What is the relationship between noise floor, bandwidth, and noise figure in a receiver?
Understanding Receiver Noise Floor
The noise floor defines the weakest signal a receiver can detect. It is set by three factors: the thermal noise power density, the receiver bandwidth, and the receiver noise figure. Understanding this relationship is fundamental to every receiver design, link budget calculation, and sensitivity specification.
Thermal noise power density at the standard temperature of 290 K equals kT = 4.00 × 10⁻²¹ W/Hz, which converts to -174 dBm/Hz. This is a physical constant; no receiver operating at room temperature can achieve a noise floor below this value per unit bandwidth. Wider bandwidths capture more noise power: doubling the bandwidth raises the noise floor by 3 dB.
The noise figure adds the receiver's own internal noise on top of the thermal floor. A receiver with a 6 dB noise figure has a noise floor 6 dB higher than the thermal limit. For a 10 MHz bandwidth receiver with 6 dB NF: Noise Floor = -174 + 70 + 6 = -98 dBm. Any signal weaker than -98 dBm will be below the noise and undetectable without additional processing gain.
This formula directly determines the maximum range of a communication or radar system. Reducing the noise figure by 3 dB lowers the noise floor by 3 dB, which increases the detection range by 41% in a radar system (fourth root of power) or extends communication range by 41% (square root of power, free space).
Where:
-174 dBm/Hz = kT at 290 K
B = receiver bandwidth (Hz)
NF = receiver noise figure (dB)
Example: 1 MHz BW, 3 dB NF
= -174 + 60 + 3 = -111 dBm
Frequently Asked Questions
Why is the thermal noise floor -174 dBm/Hz?
This comes from kT at 290 K: (1.38×10⁻²³ J/K)(290 K) = 4.00×10⁻²¹ W/Hz. Converting to dBm: 10×log10(4.00×10⁻²¹/0.001) = -174. This is a fundamental physical limit at room temperature.
Does the noise floor change with temperature?
Yes. The -174 dBm/Hz value assumes 290 K. At cryogenic temperatures (77 K), the thermal floor drops to -180 dBm/Hz. In hot environments (350 K), it rises to -173 dBm/Hz. For most terrestrial systems, 290 K is a valid approximation.
How does processing gain affect the noise floor?
Signal processing techniques like coherent integration, spread spectrum despreading, or matched filtering can extract signals below the noise floor. Processing gain does not change the physical noise floor but allows detection of signals below it by effectively narrowing the bandwidth after reception.