Group Delay/Chromatic Dispersion - Agilent Technologies 86038A User Manual

Optical dispersion analyzer
Table of Contents

Advertisement

Understanding the Attributes Being Measured
Agilent 86038A Optical Dispersion Analyzer, Third Edition

Group Delay/Chromatic Dispersion

Mathematically, the group delay of a component, fiber or network is the
derivative of optical phase of the transmitted signal with respect to optical
frequency. More intuitively, group delay is the time required for information
in a signal (that is, any specific point on the modulation waveform) to
travel the length of the optical path. By contrast, the phase delay is the
time required for a point on the unmodulated optical carrier wave to travel
the length of the optical path. The 86038A measures the group delay, not
the phase delay.
The group delay of most optical fiber and devices varies with wavelength, a
phenomenon referred to as chromatic dispersion. Mathematically, the
chromatic dispersion is the second derivative of the optical phase with
respect to optical frequency. In conventional terms, the chromatic
dispersion at a given wavelength is simply the local slope of the curve of
group delay versus wavelength, expressed in picoseconds per nanometer
(ps/nm). In a non-dispersive medium the group delay curve is independent
of wavelength and the group and phase delays are equal.
The chromatic dispersion coefficient at a given wavelength is defined as
the chromatic dispersion at that wavelength divided by the length of the
fiber, expressed in ps/(nm x km).
The zero-dispersion wavelength (usually called lambda zero) is the
wavelength at which the slope of the group delay versus wavelength curve
equals zero.
The dispersion slope at lambda zero is in turn the slope of the dispersion
curve at that wavelength. The slope at lambda zero is used by system
designers to estimate the dispersion at wavelengths away from lambda
zero.
Tips on Measuring Group Delay (GD)
• In the swept mode, over-sampling the GD shape offers the possibility of using
smoothing to reduce the noise. If sampling is marginal, then smoothing sacrific-
es information.
• For the most accurate GD measurements, normalize with the polarization scan
ner running (GD/CD + DGD/PDL Stepped), using a 30 Hz IF bandwidth. This al
lows any DGD and PDL effects to be removed from the normalization.
"CD/PMD" on page 97
• When making precision GD measurements, use the Real Time Port Monitor to
verify GD drift before normalizing or measuring. Fiber connectors are a frequent
cause of GD drift and must be clean to a much higher standard than commonly
needed for fiber optic measurements of insertion loss. Fibers require protection
from normal room air currents. Fibers need to be secured and allowed time to
settle after handling.
Refer to "Real-Time Port Monitor" on page
and to
"Sensitivity/Resolution" on page
Measurement Concepts
Refer to
101.
94.
35

Advertisement

Table of Contents
loading

Table of Contents