Artificial Intelligence & Data Science 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.

Multiplexing and spread spectrum

Clear, exam-focused notes on multiplexing (FDMA, TDMA, SDMA, CDMA) and spread‑spectrum (DSSS, FHSS): definitions, key derivations (processing gain), formulas, worked numerical examples, common mistakes and exam tips.

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

Multiplexing and spread spectrum are how many users share a wireless radio channel and how a single user's signal can resist interference and eavesdropping. An engineer uses these ideas when designing or analysing cellular systems (GSM, CDMA, LTE), Wi‑Fi, Bluetooth and modern 4G/5G radio resource allocation — to compute capacity, choose codes, size bandwidth and predict interference performance.

Key ideas

Start with precise definitions (bold the term at first use).

  • Multiplexing — a method to combine multiple information signals for transmission over a single physical channel so that the receiver can separate them. Assumption: channel has finite capacity (bandwidth and power limits). Multiplexing types used in wireless: Frequency Division Multiple Access (FDMA), Time Division Multiple Access (TDMA), Orthogonal Frequency Division Multiple Access (OFDMA), Spatial Division Multiple Access (SDMA) and Code Division Multiple Access (CDMA). Define each below before further use.

  • Frequency Division Multiple Access (FDMA) — each user is given a separate frequency band (channel) within the overall spectrum. Assumption: filters can isolate bands and guard bands avoid overlap. Draw: a horizontal frequency axis; draw adjacent coloured blocks labelled f1, f2, f3 with small guard bands between.

  • Time Division Multiple Access (TDMA) — users share the same frequency band but transmit in separate time slots. Assumption: common time synchronization. Draw: time axis with repeating frames; each frame divided into slots labelled user1..userN.

  • Orthogonal Frequency Division Multiple Access (OFDMA) — users are assigned subsets of orthogonal OFDM subcarriers across time–frequency resource blocks; orthogonality is maintained by tight time/frequency synchronization and a cyclic prefix to combat multipath. Used in LTE/4G/5G NR and Wi‑Fi 6/7.

  • Spatial Division Multiple Access (SDMA) — users separated by antenna beams or spatial filters; same frequency/time can be reused if spatial separation (beamforming) provides isolation. Assumption: antenna arrays and channel spatial resolution.

  • Code Division Multiple Access (CDMA) — users transmit simultaneously in same frequency/time but use different spreading codes; separation done by correlating with that code. Assumption: receiver knows codes and can correlate; codes have good auto‑ and cross‑correlation properties.

  • Spread spectrum — a family of techniques that deliberately spread the transmitted signal power across a wider bandwidth than the minimum needed for the information. Two main kinds:

    • Direct Sequence Spread Spectrum (DSSS) — each data symbol is multiplied by a high‑rate binary chip sequence (spreading code). Terms: a chip is a single bit of the spreading code; chip rate Rc is chips per second (chips/s); symbol/data rate Rb is bits per second (b/s). Define spreading factor (SF) or processing gain (PG) as SF = PG = Rc / Rb (linear). Assumption: spreading code is binary ±1 and known to receiver.
    • Frequency Hopping Spread Spectrum (FHSS) — transmitter hops the carrier frequency among a set of frequencies according to a pseudo‑random sequence. Terms: hop rate and hopset size. Two regimes are defined: slow hopping (multiple data bits per hop) and fast hopping (multiple hops per data bit). An approximate processing gain for FHSS is the ratio of the total hopset bandwidth to the per‑hop modulation bandwidth.

Precise theorem-like statement used in exams (DSSS processing gain):

  • For DSSS with spreading factor N = Rc/Rb, under additive white noise or wideband uncorrelated interference, coherent despreading (correlation with the same spreading sequence) increases the signal‑to‑noise ratio at the correlator output by approximately the linear factor N (i.e., PG = N), or in decibels PGdB = 10·log10(N). Assumptions: perfect synchronization between transmitter and receiver, ideal correlation properties (auto‑correlation ≈ N, cross‑correlation ≈ 0), interference uncorrelated with the spreading code. Derivation below.

Derivation (step by step) of SNR improvement in DSSS:

  1. Let a single data bit s(t) ∈ {+1, −1} for duration Tb = 1/Rb. Spreading code c(t) is ±1 at chip rate Rc and repeats every Tb with length N = Rc/Rb chips.
  2. Transmitted waveform x(t) = s(t)·c(t) (ignoring modulation to carrier for simplicity). Receiver multiplies x(t) by c(t) (despread) and integrates over Tb (correlator). Compute correlator output y = ∫_{0}^{Tb} r(t)·c(t) dt where r(t)=x(t)+n(t), n(t) is noise/interference.
  3. For the desired signal part: ∫ s(t)c(t)c(t) dt = ∫ s(t) dt since c(t)^2 = 1. This integrates to s·Tb (amplitude scales with Tb). The signal contribution after despreading accumulates coherently over N chips (equivalently factor N in discrete summary).
  4. For additive white noise with power spectral density N0/2, noise samples at the chip rate are uncorrelated; after despreading and integration over Tb the signal adds coherently (amplitude ∝ N, power ∝ N^2) while the noise power adds incoherently (power ∝ N). Thus SNR_out ≈ SNR_in · N. (Equivalently: signal terms add coherently whereas noise terms add incoherently, yielding a linear N processing gain in SNR.)
  5. Therefore PG (linear) = N = Rc / Rb and PG(dB) = 10·log10(N).

