11.1 Pulse & Digital Modulation Systems

11.1 Pulse & Digital Modulation Systems

Quick Answer: Pulse and digital modulation systems convert analog band-limited signals into discrete binary code and modulate them onto high-frequency carriers. Key formulas tested on the PRC ECE board exam include the Nyquist sampling rate ($f_s \ge 2 f_m$), Pulse Code Modulation bit rate ($R = n \cdot f_s$), Signal-to-Quantization-Noise Ratio ($SQNR \approx 6.02n + 1.76 \text{ dB}$ for sinusoidal signals), and M-ary modulation spectral efficiency ($B_w = S = R / \log_2 M$).

---

Fundamentals of Pulse Modulation

Pulse modulation systems fall into two distinct categories: analog pulse modulation and digital pulse modulation. In analog pulse modulation, a carrier consisting of a periodic train of rectangular pulses is modified by an analog message signal. The primary analog pulse modulation methods are:

1. Pulse Amplitude Modulation (PAM): The amplitude of individual pulses is made proportional to the instantaneous amplitude of the modulating signal. PAM signals remain analog in nature because pulse amplitudes take on an infinite continuous range of values.
2. Pulse Width Modulation (PWM): Also called Pulse Duration Modulation (PDM), the width or duration of each pulse varies in direct proportion to the modulating signal amplitude while pulse height remains constant.
3. Pulse Position Modulation (PPM): The time position of a constant-width pulse within its prescribed time slot is shifted relative to a reference clock in accordance with the modulating signal's instantaneous magnitude.

While PAM, PWM, and PPM offer basic pulse-based transmission, they remain susceptible to noise accumulated along transmission paths. Pulse Code Modulation (PCM) overcomes noise degradation by sampling, quantizing, and converting analog samples into fixed-length binary bit sequences.

---

Nyquist Sampling Theorem & Aliasing

To digitize a continuous-time signal without information loss, it must be sampled according to the Nyquist-Shannon Sampling Theorem.

Mathematical Statement

For a band-limited signal with a maximum modulating frequency component $f_m$, the minimum sampling frequency $f_s$ required to allow exact reconstruction from its samples is:

$f_s \ge 2 f_m$

The absolute minimum sampling frequency $f_{s,min} = 2 f_m$ is designated as the Nyquist rate, and the corresponding maximum time interval between samples $T_s = 1 / (2 f_m)$ is the Nyquist interval.

Aliasing & Anti-Aliasing Filters

If a signal is sampled below the Nyquist rate ($f_s < 2 f_m$), the high-frequency spectral components overlap into the fundamental baseband spectrum. This distortion phenomenon is known as aliasing. Overlapping spectral components generate false lower-frequency ghost signals that cannot be separated by downstream filtering.

In practical PRC examination problems, an anti-aliasing low-pass filter with a sharp cutoff frequency at $f_m$ is placed prior to the sampler to eliminate any frequency components above half the sampling frequency ($f_s / 2$). Additionally, real-world telecommunication systems utilize a guard band ($f_s > 2 f_m$) to allow practical filter roll-off. For example, standard commercial voice telephony (band-limited to $300 - 3400 \text{ Hz}$, approximated as $f_m = 4000 \text{ Hz}$) utilizes a standardized sampling frequency of $f_s = 8000 \text{ Hz}$ (or 8 kHz).

---

Quantization Process & SQNR

Quantization converts a continuous-amplitude sample into a discrete amplitude chosen from a finite set of predetermined levels $L$. If an analog voltage range $V_{max} - V_{min}$ is divided into $L$ equal step sizes, the quantization step size (or resolution) $\delta$ is:

$\delta = \frac{V_{max} - V_{min}}{L} = \frac{V_{p-p}}{2^n}$

where $n$ is the number of binary bits assigned per sample ($L = 2^n$).

