The Shannon-Hartley theorem sets the maximum speed at which data can be sent over a noisy channel with as few errors as you like: C = B × log₂(1 + SNR). It applies to Wi-Fi, mobile networks, cables, satellite links and radio. Enter the bandwidth and the signal-to-noise ratio to get the capacity, or enter a target rate to see the signal-to-noise ratio and the bandwidth you need.
How the calculation works
- SNR ratio = 10^(SNR in dB ÷ 10)
- Spectral efficiency = log₂(1 + SNR ratio), in bit/s per Hz
- Capacity C = bandwidth × spectral efficiency
- SNR needed for a rate R = 10 × log₁₀(2^(R ÷ B) − 1)
- Bandwidth needed = R ÷ spectral efficiency
Example
A Wi-Fi channel has a bandwidth of 20 MHz and a signal-to-noise ratio of 30 dB, which is 1,000 times. The spectral efficiency is 9.967 bit/s/Hz, so the capacity is 199.3 Mbit/s. To carry a target of 100 Mbit/s over the same bandwidth, you need only 14.91 dB of SNR. At 30 dB, 10.03 MHz would be enough.
Where it is used
Engineers use the capacity as a benchmark for modems, cellular standards, optical fibre and satellite links. It shows the trade-off between bandwidth, power and speed: a narrow band needs a high SNR to reach a given rate, while a wide band can work at a low SNR, which is the idea behind spread spectrum. It also explains why a weak signal slows your connection.
Good to know
This is a theoretical ceiling for a channel with white noise, and real systems lose speed to overhead, error correction, interference and non-ideal hardware. The SNR must be measured over the bandwidth in question. Capacity increases quickly at first as the SNR rises, then more slowly, so very high SNR brings small gains. At low SNR the capacity is nearly proportional to the power.
Related tools
For path loss and link budgets, use the free-space path loss calculator, and for the data in flight on a link, the bandwidth-delay product calculator. The Nyquist sampling calculator covers the other basic limit of digital signals.