Computer Engineering is coming soon

We're still writing and reviewing the notes for every subject, and we'll open it when every subject is ready. Meanwhile you can practise aptitude, ask Aria about any subject, or see the branches that are ready.

Telephone system and broadband

How the traditional telephone network (PSTN) works, how voice is digitized and switched, and the broadband technologies that replaced or augmented the last mile (ISDN, DSL, cable, fiber), with core formulas and exam-style examples.

Drafted with Aria, reviewed by the AiCanCode.org team · Updated . How we write and check our notes. Spotted an error? Use Give Feedback at the bottom of the page.

Why it matters

Telephone systems and broadband are the foundation of voice and Internet services. Engineers design, operate and troubleshoot the physical and switching parts (local loops, exchanges, trunks), and select or optimise last‑mile broadband (DSL, cable, fiber) for capacity, quality and cost.

Key ideas

Start from basic units and then build up systems used in real networks.

  • Telephone system (PSTN): the Public Switched Telephone Network — a global circuit‑switched network originally built for analog voice, now mostly digital in the core and hybrid at the edges. PSTN provides the service of a point‑to‑point dedicated circuit for each call for the call duration.

  • Local loop: the physical pair of copper wires (or other medium) that connects a subscriber's telephone to the local exchange (telephone office). Also called the subscriber loop or 'last mile'.

  • Switching: the process of connecting two subscribers. In telecoms we primarily had circuit switching: reserve a path (circuit) for the call duration. Modern networks also use packet switching (e.g., IP/VoIP) but PSTN core uses circuits and time‑division multiplexed links.

  • Signaling: the control messages to set up, supervise and tear down calls. Early signaling was in‑band (on the same channel as voice). Modern signaling is mostly out‑of‑band (separate channels) and uses protocols such as SS7 (Signalling System No.7) which carries call control, routing and supplementary services.

  • Analog to Digital conversion (PCM): voice is bandlimited, sampled, quantized and encoded using Pulse Code Modulation (PCM). The standard telephony sampling rate is 8 kHz (from the sampling theorem since voice is bandlimited to ≈4 kHz), with 8 bits per sample after companding (µ‑law or A‑law) giving 64 kb/s per voice channel (called a DS0). Explain to draw: draw a sinusoidal voice waveform, show sampling points at 8 kHz, quantization steps and codewords.

  • Companding: µ‑law (used in North America and Japan) and A‑law (used in Europe and India) are nonlinear quantization (compressor/expander) methods that allocate more quantization accuracy to small amplitudes, improving effective SNR for speech.

  • Multiplexing and trunks: Time‑Division Multiplexing (TDM) groups voice channels into higher speed links. Standard digital hierarchies: E1 (Europe/India) is 2.048 Mb/s composed of 32 time slots of 64 kb/s each (TS0 often used for framing and TS16 for signalling in some arrangements; 30 voice channels + signalling + framing in common configurations). T1 (North America) is 1.544 Mb/s with 24 × 64 kb/s channels (1.544 Mb/s = 24 × 64 kb/s = 1.536 Mb/s plus 8 kb/s framing; in channelized voice, T1 commonly uses robbed‑bit signalling). Draw: a table with time slots 0..31 labelled and show which are speech vs signalling.

  • ISDN: Integrated Services Digital Network — a set of standards for digital end‑to‑end services over copper. BRI (Basic Rate Interface) is 2B+D (2 bearer channels × 64 kb/s + 1 delta signalling channel 16 kb/s) = 144 kb/s for user channels; gross line rate is 192 kb/s including framing/overhead. PRI maps to an E1 or T1 trunk (E1 PRI usually 30B+D; T1 PRI is 23B+D).

  • Broadband: loosely, any service much faster than narrowband voice channels; typically refers to always‑on, high‑bit‑rate access (DSL, cable, fiber). Broadband increases the data rate of the local loop beyond 64 kb/s per call.

  • DSL family: Digital Subscriber Line technologies that use existing copper pairs to carry higher rates by using frequencies above voice. ADSL (Asymmetric DSL) provides higher download than upload by allocating more subcarriers to downstream. VDSL and VDSL2 operate at higher frequencies and shorter loops for much higher rates. DSL uses multicarrier modulation (DMT) where the line is split into many narrow subcarriers, each modulated (QAM) according to its SNR. In ADSL (G.992.1), DMT uses 4.3125 kHz subcarrier spacing with up to 256 tones; a typical band plan allocates 0–4 kHz to POTS, ~25 kHz–138 kHz to upstream, and ~138 kHz–1.1 MHz to downstream, separated by splitters/filters.

  • Cable broadband (DOCSIS): uses the hybrid fiber‑coax plant of cable TV providers. Subscribers share a coax segment; a Cable Modem Termination System (CMTS) at the headend schedules bursts from cable modems (DOCSIS standard family). Because the medium is shared, peak rates depend on contention.

  • Fiber access: Fiber To The Home (FTTH) brings optical fiber to the subscriber. Passive Optical Networks (PON) like GPON share a fiber using passive splitters; downstream is broadcast, upstream is time‑shared. Typical GPON line rates are ~2.488 Gb/s downstream and ~1.244 Gb/s upstream (with higher‑rate XG(S)‑PON variants at 10 Gb/s).

