control) to lock the carrier recovery PLL.
Adjust the dc offset so that the local oscillator phase is in the centre of the phase
deviation. This is when the phase detector output is not distorted at the top or bottom of
the waveform.
Now increase the modulation amplitude until distortion occurs at the top and bottom of the
output waveform. Move the blue probe back to the modulator output (monitor point 4) and
the yellow probe (the phasescope reference probe) to monitor point 2. Confirm that the
range of modulated phase change is near to plus and minus 90 degrees.
Note also that the relative phase between the carrier with no modulation and the local
oscillator is 90 degrees, due to the phase detector.
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
Practical 3: Demodulation of Phase Modulation using a Frequency
Demodulator
Objectives and Background
In the previous practical you used a phase demodulator to demodulate a phase
modulated signal. In this practical you will use the same generator, but demodulate it with
a frequency demodulator.
In this instance you will use a PLL, which will act as a frequency demodulator. The only
difference between the arrangement where the PLL acts as a frequency demodulator and
the arrangement where the PLL acts as a phase demodulator is the PLL loop cutoff
frequency.
To act as a phase demodulator, the loop cutoff frequency was set such that the
modulation did not pass through it, so the local oscillator provided a constant reference at
the carrier frequency. However, if the loop filter cut off is made high enough to let the
modulation pass, the local oscillator will follow the frequency variations of the carrier. The
demodulated output will then be a signal representing the frequency changes of the phase
modulated carrier.
You have already learnt that frequency is the differential of phase, therefore you might
expect the output to be the differential of the input modulation. This is indeed the case. To
convert it back to a signal representing phase the signal is passed through an integrator.
Remembering that integrators have the problem of integrating small dc offsets and
reaching saturation, also that the integrator on the hardware board has a reset on it, you
will understand the form of the output.
This practical shows, once again, the relationship between frequency and phase, the
appreciation of which is vital to the understanding of signal processing in analogue and
DSPbased systems.
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
Block Diagram
Make Connections Diagram
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
Chapter 11
Modulation and Coding Principles Phase Modulation
53230
Practical 3: Demodulation of Phase Modulation using a Frequency
Demodulator
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the IQ Modulator and the IQ Demodulator controls to half scale.
Open the oscilloscope. Set the Function Generator to Fast and select a sine wave
output. Set the Loop Filter Compensation switch and the Integrator switch to Fast.
Open the frequency counter and voltmeter and set the Function Generator Frequency to
5kHz and use the Signal Level Control to set the amplitude to 0.4 volts ac pp.
Move the phasescope main channel probe (blue) to the I carrier signal (monitor point 1).
Open the phasescope and use the Variable Phase Shift control associated with the
Carrier Source to set the IQ carrier phase difference to 90 degrees.
Move the oscilloscope Channel 1 probe (blue) to the loop filter output (monitor point 5)
and the oscilloscope Channel 2 probe (yellow) to the modulation input (monitor point 3).
You may need to adjust the dc Source associated with the dc Offset into the buffer of the
Loop Filter in order to lock the PLL.
Reduce the timebase speed until you can see the at least a few cycles of the modulation.
The input and output are sinusoidal but 90 degrees out of phase. Remember that the
differential of a sine is a cosine.
Change the Function Generator to triangle output. Note that the demodulator output is a
squarewave (in practice, this will not be perfect). This should not be unexpected.
Try a square wave as modulation. With squarewave modulation move the oscilloscope
Channel 1 probe (blue) to the integrator output (monitor point 7). You should be able to
see that it returns the signal to a squarewave. Remember that the integrator will integrate
any dc offset and then reset. Use the dc Source control to vary the offset into the
integrator to reduce the dc drift as much as possible.
Try sine and triangle modulation inputs.
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
Phase Shift Keying
Objectives
To appreciate the principle of phase shift keying and its relationship to analogue phase
modulation
To generate a twolevel (binary) phase shift keyed signal and investigate the spectrum
and bandwidth associated with it
To investigate the demodulation of an FSK signal using residual carrier
To understand the operation of the Costas Loop circuit for phase demodulation
To investigate the demodulation of 90 degree FSK signal using a Costas Loop and using
frequency multipliers
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
The Phase Locked Loop
A phase locked loop (PLL) is a sub system that enables an oscillator to be synchronized
in frequency and phase to an incoming signal. The block diagram shows the building
blocks that make up a phase locked loop.
Imagine that the voltage controlled oscillator (VCO) is oscillating near to the incoming
signal frequency. The output of the phase/frequency comparator is a signal that
represents the frequency error between the VCO and the incoming signal. This signal is
applied to the frequency control input of the VCO, which then changes its frequency to be
equal to the incoming signal. The output of the comparator then compares the phases of
the two signals and uses the VCO frequency control to match the two phases. The system
is now in lock. If either the signal or the VCO moves in phase with respect to each other
the comparator output moves the VCO so that the two are always locked together.
In fact most PLLs only use a phase comparator (detector). This is because phase
detectors, when presented with two different frequencies, produce an ac signal equal in
frequency to the difference between them. This has the effect of swinging the VCO up and
down in frequency and, as it passes the signal frequency, the loop locks.
Loop Stability
One of the problems that will almost certainly arise, unless steps are taken to stop it, is
instability. The loop relies on the system operating with negative feedback, i.e. if the VCO
moves, the polarity of the control signal brings it back. This is easily done when the
system is operating at, or near to, dc. However, a problem arises if you consider the loop
moving in response to a fast changing frequency. The control signal will contain an ac
component. All systems are subject to delays and phase shifts, which become more
significant at higher frequencies. Remembering that 180 degrees phase shift is equivalent
to inverting a signal, inevitably there is going to be a frequency at which the phase shift
round the loop is enough to cause the polarity to reverse and positive feedback will be
applied. This results in the system oscillating back and forth at the frequency which
produces the positive feedback.
There is another subtle problem that results in the design of a PLL system being more
difficult then you might imagine. Remember that we are using the frequency of an
oscillator to control its phase. Now phase is the integral of frequency and so there is
already a 90 degree shift caused by the VCO. This means that there only has to be
another 90 degrees of phase shift before instability, not 180 as you might have thought.
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
The same problem occurs, for example, in a mechanical position control system that uses
speed control, because position is the integral of speed. The phase lock loop is a control
system and exactly the same mathematics can be used to describe a PLL as is used to
describe a position control servo.
Of course, instability can only occur if there is enough gain in the loop at the problem
frequency. This is where the loop filter can solve the problem, as it reduces gain at higher
frequencies while maintaining control over phase. There will be a frequency that the loop
filter produces 90 degrees of phase shift, but the gain will be low and so instability will not
arise. The critical frequency is when the overall loop gain is 1 and the overall added phase
shift must be less than 90 degrees at this point. The amount by which it is less is called
phase margin and, in practice, should be about 45 degrees for good stable performance.
The design of the loop filter is not simple and has to be done knowing all the gains and
phase shifts in the system.
Many phase comparators produce a control signal that contains a significant amount of
high frequency energy but this is not normally a problem as it is removed by the loop filter.
In PLLs that use phase only comparators, the bandwidth of the loop filter also determines
the range over which the loop will lock on, or capture, a signal as the comparator only
generates an ac signal off lock. The range over which the loop will capture a signal is
called its capture range. The range over which the loop will remain locked, once lock is
achieved, is called the lock range. The time to achieve lock can be important and is
referred to as lock time.
Phase Comparator
A number of circuits will operate as phase comparators or detectors. A multiplier is often
used. If the two inputs of a multiplier are fed with two signals that are at the same
frequency, but with different phases, the output will comprise a twicefrequency
component and a dc component that represents the phase error. There are some
important restrictions to this, in that it will only operate over 180 degrees and has zero
output at 90 degrees, not zero. The graph shows the output voltage for such a detector
plotted against input phase difference.
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
This shows that the output repeats for 180 to 360, albeit in the other polarity.
This phase range problem is not significant as, in a properly operating loop, the gain is
such that only a small error has to occur before the VCO is corrected.
There are other types of comparator. The logic function exclusive OR is exactly the same
in action as the multiplier and is often used in digital circuits. More complex digital circuits
have advantages, such as: acting as frequency comparators as well as phase
comparators; operating over 360 degrees; and having less ac signal component in the
output. Most of these are based on circuits using D type flipflops. These D type
comparators have the disadvantage of making the loop much less tolerant of noise in the
signal. They are used in applications such as frequency synthesizers, while the multipliers
and OR gates are used in applications such as demodulators or carrier reference
recovery.
The design of PLLs is a complex compromise of performance parameters, the relative
importance of each performance parameter depends, to a large extent, on the application.
Applications
Many phase lock loops are used to recover some sort of constant frequency component to
provide a reference for a demodulator.
Another very common application is in frequency synthesizers, where an oscillator is
frequency divided to some low frequency and a PLL locks it to an external reference. By
changing the divider ratio, different frequencies that are all multiples of the reference
frequency can be generated or synthesised.
The PLL can also be used to demodulate FM as, when locked to the FM signal, the VCO
tracks the frequency modulation. Therefore the control signal to the VCO contains the
modulation, plus a dc component. This dc component can easily be removed using a high
pass filter. In this case the loop filter bandwidth must be high enough to pass the
modulation or the VCO will not be able to follow and the loop will come out of lock. In most
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
cases the output is passed through a post detection filter, which will remove any remaining
high frequency components but, because it is outside the loop, will not affect loop stability.
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
The Costas Loop
The Costas Loop provides a method of demodulating PSK signals when the phase shift is
90 degrees, which results in there being no carrier component in the modulated signal.
Another way of looking at this is examining the phase detector characteristic. The diagram
shows the output signals from an IQ demodulator with respect to the phase difference
between the local oscillator and the incoming signal.
In signals with phase shifts less than 90 degrees, the I output slope remains the same
polarity. When the shift is 90 degrees exactly, the two symbol positions are at the peaks
of the I output. This means that, if the mean phase moves the I output polarity reverses
and therefore no longer provides a steering control voltage to bring the voltage controlled
local oscillator back. The situation is worse for QPSK as the four symbols are positioned
where no coherent control voltage is produced.
In the simple Costas loop the I and Q signals are multiplied together as shown below.
180
90 0 90 180
Phase
I output
Q output
(I) X (Q)
output
180 90 0 90 180
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
Note, that now the output polarity is the same at 90 and +90, so a coherent control
voltage is produced as the mean phase moves either side of zero. Note also, that a signal
of opposite polarity is available, by moving the mean phase by 90 degrees. This results in
the loop being insensitive to control voltage inversion. If the mean phase moves by 180
degrees then the loop will lock equally well but the output data will be inverted. This phase
uncertainty has to be dealt with by other means.
The diagram shows the simple Costas loop block diagram. The Costas loop can be
thought of as a phase locked loop with a special phase detector. The loop performs both
the function of carrier lock as well as demodulation, since the I output will contain the data.
There will be a residual amount of twicecarrier frequency component present, which can
be removed by a low pass filter.
Also, there has to be a low pass filter in the control signal to the VCO. This removes any
datarate frequency components and means that the VCO follows the mean phase. This
filter has also to provide the control stability function to prevent control loop oscillation.
This is achieved by ensuring that the control loop gain has dropped below unity before
total phase shift reaches 180 degrees. This problem is compounded by the fact that we
are using VCO frequency to control VCO phase, and therefore 90 degrees of phase is
already present in the control loop.
Inspection of the simple Costas loop phase characteristic reveals that the simple loop
would not work with QPSK as the symbols would be placed on alternate polarity slopes.
The double Costas loop provides the answer, but at the expense of additional complexity.
X
X
I osc
Q osc
X VCO
control
I
Q
signal
data
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
The VCO control phase characteristic of this arrangement is shown below.
This provides a coherent phase control voltage for all four symbol positions.
Note that, in the double Costas loop, there are two limiter amplifiers and an inversion. The
function of the limiters is to make one input of the second set of multipliers switch between
positive and negative voltages, causing the multipliers to either invert or not invert the
signal at the other input. It is interesting to note that without the inversion the output
characteristic is a square wave.
X
I osc
X
VCO
control
I
Q
signal
X
Low
Pass
Filter
Low
Pass
Filter
X
Q osc
Limiter
Limiter
+
I data
Q data
output
180 90 0 90 180
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
Of course, the loop may lock in any one of four phase positions relative to the original
carrier. This means that the original I data may appear at the Q data output and vice
versa. It may also mean that the outputs are inverted. This has to be resolved by other
means.
Here is the mathematics describing the simple Costas loop:
If the VCO is locked to the incoming carrier then:
vco
=
c
only a small phase difference,
e
will be present.
Let the two outputs from the VCO be:
2cos
c
t in phase with the carrier
2sin
c
t in quadrature
The PSK signal input is:
s(t) = A cos [
c
t + ]
where is 0 or depending on whether the state of the digital input d is 1 or 1.
So, if d(t) is the state of the digital input, this signal expression can be written:
s(t) = A d(t) cos
c
t
The multiplier outputs are the products of the two inputs to each. Thus these outputs are:
[A d(t) cos
c
t][2cos
c
t] and
[A d(t) cos
c
t][2sin
c
t]
The reference channel output is used, i.e.:
v
out
= [A d(t) cos
c
t][2cos
c
t]
= 2A d(t) cos
2
c
t
Now, cos
2
x = 0.5[1 + cos 2x], so the expression for v
out
becomes:
v
out
= 2A d(t) [0.5 + 0.5cos
2
c
t]
= A d(t) + A d(t) cos 2
c
t
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
This expression has two components: a dc component dependant on the phase of the
digital input data and a component at twice the carrier frequency. This doublefrequency
component can be removed by a post detection filter.
When the loop is in lock, the VCO will be phaselocked by modulators (2) and (3), causing
it to produce an output from its f
90
terminal that leads the incoming signal by 90 degrees.
Since the VCO produces outputs which differ by 90 degrees, the reference signal from the
f
0
output will be in phase with the incoming PSK signal for, say, binary 1 and 180 degrees
out of phase for binary 0.
The multiplying action of modulator (1) will then produce a positive dc level when the
received and reference signals are in phase and a negative level when they are in
antiphase. Subsequent data recovery circuits convert the bipolar output from the Costas
Loop demodulator into data.
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
The IQ Modulator
The IQ modulator is a most useful building block in communications systems. It is
available as an integrated circuit with different models operating over a wide range of
frequencies.
It comprises two balanced modulators with their carrier inputs fed from the same source
but one shifted by 90 degrees. The two modulation inputs are available for the user. The
outputs of the two modulators are then summed.
The name IQ modulator comes from In phase and Quadrature. The term quadrature
simply means at 90 degrees.
The diagram shows the basic IQ modulator.
Provided the phase shift is 90 degrees at the carrier frequency then, in vector terms, the
output with respect to the input is shown below:
This means that the output is a signal at the carrier frequency and its phase will depend
on the values of the I and Q modulation inputs. Notice that the amplitude will also vary,
because the output is the result of summing two equal values at right angles. It is
therefore 1.414 times the value of when only one input is present.

+
+I
I
output
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
In mathematical terms the output is the a+jb (complex) sum of the I and Q modulation
inputs. The two modulation inputs can be two quite separate signals. This is how QAM is
generated.
If the output is required to be a phase vector with constant amplitude, the angle of which is
determined by a single input, that input signal has to be processed to generate suitable I
an Q signals.
Since an output is required that is a vector represented by r, where r is the required
constant radius and is the variable angle, and what we have is a+jb, this is done using
the equivalent of changing the mapping in the normal way.
i.e. for an input representing an angle of
sin
mod
M I =
cos
mod
M Q =
Where M is the magnitude of the required signal to drive the modulators.
By generating both the I and Q modulation inputs by processing a single input
representing angle, the output vector is of constant length and is driven round a circle
rather than a square.
This is, of course, a phase modulator.
Q
+Q
+I
I
output
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
In practical terms, the accuracy of the processing of all these signals depends on the
accuracy with which the 90 degree phase shifts can be maintained.
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
Practical 1: Generating Binary Phase Shift Keying
Objectives and Background
Phase Shift Keying (PSK) uses carrier phase changes to carry digital data. The data is in
the form of bits that are mapped to symbols.
The simplest form of PSK is binary phase shift keying, where binary data is mapped to
two symbols: one representing zero and the other representing a one. The two symbols
are simply two phases of the carrier.
As in all forms of phase modulation, the phase difference between the two symbols can
be anything from fractions of a degree to 180 degrees. The most efficient value from the
point of noise immunity is 180 degrees, i.e. 90 degrees. The disadvantage of this
magnitude of shift is that in order to demodulate any phase modulation a phase reference
is required, locked to the original carrier. As you will see, when the shift is 90 degrees,
the carrier is suppressed and a more complex circuit has to be used to recreate it. For this
reason shifts of less than 90 degrees are sometimes used.
In this practical you will generate binary phase shift keying by using an IQ modulator and
you will see the effect of phase shift on the magnitude of the residual carrier. The
phasescope gives you a powerful way of examining the signal.
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
Block Diagram
Make Connections Diagram
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
Practical 1: Generating Binary Phase Shift Keying
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Open the oscilloscope. Set the Function Generator to Fast and select a square wave
output.
Open the frequency counter and use it to set the Function Generator Frequency to
15kHz.
Open the voltmeter and use the Signal Level Control to set the amplitude to 0.3 volts ac
pp.
Move the phasescope main channel probe (blue) to the I carrier input (monitor point 1).
Open the phasescope and use the Variable Phase Shift control associated with the
Carrier Source to set the IQ carrier phase difference to 90 degrees.
Move the phasescope main channel (blue) to the phase modulator output (monitor point
4). Set it to Constellation display mode.
Use the Signal Level Control to adjust the amplitude of the modulation and note that the
phase shift can be varied.
Use the X expand on the oscilloscope to see individual carrier cycles. On the Y1 Channel
(blue) you should be able to see the carrier periodically switching between the two phase
states.
Open the spectrum analyser. Use the Signal Level Control to adjust the modulation such
as to give a total phase shift of about 90 degrees. You should be able to see the carrier
and the sidebands on the spectrum analyser.
Increase the modulation to give a total shift towards 180 degrees and note that the carrier
disappears at exactly 180 degrees (this is 90 degrees, with respect to the yellow
channel, which is connected to the +45 degree carrier signal). You may need to increase
the size of the spectrum analyser to see this.
