Noise waveform analyzer was developed to measure the statistical distributions such as amplitude probability distributions of noise envelope. Using this system installed in a van, some statistical distributions of automotive radio noise in lower frequency microwave bands (1-3 GHz), received while driving on main streets and expressways in Tokyo, were measured to provide fundamental data for designing mobile communication systems in this frequency band. Variation of the impulse height, received in typical driving conditions, can be expressed by a log-normal distribution and impulse emission seems to be Poisson-distributed in time for time history records as long as 40 minutes.

Process-related variation of optimal noise figures (Fo) in pulse-doped GaAs MESFET's is discussed in this paper. Fluctuation in gate length of the proposed devices is shown to be a dominant source of variation in noise parameters. The statistical distribution of the optimal noise figure is modeled by using the gaussian approximation of the distribution in gate length; the probability density function of Fo is derived. A comparison between the calculated results by the derived probability density function and the measured distribution of Fo showed good agreement.