Quantization Error and Noise

The difference between the instantaneous input sample voltage and its assigned discrete quantized level is the quantization error $e_q$. Assuming uniform quantization where error is uniformly distributed between $-\delta / 2$ and $+\delta / 2$, the mean square quantization noise power $N_q$ is:

$N_q = \frac{\delta^2}{12}$

Signal-to-Quantization-Noise Ratio (SQNR)

For a full-scale sinusoidal signal with peak amplitude $A = V_{p-p} / 2$, the signal power is $P_s = A^2 / 2 = (L \delta / 2)^2 / 2 = L^2 \delta^2 / 8$. Taking the ratio of signal power to quantization noise power yields the fundamental PCM SQNR formula:

$SQNR = \frac{P_s}{N_q} = \frac{L^2 \delta^2 / 8}{\delta^2 / 12} = 1.5 L^2 = 1.5 \times 2^{2n}$

Expressing SQNR in decibels:

$SQNR_{\text{dB}} = 10 \log_{10}(1.5) + 20 \log_{10}(2^n) \approx 6.02 n + 1.76 \text{ dB}$

This key equation demonstrates that adding 1 bit per sample increases the SQNR by approximately 6.02 dB (often referred to as the 6 dB per bit rule of thumb on PRC licensure exams).

---

Companding: A-Law and μ-Law Standards

In speech signals, low-amplitude vocal sounds occur far more frequently than high-amplitude peaks. Uniform quantization yields poor SQNR for weak signals because the fixed step size $\delta$ represents a large percentage relative to small amplitudes. To equalize SQNR across all signal levels, non-uniform quantization via companding (COMPression + EXPanding) is applied.

The analog signal is compressed prior to quantization at the transmitter and expanded after decoding at the receiver.

---

Pulse Code Modulation (PCM) Bit Rate & Bandwidth Calculations

The fundamental parameters governing a PCM transmission link are bit rate ($R$), bit duration ($T_b$), and minimum transmission bandwidth ($B_{min}$).

1. PCM Transmission Bit Rate ($R$): $R = n \cdot f_s = n \cdot 2 f_m$ where $n$ is bits per sample and $f_s$ is sampling frequency.
2. Bit Duration ($T_b$): $T_b = \frac{1}{R}$
3. Minimum Nyquist Transmission Bandwidth ($B_{min}$): Using ideal Nyquist low-pass filtering, the minimum channel bandwidth needed to transmit $R$ bits per second without intersymbol interference (ISI) is: $B_{min} = \frac{R}{2} = \frac{n f_s}{2}$

Step-by-Step Sample Board Exam Calculation

Problem: A standard analog voice signal occupying a bandwidth of $300 \text{ Hz}$ to $3400 \text{ Hz}$ is sampled at $8000 \text{ samples/sec}$ and encoded using an 8-bit PCM system. Calculate (a) the PCM output bit rate, (b) the duration of a single bit, and (c) the minimum channel bandwidth required.

Solution:

1. Calculate Bit Rate ($R$): $R = n \cdot f_s = 8 \text{ bits/sample} \times 8000 \text{ samples/sec} = 64,000 \text{ bps} = 64 \text{ kbps}$
2. Calculate Bit Duration ($T_b$): $T_b = \frac{1}{R} = \frac{1}{64,000 \text{ sec}} = 15.625 \ \mu\text{s}$
3. Calculate Minimum Channel Bandwidth ($B_{min}$): $B_{min} = \frac{R}{2} = \frac{64,000}{2} = 32,000 \text{ Hz} = 32 \text{ kHz}$

---

Digital Carrier Modulation Schemes

When transmitting binary data over bandpass channels (such as wireless radio links, coaxial networks, or satellite feeds), binary bits modulate a high-frequency sinusoidal carrier $v_c(t) = A_c \cos(2\pi f_c t + \phi)$.

  Digital Baseband Bitstream (10110...)
                 │
                 ▼
   ┌───────────────────────────┐
   │ Digital Carrier Modulator │ ◄── Sinusoidal Carrier Ac*cos(2π fc t + φ)
   └─────────────┬─────────────┘
                 │
                 ▼
  Modulated Output Signal (ASK, FSK, PSK, QAM)

1. Amplitude Shift Keying (ASK)

Binary ASK varies carrier amplitude $A_c$ between two levels. In On-Off Keying (OOK), a binary '1' transmits the carrier ($A_c$) and a binary '0' turns the carrier off ($0$). ASK is highly sensitive to channel attenuation and noise variations.

2. Frequency Shift Keying (FSK)

Binary FSK shifts carrier frequency between two discrete frequencies: mark frequency $f_m$ (for binary '1') and space frequency $f_s$ (for binary '0'). In Continuous Phase FSK (CPFSK) or Minimum Shift Keying (MSK), phase continuity is maintained across bit transitions, minimizing sideband bandwidth requirements.