With the shift at plus and minus 90 degrees, look at the oscilloscope. You should be able
to see that at symbol transitions the amplitude reduces. This means that this type of PSK
signal contains amplitude variations and therefore has to be amplified by linear amplifiers.
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
Practical 2: Demodulation of Binary Phase Shift Keying using Residual
Carrier
Objectives and Background
As you have seen in Practical 1, when the phase shift is less than 90 degrees there is a
carrier component left in the modulated signal. In order to demodulate the signal it is
necessary to have a local oscillator at a constant phase with respect to the original carrier.
The simplest way to achieve this is to use a phase lock loop to lock to the residual carrier.
Clearly, this will only work if the shift is less than 90 degrees.
The loop must not be able to follow the modulation, so the loop filter cutoff is arranged to
be well below the symbol rate.
There would be a problem if the data contained long strings of zeros or ones, which would
allow the PLL to drift off to one of the phase symbol. In practice, this is not usually a
problem as the data is usually processed before symbol mapping by using biphase
coding, for example, to ensure that there is no long term dc component.
The PLL needs to be able to track any changes in carrier frequency of local oscillator drift
but, in a well designed system, this is not usually a problem.
In this Practical you will see that a PLL can lock onto the residual carrier but that, when the
shift is increased to 90 degrees the system fails.
In the next Practical you will see two ways of providing a phase reference when the shift is
90 degrees.
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
Block Diagram
Make Connections Diagram
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
Practical 2: Demodulation of Binary Phase Shift Keying using Residual
Carrier
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Open the oscilloscope. Set the Function Generator to Fast and select a square wave
output. Set the Compensation on the Loop Filter to Slow.
Open the phasescope and use the Variable Phase Shift control associated with the
Carrier Source to set the IQ carrier phase difference to 90 degrees.
Open the voltmeter and the frequency counter. Set the modulation frequency to 15kHz
using the Frequency control on the Function Generator. Use the Signal Level Control to
set the modulation amplitude to 0.2 volt ac pp.
Move the phasescope main channel probe (blue) to the phase modulator output (monitor
point 4) and note that the phase modulation index is about 30 degrees. Set the phase
scope to constellation display mode.
On the oscilloscope adjust the timebase so you can see one or two cycles of modulation
on Channel 2 (green). You may be able to see the demodulated output on the Channel 1
trace. If you cannot, adjust the dc offset into the loop filter (using the dc Source control)
and the loop should lock.
Using the Signal Level Control, increase the modulation phase index and see what
happens as it approaches 90 degrees (180 degrees total). The loop will unlock and,
even by adjusting the dc offset (dc Source control), you will not be able to relock it. If you
reduce the modulation phase index you should be able to lock the loop again with the dc
offset control.
This shows that such a system does not work with 90 degree shifts.
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
Practical 3: Demodulation of 90 Degree Phase Shift Keying by using a
Costas Loop and by using Frequency Multipliers
Objectives and Background
In the previous Practical you saw that, when the phase shift is 90 degrees, which is the
optimum value for noise free demodulation, there is no residual carrier in the signal to
provide a reference for a phase lock loop carrier recovery system. An alternative way of
looking at this is that, when the phase reverses, the polarity of the loop control voltage
reverses, preventing it working. A further modulator can be added to reverse the signal
polarity, using the data as the control. Such a system is often referred to as a Costas
loop and is used extensively to demodulate PSK.
See the Concepts section on the Costas loop for a detailed explanation of how it works.
In this practical you will see a Costas loop operating. It is worth noting that the Costas loop
does not work with phase shifts other than 90 degrees. There are systems that use
more than two phase symbols, for example Quadrature Phase Shift Keying (QPSK), which
uses four phase states, and modifications of the Costas loop are required for these.
Another issue is that the demodulator cannot tell the difference between plus 90 degrees
and minus 90 degrees, so there is an ambiguity in the polarity of the output signal. This
means that zeros may become ones, and viceversa. The only solution to this is to
examine the data for a known pattern and invert it if necessary.
There is another method of carrier recovery that makes use of the observation that the
phase shift is multiplied if the frequency is multiplied. For example, if you have a total shift
of 180 degrees and the modulated signal is passed through a frequency multiplier, the
shift becomes 360 degrees. Now this is the same as zero degrees, which means there is a
constant phase signal. This is, of course, at twice the carrier frequency. By passing it
through a frequency divider, a constant signal at carrier frequency is produced that can be
used to lock a local oscillator. There is a similar phase ambiguity to the Costas loop as,
when a frequency is divided by two, it may be inverted. This results in inverted data that
has to be corrected as described.
Both carrier recovery methods are used in this Practical.
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
Block Diagram
Make Connections Diagram
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
Practical 3: Demodulation of 90 Degree Phase Shift Keying by using a
Costas Loop and by using Frequency Multipliers
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Open the oscilloscope. Set the Function Generator to Fast and select a square wave
output.
Set the Loop Filter Compensation to Slow.
Open the voltmeter and the frequency counter. Use the Frequency control of the Function
Generator to set the modulation frequency to 15kHz. Use the Signal Level Control to set
the modulation amplitude to 0.5 volts ac pp.
Open the phasescope and adjust the Variable Phase Shift control associated with the
Carrier Source to set the I and Q carrier phase difference to 90 degrees.
Move the phasescope probes (blue and yellow) to the local oscillator I and Q signals
(monitor points 5 and 6) and use the Variable Phase Shift control associated with the
Local Oscillator to set the phase difference to 90 degrees.
Return the phasescope reference channel (yellow) to the Q carrier (monitor point 2) and
place the main channel probe (blue) on the PSK generator output (monitor point 4).
You should be able to see PSK on the phasescope. Use the Signal Level Control to adjust
the amplitude of the modulation so the phase shift is 90 degrees (180 degrees total).
Remember, it is referenced to the +45 degree carrier.
Using the dc Source control, adjust the dc offset into the loop filter. You should be able to
lock the loop so that on the oscilloscope the output on the Channel 1 trace is the same
signal as the input on the Channel 2 trace.
Unlock the loop with the dc Source control and relock it again. Note that the polarity of the
output with respect to the input is indeed random. Use the Signal Level Control to reduce
the modulation amplitude so that the phase index reduces. Observe that, at near to 45
degrees, the Costas loop fails to work.
Return the modulation index to 90 degrees and relock the PLL.
Move the oscilloscope Channel 2 probe (green) to the Q carrier (monitor point 2) and the
Channel 1 probe (orange) to the PSK output (monitor point 4). Increase the timebase and
use the X expand so you can see some cycles of the carrier. You should be able to see
the phase of the PSK signal reversing.
Chapter 12
Modulation and Coding Principles Phase Shift Keying
53230
Move the oscilloscope Channel 1 probe (orange) to the output of the Frequency Multiplier
(monitor point 7). Note that, by frequency multiplying the signal by 2, the phase reversals
have been removed.
Return the oscilloscope probes to the modulation input (monitor point 3) and the
demodulated output (monitor point 8). Adjust the timebase so you can see the Costas
loop working.
Refer to the Make Connections diagram and remove connection 23 which will unlock the
carrier recovery loop. Add connection 30. This now synchronises the local oscillator using
the multiplied signal. Adjust the dc Source control, which now simply adjusts the local
oscillator centre frequency close to the carrier. This allows the signal from the divider to
lock the oscillator. If you remove connection 25 and replace it you should be able to see
the demodulated output change polarity randomly.
Try reducing the phase modulation index to near 45 degrees. Note that this system, like
the Costas loop, only works with 90 degree phase shift.
Chapter 13
Modulation and Coding Principles MultiState Phase Shift Keying
53230
MultiState Phase Shift Keying
Objectives
To appreciate the concepts of multiphase shift keying and the relationship between bit
and symbol rates for this method
To generate a 4phase (QPSK) and an 8phase (8PSK) phase shift keyed signal and
investigate their associated spectra and bandwidths
To demonstrate how noise affects these keying methods
To investigate the generation of BPSK and QPSK using only an IQ modulator
To investigate the demodulation of QPSK using a double Costas loop
To investigate carrier recovery using the frequency multiplication method
Chapter 13
Modulation and Coding Principles MultiState Phase Shift
Keying
53230
The IQ Modulator
The IQ modulator is a most useful building block in communications systems. It is
available as an integrated circuit with different models operating over a wide range of
frequencies.
It comprises two balanced modulators with their carrier inputs fed from the same source
but one shifted by 90 degrees. The two modulation inputs are available for the user. The
outputs of the two modulators are then summed.
The name IQ modulator comes from In phase and Quadrature. The term quadrature
simply means at 90 degrees.
The diagram shows the basic IQ modulator.
Provided the phase shift is 90 degrees at the carrier frequency then, in vector terms, the
output with respect to the input is shown below:
This means that the output is a signal at the carrier frequency and its phase will depend
on the values of the I and Q modulation inputs. Notice that the amplitude will also vary,

+
+I
I
output
Chapter 13
Modulation and Coding Principles MultiState Phase Shift Keying
53230
because the output is the result of summing two equal values at right angles. It is
therefore 1.414 times the value of when only one input is present.
In mathematical terms the output is the a+jb (complex) sum of the I and Q modulation
inputs. The two modulation inputs can be two quite separate signals. This is how QAM is
generated.
If the output is required to be a phase vector with constant amplitude, the angle of which is
determined by a single input, that input signal has to be processed to generate suitable I
an Q signals.
Since an output is required that is a vector represented by r, where r is the required
constant radius and is the variable angle, and what we have is a+jb, this is done using
the equivalent of changing the mapping in the normal way.
i.e. for an input representing an angle of
sin
mod
M I =
cos
mod
M Q =
Where M is the magnitude of the required signal to drive the modulators.
By generating both the I and Q modulation inputs by processing a single input
representing angle, the output vector is of constant length and is driven round a circle
rather than a square.
Q
+Q
+I
I
output
Chapter 13
Modulation and Coding Principles MultiState Phase Shift
Keying
53230
This is, of course, a phase modulator.
In practical terms, the accuracy of the processing of all these signals depends on the
accuracy with which the 90 degree phase shifts can be maintained.
Chapter 13
Modulation and Coding Principles MultiState Phase Shift Keying
53230
Symbol Rate and Bit Rate
The concepts of symbols, bits, symbol rate and bit rate are important terms in digital
communications.
The concept of a bit (a binary digit) should be familiar as a one or zero in a binary data
stream. The bit rate is simply the rate at which the bits change. For example, imagine a
system that digitized an audio signal at 32k samples per second, each sample being
digitized at 256 possible levels. This means each sample is an 8 bit word. In order to send
this stream over a simple link it would have to be turned into serial data. This means the
serial data stream would run at 32k x 8 = 256k bits per second. This is the bit rate. In this
example we are assuming that there is no extra data for synchronization or for error
correction.
These bits are then modulated onto the carrier in some form. In order to be modulated
they have to be converted to change some parameter of the carrier: its amplitude,
frequency or phase. In a simple system there would be only two states: off or on, one
frequency or the other, one of two phases etc. These states are called symbols.
In the simplest binary system there are only two symbols and each bit has two possible
states so the bits are directly mapped to symbols. This means that the symbol rate is
equal to the bit rate.
There is no reason why there have to be only two possible carrier states. In an amplitude
shift keying (ASK) system there could be more than two possible amplitude states, or in
phase shift keying (PSK) system there could be other possible phases than zero and 180
degrees. If there you had a PSK system with four possible states then each transmitted
data symbol can be decoded as being one of four states. Therefore, not one but two bits
can be carried per symbol. Now, if the bit rate remains the same, we only need to transmit
symbols at half the rate. In such a system the symbol rate is half the bit rate. If there were
16 symbols available then 4 bits per symbol could be carried and the symbol rate would
be one quarter the bit rate. Such systems are called Mary , where M is the number of
possible symbols, sometimes referred to as the order of the modulation scheme.
In such a system the bit rate (B) is:
M S B
2
log =
where S is the symbol rate and M the number of possible symbols.
To avoid confusion this bit rate is sometimes called the gross bit rate
It is important to remember that it is the symbol rate that is the rate at which the carrier
changes state. Therefore, it determines the occupied bandwidth.
It is clear that for a given bandwidth, the higher the order of the modulation scheme the
less bandwidth is used. However there is a penalty to be paid. When demodulated, the
higher the order of the scheme the more likely there are to be errors. This is obvious
Chapter 13
Modulation and Coding Principles MultiState Phase Shift
Keying
53230
because, for example, it is clearly easier to detect the difference between 0 and 180
degrees than zero, 90, 180, and 270.
There is another compromise to be made if error correcting data is added in that, although
adding extra data reduces the number of errors, the bit rate has to rise, with a
consequential increase in occupied bandwidth and received noise.
In order to calculate the amount of useful data that can be transmitted through a digital
system, first find the symbol rate. Then calculate the bit rate by using the number of bits
per symbol. The useful data, sometimes referred to as the payload, can then be
calculated by subtracting the extra data added for error correction, data identification and
synchronisation.
In a multiplexed system more than one data stream may be present and you may have to
find out what proportion of the data stream is allocated to a particular set of data. In very
complex systems this proportion may not even be constant!
Chapter 13
Modulation and Coding Principles MultiState Phase Shift Keying
53230
The Costas Loop
The Costas Loop provides a method of demodulating PSK signals when the phase shift is
90 degrees, which results in there being no carrier component in the modulated signal.
Another way of looking at this is examining the phase detector characteristic. The diagram
shows the output signals from an IQ demodulator with respect to the phase difference
between the local oscillator and the incoming signal.
In signals with phase shifts less than 90 degrees, the I output slope remains the same
polarity. When the shift is 90 degrees exactly, the two symbol positions are at the peaks
of the I output. This means that, if the mean phase moves the I output polarity reverses
and therefore no longer provides a steering control voltage to bring the voltage controlled
local oscillator back. The situation is worse for QPSK as the four symbols are positioned
where no coherent control voltage is produced.
In the simple Costas loop the I and Q signals are multiplied together as shown below.
180
90 0 90 180
Phase
I output
Q output
(I) X (Q)
output
180 90 0 90 180
Chapter 13
Modulation and Coding Principles MultiState Phase Shift
Keying
53230
Note, that now the output polarity is the same at 90 and +90, so a coherent control
voltage is produced as the mean phase moves either side of zero. Note also, that a signal
of opposite polarity is available, by moving the mean phase by 90 degrees. This results in
the loop being insensitive to control voltage inversion. If the mean phase moves by 180
degrees then the loop will lock equally well but the output data will be inverted. This phase
uncertainty has to be dealt with by other means.
The diagram shows the simple Costas loop block diagram. The Costas loop can be
thought of as a phase locked loop with a special phase detector. The loop performs both
the function of carrier lock as well as demodulation, since the I output will contain the data.
There will be a residual amount of twicecarrier frequency component present, which can
be removed by a low pass filter.
Also, there has to be a low pass filter in the control signal to the VCO. This removes any
datarate frequency components and means that the VCO follows the mean phase. This
filter has also to provide the control stability function to prevent control loop oscillation.
This is achieved by ensuring that the control loop gain has dropped below unity before
total phase shift reaches 180 degrees. This problem is compounded by the fact that we
are using VCO frequency to control VCO phase, and therefore 90 degrees of phase is
already present in the control loop.
Inspection of the simple Costas loop phase characteristic reveals that the simple loop
would not work with QPSK as the symbols would be placed on alternate polarity slopes.
X
X
I osc
Q osc
X VCO
control
I
Q
signal
data
Chapter 13
Modulation and Coding Principles MultiState Phase Shift Keying
53230
The double Costas loop provides the answer, but at the expense of additional complexity.
The VCO control phase characteristic of this arrangement is shown below.
This provides a coherent phase control voltage for all four symbol positions.
Note that, in the double Costas loop, there are two limiter amplifiers and an inversion. The
function of the limiters is to make one input of the second set of multipliers switch between
positive and negative voltages, causing the multipliers to either invert or not invert the
signal at the other input. It is interesting to note that without the inversion the output
characteristic is a square wave.
X
I osc
X
VCO
control
I
Q
signal
X
Low
Pass
Filter
Low
Pass
Filter
X
Q osc
Limiter
Limiter
+
I data
Q data
output
180 90 0 90 180
Chapter 13
Modulation and Coding Principles MultiState Phase Shift
Keying
53230
Of course, the loop may lock in any one of four phase positions relative to the original
carrier. This means that the original I data may appear at the Q data output and vice
versa. It may also mean that the outputs are inverted. This has to be resolved by other
means.
Here is the mathematics describing the simple Costas loop:
If the VCO is locked to the incoming carrier then:
vco
=
c
only a small phase difference,
e
will be present.
Let the two outputs from the VCO be:
2cos
c
t in phase with the carrier
2sin
c
t in quadrature
The PSK signal input is:
s(t) = A cos [
c
t + ]
where is 0 or depending on whether the state of the digital input d is 1 or 1.
So, if d(t) is the state of the digital input, this signal expression can be written:
s(t) = A d(t) cos
c
t
The multiplier outputs are the products of the two inputs to each. Thus these outputs are:
[A d(t) cos
c
t][2cos
c
t] and
[A d(t) cos
c
t][2sin
c
t]
The reference channel output is used, i.e.:
v
out
= [A d(t) cos
c
t][2cos
c
t]
= 2A d(t) cos
2
c
t
Now, cos
2
x = 0.5[1 + cos 2x], so the expression for v
out
becomes:
v
out
= 2A d(t) [0.5 + 0.5cos
Chapter 13
Modulation and Coding Principles MultiState Phase Shift Keying
53230
2
c
t]
= A d(t) + A d(t) cos 2
c
t
This expression has two components: a dc component dependant on the phase of the
digital input data and a component at twice the carrier frequency. This doublefrequency
component can be removed by a post detection filter.
When the loop is in lock, the VCO will be phaselocked by modulators (2) and (3), causing
it to produce an output from its f
90
terminal that leads the incoming signal by 90 degrees.
Since the VCO produces outputs which differ by 90 degrees, the reference signal from the
f
0
output will be in phase with the incoming PSK signal for, say, binary 1 and 180 degrees
out of phase for binary 0.
