User`s guide

Propagation Delay & Skew Theory
The instrument asynchronously measures times between threshold crossings.
These hits are made randomly. This randomization ensures that no jitter would be masked out by a
constant sampling rate. Each hit, or time sample, is then binned into a histogram. The histogram is
complete once the number of Hits per edge has been reached. Additionally, the instrument is not
triggered. Triggered instruments can mask out jitter if the signal used to trigger is derived from the
circuit under test.
The Histogram is the most basic information that the instrument and
GigaView software will give. For the
Skew tool, histograms of four different edge relationships can be made:
CH1 CH2
Rise Æ Rise
Rise Æ Fall
Fall Æ Fall
Fall Æ Rise
The "hits per edge" determines how many time measurements will be in each histogram. Time values are
binned and displayed on the x-axis (horizontal) while the y-axis (vertical) represents the number of hits that
occurred in that bin. Basic statistical information can then be derived from the histogram: mean value, peak-
to-peak, 1-sigma (1 standard deviation), maximum and minimum values.
See Histogram Theory
The user can also see immediately if the distribution is multi-modal there is a jitter source present.
Because the time samples are randomly made, the histogram represents a probability density function.
So if the histogram of periods were bimodal with each mode having roughly equal samples, we can infer
that the probability of a short and long period are equal. This would mean that short and long periods
are alternating. For more information, see the in-depth description.
Section 4 - GigaView © WAVECREST Corporation 2005
226