Comments on orthogonality and capacity:

  • Walsh codes (orthogonal codes) of length N provide at most N mutually orthogonal codewords for synchronous CDMA; therefore in an ideal synchronous system one can support N users with orthogonal Walsh codes. In asynchronous systems or under multipath, orthogonality breaks and codes with low cross‑correlation (PN sequences) are used. On the uplink, strict power control is required to mitigate the near–far problem (strong users overwhelming weaker ones at the base station), and multipath is commonly exploited by a RAKE receiver that combines energy from resolvable paths.

Describe diagrams to draw in an answer:

  • DSSS Tx block diagram: draw data bit → multiplier (×) with spreading code → lowpass (optional) → carrier modulator → channel. Label Rb, Rc, N.
  • CDMA multiple user diagram: show multiple user streams each multiplied by different code, summed, passed to channel; at receiver show one correlator matched to desired code.

Formulas

PG (linear) = N = Rc / Rb — Rc is chip rate (chips/s), Rb is data rate (b/s). Applies to DSSS where spreading code length = N chips per symbol. Equivalently, when Bss ≈ Rc and Bd ≈ Rb, PG ≈ Bss / Bd.

PG (dB) = 10 · log10(N) — PG in decibels. Units: dB (dimensionless). Applies when converting linear PG to decibels.

Approx. DSSS occupied bandwidth B ≈ Rc — B in hertz (Hz). Applies as an engineering approximation: DSSS signal bandwidth is on order of chip rate.

Data per TDMA slot (bits) = Rphy · (Tslot − tguard) — Rphy in bits/s, Tslot in s, tguard in s. Applies to TDMA systems with guard time per slot.

User data rate in TDMA = (Data per slot) / Tframe — Units: bits/s. Tframe is frame duration in s.

Shannon capacity: C = B · log2(1 + SNR) — C in bits/s, B in Hz, SNR linear. Applies for AWGN channel, theoretical maximum under ideal coding.

Conversion dB ↔ linear: x(dB) = 10 · log10(x_linear) — dimensionless conversion. Use for power ratios.

Worked examples

Example 1 — DSSS: processing gain and SNR improvement Given: data rate Rb = 10 kb/s (10 000 b/s), chip rate Rc = 1 Mcps (1 000 000 chips/s). Assume an initial SNR at the receiver front end SNR_in = 0 dB (linear 1). Steps:

  1. Compute spreading factor N = Rc / Rb = 1 000 000 / 10 000 = 100. Formula: N = Rc / Rb
  2. Convert PG to dB: PGdB = 10 · log10(100) = 20 dB.
  3. Approximate occupied bandwidth B ≈ Rc = 1 000 000 Hz = 1 MHz.
  4. SNR after despreading (linear) ≈ SNR_in · N = 1 · 100 = 100 ⇒ in dB 20 dB. Final answers: Processing gain = 100 (20 dB); approximate DSSS bandwidth ≈ 1 MHz; SNR after despreading = 100 (20 dB).

Example 2 — TDMA: slots, guard time and user data rate Given: physical channel rate Rphy = 1 Mbps = 1 000 000 bits/s, frame duration Tframe = 10 ms = 0.01 s, number of slots per frame = 8. Guard time per slot tguard = 50 µs = 50 × 10^−6 s. Steps:

  1. Slot duration Tslot = Tframe / 8 = 0.01 / 8 = 0.00125 s = 1.25 ms.
  2. Useful slot time = Tslot − tguard = 0.00125 − 0.00005 = 0.00120 s = 1.20 ms.
  3. Bits per slot = Rphy · useful slot time = 1 000 000 · 0.00120 = 1 200 bits.
  4. User data rate = bits per slot / Tframe = 1200 / 0.01 = 120 000 bits/s = 120 kb/s. Final answer: Each TDMA user gets 120 kb/s of payload.

Common mistakes

  • Treating Walsh orthogonality as preserved under multipath or asynchronous arrival — in practice orthogonality is lost unless timing is perfect.
  • Forgetting guard time in TDMA capacity calculations or treating guard time as part of payload.
  • Assuming PG directly increases channel capacity (bits/s). PG improves interference tolerance (SNR after despreading) but does not increase raw Shannon capacity without accounting for bandwidth increase.

For your exam

Question types produced by this topic: definitions (explain DSSS, FHSS, CDMA, PG), short notes (compare FDMA/TDMA/CDMA/SDMA), derivations (derive PG = Rc/Rb and show SNR improvement after despreading), numerical problems (compute PG, bandwidth, TDMA slot capacity), and diagrams (draw DSSS Tx/Rx block diagrams and TDMA frame structures).

What a full‑marks answer includes:

  • Clear definitions of each term used (e.g., chip, Rc, Rb, PG).
  • Assumptions stated (synchronization, AWGN, ideal correlation) when deriving results.
  • Step‑by‑step derivation for processing gain with equations and short justification why signal adds coherently and noise incoherently.
  • Correct units and numerical substitution for numericals; final numeric answers labelled and with units.
  • Diagrams described or sketched with labels for block inputs/outputs, rates and durations.

Quick check

  1. Define processing gain for DSSS in one line.
  2. If Rc = 2 Mcps and Rb = 50 kb/s, what is PG (linear) and in dB?
  3. In TDMA, if frame duration = 20 ms and there are 10 slots, what is slot duration in ms? Answers: PG = Rc/Rb; PG = 2 000 000 / 50 000 = 40 → 10·log10(40) ≈ 16.0 dB; slot duration = 20 ms / 10 = 2 ms.

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

Stuck on something here?