The multiplying action of modulator (1) will then produce a positive dc level when the
received and reference signals are in phase and a negative level when they are in
antiphase. Subsequent data recovery circuits convert the bipolar output from the Costas
Loop demodulator into data.
Chapter 13
Modulation and Coding Principles MultiState Phase Shift
Keying
53230
Symbol Space and Noise
Symbols represent data in terms of a particular state of a carrier characteristic. This may
be amplitude, frequency, phase or a combination of more than one.
Each symbol is one state. For example, phase = 0 or phase =+90; amplitude = 1 or 0.25;
etc.
How many symbols there are depends on the modulation scheme. If there is only one
symbol then the carrier carries no data. Two symbols means that the carrier can only be in
one of two states, i.e. it is a binary system such as BPSK or FSK. As the number of
available symbols increases, the amount of data per symbol increases. See the Concept
section on Symbol Rate for more on this.
An interesting way to think about this is that symbols occupy symbol space. Symbol
space is the total range of values that could exist. For phase, this would be 180 to +180
degrees. For amplitude, it would be zero to maximum amplitude. For frequency, it would
be the total allowable frequency shift. The more symbols there are, the closer together
they are in symbol space and thus the smaller the distance between them. This becomes
important if noise is present and we wish to determine which symbol was sent.
In any system, each symbol has an ideal position. For example, in a QPSK system the
ideal positions would be 0, 90, +90 and +180 degrees. Now, due to inherent inaccuracies
in the generator, they might be slightly offset from those positions and, in the presence of
noise, the symbol would vary in position round its ideal position. An additional problem is
that, for many systems, a carrier reference has to be recreated by the demodulator and
this process may not be perfect. It might be that there is a constant symbolspace error or,
worse, there may be jitter.
Now these imperfections do not cause a problem until the deviation from ideal is more
than half the distance to the next symbol space. It would then be misinterpreted as a
different symbol. This causes errors in the data. The important thing to note is that the
more symbols there are, the less space there is between them and the smaller the amount
of displacement that can be tolerated without error. This can be made worse by the fact
that, in higher order systems, each symbol is carrying more than one bit of data, which
makes a single symbol error worse.
The diagram below shows a phase system in which the ideal positions of the symbols are
marked, together with circles representing the possible symbol displacement due to noise.
It is easy to see that for a two phase system (blue) demodulation would be easy, with a
four phase system (blue +red) the demodulator would have to be better and for an eight
phase system (including green) correct demodulation would be impossible.
angle
Chapter 13
Modulation and Coding Principles MultiState Phase Shift Keying
53230
There are many ways to reduce the probability of an error, such as error correction, or the
use of very sophisticated demodulators.
The concept of symbol space is important in understanding the problems.
Chapter 13
Modulation and Coding Principles MultiState Phase Shift
Keying
53230
Practical 1: Generation and Characteristics of 4psk (QPSK) and 8psk
Objectives and Background
In Binary Phase Shift Keying (BPSK), a binary bit stream can be mapped directly to two
phase symbols. In this Practical you will see that a bit stream can be mapped to more than
two symbols. If four symbols are used the carrier can take one of four phases. This is
called Quadrature Phase Shift Keying (QPSK).
This means that the symbols are generated from one bitstream then the symbol rate is
halved. See the Concept section on Symbol Rate for more on this.
If eight phases are used then the symbol rate is reduced further for the same bit rate. This
is called 8 PSK. There are many ways to map the bits to symbols. There is no best way,
although it is often important to make sure that each symbol is used equally often; this
helps a demodulator extract a phase reference.
The method used to generate the signals in this Practical utilises an IQ modulator, an N
level data source and a sine/cosine converter. To generate BPSK or QPSK only, the IQ
modulator can be driven by binary signals, as you will see in Practical 2. To generate 8
PSK, which has intermediate 45 degree angles, the sine/cosine circuit is needed. In a real
system these would be generated by the symbol mapper, but the analogue processing
block is used here to show more clearly how it works and to show its relationship to
analogue phase modulation.
The diagram shows the phases present in a PSK signal. The blue dots show a BPSK
signal, the red dots show the extra phases in QPSK and the green dots show the further
phases in an 8PSK signal.
Note that the radius of the circle represents the amplitude, and is constant.
0
90
90
180
angle
Chapter 13
Modulation and Coding Principles MultiState Phase Shift Keying
53230
You will also see, by using the phasescope, how the addition of noise moves the phase
positions from their ideal values and how this could cause symbol overlap, depending on
the number of symbols. Look at the Concept section on Symbols and Noise for more on
this topic.
Chapter 13
Modulation and Coding Principles MultiState Phase Shift
Keying
53230
Block Diagram
Make Connections Diagram
Chapter 13
Modulation and Coding Principles MultiState Phase Shift Keying
53230
Chapter 13
Modulation and Coding Principles MultiState Phase Shift
Keying
53230
Practical 1: Generation and Characteristics of 4psk (QPSK) and 8psk
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the Signal Level Control to minimum, for zero modulation input.
Set the Amplitude control of the Noise Generator and the Variable Phase Shift control
in the Transmission Channel block to minimum.
Set the IQ Modulator controls to half scale.
Open the phasescope and use it, together with the Variable Phase Shift control
associated with the Carrier Source, to set the IQ carrier phase shift difference to 90
degrees.
Move the phasescope signal input probe (blue) to the output of the phase modulator
(monitor point 6). Set the phasescope to Constellation and switch the Persistence on.
Turn up the modulation, using the Signal Level Control, until the symbol phase shift is
180 degrees. Use the Phi Offset control on the phasescope to align the two symbols with
the 0 and 180 degree axis. This constellation shows BPSK.
Use the buttons at the bottom of the block diagram to switch the data to QPSK and note
the additional symbols at about 90 and +90 degrees (it may be necessary to finetune
the IQ Modulator controls to achieve an optimum result).
Now change to 8psk and measure the phase shift between the new symbols. Note that
the radii for the different symbols is approximately the same.
Open the oscilloscope and look at the data feeding the modulators (monitor points 4 and
5). Check the number of levels for each modulation format. Remember that one trace
represents change in quadrature, the other change inphase. Try and relate the I and Q
modulating waveforms with the constellation (ignore any transient spikes associated with
the waveforms). Close the oscilloscope.
Return to BPSK and turn up the noise on the transmission channel using the noise
Amplitude control. Observe the effect of different levels of noise on the constellation.
Change the modulation format and note how the area on the constellation diagram
occupied by each symbol merges at a different level for each format. Increase the size of
the phasescope to see this more clearly.
Chapter 13
Modulation and Coding Principles MultiState Phase Shift Keying
53230
Practical 2: Generating QPSK using only an IQ Modulator
Objectives and Background
In Practical 1, you saw how BPSK, QPSK and 8PSK can be generated using an IQ
modulator and a sine/cosine converter. As the expression for generating a constant
amplitude signal with phase from the I and Q modulator inputs is:
sin
mod
A I =
cos
mod
A Q =
For angles of 0, 90, 90, and 180 degrees, the values of sine and cos are either 0, 1, 0 or
1. This means that, for BPSK, you only need the I channel of the IQ modulator with levels
1 and 1, with the Q channel set to zero.
To generate QPSK, the I channel would again contain 1 and 1, while the Q channel also
contains 1 and 1.
This means that you do not need the sine/cosine processor block.
In the Practical you will see this simple configuration in action.
Chapter 13
Modulation and Coding Principles MultiState Phase Shift
Keying
53230
Block Diagram
Make Connections Diagram
Chapter 13
Modulation and Coding Principles MultiState Phase Shift Keying
53230
Chapter 13
Modulation and Coding Principles MultiState Phase Shift
Keying
53230
Practical 2: Generating QPSK using only an IQ Modulator
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set both Signal Level Controls to minimum, for zero modulation on both I and Q
channels.
Set the IQ Modulator controls to half scale.
Open the phasescope and use it, together with the IQ Carrier Source Variable Phase
Shift control to set the IQ carrier phase difference to 90 degrees.
Move the phasescope signal input probe (blue) to the modulator output (monitor point 5).
Switch to Constellation display and switch on the trace Persistence.
Turn up the I channel modulation level (using the upper Signal Level Control) and see
the twolevel signal containing symbols with phase shift of 180 degrees. Use the Phi
Offset control on the phasescope to rotate the constellation to line up with the I axis on
the display (vertical axis).
Open the oscilloscope and see if either, or both, the I or Q channel has data on it.
Turn up the Q channel modulation level (using the lower Signal Level Control) and see
that the constellation now has four points. When the amplitudes in the two channels are
the same, the points will be equidistant, forming a square pattern on the phasescope.
Chapter 13
Modulation and Coding Principles MultiState Phase Shift Keying
53230
Practical 3: Demodulation of QPSK using a Double Costas Loop
Objectives and Background
In this practical you will demodulate QPSK by using a double Costas loop to generate a
carrier reference. The operation of the Costas loop is explained in the Concepts section.
Sample data is generated by the microprocessor and an IQ modulator is used to generate
QPSK. The two outputs from the demodulator are monitored and compared with the input
data for the four different phase lock phase offsets.
Chapter 13
Modulation and Coding Principles MultiState Phase Shift
Keying
53230
Block Diagram
Make Connections Diagram
Chapter 13
Modulation and Coding Principles MultiState Phase Shift Keying
53230
Chapter 13
Modulation and Coding Principles MultiState Phase Shift
Keying
53230
Practical 3: Demodulation of QPSK using a Double Costas Loop
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the I and Q modulation Signal Level Controls to half scale.
Set the IQ Modulator and IQ Demodulator controls to half scale.
Set the two pairs of signal Multiplier controls to half scale.
Set the dc Source control to half scale.
Set the Compensation switch associated with the Loop Filter to Slow.
Open the phasescope and use the Variable Phase Shift control associated with the
Carrier Source to set the IQ carrier phase difference to 90 degrees.
Move the phasescope signal probe (blue) and the reference probe (yellow) to the outputs
of the Local Oscillator phase shifting circuit (monitor points 4 and 5, respectively) and
use the associated Variable Phase Shift control to set the IQ local oscillator phase
difference to 90 degrees.
Move the phasescope reference probe (yellow) back to the +45 degree carrier signal
(monitor point 2) and the signal probe (blue) to the QPSK output (monitor point 3). Switch
to Constellation mode and Persistence on. Use the Signal Level Controls to obtain a
QPSK constellation comprising four symbols at the corners of a square.
Note: You can use the Phi Offset control on the phasescope to rotate the display and the
Signal Level Controls to bring the constellation points to the corners of the display square,
which are at a radius of 0.707.
Now move the phasescope signal input (blue) to the +45 VCO signal (monitor point 5).
Switch off Constellation mode and switch off Persistence.
Use the dc Source control to lock the loop. Lock is indicated by the phasescope showing a
constant phase between the original carrier and the demodulator reference supplied by
the VCO. There will be a few degrees of phase noise on the VCO.
Unlock the loop by temporarily disconnecting the input to the Local Oscillator (connection
23 on the Make Connections diagram) and reconnect to lock it again. Do this several
times and you will see that the VCO will lock arbitrarily in one of four positions. You may
need to finetune the dc Source control to lock the loop. The position at which the VCO
locks is random.
Make sure the loop is locked and open the oscilloscope. You should see two data streams
on the two channels. Move the oscilloscope Channel 1 probe (orange) to sample I data
Chapter 13
Modulation and Coding Principles MultiState Phase Shift Keying
53230
from the generator (monitor point 6). You may be able to see that the same data is being
demodulated (green and orange traces the same, but may be inverted and/or reversed).
If you unlock and relock the loop and watch the phasescope, you should be able to see
how I and Q data become reversed and inverted depending on the phase of the VCO
reference. Try this several times, until you understand what is happening.
Chapter 13
Modulation and Coding Principles MultiState Phase Shift
Keying
53230
Practical 4: Carrier Recovery Using Frequency Multiplication
Objectives and Background
In this Practical an alternative to the double Costas loop carrier recovery system will be
used. This method is simpler but, in general, has inferior performance.
The method simply takes the QPSK signal and multiplies the frequency by four. This
results in a fourtimes frequency component, but the phase of this component is the same
for each symbol. This frequency signal is then used to lock the carrier oscillator. This
method used to lock the oscillator is called injection locking and simply means that a
small amount of an external signal, near to the oscillator free running frequency, is applied
to the oscillator, forcing it to lock to the external signal. In this instance the injection
frequency is four times the oscillator frequency. In practice, injection locking is not
considered reliable enough and more complex means are used to frequency divide and
lock the local oscillator.
Of note is the fact that the oscillator may lock onto any one of the four injection frequency
cycles per oscillator cycle. Therefore, the phase of the recovered carrier can be in any one
of four phases. This is exactly the same problem as the double Costas loop.
This method is well suited to higher order PSK such as 8PSK. This is achieved by
multiplying and dividing the signal by eight. As the PSK order increases, other methods
become more and more complicated.
Chapter 13
Modulation and Coding Principles MultiState Phase Shift Keying
53230
Block Diagram
Make Connections Diagram
Chapter 13
Modulation and Coding Principles MultiState Phase Shift
Keying
53230
Chapter 13
Modulation and Coding Principles MultiState Phase Shift Keying
53230
Practical 4: Carrier Recovery Using Frequency Multiplication
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the I and Q data Signal Level Controls to half scale.
Set the IQ Modulator and IQ Demodulator controls to half scale.
Set the dc Source control, which provides the VCO control voltage, to half scale.
Open the phasescope and use the Variable Phase Shift control associated with the
Carrier Source to set the IQ carrier phase difference to 90 degrees.
Move the phasescope signal probe (blue) and the reference probe (yellow) to the outputs
of the Local Oscillator phase shifting circuit (monitor points 4 and 5, respectively) and
use the associated Variable Phase Shift control to set the IQ local oscillator phase
difference to 90 degrees.
Move the phasescope reference probe (yellow) back to the +45 carrier (monitor point 2)
and the signal probe (blue) to the QPSK generator output (monitor point 3). Switch to
Constellation mode and switch Persistence on. Using the two Signal Level Controls,
adjust the data amplitudes to give a square QPSK constellation.
Note: You can use the Phi Offset control on the phasescope to rotate the display and the
Signal Level Controls to bring the constellation points to the corners of the display square,
which are at a radius of 0.707.
Move the signal probe (blue) to the +45 local oscillator signal (monitor point 5). Switch off
Persistence and Constellation modes.
Use the dc Source control to carefully adjust the VCO control voltage. You should be able
to make the oscillator lock to the carrier, although this can be quite difficult. Try unlocking
and then relocking the VCO. Note that it will lock in any one of four phases.
Open the oscilloscope. You should be able to see the recovered data on the two
channels. Move the Channel 2 probe (green) to sample the data (monitor point 6) and
confirm that the data is recovered. In the same way as the Costas loop, the various lock
phases of the local oscillator result in reversal of I and Q data and/or polarity inversion.
Chapter 14
Modulation and Coding Principles Quadrature Amplitude Modulation (QAM)
53230
Quadrature Amplitude Modulation (QAM)
Objectives
To appreciate the concept of Quadrature Amplitude Modulation (QAM) and how it can be
represented by a constellation diagram
To understand the differences between 16, 64 and 256 QAM
To generate 16, 64 and 256 QAM and examine their constellations
To examine the effect of amplitude and phase noise on the different QAM forms
To understand how an IQ demodulator can be used to demodulate QAM and the problem
of carrier phase reference
Chapter 14
Modulation and Coding Principles Quadrature Amplitude
Modulation (QAM)
53230
The IQ Modulator
The IQ modulator is a most useful building block in communications systems. It is
available as an integrated circuit with different models operating over a wide range of
frequencies.
It comprises two balanced modulators with their carrier inputs fed from the same source
but one shifted by 90 degrees. The two modulation inputs are available for the user. The
outputs of the two modulators are then summed.
The name IQ modulator comes from In phase and Quadrature. The term quadrature
simply means at 90 degrees.
The diagram shows the basic IQ modulator.
Provided the phase shift is 90 degrees at the carrier frequency then, in vector terms, the
output with respect to the input is shown below:
This means that the output is a signal at the carrier frequency and its phase will depend
on the values of the I and Q modulation inputs. Notice that the amplitude will also vary,

+
+I
I
output
Chapter 14
Modulation and Coding Principles Quadrature Amplitude Modulation (QAM)
53230
because the output is the result of summing two equal values at right angles. It is
therefore 1.414 times the value of when only one input is present.
In mathematical terms the output is the a+jb (complex) sum of the I and Q modulation
inputs. The two modulation inputs can be two quite separate signals. This is how QAM is
generated.
If the output is required to be a phase vector with constant amplitude, the angle of which is
determined by a single input, that input signal has to be processed to generate suitable I
an Q signals.
Since an output is required that is a vector represented by r, where r is the required
constant radius and is the variable angle, and what we have is a+jb, this is done using
the equivalent of changing the mapping in the normal way.
i.e. for an input representing an angle of
sin
mod
M I =
cos
mod
M Q =
Where M is the magnitude of the required signal to drive the modulators.
By generating both the I and Q modulation inputs by processing a single input
representing angle, the output vector is of constant length and is driven round a circle
rather than a square.
Q
+Q
+I
I
output
Chapter 14
Modulation and Coding Principles Quadrature Amplitude
Modulation (QAM)
53230
This is, of course, a phase modulator.
In practical terms, the accuracy of the processing of all these signals depends on the
accuracy with which the 90 degree phase shifts can be maintained.
Chapter 14
Modulation and Coding Principles Quadrature Amplitude Modulation (QAM)
53230
Symbol Rate and Bit Rate
The concepts of symbols, bits, symbol rate and bit rate are important terms in digital
communications.
The concept of a bit (a binary digit) should be familiar as a one or zero in a binary data
stream. The bit rate is simply the rate at which the bits change. For example, imagine a
system that digitized an audio signal at 32k samples per second, each sample being
digitized at 256 possible levels. This means each sample is an 8 bit word. In order to send
this stream over a simple link it would have to be turned into serial data. This means the
serial data stream would run at 32k x 8 = 256k bits per second. This is the bit rate. In this
example we are assuming that there is no extra data for synchronization or for error
correction.
