Digital sampling of dynamic signals

Mallinath Ogirala

SE22UMEE014

Objective:
The objective of this lab is to convert an MP3 signal into an analog signal, then analyze its
frequency, amplitude, and waveform pattern. Using a personal DAQ (digital data acquisition)
device, we will sample signals generated by waveforms from MP3 files after they are converted
to analog electrical signals. Additionally, we will examine factors that affect changes in these
analog signals, such as noise, the Nyquist limit, and aliasing.

Materials and Apparatus:

I. Oscilloscope

II. laptop

III. Probes, audio jack and BNC cables

IV. DAQ module connection

V. Notepad and pen

Introduction: This lab will focus on learning how to use a personal DAQ (digital data
acquisition) device to sample signals generated by waveforms stored in MP3 format and
converted into analog electrical signals. We will also explore concepts such as samples, sample
rate, sample period, amplitude, and frequency. Additionally, we aim to understand Nyquist
frequency (𝑓𝑛) and aliasing (𝑓).

Nyquist Frequency (𝒇𝒏):

Frequency refers to the number of times a repeating event occurs per unit time. The Nyquist
frequency is defined as half the sampling frequency, or 𝒇𝒏 = fs/2. If the waveform contains no
components above this frequency, it can be fully reconstructed from the sampled data, provided
there are no measurement errors. This principle is known as the Nyquist Sampling Theorem.
Simply put, the Nyquist frequency is half the sampling frequency.

Aliasing (𝒇):

Aliasing occurs when a signal is sampled at a rate that is too low to capture its variations
accurately, causing distinct signals to become indistinguishable or misrepresented in the sampled
data. This effect appears when the waveform frequency differs from the Nyquist frequency (𝒇 >
𝒇𝒏). To mitigate aliasing, techniques such as low-pass filters are used, which only allow
frequencies below a certain cut-off.

Sampling/DAQ Terminology:

I. Sample: A single measurement taken by the DAQ from one channel at a specific instant.

II. Sampling Period: The time interval between two consecutive samples.

III. Sampling Rate: The inverse of the sampling period, with units of Samples per second (S/s) or
Hz.

IV. Record: A sequence of continuous samples collected by the DAQ.

V. Record Length: The total number of samples in a record, usually measured in Samples (S),
with **Record Duration** representing the time from the first to the last sample in that record

Results and discussions:

From the data collected in the experiment, we observed multiple frequencies with various
readings for the given track; however, we have only recorded the maximum frequency.

Track 1:

Wave form sine

frequency 100
Time period(s) 0.01

The data above pertains to Task 1, with a sample rate of 20,000 and a total of 10,000 samples.
The frequency is 100 Hz, and the peak-to-peak time is 0.01 seconds. We observed a sine wave in
the power spectrum graph, which shows frequency modulation at 100 Hz.

Track 2:

Wave form sine

frequency 1000
Time period(s) 0.001
The data above relates to Task 2, with a sample rate of 20,000 and 10,000 samples in total. The
frequency is 1000 Hz, and the peak-to-peak time is 0.001 seconds. We observed a sine wave in
the power spectrum graph, which indicates frequency modulation at 1000 Hz.

Track 3:

Wave form sine

frequency 8000
Time period(s) 0.00125

The data above pertains to Task 3, with a sample rate of 20,000 and 10,000 samples in total. The
frequency is 8000 Hz, and the peak-to-peak time is 0.00125 seconds. In the power spectrum
graph, we observed a sine wave showing frequency modulation at 8000 Hz.

Track 4:

Wave form triangular

frequency 1000
Time period(s) 0.001

The data above is for Task 4, with a sample rate of 20,000 and a total of 10,000 samples. In the
power spectrum, we observed frequency changes at 1000 Hz, 3000 Hz, 5000 Hz, 7000 Hz, and
9000 Hz, with a peak-to-peak time of 0.001 seconds. We observed a triangular wave in the
power spectrum graph, which shows frequency modulation at 1000 Hz.

