How does interference and noise impact Wi-Fi performance?
Updated: Jul 30, 2018
Even the best designed network can be taken down by interference and noise. If signal strength, channel overlap, and other Wi-Fi parameters appear fine, then perhaps interference is to blame for poor performance.
What is the Theoretical Noise Floor?
The noise floor is the measurement of the signal created from the sum of all noise sources (Intentional and Unintentional Radiators). This may include thermal noise, cosmic noise (the Big Bang), atmospheric noise (thunderstorms), and other natural sources. In 1926, John B. Johnson and Harry Nyquist first measured this thermal noise at Bell Labs, and was thus named Johnson-Nyquist noise. Based on discoveries by Johnson and Nyquist, the thermal noise power can be calculated as:
PdBm = 10 log10 (kBTΔf x 1000)
Where the factor of 1000 represents the power given in milliwatts (rather than watts), kB is Boltzmann’s constant (1.38×10-13 J/K), T is temperature, Δf is the bandwidth in hertz over which the noise is measured. The formula can be further simplified by assuming room temperature (T=300K):
PdBm = -174 + 10 log10 (Δf)
When calculated for Wi-Fi channels we get the following results:
How about the actual measured Noise Floor?
In addition to thermal noise, interference and noise can be caused by Wi-Fi radios (both your network and other networks), cordless phones, baby monitors, video devices, Bluetooth devices, ZigBee, microwaves, RADARs, radio jammers, heavy machinery, cellular harmonics, and countless other devices.
Some of these sources may be considered Intentional Radiators (IRs) and others considered Unintentional Radiators or Incidental Radiators.
An Intentional Radiator is any device that is deliberately designed to transmit radio waves (ex. Wi-Fi APs, Bluetooth devices, baby monitors, RADARs, etc.).
An Unintentional Radiator is any device which creates radio waves unintentionally (ex. Electric motors, transformers, spurious emissions, etc.).
The measured noise floor is the sum of the thermal noise for a given bandwidth and any additional noise in the environment. In the case of a 20 MHz channel, the best possible noise floor is -101 dBm. However, due to added noise (especially in 2.4 GHz) it is often much higher.
Can a Wi-Fi adapter measure noise?
A Wi-Fi adapter is not an accurate interference / noise measurement device. The wireless NIC only transmits and receives 802.11 data and not raw ambient RF signals.
How is the noise/SNR reported by the wireless NIC if it cannot be measured? The Wi-Fi adapter vendor have come up with methods to estimate the noise floor. Vendors use sophisticated algorithms to calculate noise based on bit error rates and other factors. In other words, the Wi-Fi adapters are only capable of measuring noise as a result of 802.11 devices.
For accurate noise measurements (i.e. non-802.11 devices noise included), a Spectrum Analyzer should be used to measure all spectrum emissions in the Wi-Fi bands.