These bits are then modulated onto the carrier in some form. In order to be modulated
they have to be converted to change some parameter of the carrier: its amplitude,
frequency or phase. In a simple system there would be only two states: off or on, one
frequency or the other, one of two phases etc. These states are called symbols.
In the simplest binary system there are only two symbols and each bit has two possible
states so the bits are directly mapped to symbols. This means that the symbol rate is
equal to the bit rate.
There is no reason why there have to be only two possible carrier states. In an amplitude
shift keying (ASK) system there could be more than two possible amplitude states, or in
phase shift keying (PSK) system there could be other possible phases than zero and 180
degrees. If there you had a PSK system with four possible states then each transmitted
data symbol can be decoded as being one of four states. Therefore, not one but two bits
can be carried per symbol. Now, if the bit rate remains the same, we only need to transmit
symbols at half the rate. In such a system the symbol rate is half the bit rate. If there were
16 symbols available then 4 bits per symbol could be carried and the symbol rate would
be one quarter the bit rate. Such systems are called Mary , where M is the number of
possible symbols, sometimes referred to as the order of the modulation scheme.
In such a system the bit rate (B) is:
M S B
2
log =
where S is the symbol rate and M the number of possible symbols.
To avoid confusion this bit rate is sometimes called the gross bit rate
It is important to remember that it is the symbol rate that is the rate at which the carrier
changes state. Therefore, it determines the occupied bandwidth.
It is clear that for a given bandwidth, the higher the order of the modulation scheme the
less bandwidth is used. However there is a penalty to be paid. When demodulated, the
higher the order of the scheme the more likely there are to be errors. This is obvious
Chapter 14
Modulation and Coding Principles Quadrature Amplitude
Modulation (QAM)
53230
because, for example, it is clearly easier to detect the difference between 0 and 180
degrees than zero, 90, 180, and 270.
There is another compromise to be made if error correcting data is added in that, although
adding extra data reduces the number of errors, the bit rate has to rise, with a
consequential increase in occupied bandwidth and received noise.
In order to calculate the amount of useful data that can be transmitted through a digital
system, first find the symbol rate. Then calculate the bit rate by using the number of bits
per symbol. The useful data, sometimes referred to as the payload, can then be
calculated by subtracting the extra data added for error correction, data identification and
synchronisation.
In a multiplexed system more than one data stream may be present and you may have to
find out what proportion of the data stream is allocated to a particular set of data. In very
complex systems this proportion may not even be constant!
Chapter 14
Modulation and Coding Principles Quadrature Amplitude Modulation (QAM)
53230
The Costas Loop
The Costas Loop provides a method of demodulating PSK signals when the phase shift is
90 degrees, which results in there being no carrier component in the modulated signal.
Another way of looking at this is examining the phase detector characteristic. The diagram
shows the output signals from an IQ demodulator with respect to the phase difference
between the local oscillator and the incoming signal.
In signals with phase shifts less than 90 degrees, the I output slope remains the same
polarity. When the shift is 90 degrees exactly, the two symbol positions are at the peaks
of the I output. This means that, if the mean phase moves the I output polarity reverses
and therefore no longer provides a steering control voltage to bring the voltage controlled
local oscillator back. The situation is worse for QPSK as the four symbols are positioned
where no coherent control voltage is produced.
In the simple Costas loop the I and Q signals are multiplied together as shown below.
180
90 0 90 180
Phase
I output
Q output
(I) X (Q)
output
180 90 0 90 180
Chapter 14
Modulation and Coding Principles Quadrature Amplitude
Modulation (QAM)
53230
Note, that now the output polarity is the same at 90 and +90, so a coherent control
voltage is produced as the mean phase moves either side of zero. Note also, that a signal
of opposite polarity is available, by moving the mean phase by 90 degrees. This results in
the loop being insensitive to control voltage inversion. If the mean phase moves by 180
degrees then the loop will lock equally well but the output data will be inverted. This phase
uncertainty has to be dealt with by other means.
The diagram shows the simple Costas loop block diagram. The Costas loop can be
thought of as a phase locked loop with a special phase detector. The loop performs both
the function of carrier lock as well as demodulation, since the I output will contain the data.
There will be a residual amount of twicecarrier frequency component present, which can
be removed by a low pass filter.
Also, there has to be a low pass filter in the control signal to the VCO. This removes any
datarate frequency components and means that the VCO follows the mean phase. This
filter has also to provide the control stability function to prevent control loop oscillation.
This is achieved by ensuring that the control loop gain has dropped below unity before
total phase shift reaches 180 degrees. This problem is compounded by the fact that we
are using VCO frequency to control VCO phase, and therefore 90 degrees of phase is
already present in the control loop.
Inspection of the simple Costas loop phase characteristic reveals that the simple loop
would not work with QPSK as the symbols would be placed on alternate polarity slopes.
X
X
I osc
Q osc
X VCO
control
I
Q
signal
data
Chapter 14
Modulation and Coding Principles Quadrature Amplitude Modulation (QAM)
53230
The double Costas loop provides the answer, but at the expense of additional complexity.
The VCO control phase characteristic of this arrangement is shown below.
This provides a coherent phase control voltage for all four symbol positions.
Note that, in the double Costas loop, there are two limiter amplifiers and an inversion. The
function of the limiters is to make one input of the second set of multipliers switch between
positive and negative voltages, causing the multipliers to either invert or not invert the
signal at the other input. It is interesting to note that without the inversion the output
characteristic is a square wave.
X
I osc
X
VCO
control
I
Q
signal
X
Low
Pass
Filter
Low
Pass
Filter
X
Q osc
Limiter
Limiter
+
I data
Q data
output
180 90 0 90 180
Chapter 14
Modulation and Coding Principles Quadrature Amplitude
Modulation (QAM)
53230
Of course, the loop may lock in any one of four phase positions relative to the original
carrier. This means that the original I data may appear at the Q data output and vice
versa. It may also mean that the outputs are inverted. This has to be resolved by other
means.
Here is the mathematics describing the simple Costas loop:
If the VCO is locked to the incoming carrier then:
vco
=
c
only a small phase difference,
e
will be present.
Let the two outputs from the VCO be:
2cos
c
t in phase with the carrier
2sin
c
t in quadrature
The PSK signal input is:
s(t) = A cos [
c
t + ]
where is 0 or depending on whether the state of the digital input d is 1 or 1.
So, if d(t) is the state of the digital input, this signal expression can be written:
s(t) = A d(t) cos
c
t
The multiplier outputs are the products of the two inputs to each. Thus these outputs are:
[A d(t) cos
c
t][2cos
c
t] and
[A d(t) cos
c
t][2sin
c
t]
The reference channel output is used, i.e.:
v
out
= [A d(t) cos
c
t][2cos
c
t]
= 2A d(t) cos
2
c
t
Now, cos
2
x = 0.5[1 + cos 2x], so the expression for v
out
becomes:
v
out
= 2A d(t) [0.5 + 0.5cos
Chapter 14
Modulation and Coding Principles Quadrature Amplitude Modulation (QAM)
53230
2
c
t]
= A d(t) + A d(t) cos 2
c
t
This expression has two components: a dc component dependant on the phase of the
digital input data and a component at twice the carrier frequency. This doublefrequency
component can be removed by a post detection filter.
When the loop is in lock, the VCO will be phaselocked by modulators (2) and (3), causing
it to produce an output from its f
90
terminal that leads the incoming signal by 90 degrees.
Since the VCO produces outputs which differ by 90 degrees, the reference signal from the
f
0
output will be in phase with the incoming PSK signal for, say, binary 1 and 180 degrees
out of phase for binary 0.
The multiplying action of modulator (1) will then produce a positive dc level when the
received and reference signals are in phase and a negative level when they are in
antiphase. Subsequent data recovery circuits convert the bipolar output from the Costas
Loop demodulator into data.
Chapter 14
Modulation and Coding Principles Quadrature Amplitude
Modulation (QAM)
53230
Quadrature Amplitude Modulation (QAM)
QAM, pronounced kwam, is a hybrid of amplitude and phase modulation. It can be
imagined as two amplitude shift carriers with their phases separated by 90 degrees. This
results in a constellation of symbols arranged at different phase angles and amplitudes,
equally spaced and within a square. The total number of symbols is the product of the
number of levels in each of the orthogonal carriers. For example if there were 8 levels in
each carrier then the total number of symbols is 64. This is referred to as 64QAM. The
diagram shows the constellation of 64 QAM, together with that of 8PSK for comparison.
There are several points to note. Firstly, there are many more symbols than could be
accommodated by simply using phase. This implies greater modulation efficiency but,
obviously, at the expense of symbol spacing and hence noise immunity. Secondly, the
amplitude of the signal varies, not only at symbol transitions, as in PSK, but from symbol
to symbol.
QAM is used in 256, 64 and 16 modes. Interestingly, 4QAM is actually QPSK. In simple
QAM, the number of symbols has to be an integer square.
The greater number of available symbols, and its wide range of modes, means QAM is
used extensively in highcapacity, noisefree systems, such as pointtopoint microwave
links. It is also used in some varieties of terrestrial digital television.
QAM is demodulated in a similar way to QPSK, in as much as an IQ modulator recovers
the multilevel data on the two carriers. This relies on successful recovery of a local carrier
and, like any system which is symmetrical about zero, has no residual carrier to lock to.
The problem with QAM is that the number of phases in the signal is equal to the number
of symbols, so techniques such as Costas loops and frequency multiplication are
impractical. There are various methods used, depending in complexity and the
requirement for noise immunity. The simplest method uses an amplitude demodulator to
detect the amplitude peaks representing the corners of the constellation. This is then
used to gate a control signal from a double Costas loop. This is equivalent to dealing with
the signal as if it were a QPSK signal with symbols at the four corners. Although it works,
this only uses a small proportion of the total signal power and hence is not particularly
64
QAM
8PSK
Chapter 14
Modulation and Coding Principles Quadrature Amplitude Modulation (QAM)
53230
effective. Other methods involve creating a local QAM signal and trying to find the best
match for amplitude and phase.
Hierarchical QAM
In standard QAM the symbols are equally spaced. In hierarchical QAM symbols are
clustered. The purpose of this is that, under noisy conditions, some of the bits can be
detected after others have become too noisy. This can be used to produce a more
graceful degradation than the all or nothing characteristic of digital signals. Careful
symbol mapping is used so that, for example, the more significant data bits are carried in
the wide spaced symbol clusters. The diagram shows and example of hierarchical QAM.
Note also that the number of available symbols is reduced and that the less significant bits
have less space between them.
When the signal is noisy it is possible to tell which of the four symbol groups is being sent
but not which symbol within the group. This would mean that the system would operate
like QPSK under noisy conditions, rather than fail completely. Providing that the symbol
mapping was chosen carefully, the lower rate data would provide useful information.
Other Types of QAM
There are a few other variations of QAM. One variation is to remove the corner symbols in
order to reduce the amplitude peaks. This makes amplifying the signal easier and makes it
less likely that the signal will be distorted, which would result in extra sidebands being
produced. Here is an example of a QAM constellation with reduced amplitude peaks.
Hierarchical
QAM
Chapter 14
Modulation and Coding Principles Quadrature Amplitude
Modulation (QAM)
53230
As you can see, one of the advantages of QAM is its versatility
Chapter 14
Modulation and Coding Principles Quadrature Amplitude Modulation (QAM)
53230
Practical 1: Generation and Characteristics of QAM
Objectives and Background
In this Practical you will generate 256QAM, 64QAM and 16QAM and examine the
constellation. You will also add noise to the signal and see how noise fills the intersymbol
space. If you are not familiar with QAM you should read the Concepts section where the
different forms are described in detail.
Chapter 14
Modulation and Coding Principles Quadrature Amplitude
Modulation (QAM)
53230
Block Diagram
Make Connections Diagram
Chapter 14
Modulation and Coding Principles Quadrature Amplitude Modulation (QAM)
53230
Chapter 14
Modulation and Coding Principles Quadrature Amplitude
Modulation (QAM)
53230
Practical 1: Generation and Characteristics of QAM
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the two data Signal Level Controls to twothirds of full scale.
Set the IQ Modulator controls to half scale.
Set the noise Amplitude control associated with the Transmission Channel block to
minimum.
Open the phasescope and use the Variable Phase Shift control associated with the
Carrier Source to set the IQ carrier phase difference to 90 degrees.
Move the phasescope signal probe (blue) to the noise channel output (monitor point 3).
Set to Constellation mode, Persistence to on and HiPersist to on.
The phasescope should show a constellation of 16 points. Use the Phi Offset control on
the phasescope to rotate the pattern to line up with the square on the phasescope
graticule. Adjust the data Signal Level Controls to achieve an equal sided square. You
may also need to adjust the lower balance control on each channel of the IQ Modulator to
centralise the pattern.
Use the button on the block diagram to change to 64QAM and notice that there are now
64 symbols.
Use the button on the block diagram to switch to 256QAM and you should be able to see
the 256 symbols. You will need to increase the size of the phasescope to see this clearly.
Open the oscilloscope and observe the traces as you change back to 16QAM. You
should see that the I and Q data is 16 level for 256QAM, 8 level for 64QAM and 4 level
for 16QAM.
Select XY Mode on the oscilloscope. Adjust the sampling rate on the oscilloscope and
you should be able to see a similar constellation display to that on the phasescope. Why
do you think this is so?
Close the oscilloscope.
Adjust the Amplitude control associated with the Transmission Channel block to increase
the noise amplitude and see how the positions of the symbols are modified by noise. Each
symbol becomes clustered around its ideal position.
Increase the size of the phasescope display and set a noise amplitude that just gives
symbol separation. Now change from 16QAM to 64QAM. You should see that data
recovery would be impossible.
Chapter 14
Modulation and Coding Principles Quadrature Amplitude Modulation (QAM)
53230
Chapter 14
Modulation and Coding Principles Quadrature Amplitude
Modulation (QAM)
53230
Practical 2: The Effect of Amplitude and Phase Noise on QAM
Constellations
Objectives and Background
It is important that you can recognise how an ideal QAM constellation is modified by noise.
You have seen how simply adding noise the signal causes the constellation points to
become circular patches, as each symbol is displaced by a random distance from the
ideal.
You should have also appreciated that the constellation displays each point by
representing its amplitude by the radius (length of the phasor) and its phase by the
position on the circle (angle of the phasor). The fact that adding noise causes the symbols
to become patches tells us that the noise is present in both amplitude and phase.
There are situations when noise is present only in amplitude, or in phase, and not both.
Most likely, these would represent fault conditions, so it is useful to be able to recognise
them.
In this Practical you will first add amplitude and then phase noise to a 16QAM signal.
Amplitude noise is added by multiplying the QAM signal with noise plus a dc offset. This
has the effect of amplitude modulating the signal with noise. You will use the
phasescope to examine the result.
Phase noise is added by simply adding noise to the frequency control input of the carrier
oscillator. At small amplitudes this has the effect of moving the oscillator phase either side
of its natural phase. An interesting problem arises when we want to see the result. The
reference normally used for the phasescope is the carrier so, if both the signal and the
reference have the same phase, you will not see anything. This problem is solved by using
a multiplier, a lowpass filter and VCO to lock to the carrier. The time constant of the low
pass filter is such that it does not follow the phase noise and thus provides a steady
reference for the phasescope.
Chapter 14
Modulation and Coding Principles Quadrature Amplitude Modulation (QAM)
53230
Block Diagram
Make Connections Diagram
Chapter 14
Modulation and Coding Principles Quadrature Amplitude
Modulation (QAM)
53230
Chapter 14
Modulation and Coding Principles Quadrature Amplitude Modulation (QAM)
53230
Practical 2: The Effect of Amplitude and Phase Noise on QAM
Constellations
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the modulation Signal Level Controls to twothirds of full scale.
Set the noise Amplitude control associated with the Transmission Channel block to
minimum.
Set the pair of controls in both Multiplier blocks to half scale.
Set the IQ Modulator controls to half scale.
Set the Loop Filter Compensation switch to Slow.
Open the phasescope and use the Variable Phase Shift control associated with the
Carrier Source to set the IQ carrier phase difference to 90 degrees.
Check that the phaselocked loop is in lock by moving the phasescope signal probe (blue)
to the VCO output (monitor point 4) and then adjusting the dc Source control, if
necessary, until a steady phasescope display is obtained.
On the phasescope, switch on Constellation mode and Persistence. Set Hi Persist to
on.
Move the phasescope signal probe (blue) to the QAM noise multiplier output (monitor
point 3). Adjust the lower Multiplier offset control (the lower control in the lower Multiplier
block) through its full range until you can see a 16QAM constellation with a maximum
radius of about 0.7. Increase the size of the phasescope so you can see it better.
Use the phasescope Phi Offset control to rotate the constellation so it is square with the
graticule. You may also need to adjust the lower balance control on each channel of the
IQ Modulator to centralise the pattern.
Now add some amplitude noise by turning up the Amplitude control associated with the
Transmission Channel block. You should be able to see that noise is being added in only
one plane in the constellation (along the directions of the lengths of the phasors).
Turn the noise back tozero.
Refer to the Make Connections diagram and remove connection 11 and add connection
20.
Move the phasescope reference probe (yellow) to monitor the VCO output (monitor point
4). You will have to readjust the Phi Offset control on the phasescope to bring the display
square, again.
Chapter 14
Modulation and Coding Principles Quadrature Amplitude
Modulation (QAM)
53230
Now, turn the noise back up and you should be able to see the noise being added in the
other plane (in arcs about the same radii).
If you can fully explain these two effects then you have understood how QAM works and
what the constellation represents.
Chapter 14
Modulation and Coding Principles Quadrature Amplitude Modulation (QAM)
53230
Practical 3: Demodulation of QAM
Objectives and Background
In this Practical you will see the principles by which QAM can be demodulated. The basic
building block is the IQ demodulator.
A problem arises when trying to derive a phase reference for the local oscillator. 256QAM
has 256 phases and, therefore, trying to find a constant phase reference and maintaining
it can be rather difficult. Even 16QAM has 16 phases, four times as many as QPSK.
One method of solving this problem is by using a double Costas loop, but gating the
control signal in such a way as to only steer the oscillator at the four corners of the
constellation. This means that QAM is treated like QPSK. The gating is done on signal
amplitude, as the four corners are the points at which the instantaneous amplitude is a
maximum. However, there are real problems with this, as you are actually only using a
small fraction of the signal power to derive the carrier, so it is very unstable.
