The document describes an experiment demonstrating pulse code modulation (PCM) using an analog-to-digital converter (ADC) and digital-to-analog converter (DAC). The experiment showed how the ADC sampling rate must be at least twice the analog signal frequency to avoid aliasing. It also showed that a low-pass filter can smooth the DAC's staircase output into a representation of the original analog signal. The conclusions were that PCM can digitize analog signals for digital communication, with ADC and DAC performing the encoding and decoding, and that the filter output retains the analog input frequency regardless of the sampling rate.
1. NATIONAL COLLEGE OF SCIENCE & TECHNOLOGY
Amafel Bldg. Aguinaldo Highway Dasmariñas City, Cavite
EXPERIMENT 2
DIGITAL COMMUNICATION OF ANALOG DATA USING
PULSE-CODE MODULATION (PCM)
Tagasa, Jerald A. September 20, 2011
Signal Spectra and Signal Processing/BSECE 41A1 Score:
Engr. Grace Ramones
Instructor
2. Objectives:
Demonstrate PCM encoding using an analog-to-digital converter
(ADC).
Demonstrate PCM encoding using an digital-to-analog converter
(DAC)
Demonstrate how the ADC sampling rate is related to the
analog signal frequency.
Demonstrate the effect of low-pass filtering on the decoder
(DAC) output.
4. Data Sheet:
Materials
One ac signal generator
One pulse generator
One dual-trace oscilloscope
One dc power supply
One ADC0801 A/D converter (ADC)
One DAC0808 (1401) D/A converter (DAC)
Two SPDT switches
One 100 nF capacitor
Resistors: 100 Ω, 10 kΩ
Theory
Electronic communications is the transmission and reception of
information over a communications channel using electronic
circuits. Information is defined as knowledge or intelligence
such as audio voice or music, video, or digital data. Often the
information id unsuitable for transmission in its original form
and must be converted to a form that is suitable for the
communications system. When the communications system is
digital, analog signals must be converted into digital form
prior to transmission.
The most widely used technique for digitizing is the analog
information signals for transmission on a digital communications
system is pulse-code modulation (PCM), which we will be studied
in this experiment. Pulse-code modulation (PCM) consists of the
conversion of a series of sampled analog voltage levels into a
sequence of binary codes, with each binary number that is
proportional to the magnitude of the voltage level sampled.
Translating analog voltages into binary codes is called A/D
conversion, digitizing, or encoding. The device used to perform
this conversion process called an A/D converter, or ADC.
An ADC requires a conversion time, in which is the time required
to convert each analog voltage into its binary code. During the
ADC conversion time, the analog input voltage must remain
constant. The conversion time for most modern A/D converters is
short enough so that the analog input voltage will not change
during the conversion time. For high-frequency information
signals, the analog voltage will change during the conversion
time, introducing an error called an aperture error. In this
5. case a sample and hold amplifier (S/H amplifier) will be
required at the input of the ADC. The S/H amplifier accepts the
input and passes it through to the ADC input unchanged during
the sample mode. During the hold mode, the sampled analog
voltage is stored at the instant of sampling, making the output
of the S/H amplifier a fixed dc voltage level. Therefore, the
ADC input will be a fixed dc voltage during the ADC conversion
time.
The rate at which the analog input voltage is sampled is called
the sampling rate. The ADC conversion time puts a limit on the
sampling rate because the next sample cannot be read until the
previous conversion time is complete. The sampling rate is
important because it determines the highest analog signal
frequency that can be sampled. In order to retain the high-
frequency information in the analog signal acting sampled, a
sufficient number of samples must be taken so that all of the
voltage changes in the waveform are adequately represented.
Because a modern ADC has a very short conversion time, a high
sampling rate is possible resulting in better reproduction of
high0frequency analog signals. Nyquist frequency is equal to
twice the highest analog signal frequency component. Although
theoretically analog signal can be sampled at the Nyquist
frequency, in practice the sampling rate is usually higher,
depending on the application and other factors such as channel
bandwidth and cost limitations.
In a PCM system, the binary codes generated by the ADC are
converted into serial pulses and transmitted over the
communications medium, or channel, to the PCM receiver one bit
at a time. At the receiver, the serial pulses are converted back
to the original sequence of parallel binary codes. This sequence
of binary codes is reconverted into a series of analog voltage
levels in a D/A converter (DAC), often called a decoder. In a
properly designed system, these analog voltage levels should be
close to the analog voltage levels sampled at the transmitter.
