Signals and Transmission Fundamentals

Introduction

Data communication depends on the successful representation and transfer of information. Text, audio, images, sensor readings, and computer files are ultimately conveyed through physical or wireless media as signals. A signal is a time-varying physical quantity—commonly voltage, current, light intensity, or electromagnetic radiation—that represents information.

Understanding signals requires more than distinguishing analog from digital. Engineers must analyze amplitude, frequency, phase, spectrum, bandwidth, bit rate, attenuation, distortion, noise, delay, and jitter. These properties determine whether a communication system can deliver information accurately and at an acceptable rate.

Two fundamental limits guide transmission design. Nyquist’s noiseless-channel result relates bandwidth to the maximum signaling rate, while Shannon’s noisy-channel theorem relates bandwidth and signal-to-noise ratio to the theoretical maximum data rate. These are capacity limits, not guaranteed performance levels.

Learning Objectives

After completing this chapter, students should be able to:

• distinguish analog and digital signals;
• explain amplitude, period, frequency, phase, wavelength, spectrum, and bandwidth;
• relate bit rate, symbol rate, and signal levels;
• analyze attenuation, distortion, noise, delay, and jitter;
• apply Nyquist and Shannon capacity equations;
• interpret transmission impairments in practical systems;
• evaluate design trade-offs involving bandwidth, power, reliability, and data rate.

1. Signals as Information-Bearing Quantities

A signal is a function of time, often written as s(t). Its instantaneous value indicates the physical state of the transmission medium. For example, a copper cable may carry a changing voltage, an optical fiber may carry changing light intensity, and a radio system may radiate a changing electromagnetic field.

Signals may be classified in several ways:

• Analog signal: varies continuously in time and amplitude within a range.
• Digital signal: takes discrete, distinguishable values or levels. Time may also be discrete when samples are used.
• Periodic signal: repeats after a fixed interval.
• Aperiodic signal: does not repeat exactly.
• Baseband signal: occupies frequencies beginning near zero and is transmitted without shifting the entire spectrum to another carrier band.
• Passband signal: is centered around a carrier frequency and is commonly used in radio, cellular, satellite, and broadband systems.

The analog/digital distinction describes the signal representation; it does not by itself identify the medium or modulation method. Digital information can be transmitted using an analog passband waveform, and analog information can be sampled and transmitted digitally.

2. Sinusoidal Signals and Their Parameters

A fundamental periodic signal is the sinusoid:

Here, A is peak amplitude, measured in volts, amperes, or another physical unit; f is frequency in hertz (cycles per second); t is time in seconds; and φ is phase angle in radians. The expression assumes a sinusoidal waveform with constant amplitude and frequency.

Amplitude

Amplitude is the magnitude of a signal relative to a reference level. It may be specified as peak amplitude, peak-to-peak amplitude, or root-mean-square (RMS) amplitude. For a sinusoid with peak amplitude A,

This relationship applies only to a sinusoidal waveform with zero average value. RMS amplitude is useful because it relates to average power in resistive systems.

Period and Frequency

The period T is the time required for one cycle. Frequency is the number of cycles per second:

Frequency is measured in hertz. A signal with a period of 2 microseconds has frequency 500 kHz.

Phase

Phase describes the horizontal displacement of a sinusoid relative to a reference. Two signals at the same frequency can differ in phase even when their amplitudes are equal. Phase differences are central to phase modulation, coherent detection, multipath analysis, and synchronization.

Wavelength

For a wave traveling at velocity v, wavelength is:

Here, λ is wavelength in meters. In vacuum, electromagnetic waves travel approximately at 3 × 10⁸ m/s; in cables and optical fibers, the propagation velocity is lower. The formula assumes a uniform propagation velocity at the frequency being considered.

3. Spectrum and Bandwidth

A waveform can be represented in the time domain or frequency domain. The time-domain view shows how the signal changes over time. The frequency-domain view shows the sinusoidal components required to construct the signal. The set of frequency components is called the signal spectrum.

A pure sinusoid has one ideal frequency component. A pulse or speech waveform contains a range of frequencies. The absolute bandwidth is the difference between the highest and lowest significant frequencies:

Bandwidth B is measured in hertz. “Significant” depends on the application because real signals often have components that decline gradually rather than stopping abruptly.