Another problem is that the corner symbols only occur occasionally, even on randomised
data. Several are needed to get a sensible control signal. If the carrier and local oscillators
are not extremely stable then they will drift apart between control signal updates. Even if
they drift only 90 degrees they will lock to another phase, due to the QPSK phase
uncertainty problem. On the workboard, the oscillators use lumped inductors and
capacitors to form the oscillation circuit. Such a circuit is not good enough to achieve the
stability required. To make such a system work correctly, crystal controlled oscillators
would have to be used.
In any case, this is not a method that is used in most practical applications. The methods
used are quite complex and are normally implemented in DSP. Their scope is outside
these assignments.
So that you can see how the IQ demodulator recovers the data, this workboard uses a
phase lock loop fed from the original carrier. The important points to note are what
happens when the local oscillator is unlocked and what happens when the local oscillator
has a phase offset.
Chapter 14
Modulation and Coding Principles Quadrature Amplitude
Modulation (QAM)
53230
Block Diagram
Make Connections Diagram
Chapter 14
Modulation and Coding Principles Quadrature Amplitude Modulation (QAM)
53230
Chapter 14
Modulation and Coding Principles Quadrature Amplitude
Modulation (QAM)
53230
Practical 3: Demodulation of QAM
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the two modulation data Signal Level Controls to twothirds of full scale.
Set the IQ Modulator and IQ Demodulator controls to half scale.
Set the dc Source control to half scale.
Set the upper Multiplier block controls to half scale.
Set the Loop Filter Compensation switch to Slow.
Open the phasescope and use the Variable Phase Shift control associated with the
Carrier Source to make sure that the carrier IQ phase difference is 90 degrees.
Move the phasescope signal (blue) and reference (yellow) probes to monitor the 45
degree outputs of the VCO (monitor points 4 and 5, respectively). Adjust the Variable
Phase Shift control associated with the Local Oscillator block to give a 90 degree phase
difference.
Move the phasescope reference probe (yellow) back to the carrier +45 degree signal
(monitor point 2) and move the phasescope signal probe (blue) to the QAM modulator
output (monitor point 3).
On the phasescope, switch to Constellation mode and switch Persistence on. Set Hi
Persist to on.
Using the buttons at the bottom of the block diagram, set the modulation to 16QAM and
adjust the Signal Level Controls for an equal square QAM signal with a maximum radius of
about 0.7. Use the Phi Offset control on the phasescope to rotate the constellation so it is
square with the graticule. You may also need to adjust the lower balance control on each
channel of the IQ Modulator to centralise the pattern.
Open the oscilloscope and see that there is data being recovered on both I and Q
channels. Switch the oscilloscope to XY Mode and you should see a constellation. You
may need to reduce the Signal Level Controls in order to see all the points of the
constellation. If you adjust the dc Source control you can move the phase of the local
oscillator reference. You can see how the constellation rotates.
If you adjust the carrier and local oscillator IQ phase differences (using the relevant
Variable Phase Shift controls) you can also see that it changes the square to a trapezoid.
Notice that the best constellation does not occur exactly when the measured oscillator
phase differences are 90 degrees. This is because small phase offsets are added by the
Chapter 14
Modulation and Coding Principles Quadrature Amplitude Modulation (QAM)
53230
nonideal components in the electronics. These effects have to be compensated for in a
real system.
Refer to the Make Connections diagram and remove connection 17. This unlocks the local
oscillator. Adjust the dc Source control and somewhere near its mid position you will be
able to see that the whole constellation is rotating. This is the display you will see when
carrier lock has not been achieved.
Chapter 15
Modulation and Coding Principles Binary Pulse Code Modulation
53230
Binary PCM
Objectives
To understand the term Pulse Code Modulation (PCM) in digital data transmission
To understand the term Bit Error Rate and investigate how it is affected by signaltonoise
ratio
To appreciate that the quality of digital data does not degrade linearly with received signal
tonoise ratio
Chapter 15
Modulation and Coding Principles Binary Pulse Code
Modulation
53230
Pulse Code Modulation (PCM)
Pulse code modulation is a term used to describe a form of coding developed in the early
days of digital data transmission. It was used extensively on telephony systems.
Strictly speaking, it is not a form of modulation but a coding system.
It is generally taken to mean a method of mapping digital words to binary digits. The
binary digits are referred to as pulses in their baseband form. In simple PCM a word is
represented by bits having weights of 1, 2, 4, 8, 16 etc. These bits are then sent serially as
pulses. It is arbitrary whether the most or least significant bit is sent first as long as both
ends of a link use the same convention. A whole word, i.e. a fixed number of pulses, is
sometimes referred to as a frame. It is clearly most important that at the receiver the
correct bit has the correct weight applied to it. In some cases additional pulses are added
to achieve frame synchronisation.
The pulses representing the bits can use any binary modulation scheme for transmission
such as ASK, BPSK, or FSK.
Modern more efficient systems such as QAM have replaced PCM but the term remains
and is sometimes confusingly applied to any form of digital transmission that maps binary
data to weighted bits.
Chapter 15
Modulation and Coding Principles Binary Pulse Code Modulation
53230
Bit Error Rate Measurement (BER)
Bit error rate is a quantitative measure of the quality of a digital transmission link. As all
data is converted to bits, even though these bits may be mapped in different ways to
symbols in the modulation scheme it is possible to determine if any bit is the same state
as that transmitted.
The bit error rate is simply a comparison of the number of bits transmitted to the number
of bits received in error. It makes no assessment of the significance of the bit error, for
example, the weighing of the bit. If error correction is in use, the BER is usually measured
before error correction is applied. The reason for this is that error correction sometimes
takes account of the data content and that BER is used to asses the performance of the
physical link.
There are several ways of expressing BER, one common way is the number of bits that
have to be sent before one error occurs. This is in the form of 1 in 10
6
meaning that 1 bit
in one million will be in error.
In the Practicals a percentage display format is used.
Chapter 15
Modulation and Coding Principles Binary Pulse Code
Modulation
53230
Practical 1: PCM Baseband Signal
Objectives and Background
In this practical you will see an example of a PCM baseband signal as a series of pulses
representing an eight bit binary word, transmitted serially.
The source for the binary signal is either an analogue voltage passed through an A/D
converter or a number increasing from zero to 255 at a rate of about one per second.
The sync pulse used to trigger the oscilloscope is spaced so there is space for three
frames between them. It is possible to select only to have one frame active so you can
see the format more easily.
Chapter 15
Modulation and Coding Principles Binary Pulse Code Modulation
53230
Block Diagram
Make Connections Diagram
Chapter 15
Modulation and Coding Principles Binary Pulse Code
Modulation
53230
Chapter 15
Modulation and Coding Principles Binary Pulse Code Modulation
53230
Practical 1: PCM Baseband Signal
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the Signal Level Control (for the sync pulse attenuator) to minimum.
Set the dc Source control to full scale.
Ensure that the ADC input button at the bottom of the Practical screen has been selected.
This applies a voltage, from the dc Source control, to the ADC input.
Open the oscilloscope and adjust the Signal Level Control to give a stable trace. This
should be with about 0.5 volts peak to peak (yellow trace).
Note the output on the oscilloscope and adjust the dc Source control to change the A/D
input to see the binary code changing (blue trace).
Click on the ADC Group button at the bottom of the Practical screen. You should see the
A/D derived binary code repeated three times between the sync pulses.
Change to Digital Ramp input and see the binary number change through all the codes
from zero to all ones.
Chapter 15
Modulation and Coding Principles Binary Pulse Code
Modulation
53230
Practical 2: PCM BPSK Link
Objectives and Background
In this practical you will use the pcm format to send data over a bpsk link and asses the
effect that the signaltonoise ratio has on Bit error rate. This practical does not use
carrier phase recovery but this is not an unreasonable compromise as a good clock
recovery system would be still able to obtain a reasonably accurate carrier clock when the
signal was so poor that data recovery was almost useless.
You will use the phasescope in constellation mode to see when noise is added how the
symbols are moved from their ideal positions but errors only occur when symbols overlap
into the other symbols space.
The spectrum analyser is used to measure the signaltonoise ratio at the demodulator.
Chapter 15
Modulation and Coding Principles Binary Pulse Code Modulation
53230
Block Diagram
Make Connections Diagram
Chapter 15
Modulation and Coding Principles Binary Pulse Code
Modulation
53230
Chapter 15
Modulation and Coding Principles Binary Pulse Code Modulation
53230
Practical 2: PCM BPSK Link
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the sync pulse attenuator (middle Signal Level Control) for minimum output (fully
counterclockwise).
Set the modulation attenuator (upper Signal Level Control) to half scale.
Set the noise Amplitude in the Transmission Channel block to minimum (fully counter
clockwise).
Ensure that the ADC input button is selected (at the bottom of the Perform Practical
screen).
Open the oscilloscope and adjust the sync pulse attenuator to give a stable trace (middle
Signal Level Control). This should be with about 0.5 volts peak to peak.
Note the Pulse Code Modulation (PCM) data being applied to the multiplier.
Open the phasescope and use the upper Signal Level Control to adjust the two bspk
symbols to have a radius of approximately 0.6 (the height of the inner square on the
phasescope).
Move the oscilloscope channel one probe (blue probe) to the output of the multiplier
(monitor point 4) and adjust both I channel balance controls (the upper two controls in the
IQ Modulator block) for minimum amplitude variation. You will be able to see the abrupt
phase transitions as faint black lines.
Check the signals at monitor points 5 and 6 with the oscilloscope to confirm that the signal
is being transmitted correctly and the data is being recovered.
Open the spectrum analyser and note that the signaltonoise ratio is in excess of 50dB.
Identify the controls to the left of the Micro Controller and A/DD/A block. Set the A/D2
Amplitude control to 2/3 full scale and the A/D2 Offset control to half scale.
Open the frequency counter, which is set to BER (Bit Error Rate) mode. The BER should
read zero. This indicates that all the received symbols correspond with those sent.
Now, increase slowly the noise Amplitude control in the Transmission Channel block and
note that the signaltonoise ratio on the spectrum analyser decreases and the symbol
positions on the phasescope become displaced from ideal. The average of all the symbol
positions is the ideal position.
Chapter 15
Modulation and Coding Principles Binary Pulse Code
Modulation
53230
Use the noise Amplitude control to set the signaltonoise ratio (SNR) to about 20dB. This
is made easier by selecting the Average function on the spectrum analyser, although any
adjustments now take longer to settle.
Note that, although the symbols on the phasescope have considerable noise on them, the
BER should be still zero. Remember that the signaltonoise ratio has been reduced by
30dB, i.e. a power ratio of 1000.
Increase the noise amplitude more. You should find that suddenly, around 15dB SNR, the
BER starts to rise. Increase the noise to maximum, which should result in a signaltonoise
ratio of about 10dB (you may have to reduce the modulation amplitudethe upper Signal
Level Controlto achieve this). You should see that the BER has degraded to around 3 to
4 percent errors, which, in a real system would be considered rather poor.
Set the signaltonoise ratio to give a BER of about 1 percent. Use the button at the
bottom of the Perform Practical screen to change the PCM input to Digital Ramp. Note
that, as you would expect, the actual data content has no effect on the BER.
To see what happens at lower than 10dB SNR, you can decrease the amplitude of the
modulation signal (use the upper Signal Level Control). The BER becomes degraded very
quickly and you can see that it is impossible to see on the phasescope where the correct
symbol should be.
Use the upper Signal Level Control to reduce the modulation signal to zero. The BER
meter displays O/R (over range) at 50 percent. Note that, even on random data, the worst
the BER meter can display is 50 percent. This is due to random chance, as there are only
two possible symbols in BPSK.
Chapter 16
Modulation and Coding Principles BiPhase Data Format
53230
BiPhase Data Format
Objectives
To understand the concepts of split phase, or biphase data formats of digital signals
To generate biphase formatted data and examine its waveform and spectrum
To decode biphase formatted data and appreciate the problem of phase uncertainty
Chapter 16
Modulation and Coding Principles BiPhase Data Format
53230
More information on Symbol Mapping
Symbol mapping is the process of taking digital data and allocating symbols to represent
it.
In a simple binary system, like BPSK and a simple binary data stream, the process is
obvious and simple: one binary state is mapped to one symbol and the other binary state
to the other. The only decision the designer has to make is which way round they are. In
most cases this would be arbitrary.
Once the modulation scheme becomes more complex, mapping becomes more complex.
Examine the example of QPSK. This has four symbols. If the data is simply a binary
stream, it only has two states. One method would be to take alternate bits and map them
to the alternate I and Q data. Exactly in what order this is done is, for the most part, again
arbitrary but, of course, the receive system has to be designed to the same mapping, or
the data will be jumbled up when it is demodulated.
Higher order systems require more complex mapping. Look at 256QAM (256 state
quadrature amplitude modulation). This has 4 bits on the I channel and 4 bits on the Q
channel for each symbol. If the data was made up of 8bit words, then the whole word
could be represented by one symbol. One mapping system might place all the high order
bits in the I channel and all the low order bits in the Q channel. This would work but,
unless the data was randomised in some way, could result in carrier recovery being
difficult. Other combinations have advantages and disadvantages. Other considerations
may include the fact that synchronisation data or error correcting data has been added.
If the system is even more complex, like hierarchical QAM, then the mapping is very
complex.
Chapter 16
Modulation and Coding Principles BiPhase Data Format
53230
Split Phase Data
One of the problems with many simple data formats is that they contain a dc component,
or have no transitions when the data is all 1s or all 0s. This creates problems, particularly
when modulated onto a carrier for transmission.
Split phase data overcomes theses problems by representing the data bits by transitions
rather than levels. A 1 is represented by a negative (high to low) change and a 0 by a
positive (low to high) change. This is sometimes called biphaselow or splitphase
low (spl) It could equally well be the other way round and would be called splitphase
high (sph).
This type of data stream contains transitions at the bit rate, but they are of opposite
polarity, depending on the data. If the transitions are put through a differentiator and then
rectified, pulses are produced that are at twice the bit rate. This can be used to generate a
clock at bit rate. However, it has to be divided by two, which creates a phase uncertainty
that, depending on the type of detector used, can result in false or inverted output data.
This can only be detected by using techniques such as sync word recognition.
Biphase coding is very effective but at the expense of some complexity in coding and
decoding. However, this is not usually an issue when using modern hardware solutions.
The diagram below shows biphaselow coded data.
This shows a unipolar form but it could equally well be bipolar. Since this form of coding is
usually used to create modulation symbols, in itself this is unimportant.
Chapter 16
Modulation and Coding Principles BiPhase Data Format
53230
Chapter 16
Modulation and Coding Principles BiPhase Data Format
53230
Practical 1: Generating BiPhase Data
Objectives and Background
In this practical you will generate biphase data and use the oscilloscope and spectrum
analyser to examine the signal.
Chapter 16
Modulation and Coding Principles BiPhase Data Format
53230
Block Diagram
Make Connections Diagram
Chapter 16
Modulation and Coding Principles BiPhase Data Format
53230
Chapter 16
Modulation and Coding Principles BiPhase Data Format
53230
Practical 1: Generating BiPhase Data
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the Signal Level Control to minimum.
Open the oscilloscope. The upper trace (blue) is NRZ data and the lower trace (yellow) is
the synchronisation signal.
Notice that, without any synchronising signal, the data is continuously moving in time.
Increase the Signal Level Control until the synchronisation is achieved.
Note the two signals. Confirm that the sequence of 1s and 0s, starting at the sync pulse,
is:
1 0 1 1 0 1 0 0 1 1 1 1 0 0 1 1 0 1 0 0 1 0 0 1
You may need to increase the size of the oscilloscope and decrease its timebase to see
all the data clearly.
Transfer the oscilloscope Channel 1 probe (blue) to the output of the BiPhase Coder
block (monitor point 4). Observe the biphase coded waveform.
Open the spectrum analyser and observe the spectrum of the biphase data.
By moving the spectrum analyser signal probe (blue) back and forth between the input
data (monitor point 1) and the coded data (monitor point 4), compare the two spectra.
Note that the biphase coded data has a higher bandwidth. You should be able to answer
why that is.
Close the spectrum analyser.
Now, use the oscilloscope to compare the input and output data of the Coder by moving
the oscilloscope Channel 2 probe (yellow) to the input of the Coder (monitor point 1).
Ensure that the Channel 1 probe is on the output point (monitor point 4).
Expand the timebase of the oscilloscope to give a clearer display. Ignore the transient
output spikes that are associated with the input data transitions these are due to minor
circuit imperfections and are not significant.
Satisfy yourself that the output data pattern is as would be expected. In relation to the
input data, note where the output data changes state. Refer to the Concept section on
Split Phase Coding to verify its operation.
Now, use the button at the bottom of the block diagram to switch to All 1. Note that the
output data is still present in the form of a square wave. Move the Channel 2 (yellow)
Chapter 16
Modulation and Coding Principles BiPhase Data Format
53230
probe to the Clock output (monitor point 2) and see that the output data is at the same
frequency as the clock.
Use the button at the bottom of the block diagram to switch to All 0. Note that the output
data is still a square wave, but that it is now in antiphase. Switch back and forwards
between All 1 and All 0 to verify this.
Chapter 16
Modulation and Coding Principles BiPhase Data Format
53230
Practical 2: Decoding BiPhase Data
Objectives and Background
In this practical you will use the biphase generator that you investigated in Practical 1 to
encode a data stream and a decoder block to recover the clock and hence the data. You
will notice that the data can be inverted due to the phase uncertainty introduced by the
dividebytwo block in the decoder.
Chapter 16
Modulation and Coding Principles BiPhase Data Format
53230
Block Diagram
Make Connections Diagram
Chapter 16
Modulation and Coding Principles BiPhase Data Format
53230
Chapter 16
Modulation and Coding Principles BiPhase Data Format
53230
Practical 2: Decoding BiPhase Data
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the Signal Level Control to minimum.
Open the oscilloscope. The upper trace (blue) is BiPhase Coded data and the lower trace
(yellow) is the synchronisation signal.