Because the sequence of binary codes applied to the DAC input
represent a series of dc voltage levels, the output of the DAC
has a staircase (step) characteristic. Therefore, the resulting
DAC output voltage waveshape is only an approximation to the
original analog voltage waveshape at the transmitter. These
steps can be smoothed out into an analog voltage variation by
6. passing the DAC output through a low-pass filter with a cutoff
frequency that is higher than the highest-frequency component in
the analog information signal. The low-pass filter changes the
steps into a smooth curve by eliminating many of the harmonic
frequency. If the sampling rate at the transmitter is high
enough, the low-pass filter output should be a good
representation of the original analog signal.
In this experiment, pulse code modulation (encoding) and
demodulation (decoding) will be demonstrated using an 8-bit ADC
feeding an 8-bit DAC, as shown in Figure 2-1. This ADC will
convert each of the sampled analog voltages into 8-bit binary
code as that represent binary numbers proportional to the
magnitude of the sampled analog voltages. The sampling frequency
generator, connected to the start-of conversion (SOC) terminal
on the ADC, will start conversion at the beginning of each
sampling pulse. Therefore, the frequency of the sampling
frequency generator will determine the sampling frequency
(sampling rate) of the ADC. The 5 volts connected to the VREF+
terminal of the ADC sets the voltage range to 0-5 V. The 5 volts
connected to the output (OE) terminal on the ADC will keep the
digital output connected to the digital bus. The DAC will
convert these digital codes back to the sampled analog voltage
levels. This will result in a staircase output, which will
follow the original analog voltage variations. The staircase
output of the DAC feeds of a low-pass filter, which will produce
a smooth output curve that should be a close approximation to
the original analog input curve. The 5 volts connected to the +
terminal of the DAC sets the voltage range 0-5 V. The values of
resistor R and capacitor C determine the cutoff frequency (fC)
of the low-pass filter, which is determined from the equation
Figure 23–1 Pulse-Code Modulation (PCM)
7. XSC2
G
T
A B C D
S1 VCC
Key = A 5V
U1
Vin D0
S2
D1
V2 D2
D3 Key = B
2 Vpk D4
10kHz
D5
0° Vref+
D6
Vref-
D7
SOC VCC
OE EOC 5V
D0
D1
D2
D3
D4
D5
D6
D7
ADC
V1 Vref+ R1
VDAC8 Output
5V -0V Vref- 100Ω
200kHz
U2
R2
10kΩ C1
100nF
In an actual PCM system, the ADC output would be transmitted to
serial format over a transmission line to the receiver and
converted back to parallel format before being applied to the
DAC input. In Figure 23-1, the ADC output is connected to the
DAC input by the digital bus for demonstration purposes only.
PROCEDURE:
Step 1 Open circuit file FIG 23-1. Bring down the
oscilloscope enlargement. Make sure that the
following settings are selected. Time base (Scale =
20 µs/Div, Xpos = 0 Y/T), Ch A(Scale 2 V/Div, Ypos
= 0, DC) Ch B (Scale = 2 V/Div, Ypos = 0, DC),
Trigger (Pos edge, Level = 0, Auto). Run the
simulation to completion. (Wait for the simulation
to begin). You have plotted the analog input signal
(red) and the DAC output (blue) on the
oscilloscope. Measure the time between samples (TS)
on the DAC output curve plot.
TS = 4 µs
Step 2 Calculate the sampling frequency (fS) based on the
time between samples (TS)
fS = 250 kHz
Question: How did the measure sampling frequency compare with
the frequency of the sampling frequency generator?
Both frequency have difference of 50 kHz.
8. How did the sampling frequency compare with the analog input
frequency? Was it more than twice the analog input frequency?
The sampling frequency is 20 times higher. It is
more than twice the analog input frequency.
How did the sampling frequency compare with the Nyquist
frequency?
The Nyquist frequency is higher. Nyquist is 6.28
times more than the sampling frequency.
Step 3 Click the arrow in the circuit window and press the A
key to change Switch A to the sampling generator output.
Change the oscilloscope time base to 10 µs/Div. Run the
simulation for one oscilloscope screen display, and then
pause the simulation. You are plotting the sampling
generator (red) and the DAC output (blue).
Question: What is the relationship between the sampling
generator output and the DAC staircase output?
Both outputs are both in digital
Step 4 Change the oscilloscope time base scale to 20 µs/Div.
Click the arrow in the circuit window and press the A
key to change Switch A to the analog input. Press the B
key to change the Switch B to Filter Output. Bring down
the oscilloscope enlargement and run the simulation to
completion. You are plotting the analog input (red) and
the low-pass filter output (blue) on the oscilloscope
Questions: What happened to the DAC output after filtering? Is
the filter output waveshape a close representation of the analog
input waveshape?
The output became analog after filtering. Yes it is
close representation.
Step 5 Calculate the cutoff frequency (fC) of the low-pass
filter.
fC = 15.915 kHz
Question: How does the filter cutoff frequency compare with the
analog input frequency?
