Method speeds characterization for PMD compensation


Researchers at the Massachusetts Institute of Technology (MIT) have developed a new way of determining the polarization-mode-dispersion properties of fiber so that they can be compensated for in real time. Simple measurements taken from a polarimeter are computer-processed using new MIT algorithms to determine the required compensation without relying on feedback.

Polarization-mode dispersion (PMD) represents an important restriction on the speed of many fiber-based communications systems because it increases the effective width of a laser pulse by having different light with different polarizations travel at different speeds. The phenomena that cause this broadening change over time and, because they are complex, are difficult to compensate for quickly enough using a simple feedback loop. The MIT system, therefore, uses feed-forward compensation only.

The basic idea is that both the state and degree of polarization are measured across the entire pulse at the fiber output. The output pulse will effectively consist of two parts, one for each of the two principle states of polarization. How these two states have been distorted by the PMD can be inferred by the degree and state of polarization of the average. For example, the lower the overall degree of polarization, the more evenly (in terms of energy) the two pulses have been able to propagate: they cancel each other out. By using two algorithms the group has developed, Hermann Haus and his colleagues can calculate the principal states of polarization from the polarimeter information.

In the experimental system (see figure), a liquid-crystal polarization scrambler is used so that the input to the differential-group-delay emulator (a device that mimics the effect of a PMD fiber) varies quickly over time. This way, the polarimeter can take many data points for any particular level of dispersion (because it changes more slowly), and thus, statistically, find the principal states of polarization. This information can then be used to control polarization and differential-group-delay compensators. Though the system cannot determine the PMD sign, a simple, fast feedback element could be used, say the researchers, to ensure appropriate operation.

For more information, contact P. C. Chou at

Sunny Bains


  1. P. C. Chou, J. M. Fini, and H. A. Haus, IEEE Phot. Tech. Lett. 13 (6), 568 (June 6, 2001).
More in Home