Frequency Reuse: The Mathematical Heart of Cellular Capacity

If the cellular concept is the heart of mobile networks, then frequency reuse is its lifeblood. This elegant mathematical framework determines how the same precious frequencies can be used repeatedly across a coverage area while keeping interference below acceptable levels.

Frequency reuse is what transforms 50 MHz of spectrum from supporting a few hundred calls into supporting millions of simultaneous connections. It's the reason your mobile phone works in a stadium with 80,000 other users, each expecting instant connectivity.

This page explores the mathematical foundations of frequency reuse, the engineering trade-offs involved in reuse planning, and how modern networks have evolved beyond traditional reuse patterns to extract maximum capacity from limited spectrum.

The fundamental insight:

Radio signals weaken with distance. If you transmit on a frequency in one location, that signal becomes undetectable at some distance away. Beyond that distance, the same frequency can be reused by another transmitter without interference.

Formal definition:

Frequency reuse is the assignment of radio channels (frequencies) to cells in such a way that:

1. No two adjacent cells use the same frequencies
2. Cells using identical frequencies are separated by enough distance that interference is acceptable
3. The pattern repeats systematically across the entire coverage area

The reuse factor (or cluster size) N is the number of distinct channel groups needed before the pattern repeats. If you have 280 channels and N=7, each cell gets 280/7 = 40 channels.

*Assuming 280 total channels in the system.

A cluster is a group of cells that collectively use all available frequencies exactly once. The entire coverage area is covered by repeating this cluster pattern. Understanding cluster geometry is essential for frequency planning.

Mathematical constraint on cluster size:

Not every number can be a valid cluster size. For hexagonal cells to tessellate properly with frequency reuse, the cluster size N must satisfy:

N = i² + i×j + j²

Where i and j are non-negative integers. This formula arises from the geometry of hexagonal tessellation.

Finding co-channel cells:

Given the (i, j) values, you can locate co-channel cells (cells using the same frequencies) using this rule:

1. Start at any cell
2. Move i cells in any direction
3. Turn 60° counter-clockwise
4. Move j cells in that new direction
5. You've arrived at a co-channel cell

This pattern repeats infinitely, covering the entire network. Every cell has exactly 6 first-tier co-channel neighbors arranged symmetrically around it.

The reuse distance (D) is the minimum distance between cells using the same frequency. This distance determines the signal-to-interference ratio (SIR) experienced by users, which directly affects call quality and data throughput.

Reuse distance formula:

For hexagonal cells with cell radius R and cluster size N:

D = R × √(3N)

This fundamental relationship connects cluster size to physical separation between co-channel cells.

Co-channel interference analysis:

The signal-to-interference ratio (SIR) determines communication quality. A mobile user receives:

• Desired signal (S) from their serving base station
• Interfering signals (I) from co-channel base stations

For a user at the edge of their cell (worst case), the SIR can be approximated:

SIR ≈ (D/R)^γ / 6

Where γ is the path loss exponent (typically 3-5 in mobile environments). The factor of 6 accounts for the six first-tier co-channel interferers.

For N=7 and γ=4:

• D/R = 4.58
• SIR ≈ (4.58)⁴ / 6 ≈ 440 / 6 ≈ 73.3 (about 18.6 dB)

18.6 dB SIR is adequate for voice but marginal for high-speed data.

Translating reuse theory into real network configurations requires careful frequency planning. This process assigns specific frequencies to each cell while managing both co-channel interference (same frequency) and adjacent-channel interference (nearby frequencies).

The frequency planning process:

Adjacent-channel interference:

Beyond co-channel interference, adjacent-channel interference (ACI) occurs when signals on nearby frequencies 'leak' into each other due to imperfect filtering. Frequency planning must also consider:

• Adjacent channel separation — Don't assign channels N and N+1 to neighboring cells
• Guard bands — Unused frequencies between channel groups
• Filter characteristics — Receiver filter selectivity determines required separation

These constraints further complicate frequency planning, sometimes requiring larger effective cluster sizes or accepting some capacity reduction.