A channel also has a bandwidth, determined by physical properties, filters, regulations, and engineering design. Reliable transmission requires the signal spectrum to fit sufficiently within the channel bandwidth. Filtering a signal reduces unwanted frequencies but may also remove information or create waveform distortion.

Fourier Perspective

Fourier analysis states, broadly, that suitable signals can be represented as sums of sinusoids. A square wave, for example, requires a fundamental frequency and odd harmonics. An ideal square wave has infinitely many harmonics, but practical systems transmit only a finite number. The fewer harmonics preserved, the more rounded the received edges become.

This explains why a digital waveform is not “purely digital” in the physical medium. A sequence of voltage levels has a spectrum and is subject to bandwidth limitation, ringing, and intersymbol interference.

Figure: A finite-duration pulse contains a range of frequency components rather than one frequency.

4. Digital Signals, Bits, Symbols, and Baud

A bit is a binary information unit, usually represented as 0 or 1. A symbol is one signal state transmitted during a signaling interval. If a modulation scheme uses M distinct signal states, each symbol can represent:

If the symbol rate is S symbols per second, the ideal uncoded bit rate is:

where R_b is bit rate in bits per second. This assumes all signal states are valid, symbols are equally useful, and coding overhead is excluded. In practice, framing, error-control coding, synchronization, and protocol overhead reduce useful throughput.

Baud means symbols per second, not bits per second. With binary signaling, one symbol may carry one bit, so baud and bit/s have the same numerical value. With multilevel signaling, one symbol can carry several bits and the bit rate exceeds the baud rate.

Analog and Digital Representation

Analog information, such as speech, changes continuously. Digital systems commonly sample it at regular intervals, quantize each sample, and encode the quantization result. Sampling and quantization introduce requirements and errors. A continuous-time analog waveform is not automatically recovered perfectly merely because it is called digital.

For a low-pass signal whose highest frequency is f_max, ideal sampling requires:

This is the Nyquist sampling condition for band-limited signals under ideal assumptions. Sampling below this rate can produce aliasing, in which higher-frequency components appear as false lower frequencies. Practical systems sample above the minimum and use anti-aliasing filters.

5. Transmission Impairments

A received signal differs from the transmitted signal because the channel is imperfect. The major impairments are attenuation, distortion, noise, delay, and jitter.

Attenuation

Attenuation is the reduction in signal power or amplitude as distance increases. It results from resistance, absorption, radiation, connector losses, and other mechanisms. Power gain or loss is commonly expressed in decibels:

For voltage measurements across equal impedances, the equivalent expression is:

The power formula requires power quantities. The voltage formula assumes equal reference impedances; otherwise power depends on impedance as well. A negative value denotes loss, while a positive value denotes gain.

A transmission path with several cascaded sections has total loss equal to the sum of their decibel losses. This follows because power ratios multiply while logarithms convert multiplication into addition.

Distortion

Distortion occurs when different frequency components experience unequal attenuation or phase shift. The received waveform then differs in shape from the transmitted waveform even without random noise. Amplitude distortion changes component magnitudes; phase or delay distortion changes their timing relationships.

Intersymbol interference (ISI) is a critical digital consequence of distortion. Energy from one symbol spreads into neighboring symbol intervals, making the receiver’s decision less reliable. Causes include limited bandwidth, multipath propagation, and inadequate equalization.

Noise

Noise is unwanted random energy added to a signal. Common categories include:

• Thermal noise: caused by random motion of charge carriers and present in electronic components.
• Induced noise: introduced by motors, power lines, switching equipment, and electromagnetic coupling.
• Crosstalk: unwanted coupling between adjacent circuits or channels.
• Impulse noise: short-duration, high-amplitude disturbances such as electrical switching events.

Signal-to-noise ratio (SNR) compares signal power with noise power:

Its decibel form is:

SNR is dimensionless as a ratio. A larger SNR generally permits more reliable detection, although interference, nonlinearities, coding, and receiver design also matter.

Propagation Delay

Propagation delay is the time required for a signal to travel through a medium:

Here, d is distance and v is propagation velocity. Propagation delay is distinct from transmission delay, which is the time to place all packet bits onto the link:

Here, L is packet length in bits and R is link bit rate in bits per second. Processing and queuing delays may also contribute to end-to-end delay.

Jitter

Jitter is variation in the timing of signal transitions or packet arrivals. A constant delay is often manageable; variable delay is more disruptive. In voice and video, excessive packet-arrival jitter causes gaps or requires larger playback buffers. In high-speed digital links, timing jitter can cause sampling at the wrong instant and increase bit errors.

