User`s guide

Random Data No Marker Theory
This method is not as accurate as known pattern with a marker when the DCD/ISI component is large
because the histogram of measurements at each multiple of the UI is comprised of different bit sequences.
This could potentially increase the standard deviation for each multiple UI measurement, which in turn could
inflate the autocorrelation function of RJ and PJ. TailFit is used to determine RJ. This method takes longer
than known pattern with marker or random data with bit clock because a sufficient statistical sample must be
taken for each UI in the data pattern. Acquiring a statistically significant data set is purely a random event
for each UI and each UI may not be weighted equally in the data stream.
Section 4 - GigaView © WAVECREST Corporation 2005
162