Statistical characterization of digital communication channels

Bit-error rate and eye diagram measurements have been the methods of choice in routine characterization of digital communication channels. The bit-error rate, however, does not provide insight into the digital channel's effect on transmitted signals, and in many situations the underlying assumptions behind eye diagram interpretation may not be justifiable. In order to supplement the conventional testing methods and improve upon their shortcomings, a general measurement tool has been developed to obtain direct statistical distributions of pulse shape parameters (distributions of the leading and trailing edge timing jitters and slopes, and the pulse peak heights) in addition to collections of severely degraded waveforms. The statistical distribution function contains a nearly complete picture of the noise process, which, in turn, reveals the mechanism of the noise generation process. Using this test setup, Gaussian time jitters of < 10 ps (FWHM) have been successfully measured.