Increase the Signal Level Control until the synchronisation is achieved.
Note the two signals. Confirm that the sequence of 1s and 0s, starting at the sync pulse,
is:
1 0 1 1 0 1 0 0 1 1 1 1 0 0 1 1 0 1 0 0 1 0 0 1
You may need to increase the size of the oscilloscope and decrease its timebase to see
all the data clearly.
Transfer the oscilloscope Channel 1 probe (blue) to the output of the BiPhase Decoder
block (monitor point 8). Observe the decoded waveform.
Adjust the Bit rate lock control on the BiPhase decoder to achieve sync.
Move the oscilloscope Channel 1 probe (blue) to the output of the Binary Source (monitor
point 1) and compare the waveform with that on the Data+ output. You can move the
probe back and forth between monitor points 1 and 8 to do this.
Confirm, also, that output from the Data output (monitor point 7) is the complement of the
Data+ output.
Use the button at the bottom of the block diagram to change the data to All 1 and
observe the Data+ output.
Change to All 0 and observe the output.
Now, adjust the Bit rate lock control so that lock is lost and then readjust it to gain lock,
again. Observe the state of the Data+ output. Repeat this procedure several times to see
the effect of the phase uncertainty when lock is achieved.
Change back to Sequence and repeat the procedure of the last paragraph to see the data
inversion due to this uncertainty.
Chapter 16
Modulation and Coding Principles BiPhase Data Format
53230
Chapter 17
Modulation and Coding Principles Alternate Mark Inversion
53230
Alternate Mark Inversion
Objectives
To understand the concept of the alternate mark inversion (AMI) coding method and its
advantages and disadvantages
To generate AMI data and to compare it with RZ and NRZ formatting
To determine the difference in dc component of RZ, NRZ and AMI signals
Chapter 17
Modulation and Coding Principles Alternate Mark
Inversion
53230
More information on Symbol Mapping
Symbol mapping is the process of taking digital data and allocating symbols to represent
it.
In a simple binary system, like BPSK and a simple binary data stream, the process is
obvious and simple: one binary state is mapped to one symbol and the other binary state
to the other. The only decision the designer has to make is which way round they are. In
most cases this would be arbitrary.
Once the modulation scheme becomes more complex, mapping becomes more complex.
Examine the example of QPSK. This has four symbols. If the data is simply a binary
stream, it only has two states. One method would be to take alternate bits and map them
to the alternate I and Q data. Exactly in what order this is done is, for the most part, again
arbitrary but, of course, the receive system has to be designed to the same mapping, or
the data will be jumbled up when it is demodulated.
Higher order systems require more complex mapping. Look at 256QAM (256 state
quadrature amplitude modulation). This has 4 bits on the I channel and 4 bits on the Q
channel for each symbol. If the data was made up of 8bit words, then the whole word
could be represented by one symbol. One mapping system might place all the high order
bits in the I channel and all the low order bits in the Q channel. This would work but,
unless the data was randomised in some way, could result in carrier recovery being
difficult. Other combinations have advantages and disadvantages. Other considerations
may include the fact that synchronisation data or error correcting data has been added.
If the system is even more complex, like hierarchical QAM, then the mapping is very
complex.
Chapter 17
Modulation and Coding Principles Alternate Mark Inversion
53230
Alternate Mark Inversion Coding
Alternate mark inversion (AMI) is a coding method that represents a 1 by a pulse and a
0 by no pulse. The polarity of the pulses is inverted on alternate 1 s.
The advantage of such a system is that the overall dc component is zero, as there are by
definition an equal number of positive and negative pulses. The problem is that if there is
a long sequence of zeros there are no transitions and bit synchronization may be lost.
AMI is shown in the diagram below.
There are various modified versions of AMI that detect long sequences of zeros and add
violation bits. This prevents any clock recovery system from failing due to lack of
transitions. There are various forms of these codes, one being HDB3.
Chapter 17
Modulation and Coding Principles Alternate Mark
Inversion
53230
Practical 1: Generating AMI Coding
Objectives and Background
In this practical you will see generation of Alternate Mark Inversion (AMI) coding and use
the oscilloscope to compare it with other data formats.
Chapter 17
Modulation and Coding Principles Alternate Mark Inversion
53230
Block Diagram
Make Connections Diagram
Chapter 17
Modulation and Coding Principles Alternate Mark
Inversion
53230
Chapter 17
Modulation and Coding Principles Alternate Mark Inversion
53230
Practical 1: Generating AMI Coding
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the Signal Level Control to minimum.
Open the oscilloscope. The upper trace (blue) is the data and the lower trace (yellow) is
the synchronising signal.
Notice that, without any synchronising signal, the data is continuously moving with time.
Increase the Signal Level Control until synchronisation is achieved.
Note the two signals. The Bipolar NRZ sequence, starting at the sync pulse, should be:
1 0 1 1 0 1 0 0 1 1 1 1 0 0 1 1 0 1 0 0 1 0 0 1
You may need to increase the size of the oscilloscope and decrease its timebase to see
all the data clearly.
Use the button at the bottom of the block diagram to switch to Bipolar RZ. Confirm the
coded data.
Now switch to AMI and note how a sequence of ones is represented by an alternate
sequence of +1 and 1. Note also that a sequence of zeros generates no transitions. Test
this using the All 1 and All 0 buttons to give these data sequences.
Open the spectrum analyser and compare the dc component of NRZ, RZ and AMI. Switch
the Alias Hi on to prevent the analyser switching ranges.
Chapter 17
Modulation and Coding Principles Alternate Mark
Inversion
53230
Chapter 18
Modulation and Coding Principles Word Synchronisation
53230
Word Synchronisation
Objectives
To understand the necessity for word synchronisation in digital data transmission
To investigate how to achieve synchronism using the insertion of a synchronisation word
Chapter 18
Modulation and Coding Principles Word Synchronisation
53230
Word Synchronisation
In a digital communication receiver system there are various timing references that have
to be synchronized in order to recover the data successfully. If the system utilizes a carrier
of some description, such as does a radio system, then a local carrier reference has to be
generated and synchronized to the original carrier. Following that, a clock at the original
symbol rate has to be synchronized in order to recover a data stream. This data stream
has to be divided up into words, with a clock operating at word rate, and synchronized
such that the correct bits are placed in the correct place within each word.
There may be other issues, such as data polarity uncertainties or, in the case of a high
order modulation scheme, some uncertainty where the bits are. Finally, in some cases,
collections of words have to be divided into frames, which may, or may not, have error
correcting data associated with them.
One of the most common methods of achieving word synchronization is to insert, at
regular intervals, a known set of words. The data is then passed through a set of registers
and tested for the known word or words. Once the pattern is located, then the location of a
word within the bit stream is known. Further words can be located simply by the bit clock.
Although this procedure sounds fairly simple, it has a number of associated problems. The
first is: how do you guarantee that this particular synchronizing sequence will not appear
elsewhere in the data by chance? The answer is that, unless precautions are taken, you
cannot.
Usually, the sync sequence is several words long and can be made to be an unlikely
combination. It can be arranged to make sure that this sequence does not occur in the
data, but this will be at the expense of more preprocessing. Also, adding these sync
words reduces the available bits for data, so it would not be a good idea to make them too
long.
Another issue is: how often to you send them? If they are sent very often, sync can be
achieved quickly and, if sync is lost due to data loss, then reacquired quickly. The more
often sync is sent then, again, the available space for data is reduced.
Sync words can be conveniently sent at the beginning of frames, so that they can be used
for both word and frame sync. The options and methods used are many and various, often
depending on the system. The method chosen on a link that has short bursts of data that
requires quick acquisition under fairly noisy conditions might be different from that chosen
on a system that operates continuously in a fairly noisefree environment.
Chapter 18
Modulation and Coding Principles Word Synchronisation
53230
Practical 1: Word Synchronisation
Objectives and Background
In this Practical you will see a very simple system where a sync word is added to a
sequence at regular intervals. See the Concepts section for an explanation of the
concepts and the associated problems.
The data used is a 24 bit sequence comprising 3 eightbit words. The final word is
changed to a sync word at regular intervals.
This would not be satisfactory in a real system but will illustrate the concept. You can
switch the sequence to all ones and all zeros and still see the sync words being inserted.
Chapter 18
Modulation and Coding Principles Word Synchronisation
53230
Block Diagram
Make Connections Diagram
Chapter 18
Modulation and Coding Principles Word Synchronisation
53230
Chapter 18
Modulation and Coding Principles Word Synchronisation
53230
Practical 1: Word Synchronisation
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the Signal Level Control to minimum.
Open the oscilloscope. The upper trace (blue) is the data and the lower trace (yellow) is
the synchronising signal.
Notice that, without the synchronising signal, the data is continuously moving with time.
Increase the Signal Level Control until synchronisation is achieved.
Note the two signals. The Bipolar NRZ sequence, starting at the sync pulse, should be:
1 0 1 1 0 1 0 0 1 1 1 1 0 0 1 1 0 1 0 0 1 0 0 1
You may need to increase the size of the oscilloscope and decrease its timebase to see
all the data clearly.
Now switch to Sequence+Sync and note that the final word changes at intervals to the
sync word. The sync word is :
1 0 1 0 1 0 1 0
It may not appear at regular intervals due to the fact that the insertion rate and the
oscilloscope retrace rate are not synchronised.
Switch to All 1 and All 0 and note that the sync sequence is still inserted.
Chapter 18
Modulation and Coding Principles Word Synchronisation
53230
Chapter 19
Modulation and Coding Principles Uncoded Binary Data Formats
53230
Uncoded Binary Data Formats
Objectives
To understand the concepts of return to zero (RZ) and nonreturn to zero (NRZ) coding of
digital signals and to examine the spectra and bandwidth of each form
To generate both unipolar and bipolar RZ and NRZ signals and examine their waveforms
and spectra
To examine the dc component and bitrate components of such signals
Chapter 19
Modulation and Coding Principles Uncoded Binary Data
Formats
53230
Sampling
Signals in the real the world are analogue. In a digital communications system the first
process is to turn these analogue signals into digital format.
The signals could be anything: speech, television or representing the pH of a liquid, for
example. However, the common factor linking analogue signals is that they are time
continuous. This means that they are varying in time in a smooth manner. The diagram
shows a typical time continuous varying signal.
A digital signal is a series of discrete numbers that describes the signal, where each
number represents the signal at a particular point in time. This means that analogue signal
has to be sampled at various points in time and each value converted to a digital
number. This concept of sampling is very important to understand.
In order for the digital signal to be useful, three further factors have to be considered:
the sampling has to be regular;
the time interval between samples has to be short enough to follow the fastest changes in
the analogue signal;
in a digital signal not only is the time domain in discrete steps but so is the signal itself.
For example a signal may be represented by zero to fifteen amplitude states, which might
mean that some of the finer detail may be lost. The number of steps to which the signal is
digitised is an important consideration.
The terms used to describe these digitising parameters are:
Time
Signal
Chapter 19
Modulation and Coding Principles Uncoded Binary Data Formats
53230
the rate at which the signal is sampled regularly is called the sampling rate;
the number of levels in the digital signal is called the resolution;
the resolution is often a power of two as this represents steps in the number of bits in a
binary system.
For example 16 levels requires 4 bits and 256 levels requires 8 bits.
The following diagram shows the same signal but sampled and digitised to 8 levels
Note that the output steps between the available levels and is timed at the sampling
points. Note also that some of the detail of the signal has been lost due to both the lack of
resolution and the low sampling rate. In a digital system the choice of resolution and
sampling rate must be made very carefully.
If the sampling rate is far too low, then the wrong waveshape can be produced from time
repetitive signals. This effect is called aliasing and is described in another Theory section.
There are several methods of implementing both the analogue to digital process and the
digital to analogue process and these are described in another Theory section.
Time
3 Si
Sampling
points
Available
levels
Digitised
output
Chapter 19
Modulation and Coding Principles Uncoded Binary Data
Formats
53230
Data Formats
In dealing with binary signals you have accepted that they are represented by electrical
signals. One common convention for this is by simply representing a binary one by a
particular voltage and a zero by zero voltage. In this simple case, when two ones follow
each other the electrical signal simply stays at the same voltage. The diagram shows this
sort of signal.
This sort of format is referred to as Non Return to Zero or NRZ. This name results from
the fact that the signal does not return to zero voltage at the end of each bit. It is the most
common format inside a digital electronic circuit.
This format is more correctly called unipolar NRZ as the voltage is positive and zero.
The same data in bipolar NRZ would look like this:
Chapter 19
Modulation and Coding Principles Uncoded Binary Data Formats
53230
In this format a one is represented by +V and a zero by V. Obviously, in this case the
signal amplitude is twice as large but that need not be the case if the voltage were
reduced. The important fact to appreciate is that bipolar NRZ can be generated from
unipolar NRZ by simply adding a dc offset. The only case where this is important is when
direct cable transmission is used such as the original telegraph systems.
Another format is called Return to Zero or RZ. In this format only half the bit period is
used to carry each bit, the second half is always a zero. The unipolar and bipolar forms
are different and not simply related by a dc offset.
The same data in unipolar RZ is shown below.
Bipolar RZ is shown here.
Chapter 19
Modulation and Coding Principles Uncoded Binary Data
Formats
53230
You might wonder why these different formats exist. The answer lies in the complexity or
ease of decoding the data. In order to decode NRZ, a clock must be available that shows
where each bit starts and finishes. This clock might be available within, for example, an
electronic circuit. If NRZ is used to transmit data over a link then the clock has to be
regenerated at the receiving end. This is rather difficult for several reasons. One of these
is that if a long sequence of ones or zeros is transmitted no changes occur in the data and
so there is nothing with which to synchronise a clock recovery circuit. There is also the
problem that if the dc level drifts then the two levels will be wrongly decoded. Any form of
NRZ or RZ will only work if the transmission path has a frequency response including dc.
In most cases this is not so.
The frequency components present in each of the formats is demonstrated in the
practical.
Note that NRZ has a maximum bit rate equal to half the symbol rate. RZ has a frequency
component equal to the symbol rate. On the other hand, providing that there are no very
long sequences of ones or zeros, clock recovery from RZ is more feasible.
Chapter 19
Modulation and Coding Principles Uncoded Binary Data Formats
53230
More information on Symbol Mapping
Symbol mapping is the process of taking digital data and allocating symbols to represent
it.
In a simple binary system, like BPSK and a simple binary data stream, the process is
obvious and simple: one binary state is mapped to one symbol and the other binary state
to the other. The only decision the designer has to make is which way round they are. In
most cases this would be arbitrary.
Once the modulation scheme becomes more complex, mapping becomes more complex.
Examine the example of QPSK. This has four symbols. If the data is simply a binary
stream, it only has two states. One method would be to take alternate bits and map them
to the alternate I and Q data. Exactly in what order this is done is, for the most part, again
arbitrary but, of course, the receive system has to be designed to the same mapping, or
the data will be jumbled up when it is demodulated.
Higher order systems require more complex mapping. Look at 256QAM (256 state
quadrature amplitude modulation). This has 4 bits on the I channel and 4 bits on the Q
channel for each symbol. If the data was made up of 8bit words, then the whole word
could be represented by one symbol. One mapping system might place all the high order
bits in the I channel and all the low order bits in the Q channel. This would work but,
unless the data was randomised in some way, could result in carrier recovery being
difficult. Other combinations have advantages and disadvantages. Other considerations
may include the fact that synchronisation data or error correcting data has been added.
If the system is even more complex, like hierarchical QAM, then the mapping is very
complex.
Chapter 19
Modulation and Coding Principles Uncoded Binary Data
Formats
53230
Practical 1: Comparing NRZ and RZ in both Bipolar and Unipolar Forms
Objectives and Background
In this practical you will generate RZ and NRZ in both unipolar and bipolar forms and
examine the waveforms on the oscilloscope. The spectrum analyser will be used to
identify the different frequency components relative to the bit rate.
The different signals are generated by using the onboard microprocessor and the D/A
converters. A repeating 24bit sequence is used, so you can see clearly what is
happening. Buttons control the output format and also, if the output stream is the 24bit
sequence, all ones or all zeros.
In order to synchronise the oscilloscope, a signal containing a pulse at the start of the
sequence is available from the 1 bit Output of the processor.
The time calibration marks have been set so that there is a line every bit period.
Chapter 19
Modulation and Coding Principles Uncoded Binary Data Formats
53230
Block Diagram
Make Connections Diagram
Chapter 19
Modulation and Coding Principles Uncoded Binary Data
Formats
53230
Chapter 19
Modulation and Coding Principles Uncoded Binary Data Formats
53230
Practical 1: Comparing NRZ and RZ in both Bipolar and Unipolar Forms
Perform Practical
Use the Make Connections diagram to show the required connections on the hardware.
Set the Signal Level Control to minimum.
Open the oscilloscope. The upper trace (blue) is NRZ data and the lower trace (yellow) is
the synchronisation signal.
Notice that, without any synchronising signal, the data is continuously moving in time.
Increase the Signal Level Control until the synchronisation is achieved.
Note the two signals. Confirm that the sequence of 1s and 0s, starting at the sync pulse,
is:
1 0 1 1 0 1 0 0 1 1 1 1 0 0 1 1 0 1 0 0 1 0 0 1
You may need to increase the size of the oscilloscope and decrease its timebase to see
all the data clearly.
Now, use the button at the bottom of the block diagram to switch to Bipolar NRZ. Note
that the data now switches between plus and minus 0.45 volts but the general shape is
the same.
Now select Unipolar RZ data button. Notice that a 1 is represented by a positive pulse
during the first half of the bit period.
Now select Bipolar RZ. A 1 is still a positive pulse but now a 0 is a negative pulse
during the first half of the bit period.
Chapter 19
Modulation and Coding Principles Uncoded Binary Data
Formats
53230
Modulation and Coding Principles Appendices
53230
Using the Test Equipment
General Notes
Any of the instruments can be resized and moved at any time using conventional drag
anddrop mouse techniques. If you make an instrument small enough then only the
display area will be shown; you must increase its size again in order to restore the
controls. If you close any of the instruments and open them again they will return to their
default settings. Each instrument has a Defaults button which returns the equipment to its
default settings (equivalent to closing and reopening the instrument). If you want to return
all the instruments (and any other resource windows) to their default size and position
simply click the Auto Position button in the assignment side bar.
