Data Communication Fundamentals
Open in the Padhaiko app2.1 Analog and Digital Data
To cross a physical medium, data must be converted into electromagnetic signals. Both data and signals can be either analog or digital, and signals vary in structural complexity, dimensionality, and channel usage.
Data, Signals, and Signal Dimensions
- Analog Data vs. Digital Data: Analog data varies continuously over an interval (e.g., speech pressure, ambient temperature), whereas digital data takes discrete, countable states (e.g., text characters, binary numbers).
- Simple (Basic) vs. Composite Signals: A simple analog signal is a single sine wave that cannot be decomposed into simpler sinusoids. A composite signal is a weighted sum of multiple sine waves of different frequencies, amplitudes, and phases.
- Single-Dimensional vs. Multi-Dimensional Signals: A 1D signal is a function of a single independent variable, typically time, e.g., . A multi-dimensional signal depends on two or more independent variables, e.g., a grayscale image or a video sequence .
- Single-Channel vs. Multi-Channel Transmission: A single-channel signal carries one information stream over one path, whereas multi-channel systems carry multiple independent streams concurrently (via separate wires/frequencies or multiplexing).
Elementary Building-Block Signals
Five standard test signals appear repeatedly in communication analysis (DC Ch2):
| Signal | Notation | Mathematical Definition | Physical Role |
|---|---|---|---|
| Unit Step | Models a signal switched on at | ||
| Unit Impulse (Dirac) | for , | Ideal instantaneous pulse; flat spectrum | |
| Unit Ramp | Linearly growing signal from | ||
| Signum | Polarity indicator () | ||
| Sinc (Sampling) | (with ) | Fourier transform of a rectangular pulse |
2.2 Analog Signals, Periodic and Aperiodic Signals, and Signal Characteristics
An analog signal has infinitely many amplitude values over a continuous range. Signals are classified across five orthogonal axes (DC Ch2).
Continuous-Time vs. Discrete-Time Signals
- A continuous-time (CT) signal is defined at every real instant (e.g., raw acoustic voltage from a microphone).
- A discrete-time (DT) signal is defined only at integer time indices , usually obtained by sampling a CT signal every seconds.
Periodic vs. Aperiodic (Non-Periodic) Signals
Periodic and Aperiodic Signals
A continuous-time signal is periodic if there exists a positive constant such that:
The smallest positive value satisfying this condition is the fundamental period, and is the fundamental frequency. A signal that does not repeat over any finite period is aperiodic (non-periodic).
- Sum of two periodic signals: If where has period and has period , then is periodic if and only if the ratio is a rational number ( in lowest terms). Its fundamental period is .
Time-Domain and Frequency-Domain Characteristics of Periodic Signals
A general sinusoidal analog signal is written in the time domain as:
| Domain | Parameter | Symbol & Formula | Meaning |
|---|---|---|---|
| Time Domain | Peak Amplitude | Maximum displacement from the zero axis | |
| Time Domain | Period & Frequency | , | Duration of one cycle (); cycles per second () |
| Time Domain | Phase | (rad or degrees) | Horizontal shift of the waveform relative to |
| Time Domain | Wavelength | (free space: ) | Spatial distance covered in one full cycle () |
| Frequency Domain | Frequency Spectrum | Plot of and vs. | Shows which discrete harmonic frequencies form the signal |
| Frequency Domain | Bandwidth | Total frequency span occupied by a composite signal () |
Additional Signal Classifications
- Deterministic vs. Non-Deterministic (Random) Signals: A deterministic signal has no uncertainty and can be described exactly by a mathematical formula (e.g., ). A non-deterministic (random) signal exhibits unpredictable values (e.g., thermal noise, unpredicted speech) and is modeled probabilistically.
- Even and Odd Signals:
- Even: (symmetric about the vertical axis, e.g., , ).
- Odd: (antisymmetric about the origin, e.g., , ).
- Every signal decomposes uniquely into even and odd parts:
- Causal, Non-Causal, and Anti-Causal Signals:
- Causal: for all (e.g., ).
- Anti-causal: for all (e.g., ).
- Non-causal: Nonzero for both and (e.g., ).