6. Nyquist Noiseless-Channel Limit

For a noiseless channel with bandwidth B hertz and M discrete signal levels, the maximum theoretical bit rate is:

This is the Nyquist bit-rate limit. It assumes an ideal noiseless channel, perfect synchronization, and appropriately shaped signals. It describes a theoretical maximum, not the guaranteed rate of a real network.

For binary signaling, M = 2, so:

Increasing the number of signal levels can increase the theoretical rate without increasing bandwidth, but the levels become closer together and more vulnerable to noise and distortion.

Worked Example 1: Nyquist Limit

A noiseless channel has bandwidth 3 kHz and uses eight signal levels. Since M = 8, log₂(8) = 3. Therefore:

The theoretical maximum is 18 kbit/s. Real systems achieve less because channels are not noiseless and because overhead and implementation constraints consume capacity.

7. Shannon Noisy-Channel Capacity

For a band-limited channel with additive white Gaussian noise, Shannon’s capacity is:

Here, C is the theoretical maximum error-free data rate in bit/s, B is bandwidth in hertz, and SNR is the linear signal-to-noise power ratio—not its decibel value. The formula assumes a specified bandwidth, statistically characterized noise, arbitrarily effective coding, and sufficiently long codewords. It does not specify a practical modulation or coding scheme.

If SNR is given in decibels, convert it first:

Worked Example 2: Shannon Capacity

A channel has bandwidth 3 kHz and SNR of 30 dB. Convert SNR:

Then:

Thus, capacity is approximately 29.9 kbit/s under the stated assumptions. A practical system must operate below this limit to achieve a useful error rate.

Nyquist and Shannon Together

Nyquist emphasizes bandwidth and the number of distinguishable signal levels in an ideal channel. Shannon emphasizes bandwidth and SNR in a noisy channel. A real design must satisfy both conceptual constraints: increasing levels alone cannot overcome noise indefinitely, and increasing bandwidth may increase noise power depending on the system.

8. Transmission Media and Design Trade-offs

Guided media include twisted-pair cable, coaxial cable, and optical fiber. Unguided media include terrestrial radio, microwave, satellite, and infrared links.

Twisted pair is inexpensive and widely used, but attenuation, crosstalk, and electromagnetic interference constrain distance and rate. Coaxial cable provides better shielding and higher bandwidth than ordinary twisted pair, but is less flexible. Optical fiber offers very high bandwidth, low attenuation over long distances, and immunity to electromagnetic interference, although optical components and installation can be more specialized. Wireless systems provide mobility but face interference, fading, spectrum regulation, and shared-medium contention.

Designers trade among:

• bandwidth and equipment cost;
• transmit power and battery life;
• data rate and reliability;
• distance and repeater requirements;
• latency and buffering;
• complexity of modulation, equalization, and coding;
• regulatory limits and electromagnetic compatibility.

Equalizers compensate for predictable channel distortion. Filters restrict unwanted frequencies. Repeaters regenerate or reshape signals. Error-detecting and error-correcting codes add redundancy so that receivers can detect or correct some errors. These techniques improve reliability but consume bandwidth or processing resources.

9. Applied Case: Industrial Sensor Link

Background

Consider an illustrative case in which a manufacturing plant must connect temperature and vibration sensors to a monitoring server. Each sensor produces 12 kbit/s of useful data. The cable route is 2 km long, and the environment contains motors and variable-frequency drives that generate electromagnetic interference.

Problem

The plant initially proposes a low-cost copper link operating at 100 kbit/s. Test measurements show substantial high-frequency noise and occasional timing variation. The engineering team must determine whether the proposed link can support the sensors reliably.

Analysis

The aggregate useful data rate for eight sensors is:

This leaves only 4 kbit/s below the nominal link rate, before framing, addressing, acknowledgments, error-control coding, and retransmissions. Therefore, the nominal rate is inadequate even if the physical channel could maintain it.

The team measures an SNR of 18 dB over the cable’s usable bandwidth of 80 kHz. Converting SNR:

The Shannon upper bound is:

This bound does not guarantee a 481 kbit/s implementation, but it indicates that the channel’s measured SNR and bandwidth are not the only obstacles. Protocol overhead, impulsive interference, cable installation, and timing recovery must be addressed.