3. Phase Shift Keying (PSK)

PSK alters carrier phase $\phi$ while maintaining constant amplitude $A_c$, providing superior noise immunity over ASK.

• Binary PSK (BPSK): Uses 2 phase states separated by $180^\circ$ ($0^\circ$ for '1', $180^\circ$ for '0'). Each symbol represents 1 bit ($N=1$).
• Quadrature PSK (QPSK): Uses 4 phase states separated by $90^\circ$ ($45^\circ, 135^\circ, 225^\circ, 315^\circ$). Each symbol encodes $N=2$ bits (dibit: 00, 01, 11, 10).
• 8-PSK: Uses 8 phase states separated by $45^\circ$. Each symbol encodes $N=3$ bits (tribit).

4. Quadrature Amplitude Modulation (QAM)

QAM simultaneously modulates both amplitude and phase of two orthogonal carriers (in-phase $I$ and quadrature $Q$) operating at the same carrier frequency $f_c$:

$s(t) = I(t) \cos(2\pi f_c t) - Q(t) \sin(2\pi f_c t)$

Common QAM orders include 16-QAM (4 bits/symbol), 64-QAM (6 bits/symbol), and 256-QAM (8 bits/symbol), widely used in 4G/5G mobile networks, Wi-Fi (802.11ac/ax), and digital cable TV (DOCSIS).

---

Modulation Performance Comparison

The table below summarizes digital carrier modulation parameters frequently asked in licensure board exams:

---

Constellation Diagrams & Symbol Mapping

A constellation diagram plots digital modulation symbols on a two-dimensional Cartesian plane where the horizontal axis represents the In-Phase ($I$) component and the vertical axis represents the Quadrature ($Q$) component.

• BPSK Constellation: Consists of 2 points located at $(+A, 0)$ and $(-A, 0)$ on the $I$-axis.
• QPSK Constellation: Consists of 4 points located in the four quadrants at $(\pm A/\sqrt{2}, \pm A/\sqrt{2})$, with equal distance from the origin.
• 16-QAM Constellation: Displays a rectangular grid of 16 points ($4 \times 4$ array) combining 3 distinct amplitude levels and 12 phase angles.

To minimize bit error rates (BER), adjacent constellation points are mapped using Gray Coding, ensuring that neighboring symbol decisions differ by only 1 single bit position.

---

Line Coding Techniques

Line coding converts binary bit sequences into baseband voltage waveforms suitable for physical metallic or optical channels. Key selection criteria include self-clocking capability, DC balance (zero net DC offset), and spectral occupancy.

Bitstream:     1           0           1           1           0
         ┌───┐       ┌───────┐   ┌───┐       ┌───┐       ┌───────┐
NRZ-L:   │   │       │       │   │   │       │   │       │       │
       ──┘   └───────┴───────┴───┘   └───────┘   └───────┴───────┴──

         ┌───┐   ┌───┐       ┌───┐   ┌───┐   ┌───┐       ┌───┐   ┌───┐
Manch:   │   └───┤   │───────┤   └───┤   └───┤   │───────┤   └───┤   │
       ──┘       └───┘       └───┘       └───┘   └───┘       └───┘   └───

1. Non-Return-to-Zero Level (NRZ-L): Binary '1' is represented by one voltage level (e.g., $+V$), and binary '0' by another (e.g., $-V$). Lacks self-clocking and contains high DC content.
2. Non-Return-to-Zero Invert (NRZ-I): Transition at the boundary of a bit period indicates binary '1'; no transition indicates binary '0'. Provides differential encoding.
3. Return-to-Zero (RZ): Signal returns to zero volts halfway through each bit period. Ensures timing synchronization but doubles required channel bandwidth.
4. Alternate Mark Inversion (AMI): A bipolar encoding where binary '0' is zero volts, while binary '1' alternates between $+V$ and $-V$. Eliminates DC content and enables single-bit error detection.
5. Manchester Code: Binary '1' is represented by a high-to-low voltage transition at the midpoint of the bit interval; binary '0' by a low-to-high transition. Guarantees clock recovery for every bit but requires double the minimum bandwidth ($B = R$) compared to NRZ.