Radio noise measures

The character of radio noise

Radio noise can be from a range of sources. Many of those sources (eg galactic noise, thermal noise) are broadband noise, they have a constant power density vs frequency, or are approximately so over a frequency range of interest.

Whilst with narrowband signals we can easily speak of the total power in the signal, many noise sources are broadband and it is inappropriate / irrelevant to attempt to put a figure on total power.

The simplest and very common case of broadband noise is noise that has constant power density vs frequency. That is to say that the power (in Watts) per Hz of bandwidth is the same at all frequencies. This implies that when such broadband noise is passed through a narrow filter, the total power transmitted is proportional to the filter bandwidth.

Some types of ambient radio noise (eg galactic noise, and often also man made noise) are broadband noise. If you cannot identify spectral peaks when tuning across a quiet band, then the noise is broadband noise (often termed white noise). White Noise means broadband noise with a constant power / bandwidth density... but hams use the term liberally and loosely.

Specifying the 'amount' of radio noise


As implied earlier, the power captured in a measurement depends on the measurement bandwidth (MBW).

So if an absolute power measurement is stated, it must be for a given MBW. For example, noise power could be stated as 1µW in 10kHz bandwidth. Because power is proportional to bandwidth, then the power in a different bandwidth B can be calculated easily by multiplying by B/MBW. In the case of logarithmic power units (eg dBm), the calculations must take account of the units.


The intensity of electromagnetic waves in space at a distance from the source can be specified in terms of:

  • the electromagnetic power flux density, the rate of flow of electromagnetic energy per unit area (W/m2);
  • the electric field strength (V/m); or
  • the magnetic field strength (A/m)

Again, for broadband noise, the stated power flux density must be for a given MBW. For example, noise power could be stated as 1µW/m2 in 10kHz bandwidth. Because power is proportional to bandwidth, then the power in a different bandwidth B can be calculated easily by multiplying by B/MBW. In the case of logarithmic power units (eg dBm/m2), the calculations must take account of the units.

Field Strength is somewhat similar to power in that the MBW must be specified to be meaningful, but the units are proportional to square root of power so adjustment of field strength to a different bandwidth B can be calculated by multiplying by (B/MBW)^0.5. Again, in the case of logarithmic field strength units (eg dBµV/m), the calculations must take account of the units.

Ambient Equivalent Noise Temperature

Another option is to specify the noise power density of the source by specifying the absolute temperature (in Kelvins) of an equivalent resistor that generates the same noise power density. (Remember that the noise power generated in a resistance Pn=KTB, and the noise power density NPD=KT where K is Boltzmann's constant.) For example, the radio noise from the sky at a particular direction and band might be specified as 4K (and would be written as ta=4K).

Ambient Noise Factor, Noise Figure

The expected value of galactic noise on the 20m band is something like 50,000K. Whilst noise temperature is really useful for quiet microwave bands, it is unwieldy for noisy HF bands. Noise power density can also be specified relative to the noise in the equivalent resistor at reference temperature t0=290K (fa), or in decibels as 10 times the logarithm of fa written as Fa in dB.

Fig 2: Ambient noise temperature vs ambient noise figure

Fig 2 shows the relationship between ambient Noise Figure (Fa) and ambient noise temperature (Ta).

For example, an ambient Noise Figure of 6dB is an ambient Noise Temperature of 1155K.

Note that ta and Fa are independent of bandwidth. Note also the notation, capitals (eg Fa ) mean logarithmic units (eg dB), lower case (ta, fa) are ordinary linear units or ratios, though T is also used  for Kelvins. Some literature uses variously tam, Tam, and , Fam.

How to measure ambient noise

One way to measure ambient noise is to compare the noise power received from a known antenna with the noise power received from some other known source (a known resistor at a known temperature for example). The receiver instrument's own noise must be known and factored into calculations.

For example, if using a receiver with an equivalent noise temperature of 170K (NF=2dB),the noise power received from an impedance matched lossless antenna system was say, 6dB higher than the noise from a termination at 290K, the ambient noise temperature is given by ta=10^(6/10)*(170+290)-170=1661K. Fa would be 7.6dB.

Note that the measured noise rise is not of itself an absolute measure of the ambient noise because of the contribution of the receiver's internal noise which varies from receiver to receiver and must be factored into the calculation as in the previous paragraph. Also, the example above assumed a lossless antenna, in the real world, the antenna system would have some loss and the calculated value would need to be adjusted for that loss.

A more flexible method of measuring ambient noise is described at Ambient noise calculator, including an online calculator to perform the calculations.

To convert receiver sensitivity from one metric to another, see Receiver sensitivity metric converter.

Why are S meter readings usually meaningless

S meter measurements have a number of disadvantages:

  • S meters in SSB communications receivers are usually operated from AGC voltage, and most designs defer AGC operation until the input signal is something like 20dB above the receiver internal noise (meaning no S meter deflection on weak signals);
  • S meters are rarely accurately calibrated;
  • effective noise bandwidth is a variable, often unknown, and in many receivers variable from the front panel;
  • the response of the meter is not true average power as desired, or even close to it, but typically closer to quasi peak responding so not good for power measurement on impulse noise; and
  • receiver systems with preamps or converters preceding the main receiver introduce further gain and noise and impedance matching variables.



Version Date Description
1.01 28/05/2008 Initial.

V1.05 01/04/09 16:00:28 -0700 .


© Copyright: Owen Duffy 1995, 2021. All rights reserved. Disclaimer.