Approach and Solution

The engineers select shielded twisted pair with proper grounding, reduce the signaling bandwidth to avoid the most disturbed frequency region, and use a robust multilevel modulation scheme with forward-error correction. They also install surge protection, use differential signaling, and configure a modest receive buffer to absorb jitter. A link budget and laboratory stress tests are completed before deployment.

Outcome

The upgraded design provides sufficient margin for useful traffic and overhead. Monitoring remains reliable during motor startup, although the system records occasional corrected errors. The design also includes alarms for rising error rates and a maintenance procedure for checking shielding and connectors.

Learning Points

The case illustrates that nominal bit rate is not throughput, Shannon capacity is an upper bound rather than a product specification, and environmental noise must be measured rather than assumed. Reliable transmission depends on the complete physical and protocol design.

10. Professional, Ethical, and Regulatory Considerations

Transmission engineers must not report theoretical capacity as guaranteed service performance. Measurements should state bandwidth, SNR definition, test conditions, error-rate criteria, and whether overhead is included. Misrepresenting performance can affect safety, contracts, and public trust.

Wireless systems must comply with national spectrum allocations, transmit-power limits, electromagnetic compatibility rules, and exposure requirements. Critical infrastructure should use adequate redundancy, monitoring, authentication, and secure management interfaces. Although signal analysis is primarily a physical-layer topic, insecure configuration or unauthenticated control traffic can turn a reliable link into a security vulnerability.

Engineers should also consider accessibility, energy consumption, e-waste, and maintainability. Choosing a higher-capacity technology is not automatically sustainable if it increases unnecessary power use or makes repair impossible.

11. Contemporary Developments

Modern communication systems combine sophisticated modulation, adaptive coding, multiple antennas, channel estimation, and digital signal processing. Adaptive systems alter coding rate, modulation order, transmit power, or bandwidth in response to channel conditions. Optical networks use wavelength-division multiplexing to carry separate channels at different optical wavelengths. Wireless systems exploit spatial diversity and beamforming, while low-power sensor networks prioritize energy efficiency over peak bit rate.

Machine-learning methods are increasingly used for channel estimation, modulation classification, interference detection, and predictive maintenance. However, data-driven methods do not remove physical limits. They require representative training data, careful validation, explainable operating boundaries, and safeguards against failure under unusual channel conditions.

Best Practices

1. Define whether a rate is raw bit rate, net throughput, or application payload rate.
2. Convert decibel SNR values to linear ratios before using Shannon’s equation.
3. Distinguish symbol rate from bit rate.
4. Include attenuation, noise, distortion, delay, and jitter in link analysis.
5. Use measurements under realistic environmental and loading conditions.
6. Leave engineering margin below theoretical limits.
7. Document assumptions, units, test instruments, and error-rate criteria.
8. Use filtering, shielding, grounding, equalization, coding, and buffering appropriately.
9. Treat regulatory compliance and security as design requirements, not afterthoughts.

Chapter Summary

Signals are physical representations of information. Their principal parameters include amplitude, frequency, period, phase, wavelength, spectrum, and bandwidth. Analog signals vary continuously, whereas digital systems represent information using discrete levels, although both ultimately travel as physical waveforms.

Bit rate depends on symbol rate and the number of bits represented by each symbol. Channel impairments include attenuation, distortion, noise, propagation delay, transmission delay, and jitter. Decibels express gain or loss logarithmically, while SNR compares signal power with noise power.

Nyquist’s equation, R_max = 2B log₂(M), gives the maximum theoretical rate for a noiseless bandwidth-limited channel with M levels. Shannon’s equation, C = B log₂(1 + SNR), gives the theoretical error-free capacity of a noisy channel under specified assumptions. Neither equation guarantees practical throughput. Effective design requires measurement, coding, filtering, synchronization, suitable media, engineering margin, and attention to regulatory and security requirements.

Key Terms

Signal: A physical quantity that varies to represent information.

Analog signal: A continuously varying signal in time and amplitude.

Digital signal: A signal represented by discrete states or levels.

Amplitude: Magnitude of a signal relative to a reference.

Frequency: Number of cycles per second, measured in hertz.

Period: Time required for one complete cycle.

Phase: Relative position of a periodic waveform in its cycle.

Wavelength: Distance occupied by one cycle of a propagating wave.

Spectrum: Distribution of signal energy across frequency.

Bandwidth: Difference between upper and lower significant frequencies.

Bit rate: Number of bits transmitted per second.