They have difference of approximately 6 kHz.
Step 6 Change the filter capacitor (C) to 20 nF and calculate
the new cutoff frequency (fC).
fC = 79.577 kHz
Step 7 Bring down the oscilloscope enlargement and run the
simulation to completion again.
Question: How did the new filter output compare with the
previous filter output? Explain.
9. It is almost the same.
Step 8 Change the filter capacitor (C) back to 100 nF. Change
the Switch B back to the DAC output. Change the
frequency of the sampling frequency generator to 100
kHz. Bring down the oscilloscope enlargement and run the
simulation to completion. You are plotting the analog
input (red) and the DAC output (blue) on the
oscilloscope screen. Measure the time between the
samples (TS) on the DAC output curve plot (blue)
TS = 9.5µs
Question: How does the time between the samples in Step 8
compare with the time between the samples in Step 1?
The time between the samples in Step 8 doubles.
Step 9 Calculate the new sampling frequency (fS) based on the
time between the samples (TS) in Step 8?
fS=105.26Hz
Question: How does the new sampling frequency compare with the
analog input frequency?
It is 10 times the analog input frequency.
Step 10 Click the arrow in the circuit window and change
the Switch B to the filter output. Bring down the
oscilloscope enlargement and run the simulation again.
Question: How does the curve plot in Step 10 compare with the
curve plot in Step 4 at the higher sampling frequency?
Is the curve as smooth as in Step 4? Explain why.
Yes, they are the same. It is as smooth as in Step 4.
Nothing changed. It does not affect the filter.
Step 11 Change the frequency of the sampling frequency generator
to 50 kHz and change Switch B back to the DAC output.
Bring down the oscilloscope enlargement and run the
simulation to completion. Measure the time between
samples (TS) on the DAC output curve plot (blue).
TS = 19µs
Question: How does the time between samples in Step 11 compare
with the time between the samples in Step 8?
It doubles.
Step 12 Calculate the new sampling frequency (fS) based on the
time between samples (TS) in Step 11.
fS=52.631 kHz
Question: How does the new sampling frequency compare with the
analog input frequency?
The new sampling frequency is 5 times the analog input.
10. Step 13 Click the arrow in the circuit window and change the
Switch B to the filter output. Bring down the
oscilloscope enlargement and run the simulation to
completion again.
Question: How does the curve plot in Step 13 compare with the
curve plot in Step 10 at the higher sampling frequency?
Is the curve as smooth as in Step 10? Explain why.
Yes, nothing changed. The frequency of the sampling
generator does not affect the filter.
Step 14 Calculate the frequency of the filter output (f) based
on the period for one cycle (T).
T=10kHz
Question: How does the frequency of the filter output compare
with the frequency of the analog input? Was this
expected based on the sampling frequency? Explain why.
It is the same. Yes, it is expected.
Step 15 Change the frequency of the sampling frequency generator
to 15 kHz and change Switch B back to the DAC output.
Bring down the oscilloscope enlargement and run the
simulation to completion. Measure the time between
samples (TS) on the DAC output curve plot (blue)
TS = 66.5µs
Question: How does the time between samples in Step 15 compare
with the time between samples in Step 11?
It is 3.5 times higher than the time in Step 11.
Step 16 Calculate the new sampling frequency (fS) based on the
time between samples (TS) in Step 15.
fS=15.037 kHz
Question: How does the new sampling frequency compare with the
analog input frequency?
It is 5 kHz greater than the analog input frequency.
How does the new sampling frequency compare with the Nyquist
frequency?
It is 6.28 times smaller than the Nyquist frequency.
Step 17 Click the arrow in the circuit window and change
the Switch B to the filter output. Bring down the
oscilloscope enlargement and run the simulation to
completion again.
Question: How does the curve plot in Step 17 compare with the
curve plot in Step 13 at the higher sampling frequency?
They are the same.
11. Step 18 Calculate the frequency of the filter output (f) based
on the time period for one cycle (T).
f=10kHz
Question: How does the frequency of the filter output compare
with the frequency of the analog input? Was this
expected based on the sampling frequency?
They are the same. For sampling frequency of 15.037 kHz,
it is expected to have same outputs.
12. CONCLUSION:
I conclude, that analog signal can be digitize for digital
communication. One way is the PCM. ADC and DAC are used for
encoding and decoding of PCM.
The ADC provides the sampling frequency. The sampling
frequency is inversely proportional to the sampling time of the
DAC output. The staircase output is the output generated by the
DAC. It is digital signal like the sampling pulse.
The filter frequency is the frequency of the analog input
frequency. The cutoff frequency is inversely proportional to the
capacitance and remain constant as the sampling frequency changes.