  • Capacity limits: The Nyquist and Shannon formulas (see Formulas) give theoretical limits based on bandwidth and signal‑to‑noise ratio. In practice line attenuation, crosstalk and interference and equipment constraints reduce attainable rate.

  • Basic network elements to draw for an exam: (1) PSTN block diagram: subscriber — local loop — local exchange — trunk network — tandem/long‑distance switch — destination exchange — local loop — subscriber. Label analog/digital conversion and multiplexers. (2) DSL spectrum split: voice band (0–4 kHz), upstream band, downstream band with many subcarriers. (3) TDM frame for E1: 32 slots × 8 bits × 8000 frames/s = 2.048 Mb/s.

Formulas

f_s >= 2 · f_max — Sampling theorem (Nyquist sampling). f_s in Hz, f_max the highest frequency component of the signal in Hz. Applies to bandlimited analog signals.

R_pcm = f_s · n — PCM bit rate. R_pcm in bits/s, f_s in samples/s, n bits per sample. For standard telephony: f_s = 8000 Hz, n = 8 bits ⇒ 64 000 b/s.

C_Nyquist = 2 · B · log2(M) — Nyquist (noiseless discrete levels) capacity in bits/s, where B is bandwidth in Hz and M is number of discrete signal levels. Applies to an ideal noiseless channel.

C_Shannon = B · log2(1 + SNR) — Shannon capacity in bits/s, B in Hz, SNR is linear (not dB). Gives maximum achievable bit rate with arbitrarily low error probability for given bandwidth and SNR. Use SNR_linear = 10^(SNR_dB/10) to convert from decibels.

R_TDM = N_ch · R_ch — Gross bit rate of a TDM link. N_ch is number of channels, R_ch is bit rate per channel (e.g., 64 kb/s per voice channel). Units bits/s.

E1_rate = 32 · 8 · 8000 = 2 048 000 — E1 total bit rate in bits/s. Here 32 timeslots × 8 bits per timeslot × 8000 frames/s.

Worked examples

  1. PCM voice channel bit rate (standard telephony)
  • Given: sampling frequency f_s = 8000 samples/s, bits per sample n = 8.
  • Formula: R_pcm = f_s · n.
  • Calculation: R_pcm = 8000 · 8 = 64 000 bits/s.
  • Final answer: 64 kb/s (one DS0 voice channel).
  1. E1 framing and voice channels
  • Given: an E1 frame has 32 timeslots, 8 bits each, 8000 frames per second.
  • Formula: E1_rate = 32 · 8 · 8000.
  • Calculation: E1_rate = 32 · 8 · 8000 = 2 048 000 bits/s.
  • If TS0 used for framing and TS16 used for signalling, voice channels = 30.
  • Voice capacity: 30 · 64 000 = 1 920 000 bits/s (plus framing and signalling overhead).
  • Final answer: E1 carries 2.048 Mb/s total; typically 30 simultaneous voice calls (30 × 64 kb/s = 1.92 Mb/s) plus signalling and framing.
  1. Shannon capacity example
  • Given: a copper link has usable bandwidth B = 4000 Hz and measured SNR = 30 dB. Convert SNR to linear: SNR_linear = 10^(SNR_dB/10) = 10^(30/10) = 1000.
  • Formula: C = B · log2(1 + SNR).
  • Calculation: C = 4000 · log2(1 + 1000) ≈ 4000 · log2(1001) ≈ 4000 · 9.967 ≈ 39 868 bits/s.
  • Final answer: ≈ 39.9 kb/s theoretical maximum for that 4 kHz band given SNR=30 dB. (In practice digital coding and wider spectra let DSL exceed voice band limits.)

Check your arithmetic: 10^(30/10) = 1000; log2(1001) ≈ ln(1001)/ln2 ≈ 6.908/0.693 ≈ 9.968 — calculation consistent.

Common mistakes

  • Confusing DS0 (64 kb/s) with E1/T1 total rates and counts (E1 = 32 slots; T1 = 24 channels). Always show the arithmetic.
  • Assuming DSL downstream = upstream (ADSL is asymmetric).
  • Treating cable broadband as point‑to‑point when it is shared; capacity calculations must consider contention.

For your exam

Questions usually ask for definitions (PSTN, PCM, local loop), derivations (show why 8 kHz sampling and 8 bits → 64 kb/s), block diagrams (draw PSTN and DSL spectrum split), comparisons (copper DSL vs cable vs fiber for bandwidth/latency), and numerical problems using Nyquist/Shannon and TDM capacity.

Full‑marks answers include: clear definitions (bold/explicit), a labelled diagram where requested, the exact formula used, step‑by‑step arithmetic with units, and a short sentence stating assumptions and practical limitations (e.g., noise, channel attenuation, contention).

Quick check

  1. What is the standard PCM sample rate used for telephone voice? 2) How many 64 kb/s channels are carried in an E1 link when two timeslots are reserved for framing/signalling? 3) State the Shannon capacity formula and explain what SNR must be given as.

Answers: 1) 8000 samples/s (8 kHz). 2) 30 channels (32 − 2 = 30). 3) C = B · log2(1 + SNR), where SNR is linear (not in dB) and B is bandwidth in Hz.

Finished this topic? Mark it so your progress, study plan and readiness keep up.

Stuck on something here?