Symbol rate: Number of signal symbols transmitted per second.

Attenuation: Reduction in signal strength during transmission.

Distortion: Change in waveform shape caused by unequal channel effects.

Noise: Unwanted energy added to a transmitted signal.

SNR: Ratio of signal power to noise power.

Propagation delay: Time for a signal to travel through a medium.

Jitter: Variation in signal or packet timing.

Nyquist limit: Theoretical rate limit for a noiseless bandwidth-limited channel.

Shannon capacity: Theoretical maximum reliable rate for a noisy channel.

Review Questions

Short Answer Questions

1. What is the difference between an analog signal and a digital signal?
2. Define frequency, period, and phase.
3. Why is baud not necessarily equal to bit rate?
4. What is signal-to-noise ratio, and why is it important?
5. Distinguish propagation delay from transmission delay.

Long Answer Questions

1. Explain how a time-domain waveform can be analyzed using its frequency spectrum. Include the significance of bandwidth.
2. Describe attenuation, distortion, noise, delay, and jitter, and explain how each can affect communication quality.
3. Derive or explain the practical meaning of the Nyquist noiseless-channel equation.
4. Explain Shannon capacity, including its assumptions and limitations in real systems.
5. Compare twisted pair, coaxial cable, optical fiber, and wireless media in terms of bandwidth, interference, cost, and applications.

Analytical / Application Questions

1. A noiseless 4 kHz channel uses 16 signal levels. Calculate its Nyquist maximum bit rate and explain why a practical system may achieve less.
2. A channel has 10 kHz bandwidth and SNR of 20 dB. Convert SNR to a linear ratio and calculate Shannon capacity.
3. A network carries ten sensors, each producing 15 kbit/s, over a link rated at 160 kbit/s. Evaluate whether the link has adequate nominal capacity after allowing for 20% protocol and error-control overhead. Justify your conclusion.

Multiple Choice Questions with Answers

Question 1

Which parameter specifies the number of cycles completed by a periodic signal per second?

A. Amplitude

B. Frequency

C. Phase

D. Wavelength

Correct Answer: B. Frequency

Question 2

A channel has bandwidth 3 kHz and uses eight signal levels in a noiseless environment. What is its Nyquist maximum bit rate?

A. 6 kbit/s

B. 9 kbit/s

C. 18 kbit/s

D. 24 kbit/s

Correct Answer: C. 18 kbit/s

Question 3

Which condition is required when applying Shannon’s capacity equation to a value expressed in decibels?

A. Multiply the dB value by bandwidth

B. Convert the dB value to a linear SNR

C. Divide the dB value by the number of symbols

D. Treat the dB value as the number of signal levels

Correct Answer: B. Convert the dB value to a linear SNR

Question 4

What is the principal effect of attenuation?

A. Reduction in signal strength

B. Creation of additional symbols

C. Increase in propagation velocity

D. Conversion of digital data into analog data

Correct Answer: A. Reduction in signal strength

Question 5

A modulation scheme uses 16 distinct signal states. How many bits can one symbol represent, ignoring coding overhead?

A. 2

B. 4

C. 8

D. 16

Correct Answer: B. 4

Question 6

Which impairment describes variation in the timing of signal transitions or packet arrivals?

A. Attenuation

B. Crosstalk

C. Jitter

D. Wavelength

Correct Answer: C. Jitter

Question 7

A packet contains 12,000 bits and is transmitted over a 2 Mbit/s link. What is its transmission delay?

A. 0.006 s

B. 0.012 s

C. 0.024 s

D. 6 s

Correct Answer: A. 0.006 s

Question 8

Which statement about Shannon capacity is correct?

A. It guarantees the practical throughput of every modem

B. It applies only to noiseless channels

C. It is a theoretical upper bound under stated channel assumptions

D. It is determined only by the number of signal levels

Correct Answer: C. It is a theoretical upper bound under stated channel assumptions

Question 9

Intersymbol interference occurs when:

A. Noise power is exactly zero

B. Energy from one symbol overlaps neighboring symbol intervals

C. A signal has only one frequency component

D. Propagation velocity equals the speed of light

Correct Answer: B. Energy from one symbol overlaps neighboring symbol intervals

Question 10

For a sinusoid with peak amplitude A, which expression gives its RMS amplitude under the stated sinusoidal assumption?

A. A

B. 2A

C. A²

D. A/√2

Correct Answer: D. A/√2