Some instruments allow you to place a cursor (by clicking the mouse) at any position on
their display; the cursor reveals information regarding the point at which it is located. You
will have to reactivate this cursor each time you change the settings, size or position of the
instrument.
The Oscilloscope
The Discovery oscilloscope has many of the functions that you would find on a
conventional or computerdriven scope. Its fundamental purpose is to show varying
waveforms plotted against time. It is a dual trace scope, which means that it can display
two separate waveforms at the same time.
The Y (voltage) axis is set to a default value by the practical for each channel, but you
may change it by using the + button for more volts/div and the  button for less volts/div.
Either only one channel can be displayed or both channels. The Y2 Show tick box
determines whether the second channel is shown. In twochannel mode, if the Overlay
box is ticked, the two traces are superimposed on the same scale as for one trace. If
Overlay is not ticked the display area is divided into two and each trace is displayed half
Modulation and Coding Principles Appendices
53230
size. The Y1 dc and Y2 dc tickboxes determine if the inputs are dc coupled or not (ac
coupled). If the signal has a large dc offset then ac coupling can be useful.
The X (time) axis is set to a default value by the practical but you may change it by using
the ^ button for a faster timebase and the v button for a slower timebase. The <> and ><
buttons provide a means of further expanding the trace if the highest, or lowest, timebase
is in use. If you have the X scale expanded and select a lower timebase speed then the X
scale automatically returns to its default setting.
An antialias feature automatically switches the timebase speed up if you select a rate
that may produce a misleading display due to aliasing. If this feature has increased the
timebase rate then the ^ button is coloured red.
The oscilloscope can also be operated in XY mode, where data from channel 1 is in the
vertical axis and data from channel 2 is in the horizontal axis. Because the oscilloscope is
a digital sampling scope, in XY mode the time base settings are still relevant and
determine the sampling rate for both channels. Also in XY mode the traces have
persistence and stay on the screen longer than one trace refresh.
Note that you can switch off the antialiasing feature from the main laboratory screen.
Triggering takes place when the selected trace crosses the zero volt level. If the Y2 Trig
box is ticked, then the trigger source is Channel 2. Otherwise, Channel 1 is used. The Neg
trig box enables only negative transitions to trigger the scope. Normally only positive ones
do.
If the signal has a large dc offset, ac triggering can be useful.
You can return to the default settings by pressing the Default button. The Auto Position
button on the Discovery laboratory window moves all the test instruments back to their
default positions and sizes on the screen but does not affect their settings.
A cursor is available to make more accurate measurements. Left click on the display area
to activate it. The green cursor can be moved to anywhere on a waveform. Move the
mouse away and back into it to allow a tooltip window to open with the measurement data
displayed for that point.
You have to reactivate the cursor if you change the settings, size or position of the
oscilloscope.
By right clicking on the display an options box appears. The options available are:
Print Display Sends image to the default printer.
Export Display to File Opens a window enabling the name for the file you wish to use to
be entered and the directory where to save the file can be selected.
Export Display to File (reverse colours)  Opens a window enabling the name for the file
you wish to use to be entered and the directory where to save the file can be selected.
Modulation and Coding Principles Appendices
53230
The Spectrum Analyser
The spectrum analyser enables you to look at signals in the frequency domain. In
common with many modern test instruments, it uses DSP to transform time domain data
into frequency domain data. The mathematics to do this is called a Fourier transform.
The Y (amplitude) scale is calibrated in Decibels relative to an arbitrary dotted line near to
the top of the screen. The dB scale is linear and the number of dB per division is shown in
the box. The Y (amplitude) axis is set to a default value by the practical, but you may
change it by using the + button or the  button to change the Ref dB value higher or lower.
The minimum level that you can see is determined by the assignment, and ultimately by
the noise in the system.
The analyser has the capability of showing two channels at the same time. Click Ch2
Show button to show channel 2 as well as channel 1.
The X (frequency) axis is calibrated in MHz, kHz or Hz per division, as appropriate. The
default scale is set by the practical but you may change it by using the Higher Frequency
and Lower Frequency buttons.
The antialias feature will operate if you try to set the frequency too low. The Higher
Frequency button is shown red if this feature has increased the frequency. Note that if a
new frequency component appears such as noise, the antialias feature may operate
suddenly. The Alias Hi tickbox allows you to increase the threshold at which the antialias
feature operates. This allows signals to be examined that have larger amounts of
harmonic content. The default setting for this is off.
A cursor is available to make more accurate measurements. Left click on the display area
to activate it. The green cursor can be moved to anywhere on a waveform. Move the
Modulation and Coding Principles Appendices
53230
mouse away and back into it to allow a tooltip window to open with the measurement data
displayed for that point.
You will have to reactivate the cursor if you change the settings, size or position of the
spectrum analyser.
By right clicking on the display an options box appears. The options available are:
Print Display Sends image to the default printer.
Export Display to File Opens a window enabling the name for the file you wish to use to
be entered and the directory where to save the file can be selected.
Export Display to File (reverse colours)  Opens a window enabling the name for the file
you wish to use to be entered and the directory where to save the file can be selected.
The Phasescope
The Phasescope is a special instrument that compares two signals in phase and
amplitude (magnitude). The two signals are referred to as the reference and the input. The
display is in polar format, i.e. the phase is in the form of a circle and the amplitude as the
radius. The use of a circle is possible because phase is a continuous function repeating
every 360 degrees. The display can be seen as Polar, as the one orthogonal axis
represents the real component and other the imaginary part. The convention here is that
the real axis is the X axis, which means that zero degrees is straight up or at 12 on a clock
face. +90 degrees is at 3 on the clock face and 90 at 9. It is important to note that in
terms of phase +180 degrees is the same as 180 degrees.
The radius scale has one circle at radius = 1 (the outermost circle) i.e. the two signals are
of the same amplitude. Further inner circles are at 0.707, 0.5 and 0.25.
The circle at 0.5 has a square associated with it, the corners of which are at 0.707. This
represents the case when two orthogonal vectors of amplitude = 0.5 are added.
Modulation and Coding Principles Appendices
53230
In some cases only the phase is of interest so, if you click the Phase Only box, the radius
is set to 1.
The conventional display is that of a vector i.e. a line joining the point to the centre.
However, in some cases it is much easier to interpret the display if only a point is drawn.
Where the amplitude and phase is varying between discrete values they are shown as a
pattern of dots resembling stars, hence the term constellation display. This mode can be
selected by ticking the Constellation box. In constellation mode, the persistence of the
display can be varied. By selecting the Persistence tick box, traces stay on the screen for
a number of trace refreshes before being removed. By selecting Hi Persist this time is
extended.
If the two signals are of different frequencies the result is a continuously rotating vector,
rotating at a rate equal to the difference in frequency. The direction depends on the sign of
the frequency difference. If the rate is fairly fast, the instrument may only be able to show
a limited number of discrete values.
In many cases the reference input will not be at exactly zero degrees with respect to the
theoretical zero degrees of the input signal. This causes the display to be rotated. In some
cases this may be important to know, but where it is not the Phi Offset control gives the
ability to rotate the display for easier interpretation.
The coloured indicator (Ref Ch) to the top left of the display tells you which probe is being
used as the reference channel.
A cursor is available to enable more accurate measurement. Click the display to use it.
By right clicking on the display an options box appears. The options available are:
Print Display Sends image to the default printer.
Export Display to File Opens a window enabling the name for the file you wish to use to
be entered and the directory where to save the file can be selected.
Export Display to File (reverse colours)  Opens a window enabling the name for the file
you wish to use to be entered and the directory where to save the file can be selected.
The Voltmeter
Modulation and Coding Principles Appendices
53230
The meter is simply an ac and dc voltmeter that displays the value in digital form. It can be
used in ac mode by clicking ac pp, in which case the value represents the peak to peak
value. If the waveform has a high crest factor the results can be slightly surprising. In dc
mode, if there is an ac component present, the average value is displayed.
The Frequency Counter
This has the facilities of a conventional frequency meter/counter. It will display in either
frequency or time. If the input amplitude is too low a warning message will be displayed.
Like all frequency counters, it can produce misleading results if the waveform is complex
or contains many frequencies.
Modulation and Coding Principles Appendices
53230
Discovery System Help
Although the Discovery environment is very easy to operate, these notes will help you use
all its facilities more quickly.
If there is a demonstration assignment, slider controls in the software perform functions
that would normally be performed on the hardware. In normal assignments, if the any of
hardware systems fail to initialise the system reverts to demonstration mode. This means
that none of the test equipment is available.
The Assignment Window
The assignment window opens when an assignment is launched. If you are reading this
you have already found the help button in the side bar of the assignment window!
The assignment window consists of a title bar across the top, an assignment side bar at
the righthand edge, and the main working area. By default, the overall assignment
objectives are initially shown in the main working area whenever an assignment is opened.
The assignment window occupies the entire screen space and it cannot be resized (but it
can be moved by dragging the title bar, and it can be minimised to the task bar). The title
bar includes the name of the selected assignment. The side bar contains the practicals
and any additional resources that are relevant for the selected assignment. The side bar
cannot be repositioned from the righthand edge of the assignment window. An example
of an assignment window is shown below.
Modulation and Coding Principles Appendices
53230
The precise appearance of the assignment window will depend on the skin that has been
selected by your tutor. However, the behaviour of each of the buttons and icons will
remain the same, irrespective of this.
The clock (if you have one active) at the top of the side bar retrieves its time from the
computer system clock. By double clicking on the clock turns it into a stop watch. To start
the stop watch single click on the clock, click again to stop the stop watch. Double clicking
again will return it to the clock function.
There are a number of resource buttons available in the assignment side bar. These are
relevant to the selected assignment. In general, the resources available will vary with the
assignment. For example, some assignments have video clips and some do not. However,
the Technical Terms, Help and Auto Position buttons have identical functionality in every
assignment. You can click on any resource in any order, close them again, or minimise
them to suit the way you work.
Practicals are listed in numerical order in the side bar. When you hover the mouse over a
practical button, its proper title will briefly be shown in a popup tooltip. There can be up
to four practicals in any assignment. You can have only one practical window open at any
time.
To perform a practical, leftclick on its button in the assignment side bar. The assignment
objectives, if shown in the main working area, will close, and the selected practical will
appear in its own window initially on the righthand side of the main working area, as
shown below. You can move and resize the practical window as desired (even beyond the
assignment window). However, its default size and position is designed to allow the test
equipment to be displayed down the lefthand side of the main working area without
overlapping the instructions for the practical.
Modulation and Coding Principles Appendices
53230
Again, the precise appearance of the practical window can be determined by your tutor
but the behaviour of each of the buttons and icons will remain the same, irrespective of
this. Whatever it looks like, the practical window should have icons for the test equipment,
together with buttons for Objectives & Background, Make Connections, Circuit Simulator
and Test Equipment Manuals. These resources are found in side bar, located on the right
hand edge of the practical window. The resources will depend on which practical you have
selected. Therefore not all the resources are available in every practical. If a resource is
unavailable, it will be shown greyed out. To open any resource, leftclick on its icon or
button. Note that when you close a practical window, any resources that you have opened
will close. You may open any resource at any time, provided it available during the
practical. The Circuit Simulator will only be available if you have one loaded.
Note that if the hardware is switched off, unavailable, or its software driver is not installed,
all the test equipment is disabled. However, you can open any other window. If you switch
on the hardware it will be necessary to close the assignment window and open it again to
enable the test equipment.
Resource Windows
These are windows may be moved, resized and scrolled. You may minimise or maximise
them. The system defaults to Auto Position, which means that as you open each
resource window it places it in a convenient position. Most resource windows initially place
themselves inside the practical window, selectable using tabs. Each one lays over the
previous one. You can select which one is on top by clicking the tab at the top of the
practical window. You can see how many windows you have open from the number of
tabs. If you want to see several windows at once then drag them out of the practical
window to where you wish on the screen. If you close a window it disappears from the
resources tab bar.
If you want to return all the windows to their default size and position simply click the Auto
Position button in the assignment side bar.
Make Connections Window
This movable and resizable window shows the wire connections (2mm patch leads) you
need to make on the hardware to make a practical work. Note that some of the wires
connect the monitoring points into the data acquisition switch matrix. If this is not done
correctly the monitoring points on the practical diagram will not correspond with those on
the hardware. The window opens with no connections shown. You can show the
connections one by one by clicking the Show Next button or simply pressing the space bar
on the keyboard. If you want to remove the connections and start again click the Start
Modulation and Coding Principles Appendices
53230
Again button. The Show Function button toggles the appearance of the block circuit
diagram associated with the practical.
Test Equipment
The test instruments will autoplace themselves on the left of the screen at a default size.
You may move or resize any instrument at any time. Note that below a useable size only
the screen of the instrument will be shown, without the adjustment controls. Each piece of
test equipment will launch with default settings. You may change these settings at any
time. There is an auto antialias feature that prevents you setting timebase or frequency
settings that may give misleading displays. If auto antialias has operated the button turns
red. You can turn off the antialiasing feature, but you should be aware that it may result in
misleading displays.
You may return to the default settings by pressing the Default button on each piece of test
equipment. If you wish to return all the equipment to their original positions on the left of
the screen click Auto Position on the side bar of the assignment window.
Note that if you close a piece of test equipment and open it again it returns to its default
position and settings.
If you want more information on how a piece of test equipment works and how to interpret
the displays, see the Test Equipment Manuals resource in the practical side bar.
On slower computers it may be noticeable that the refresh rate of each instrument is
reduced if all the instruments are open at once. If this is an issue then only have open the
instrument(s) you actually need to use.
Test Equipment Cursors
If you left click on the display of a piece of test equipment that has a screen, a green
cursor marker will appear where you have clicked. Click to move the cursor to the part of
the trace that you wish to measure. If you then move the mouse into the cursor a tooltip
will appear displaying the values representing that position. Note if you resize or change
settings any current cursor will be removed.
Practical Window
This window contains the instructions for performing the practical, as well as a block, or
circuit, diagram showing the circuit parts of the hardware board involved in the practical.
On the diagram are the monitoring points that you use to explore how the system works
and to make measurements. The horizontal divider bar between the instructions and the
diagram can be moved up and down if you want the relative size of the practical
Modulation and Coding Principles Appendices
53230
instruction window to diagram to be different. Note that the aspect ratio of the diagram is
fixed.
Information Buttons on Practical Diagrams
On many of the symbols on the diagram you will find a button that gives access to new
windows that provide more information on the circuit that the symbol represents. Note that
these windows are modal, which means that you can have only one open at a time and
you must close it before continuing with anything else.
A Further Information point looks like this
Probes
The practical diagram has probes on it, which start in default positions. These determine
where on the hardware the signals are being monitored.
Selecting and Moving the Probes
Probes are indicated by the coloured icons like this .
If this probe is the selected probe it then looks like this (notice the black top to the
probe). You select a probe by left clicking on it.
Monitor points look like this
If you place the mouse over a monitor point a tooltip will show a description of what signal
it is.
You can move the selected probe by simply clicking on the required monitor point. If you
want to move the probe again you do not have to reselect it. To change which probe is
selected click on the probe you want to select.
You can also move a probe by the normal draganddrop method, common to Windows
programs.
Probes and Test Equipment Traces
The association between probes and traces displayed on the test equipment is by colour.
Data from the blue probe is displayed as a blue trace. Yellow, orange and green probes
Modulation and Coding Principles Appendices
53230
and traces operate in a similar way. Which piece of test equipment is allocated to which
probe is defined by the practical.
Note that the phasescope shows the relative phase and magnitude of the signal on its
input probe using another probe as the reference. The reference probe colour is indicated
by the coloured square to the top left corner of the phasescope display.
Practical Buttons
On some practicals there are buttons at the bottom of the diagram that select some
parameter in the practical. These can be single buttons or in groups. Only one of each
button in a group may be selected at one time.
Slider Controls
Where slider controls are used you may find you can get finer control by clicking on it and
then using the up and down arrow keys on your keyboard.
Modulation and Coding Principles Appendices
53230
Modulation and Coding Principles Appendices
53230
More information on Amplifiers
Amplifiers, as their name suggests, are electronic circuits that increase the magnitude of a signal. In
fact amplifiers are one of the most common circuit blocks but they are not as simple as they might
seem.
There are several types of amplifiers:
Voltage amplifiers, where the output voltage is greater then the input voltage. Often the input
impedance is high and the output impedance low.
Current amplifiers, the output is a current. The input can be either a voltage or a current.
Power amplifiers, where the overall power available at the output is greater then that supplied at
the input. The input and output impedances can be virtually anything and need not be equal.
The signal magnification is usually referred to as gain. In order to interpret gain figures correctly it
is important to know what type of amplifier it is and what the input an output impedances are. Gain
is simply a ratio and is sometimes expressed in decibels to avoid large unmanageable numbers.
Some voltage and current amplifiers pass any dc component in the signal. These are called dc
coupled amplifiers. Some only pass the ac components and are said to be ac coupled. The overall
range of frequencies that are passed is referred to as the bandwidth.
An ideal amplifier passes the signal with no distortion of any kind. There are no such devices. A
real amplifier distorts the signal by nonlinear amplification, unequal frequency response,
differential phase changes, slew rate limitations and by adding noise. As you can now see,
amplifiers can be a significant problem in a system.
There are some cases when an amplifier is made intentionally nonlinear. The most common
example is when the gain is made very large, but the output limits at a particular amplitude. This
means that when amplitude changes are not important such as in frequency modulated signals, some
improvement in noise performance can be obtained. An amplifier operating in this manner is often
called a limiter. The output of a limiter is almost always a square wave.
One final characteristic is that for some amplifiers when operating inside their passband the output
is in phase with the input. These are called noninverting. For others, the output is intentionally 180
degrees out of phase. These are inverting amplifiers. It is sometimes necessary to invert a signal, to
subtract it from another for example, so an amplifier block may be used that has no gain but inverts
the signal.
Modulation and Coding Principles Appendices
53230
More information on Data Sources
The data sources used are of three types:
One is simply a square wave generator, which can be thought of as a constant
stream of binary ones and zeros. Because the stream is regular it is easy to see
what is happening.
