TABLE OF CONTENTS Noise is unwanted fluctuation superimposed upon a desired signal. It determines the accuracy and repeatability with which we can measure the signal. During the past few years, improvements in the performance of wireless communications systems have led to the need for tighter specification limits on noise and, thus, a better understanding of it.
Noise is an unfortunate entity that will always be present when performing measurements; for example, an amplifier's output power level is dominated by the noise of the amplifier at very low input power levels [1]. Typically, noise is undesirable, as when noise interferes with a particular parameter that one is attempting to measure, such as a current or voltage signal. In this case noise disrupts the accuracy of the measurement. However, when working with very low-level signals in wireless communications, the need to measure noise levels makes understanding noise desirable. Noise figure and phase noise are two parameters of wireless and SOC devices that warrant an understanding of the behavior of noise.
Noise, being a random process, is characterized as nondeterministic. As a result, when analyzing noise in either the time or frequency domains, statistical approaches must be used. At RF frequencies the analysis is best accomplished using frequency-domain analysis; hence, this discussion will focus on that.
The seemingly obvious approach to characterize noise in the frequency domain is simply to take the Fourier transform of the noise signal. However, this is not possible since the random noise waveform...