Track 5:

Wave form square

frequency 1000
Time period(s) 0.001

The data above pertains to Task 5, with a sample rate of 20,000 and a total of 10,000 samples. In
the power spectrum, frequency changes are observed at 1000 Hz, 3000 Hz, 5000 Hz, 7000 Hz,
and 9000 Hz, with a peak-to-peak time of 0.001 seconds. A square wave is visible in the power
spectrum graph, and although various frequencies were observed during the experiment,
frequency modulation occurs at 1000 Hz.

Track 6:
 Wave form ramp
frequency 1000
Time period(s) 0.001

The provided data for task 6 consists of 10,000 samples collected at a 20,000 Hz sample rate.
The power spectrum analysis reveals frequency modulation primarily centered around 1000 Hz,
although other frequencies are also present. The peak-to-peak time plot indicates a ramp
waveform with a duration of 0.001 seconds.

Comparision between the waves:

Square waves, with their sharp edges, demand more high-frequency components to accurately
represent their shape compared to triangle waves and ramp waves. Triangle waves, having
smoother, linear transitions, require fewer high-frequency components. Ramp waves, with their
gradual, unidirectional slope, need the least amount of high-frequency components. This
difference in high-frequency content is directly linked to the smoothness of the waveform's
transitions.

Track 7:

Wave form Noisy sine

frequency 1000
Time period(s) 0.001

The data for task 7 comprises 10,000 samples collected at a 20,000 Hz sample rate. The signal is
a noisy sine wave with a frequency of 1000 Hz and a peak-to-peak time duration of 0.001
seconds. The power spectrum analysis confirms the presence of frequency modulation centered
around 1000 Hz.

Track 8:

Wave form Product of sine

frequency 100,1000,8000
Time period(s) 0.00002

The data for task 8 consists of 10,000 samples collected at a 20,000 Hz sample rate. The signal is
a product of sine waves with frequencies of 1000 Hz, 100 Hz, and 8000 Hz. The power spectrum
analysis reveals frequency modulation at multiple points, indicating the presence of various
frequency components within the signal.
Track 9:

Wave form Amplitude modulated

frequency 1900,2100
Time period(s) 0.0005

The data for task 9 involves 10,000 samples collected at a 20,000 Hz sample rate. The signal is
an amplitude-modulated wave with carrier frequencies of 1900 Hz and 2100 Hz. The peak-to-
peak time plot was not accurately represented. The power spectrum analysis shows frequency
modulation at these two frequencies, but due to aliasing effects, the individual frequencies were
not clearly distinguishable.

The Nyquist rate, a fundamental principle in signal processing, dictates the minimum sampling
rate required to accurately capture a signal without distortion. Sampling a signal below this rate
can lead to aliasing, where high-frequency components of the signal are misinterpreted as lower
frequencies. This was evident in our experiments, where lower sampling rates resulted in
distorted waveforms and additional peaks in the power spectrum.

We also observed variations in high-frequency content among different waveforms. Square

waves, with their sharp transitions, exhibit significant high-frequency power, while triangle and
ramp waves, with smoother transitions, have lower high-frequency content. This highlights the
importance of adequate sampling rates, especially for signals with rapid changes, to accurately
represent their frequency components.

In this practical experiment, we applied theoretical concepts to real-world scenarios by

converting an MP3 audio signal into an analog waveform. We explored how factors like
frequency, waveform shape, and sampling rate influence the representation of signals. Using a
DAQ system, we analyzed the sampled data to measure amplitude, frequency, and waveform
shapes, as well as to identify common challenges in analog-to-digital conversion, such as noise,
the Nyquist limit, and aliasing artifacts.

Conclusion:

Understanding and appropriately setting DAQ parameters, such as sampling rate, sample period,
and record length, is crucial in signal analysis. Proper sampling ensures accurate waveform
representation, minimizes aliasing, and provides clear insights into the frequency content of the
signal. This is particularly important in fields like audio and telecommunications, where precise
signal analysis is essential.