- Energy and Power Signals:
| Aspect | Energy Signal | Power Signal |
|---|---|---|
| Defining Condition | Finite total energy and zero average power (, ) | Finite non-zero average power and infinite energy (, ) |
| Continuous-Time Formula | ||
| Discrete-Time Formula | ||
| Typical Examples | Aperiodic, time-limited or decaying pulses (, ) | Practical periodic signals () and |
2.3 Fourier Series, Fourier Transform, and Digital Signals
Frequency-domain analysis reveals the spectral components carried by periodic and aperiodic waveforms (DC Ch3).
Fourier Series Representation of Periodic Signals
Any practical periodic signal of fundamental period () satisfying the Dirichlet conditions (absolutely integrable over one period, finite number of maxima/minima per period, and finite number of finite discontinuities per period) admits a Fourier series expansion.
Trigonometric and Exponential Fourier Series
- Trigonometric Fourier Series:
with coefficients:
- If is even, all . If is odd, and all .
- Exponential Fourier Series:
Here is the DC component and are complex spectral coefficients giving the discrete magnitude spectrum and phase spectrum at harmonic frequencies .
- Key Properties of Continuous-Time Fourier Series: Linearity (), Time Shifting (), Time Reversal (), Conjugate Symmetry for real (), and Parseval's Power Theorem:
Fourier Transform Representation of Aperiodic Signals
As the period , discrete harmonics merge into a continuous spectrum, yielding the Fourier Transform (FT) for aperiodic signals.
Continuous-Time Fourier Transform Pair
For an aperiodic signal satisfying the Dirichlet conditions (absolutely integrable: ):
The spectrum is represented continuously by magnitude and phase .
| Time Signal | Fourier Transform | Key Property / Note |
|---|---|---|
| Flat spectrum across all frequencies | ||
| (constant DC) | Impulse at (derived via IFT) | |
| Causal decaying exponential | ||
| Two-sided decaying exponential (real, even FT) | ||
| Evaluated via | ||
| Derived from | ||
| Time Shift: | Changes phase only; magnitude unaltered | |
| Time Scaling: | Compression in time expands bandwidth | |
| Convolution: | Time-domain convolution becomes frequency multiplication |
Digital Signals and Their Characteristics
A digital signal has a finite number of discrete amplitude levels (for example, and for binary and ).
- Bit Rate (): Number of bits transmitted per second ().
- Bit Interval (): Duration of one bit, seconds.
- Bit Length (): Physical distance one bit occupies on the transmission medium, .
- Spectral Nature: By Fourier analysis, an ideal rectangular pulse train is a composite analog signal containing an infinite series of odd harmonics. Transmitting a digital signal faithfully without rounding off sharp edges requires a wide baseband bandwidth (a low-pass channel starting at ). When only a bandpass channel is available, the digital signal must first modulate a carrier (broadband transmission).
2.4 Analog and Digital Transmission, Transmission Modes, and Impairments
Analog vs. Digital Transmission and Transmission Modes
- Analog transmission sends analog signals without regard to their underlying content (analog or digital data). Signals weaken over distance and require amplifiers, which amplify both the desired signal and any accumulated noise.
- Digital transmission relies on the discrete content of the signal. Regenerative repeaters sample the incoming waveform, decide the original bit values, and transmit a fresh, noise-free pulse stream—preventing noise accumulation.
When transferring bits between devices, transmission mode describes both bit grouping and timing (pages="added"):
| Mode | Mechanism | Advantages & Trade-offs |
|---|---|---|
| Parallel | Sends bits simultaneously over separate wires | High speed over short distances (CPU buses); suffers from skew and cost over long runs |
| Serial — Asynchronous | Sends bit at a time framed by start () and stop () bits per byte | No shared clock wire needed; framing overhead (e.g., UART, RS-232) |
| Serial — Synchronous | Sends continuous bit stream in large frames with embedded clock/preamble | High efficiency, low overhead; requires clock recovery (e.g., Ethernet, HDLC) |
| Serial — Isochronous | Guarantees fixed, bounded timing between frames across the stream | Essential for real-time audio/video where jitter cannot be tolerated |
Transmission Impairments
Physical channels are imperfect. Three primary impairments—attenuation, distortion, and noise—degrade transmitted signals (DC Ch1).
Attenuation, Distortion, and Noise
- Attenuation: Loss of signal energy as it propagates through a medium (due to ohmic resistance in copper or absorption/scattering in fiber). Measured in decibels ():
Negative denotes power loss ( power); positive denotes amplification ( power). Mitigation: Amplifiers (analog) or repeaters (digital).