Classical fixed-frequency assignment served well for decades, but modern cellular technologies have evolved far beyond these static patterns. Understanding this evolution reveals how networks continue to extract more capacity from limited spectrum.

Frequency hopping:

Instead of staying on one frequency, transmissions hop rapidly across multiple frequencies according to a pseudo-random pattern. Benefits:

• Interference averaging — Brief interference affects only portions of communication
• Frequency diversity — Avoids deep fades on specific frequencies
• Interference randomization — Reduces worst-case interference scenarios

GSM uses frequency hopping; the system rapidly switches between assigned channels. This effectively averages interference, allowing tighter reuse than purely static assignment.

Dynamic channel assignment:

Rather than fixed frequency plans, channels are assigned dynamically based on current interference conditions:

• Centralized DCA — Network controller assigns channels based on measurements from multiple cells
• Distributed DCA — Each base station autonomously selects channels based on local measurements
• Hybrid approaches — Combination of fixed base assignment with dynamic overflow

DCA can achieve 15-30% capacity improvement over fixed assignment in variable traffic conditions.

Fractional Frequency Reuse (FFR):

FFR, used in LTE, divides cell area into regions with different reuse patterns:

• Cell center — Uses all frequencies with low power (reuse = 1)
• Cell edge — Uses specific frequency subset at higher power (reuse = 3 or greater)

This approach maximizes capacity for users near the tower (good SIR anyway) while protecting cell-edge users who need more interference isolation. FFR can improve cell-edge throughput by 3-5× compared to reuse-1.

The evolution from FDMA/TDMA systems (like GSM) to CDMA (3G) and OFDMA (4G/5G) fundamentally changed how frequency reuse works. These technologies enable what's often called universal frequency reuse or reuse factor = 1.

The CDMA revolution:

CDMA (Code Division Multiple Access) was the first commercial technology to achieve reuse-1. Key enablers:

1. Spread spectrum — Signals spread across wide bandwidth, appearing as noise to non-intended receivers
2. Orthogonal codes — Different users use codes that don't interfere
3. Tight power control — Signals are adjusted 800 times per second to minimize interference
4. Soft handoff — Users connect to multiple cells simultaneously at boundaries

CDMA's capacity is limited by total interference, not fixed channel count. This allows dynamic capacity sharing across cells.

OFDMA and interference coordination:

OFDMA (Orthogonal Frequency Division Multiple Access) in 4G/5G further advances reuse-1:

• Scheduler-based allocation — The network dynamically assigns resource blocks
• Inter-Cell Interference Coordination (ICIC) — Neighboring cells coordinate to avoid scheduling conflicting resources
• Enhanced ICIC (eICIC) — Time-domain coordination for heterogeneous networks
• CoMP (Coordinated Multi-Point) — Multiple cells jointly transmit to cell-edge users

These techniques enable near-optimal use of spectrum across the entire network, far exceeding classical reuse limitations.

Understanding how frequency reuse translates to network capacity is essential for network planning. Let's work through the key calculations.

Basic capacity formula:

For a frequency-reuse-based system:

Channels per cell = Total System Channels / Cluster Size

Total capacity = Channels per cell × Number of cells

Example calculation:

Consider a network with:

• 10 MHz spectrum allocation
• 200 kHz channel bandwidth (like GSM)
• 7-cell reuse pattern
• 100 cells covering a metropolitan area

Step 1: Total channels = 10,000 kHz / 200 kHz = 50 channels

Step 2: Channels per cell = 50 / 7 ≈ 7 channels/cell

Step 3: Total network capacity = 7 × 100 = 700 simultaneous calls

Step 4: With 2% blocking probability and Erlang B formula, each cell can support approximately 3 Erlangs, meaning about 36-45 active users per busy hour

Step 5: Network-wide busy hour capacity ≈ 3,600-4,500 active users