Shannon-Hartley Channel Capacity

The Shannon-Hartley theorem gives the theoretical maximum information rate of a bandwidth-limited channel with additive white Gaussian noise. The formula is C = B log2(1 + SNR), where C is capacity in bits per second, B is bandwidth in hertz, and SNR is the linear signal-to-noise power ratio. It does not describe a particular modem, code, antenna, or protocol. It defines an upper bound: no communication system can reliably exceed that capacity under the stated assumptions.

Capacity grows linearly with bandwidth but only logarithmically with SNR. Doubling bandwidth doubles capacity if SNR stays the same. Increasing SNR has diminishing returns because log2(1 + SNR) grows slowly at high SNR. This tradeoff is central to communication engineering. A system can chase more throughput by using more bandwidth, by improving SNR, or by using modulation and coding that operate closer to the theoretical limit.

This calculator accepts channel bandwidth and SNR in decibels. It converts SNR from dB to linear ratio using 10^(SNRdB / 10), computes spectral efficiency in bits per second per hertz, then multiplies by bandwidth. The result is a capacity bound, not a guaranteed payload rate. Real systems spend capacity on coding, pilots, synchronization, guard intervals, headers, retransmissions, and implementation margin.

Manual Calculation Steps

Consider a 20 MHz channel with 30 dB SNR. First convert 30 dB to a linear power ratio: 10^(30 / 10) = 1000. Spectral efficiency is log2(1 + 1000), which is about 9.967 bits per second per hertz. Multiply by 20,000,000 Hz to get about 199.3 Mbps theoretical capacity. A real radio using that channel may deliver less user throughput because of coding rate, medium access overhead, fading margin, and protocol framing.

If SNR increases from 30 dB to 33 dB, the linear ratio doubles from 1000 to about 1995, but spectral efficiency increases only from about 9.97 to about 10.96 bits per second per hertz. The extra 3 dB helps, but it does not double capacity. If bandwidth doubles from 20 MHz to 40 MHz at the same SNR, the capacity bound doubles. This is why wide channels are so attractive when spectrum is available.

Spectral Efficiency and Eb/N0

Spectral efficiency is C / B, measured in bits per second per hertz. Higher spectral efficiency means the system sends more bits through each hertz of bandwidth, usually by using higher-order modulation and stronger coding. The calculator also reports an equivalent Eb/N0 relationship at the capacity point. Eb/N0 compares energy per information bit to noise spectral density and is widely used when evaluating coding and modulation performance.

Shannon capacity assumes ideal coding over long blocks and known statistical conditions. Practical systems use finite block lengths, finite processing power, imperfect channel estimates, oscillator phase noise, nonlinear amplifiers, quantization, interference, and time-varying channels. Modern systems such as Wi-Fi, LTE, 5G, cable modems, and satellite links use adaptive modulation and coding so the data rate changes with channel quality.

Engineering Applications

Channel capacity calculations are used in link budgets, RF planning, wired communication, satellite links, acoustic modems, optical communication, and telemetry systems. They help answer whether a proposed bandwidth and SNR can theoretically support a target data rate before detailed protocol overhead is considered. If the target rate is above Shannon capacity, no implementation detail can save the design. If the target is below capacity, the question becomes whether a practical modulation, coding, and hardware architecture can approach the needed efficiency with enough margin.

The theorem is powerful because it separates physical limits from implementation quality. It does not tell an engineer how to build the modem, but it defines the ceiling. Used with SNR calculations, bandwidth allocation, antenna gain, noise figure, and regulatory constraints, Shannon-Hartley capacity becomes a first-order compass for communication-system feasibility.

Interference should not be casually folded into the formula without thought. The theorem assumes additive white Gaussian noise, while many real channels contain narrowband interferers, impulsive noise, adjacent-channel energy, multipath fading, or time-varying blockers. Engineers often convert the impairment into an effective SINR for a link budget, but that approximation depends on receiver design and waveform details. A channel with the same average SNR can perform very differently if the noise is not spectrally flat or if fading creates deep notches.

Capacity also assumes reliable communication with arbitrarily small error probability as coding block length grows. Low-latency control links, short packets, and battery-powered devices cannot always use long codes or heavy decoding. Those systems may operate far below the Shannon limit by design because latency, complexity, and energy are part of the engineering constraint set.