- Distortion: Alteration of a composite signal's shape because different frequency components experience unequal attenuation (amplitude distortion) or unequal propagation speeds (delay / phase distortion), causing symbols to smear into one another (inter-symbol interference, ISI). Mitigation: Channel equalizers.
- Noise: Unwanted random or extraneous electrical energy that corrupts the desired signal:
- Thermal (White / Johnson–Nyquist) Noise: Random thermal agitation of electrons in conductors, uniform across frequencies. Noise power and RMS noise voltage are:
where is Boltzmann's constant, is absolute temperature (), is bandwidth (), and is resistance ().
- Intermodulation Noise: Nonlinearity in a medium mixes signals at to produce spurious sums/differences (, ).
- Crosstalk: Electromagnetic coupling between adjacent wire pairs or channels.
- Impulse Noise: Short, high-amplitude spikes from lightning, motor switching, or power transients (the chief source of burst errors in digital links).
- Multipath Interference: Reflected waves arriving with different delays cause fading and ISI in wireless links.
The quality of a channel against noise is quantified by the Signal-to-Noise Ratio ():
2.5 Data Rate Limits: Channel Capacity, Nyquist Bandwidth, and Shannon Capacity
The maximum bit rate achievable over a channel depends on three parameters (DC Ch4): available bandwidth (), number of discrete signal levels (), and channel noise level ().
Nyquist Bit Rate and Shannon–Hartley Capacity
- Noiseless Channel — Nyquist Bit Rate:
where is the channel bandwidth in and is the number of discrete voltage/signal levels. Increasing raises bit rate, but packing more levels into a fixed voltage range reduces receiver noise margin and reliability.
- Noisy Channel — Shannon–Hartley Capacity Theorem (1944):
where is the linear (dimensionless) signal-to-noise power ratio. No coding or modulation scheme can achieve error-free transmission faster than .
- High-SNR Practical Approximation (valid when ):
- Combining Both Limits in Design: Shannon's formula yields the absolute upper bound imposed by noise and bandwidth. The engineer selects a practical target rate and uses Nyquist's formula () to determine the required number of signal levels .
2.6 Performance of Network
Network and channel performance is evaluated using six interrelated metrics (DC Ch4 and micro-syllabus):
| Metric | Definition & Formula | Key Distinction / Significance |
|---|---|---|
| Bandwidth | In Hz: frequency range passed by a channel. In bps: theoretical maximum bit rate of a link. | Potential capacity ceiling of the link. |
| Throughput | Actual measured rate; always link bandwidth due to congestion and end-device limits. | |
| Latency (Delay) | Total time from the first bit leaving the sender to the last bit arriving at the receiver. | |
| Propagation Time | Time for one bit to cross the medium ( in vacuum; in cable). | |
| Transmission Time | Time between the first bit and last bit entering the link; dominates for large files on slow links. | |
| Queuing & Processing | Time waiting in router buffers () + header inspection () | Variable component that grows steeply under network congestion. |
| Bandwidth–Delay Product | Number of bits that fill the physical "pipe" in transit ( for full-duplex window sizing). | |
| Jitter | Variation in end-to-end packet arrival delay | Causes audio/video glitches in real-time streams; smoothed via jitter buffers. |
| Bit Rate vs. Baud Rate | Bit rate () is bits/s; Baud rate () is signal elements (symbols)/s. | |
| Bit Error Rate (BER) | Measures channel reliability; improves as increases. |
Worked Examples
1
- Determine the energy and average power of , and classify it as an energy or power signal.
- Test whether is periodic; if so, find its fundamental period .
- Find the even and odd parts of for .
1. Energy and power of : Since for :
Because and , is an energy signal.
2. Periodicity test for : Comparing with gives and :
The ratio is a rational number. Therefore, is periodic with fundamental period:
3. Even and odd parts of on : Replacing with gives :
- , Energy signal.
- Periodic with fundamental period .
- and for .
2
Find the exponential Fourier series coefficients for the periodic signal
and determine the magnitude and phase spectra for all nonzero harmonics.
Expand using Euler's formulas:
Substituting these into :
Identify the spectral coefficients : Matching against :
- ,
- ,
- for all .
Compute magnitudes and phases :
- ,
- ; ,
- ; ,
, (, , ), and (, ).