Another is a similar binary source, but originating in a microprocessor on the
workboard. It is used to generate more complex binary signals.
The third also uses the microprocessor, but is the result of passing the output from
the microprocessor through a pair of digital to analogue converters. This means
that there are available two data source signals, each of which has more than two
levels.
These are used to generate the modulation schemes that have more than one bit
per symbol.
Modulation and Coding Principles Appendices
53230
More information on Detectors and Demodulators
The words detector and demodulator can be taken to have the same meaning.
The name detector is generally regarded as being old fashioned and is now
normally used in connection with simple demodulators for analogue amplitude and
frequency modulation.
The purpose of the demodulator is to recover the original modulating signal with
the minimum of distortion and interference. They can be very simple, or it can be
very complex to demodulate a complex modulation scheme in the presence of
noise.
Envelope Detectors
The simplest way of dealing with an amplitude modulated (AM) signal is to use a
simple halfwave rectifier circuit. If the signal is simply passed through a diode to a
resistive load, the output will be a series of halfcycle pulses at the carrier
frequency. So the diode is followed by a filter, typically a capacitor and resistor in
parallel.
The capacitor is charged by the diode almost to the peak value of the carrier cycles
and the output therefore follows the envelope of the amplitude modulation. Hence
the term envelope detector.
The time constant of the RC network is important because if it is too short the
output will contain a large component at carrier frequency. However, if it is too long
it will filter out a significant amount of the required demodulated output. A fullwave
rectifier may be used, which means the carrier component is at twice the carrier
frequency and the filter has an easier job. Usually there is sufficient difference
between the upper limit of the baseband signal and the carrier to make this
unimportant.
Product Detectors
If the AM signal is multiplied with (i.e. modulated by) a frequency equal to that of its
carrier, the two sidebands are mixed down to the original modulating frequency and the
carrier appears as a dc component.
This frequency source is referred to as a carrier insertion oscillator, a beat frequency oscillator or as a type of local
oscillator.
The mathematics of the process show that this will only happen if the mixing
frequency is equal not only in frequency to that of the carrier, but also in phase; i.e.
Modulation and Coding Principles Appendices
53230
the two signals are synchronous. This is why a product detector when used for AM
is sometimes called a synchronous detector. For AM, the effect is very similar to a
fullwave rectifier, rather than the halfwave of the envelope detector.
The output still needs a postdetection filter to remove the residual ripple but, like a
full wave rectifier, the ripple is at twice the carrier frequency and is therefore further
away from the modulation and hence easier to remove. Also, in general terms, the
product detector gives less distortion, as it eliminates the non linearity of diodes.
Generating the carrier local oscillator is easy but making sure that it is at exactly the
correct frequency is much more difficult. When the phase needs to be locked to that
of the original carrier it becomes very complex.
Modulation and Coding Principles Appendices
53230
More information on Differentiators
The differentiator is an electronic, or DSP, block that implements the mathematical
differentiating operation with respect to time.
In just the same way as with the mathematical operations, the differentiator is the
opposite of the integrator.
For a differentiator, if an input is applied that increases linearly with time the output
will be a constant value. In electronics, this function is usually achieved with an
operational amplifier that has a capacitor in its input path and a resistor in its
feedback path. In DSP it can be done by subtracting the values of a discrete time
sampled signal.
The diagrams show the action of a differentiator on an input signal comprising a
number of different linearly varying voltages.
Modulation and Coding Principles Appendices
53230
A differentiator produces an output that is proportional to the rate of change of the
input. If a square wave is applied to the input the output would theoretically be a
series of positive and negative pulses of infinite amplitude and zero duration.
Practically, because no square wave is ideal and no differentiator can respond
quickly enough, a typical output waveform may be as shown in the diagram below.
Modulation and Coding Principles Appendices
53230
A problem arises in that noise impulses can cause a differentiator to produce
unwanted output that could cause limiting of the circuit. Often the high frequency
response of a differentiator is made poor, such that the circuit does not respond to
transient noise impulses.
Modulation and Coding Principles Appendices
53230
More information on Filters
Filters are devices that pass signals of certain frequencies and block others.
They can be very simple or complex. Until relatively recently, most filters were analogue
devices but, with the increased power of digital signal processing (DSP), many filters are
now implemented digitally. The mathematics of analogue and digital filters are essentially
the same.
Sometimes, analogue filters use only passive components: like capacitors and inductors.
However, many operating at low frequency use active components: like operational
amplifiers.
Filters divide into several categories depending on the ranges of frequencies they pass.
The most common is called a lowpass filter, which passes all signals up to a certain
frequency. This frequency is called the cutoff frequency.
A high pass filter only passes signals above its cutoff frequency.
The third type is called a bandpass filter and passes signals between two limits.
As you can see from the diagram, the response of the filter does not fall to zero immediately at the cutoff frequency.
Frequency
Amplitude
low pass
bandpass
highpass
Modulation and Coding Principles Appendices
53230
The steepness of the response, called the rolloff, is determined by the complexity of the filter. Non digital filters are
ultimately limited by losses in the components. Digital filters can have almost ideal responses.
Other considerations in filter design can be:
their input and output impedances,
passband loss,
passband ripple,
signal delay and
phase response.
The design of both analogue and digital filters is a complex subject that has been
made somewhat easier by computer simulation.
Modulation and Coding Principles Appendices
53230
More information on Data Formatting
This block takes binary data and maps it to voltage levels. There are many possible
formats such as NRZ, RZ, bipolar and unipolar. Some of the formats are described
in the Concepts section of assignments dealing with those formats.
Modulation and Coding Principles Appendices
53230
More information on Frequency Dividers
The frequency divider is a block that divided the frequency of an incoming signal by an
integer. The only method available is by using digital counters. These devices can divide
by anything from 2 to an almost unlimited number. Digital counters are binary and the
simplest numbers to divide by are powers of two. Other numbers are achievable by using
decoders or presettable counters. At very high frequencies delays in such systems can
become a problem.
Digital circuits all require specific voltage levels to operate and therefore buffering and
limiting may be needed if a small amplitude analogue signal is the input.
Digital counters come with a multitude of functions such as reset and preset inputs,
overflow outputs, up or down counting and variable modulus controls. The logic families
currently available include ttl, cmos, hcmos, and acmos. The choice is usually determined
by voltage levels, power consumption and frequency range.
Modulation and Coding Principles Appendices
53230
More information on Frequency Multipliers
A frequency multiplier is simply a block that multiplies a signal frequency by an integer
number. For example, if the input was 2MHz and the device was a frequency doubler the
output would be 4MHz.
Mathematically:
F
out
= nF
in
where n is an integer.
The simplest method to achieve this is by turning the incoming signals into pulses, which
contain many harmonics, and using a bandpass filter to select the wanted harmonic. This
method works for any integer value, but the efficiency drops significantly above about 5. It
only works for narrowband signals containing components in a frequency range less than
the original centre frequency divided by the multiplication factor.
This method is usually used in high frequency applications that are usually narrow band.
An alternative method is available for frequency doublers. This uses a multiplier block with
both inputs fed from the same source. This method works well for signals containing a
wide range of frequencies.
Input
input
Pulse
Generator
Band pass filter
output
Multiplier
output
Modulation and Coding Principles Appendices
53230
General Template
Title
Write text here.
Modulation and Coding Principles Appendices
53230
More information on Integrators
The integrator is an electronic, or DSP, block that implements the mathematical
integrating operation with respect to time.
This means that for a constant input the output increases linearly with time.
In electronics, this function is usually achieved with an operational amplifier that has a
capacitor in its feedback path. In DSP it can be done by summing the values of a discrete
time sampled signal.
The diagram shows the action of an integrator on an input signal comprising a number of
different levels.
One problem arises from the fact that the output of an integrator can, in theory, reach
infinite values. This is clearly not possible in an electronic or DSP implementation. The
solution is either to limit the value at some practical level, or reset it. The reset solution
can be used if the output represents a signal that repeats: like an angle, for example. In
some situations other methods, like ensuring that the number of ones and zeros are equal
over a period of time may prevent the situation occurring.
Input
Output
Modulation and Coding Principles Appendices
53230
More information on Multipliers
You are familiar with the mathematical process of multiplication, but how does it work with two signals?
In fact, it is exactly the same process, where the instantaneous values of two
signals are mathematically multiplied together. In most cases at least one of the
signals is time varying.
Firstly, consider two constant dc voltages: one of 2 volt and the other of 3 volts. The
product is clearly 6 volts. Now if one were a sine wave of 3 volts peak to peak and
the other a constant 2 volts then the result is a sine wave of 6 volts peak to peak.
Of importance is what happens if the constant voltage is minus 2 volts. The result is
still a sine wave of peak to peak amplitude 6 volts. However, mathematically the
sine wave is now a minus sine and has therefore been reversed in phase.
Remember that, in mathematics, multiplying two negative numbers together results
in a positive number.
Of importance also is what happens when both signal are time varying. If they were
both sine waves then the result would be of the form
Output(t)=sine(Signal 1) x sine (Signal 2)
This is an amplitude modulation process and results in new frequencies being
produced The mathematics for this can be found in the Modulation Maths concept.
The terms amplitude modulation and multiplication have the same meaning, but the term multiplier sometimes describes the process
better when the objective is not that of modulation However all modulators are multipliers of some sort. In some circumstances it is not
necessary to be able to deal with both positive and negative signals; that might add unnecessary complication to the circuit. Multipliers
that can deal with both signals of both polarity are called four quadrant multipliers, or sometimes as being balanced.
Signal 2
Signal 1
X (Signal 1) X (Signal 2)
Modulation and Coding Principles Appendices
53230
When the purpose is to produce a product frequency for further processing, but not at baseband, the term mixer is often used.
When a multiplier is implemented in analogue electronics there are usually some inaccuracies in the performance, caused by physical
effects in the components. There is usually some distortion, which means that some of the carrier signal energy is transferred to the
second harmonic. This is not usually a problem, as multipliers are often followed by some sort of filter.
A more serious problem is that there is usually a small amount of dc offset on one or more of the inputs. This means that, when a zero
signal is put in, the other input is still multiplied by a small amount. The multiplier is then said to be unbalanced. This may or may not
be a problem but, when it is, a pair of balancing controls are added and adjusted to make the balance as perfect as possible.
One of the major causes of imbalance is non perfect matching of components, due to slight variations of actual component values. In
very critical applications one also has to be aware that the balance can change a small amount with temperature. One way to reduce
this effect is to have all the transistors fabricated on one silicon chip, so they are all at the same temperature.
The modulators on the hardware associated with this product have balance controls so you can adjust the balance where necessary.
Modulation and Coding Principles Appendices
53230
More information on Noise Generators
Being able to generate adjustable amounts of defined electronic noise is useful for testing
systems that will have to work in a real situation in the presence of noise.
There are several ways by which noise may be generated. One uses the inherent
properties of a semiconductor junction to produce low amplitude wideband noise. This sort
of source is used at very high frequencies.
At low frequencies a special shift register configuration can be used to generate noise
over defined bandwidths. The noise generated in this way is not truly random but, in most
cases by choosing the system parameters well, the difference is not significant. This type
of circuit is called a pseudo random binary sequence generator (PRBS) and the noise is
called pseudo random noise.
The generator on the hardware board uses a PRBS.
Modulation and Coding Principles Appendices
53230
More information on Paths
What is meant by a signal path? In a general it means a method by which an electrical signal travels from one place to
another in a system. In some cases this may simply be a direct connection and, for practical purposes, the signal at one
end of the path is the same as the other. Inside a system the signal may be processed by some circuit, such as an
amplifier, or by something more complex.
When a signal is sent some distance by radio transmission or via a cable, for
example, this is also referred to as a path or, more correctly, as a transmission
channel. Hopefully when the signal arrives it is not significantly different from that
which was transmitted. However, some changes will have taken place. It will almost
certainly be smaller in amplitude, for example.
On its journey, the signal may well have some noise added to it, and it may be
distorted in some way. Some types of distortion are subtle, such as echoes caused
by radio transmission, or by imperfections in a cable. Almost certainly there will be
more than one type of distortion present and these will depend on the transmission
medium.
Much of the processing that is applied to a signal before and after transmission,
and also the choice of modulation method used, are to combat problems induced
by the transmission channel.
In this equipment a transmission path is modelled by adding noise and perhaps a phase shift, which allow you to see many of the
effects.
Transmission
System
Path or
Channel
Reception
System
Path or
Channel
Signal
in
Noise
Phase
Shift
Signal
out
Modulation and Coding Principles Appendices
53230
More information on Phase Shift Networks
A phase shift network is a subsystem that introduces an intentional phase shift into a
signal. Adding phase shift to a signal is very easy. In fact, not adding phase shift is more
of a problem. However, adding a known phase shift is much more difficult.
An even more difficult problem is adding a phase shift that is not dependant on frequency.
Simply adding a delay to a signal will introduce a phase shift, but that shift is proportional
to frequency, because wavelength is inversely proportional to frequency.
Many, if not all, filters add phase shift but, in most, this varies wildly with frequency.
To achieve a known phase shift at a single frequency is quite easy and an analogue
network can be designed and produced with few components. In an IQ modulator, for
example, the carrier has to be shifted by 90 degrees. Since the carrier is normally at a
single frequency, this phase shift is not difficult.
Wideband analogue 90 degree phase shift networks are very difficult to design and the
results are never perfect. They can be achieved with a large number of matched, close
tolerance components and often comprise actually two networks, the outputs of which
track 90 degrees apart over quite a wide frequency range.
However, the advent of digital signal processing means that wideband 90 degree shifts
can be generated much more easily. The mathematical term for this is a Hilbert
Transform.
Modulation and Coding Principles Appendices
53230
More information on the Sine/Cosine Processor Block
This block uses analogue electronics to reproduce a mathematical function.
For an input voltage, V
in
, representing an angle, the outputs are:
Output
(sine)
= sin(V
in
)
Output
(cosine)
= cos(V
in
)
In mathematics, because both sine and cosine are functions that repeat every 360
degrees, the angle can be infinite. For example, the sine and cosines of 360+90
degrees are the same as for 90 degrees.
Here, an infinite angle would be represented by an infinite voltage and, since this is
impractical, in the Angle Generator on the workboard the inputs are limited to
represent 360 degrees. In fact, the system input is arranged to represent minus 180
degrees through zero to plus 180 degrees. For convenience the minus 180 degrees
is represented by 0.5 volts and plus +180 degrees by +0.5 volts.
The output is scaled so that +1 (sin90 for example) is represented by +0.5 volts and
1 by 0.5 volts.
In graphical terms, the transfer function looks like this.
+0.5
Modulation and Coding Principles Appendices
53230
The Angle Generator block has many uses, particularly when fed with a varying voltage that represents an angle in
phase modulation. This produces sine and cosine signals which, when fed to an IQ modulator, produces a phase
modulated signal. This works because the two orthogonal carriers in the IQ modulator need to be modulated with sine
and cosine to achieve the same effect as the mathematics of Cartesian to polar conversion.
An interesting and important observation is that if the sine/cos block input is a
sawtooth wave, varying between 0.5V and +0.5V, the output is a sine and cos
continuous function. The sharp transition from +0.5 to 0.5 represents the point
where sin(180) is the same as sin(+180)
This is an important concept when an IQ modulator is used to generate an FM
signal.
0
0.5
+0.5
Input
0.5
0
Output
cos
sin
Modulation and Coding Principles Appendices
53230
More information on Signal Sources
A signal source is usually some kind of generator that produces a signal output
with no signal input. An oscillator is an example of a signal source. Of course, the
amplitude, frequency and waveshape may take many values, or it could be simply a
constant dc voltage. Usually, we differentiate between a dc source that is used as a
power supply and that used for a signal in a circuit.
Sources may be completely autonomous, i.e. they have no inputs at all, simply an
output. However, many sources have inputs that control an output parameter such
as amplitude or frequency. These are called control inputs. Some may have
synchronising inputs that allow the output phase or frequency to be locked to an
input signal. To be regarded as a true source the output should continue when all
the control signals are removed.
Source Control
input
Control
input
Output
Modulation and Coding Principles Appendices
53230
More information on Signal Summing
Summing blocks carry out the signal processing of mathematical addition. The output signal is the mathematical sum of
the inputs. There can be any number of inputs.
Consider adding a 3 volt peak to peak sine wave and a constant 2 volt dc signal.
The sine wave signal varies between plus 1.5 volts and minus 1.5 volts. The output
will be a sine wave with a peaktopeak amplitude of 3 volts but varying between 3.5
volts and 0.5 volts.
Mathematically :
Output(t) = Signal
1
(t) + Signal
2
(t)
It is important to understand that addition is a linear process, i.e. the only
frequencies present in the output are those that were in the two input signals. This
is different to a nonlinear process such as multiplication in which new frequencies
are generated.
Signal 1
Signal 2
+
(Signal 1) + (Signal
2)
Modulation and Coding Principles Appendices
53230
More information on Voltage Controlled Oscillators
A voltage controlled oscillator (VCO) is a particular type of signal source. It is a
source the output frequency of which is a function of a control voltage applied to it.
With some VCOs it may only be possible to vary the frequency over a small range;
with others, it is sometimes possible to vary from almost zero frequency. When a
voltage equal to half the possible control voltage range is applied, the resultant
output frequency is called the centre frequency.
The total possible frequency variation is called the frequency range.
The total possible control voltage variation is called the control voltage range
The ratio of frequency range to control voltage range is called the control sensitivity
and its units are MHz, kHz or Hz per volt.
As far as possible, the source is designed so that the variation of frequency with
control voltage is constant throughout its range; however, this is not always
possible. The parameter describing this is called control linearity
The parameter describing how fast the frequency can be changed is called the
control bandwidth.
Also, as far as possible, the output amplitude should be constant over the whole
frequency range.
The Carrier Source and the Local Oscillator on the 53230 workboard are VCO
sources. They are designed such that if you do not have any control input voltage
they automatically operate at their centre frequency (nominally, 1MHz). Their
frequency range is about 100kHz for a control voltage range of about 4 volts.
VCO
Control
Output
Modulation and Coding Principles Appendices
53230
Voltage controlled oscillators are used extensively in communication systems: in
such subsystems as frequency modulators, phase locked loops and frequency
synthesisers.