3
- A signal travels from Point 1 to Point 3 through a noisy line and an intermediate amplifier at Point 2. If and , find the overall gain or loss of the system in .
- Find the RMS thermal noise voltage across an resistor at an ambient temperature of over the frequency band to ().
1. Overall gain in decibels:
Adding decibel gains across cascaded stages:
2. RMS thermal noise voltage: Convert temperature to Kelvin and find the passband width :
Substitute into :
- Overall system gain = .
- RMS thermal noise voltage .
4
- A channel has a bandwidth of and an of . Find its Shannon capacity. If we transmit at a safer rate of , how many signal levels are needed by Nyquist's formula?
- Compute the propagation time and transmission time for a image sent at over a cable where the propagation speed is .
1. Shannon capacity and required signal levels:
For using :
2. Propagation time vs. transmission time:
Because the file is large and the link rate is modest (), transmission time dominates over propagation time.
- Shannon capacity = ; requires signal levels.
- , .
Exercises
Exercises
Define energy signal and power signal. Determine the energy and power of () and classify it.
Definitions: An energy signal has finite non-zero total energy () and zero average power (). A power signal has finite non-zero average power () and infinite energy ().
Evaluate and for ():
Since and , is an energy signal.
, ; therefore () is an energy signal.
Test whether is periodic, and if so, calculate its fundamental frequency and fundamental period . Also check whether is periodic.
Periodicity of : Comparing with gives . Thus:
So is periodic with .
Periodicity of piecewise : Although (for ) and (for ) each repeat every , switches definition once at , creating a jump discontinuity from (left limit) to (right value) that never recurs at for . Hence across , so is aperiodic (non-periodic).
is periodic with and ; is aperiodic.
Find the Fourier transform of (), and use the inverse Fourier transform to show that .
Fourier transform of :
Fourier transform of unity : Since is not absolutely integrable, apply the Inverse Fourier Transform with candidate :
Hence .
and .
Describe the primary transmission impairments (attenuation, distortion, and noise). Calculate the thermal noise power density in at ().
Primary transmission impairments:
- Attenuation: Loss of signal power along a medium; expressed in and compensated by amplifiers or repeaters.
- Distortion: Waveform deformation in composite signals caused by unequal attenuation or unequal propagation speeds across frequencies (delay distortion), mitigated by equalizers.
- Noise: Unwanted energy corrupting the signal, including thermal (white) noise, intermodulation noise, crosstalk, and impulse noise.
Thermal noise power density at :
The three impairments are attenuation, distortion, and noise; at , the thermal noise power density is .
- State Nyquist's bit rate and Shannon's capacity formulas. Find the Shannon capacity of a channel with , and comment on the result.
- A channel requires a capacity of . What minimum linear (and ) is needed?
1. Capacity at and : Nyquist (noiseless): . Shannon (noisy): . For :
When signal power equals noise power (), the maximum theoretical bit rate equals the channel bandwidth ().
2. Required SNR for over :
- ( when ).
- Required ().
- A network passes an average of frames per minute, each carrying bits. Calculate its throughput.
- Compute the propagation time and transmission time for a email sent at over (). Also define jitter.
1. Throughput calculation:
2. Propagation time, transmission time, and jitter:
Here dominates completely because the message is short and the link speed is . Jitter is the variation in end-to-end arrival delay across consecutive packets, which degrades time-sensitive audio and video playback.
- Throughput = .
- , ; jitter is packet-to-packet delay variation.
- Distinguish between bit rate and baud rate. If a digital link sends using -level signaling (), find its baud rate.
- Out of transmitted bits, bits arrive inverted due to impulse noise. Calculate the Bit Error Rate (BER). Also distinguish between asynchronous and synchronous serial transmission.
1. Bit rate vs. baud rate: Bit rate () is the number of data bits per second, whereas baud rate () is the number of signal elements (symbols) per second: . For , each symbol carries bits:
2. BER and asynchronous vs. synchronous transmission:
Asynchronous transmission frames each byte individually with start/stop bits so clocks resynchronize per character (simple, higher overhead); synchronous transmission sends long frames continuously with a shared or recovered clock (low overhead, high efficiency).
- Baud rate = .
- ; asynchronous uses start/stop bits per byte whereas synchronous sends continuous clocked frames.