If you have recently looked at the specifications for polarization-mode-dispersion (PMD) test equipment, you might have noticed that some vendors quote PMD uncertainty in “weak-mode coupling” condition, while others talk about “strong-mode coupling” condition. What do these two concepts refer to? This article will explain the two types of mode coupling and show that they are as different as apples and oranges.
Let’s begin with a basic fact: light
is made of two perpendicular
polarizations, commonly named
electric and magnetic fields (or
waves) that travel into the fiber.
Add to that fact the reality that
fiber is not perfect, which is
caused by the core roundness,
pureness, etc., or the local
imperfections or stresses that
can create variation in the index
of refraction (which can be seen
as the density of the glass).
Therefore, on a given length of
fiber, one path will on average
have less stress than another,
enabling light to travel faster;
this is often referred to as the
fast axis as opposed to the slow
axis; note that these two axes are
Each fiber section can have a different fast and slow axis and the length of these sections can greatly vary depending on the type and age of the fiber. Every time a section change occurs, energy transfers from one mode (fast or slow) to something else, depending on the section that follows (known as mode coupling), and the fibers in which this occurs are usually referred to as strong-mode coupling. In some fibers, this randomness is eliminated by imposing a faster fiber path during production; for example, this can be achieved by creating index of refraction variations that are much greater than those that occur due to imperfections (referred to as polarization-maintaining fibers (PMF)). Several different PMF designs are used and most work by inducing stress in the core via a non-circular cladding cross-section or rods of other material included within the cladding. Since there is only one section with always the same fast and slow axis, these fibers are also referred to as weak-mode coupling.
Upon closer examination of the light pulse, composed of two perpendicular polarizations, it should be noted that as this light enters a fiber that has two different propagation speeds: one of the polarizations will go faster than the other, splitting the logical “1” into two subcomponents, as illustrated in figure 3.
In a weak-mode coupling fiber, such as PMF, PMD is measured as the delay at the arrival point between the two subcomponents (see figure 4). Since the delay is fixed and engineered in the fiber, it will not change with time and external factors. Wavelength change or input state of polarization change will not modify the delay between the two propagation axes, making PMD measurements easily accessible while providing a high degree of accuracy.
The first segment of the fiber acts as a local PMF section—it splits the pulse in two, as seen in
In this case, since no speed path was engineered, there is a disturbance/imperfection/impurity a few meters after this section, and a mode-coupling change occurs, which is typical of any telecom fiber.
The two output pulses in figure 5 are once again split in two and lead to the result illustrated in figure 6. As the length of the fiber continues, this process is repeated a number of times, which results in a pulse that is much broader with a random distribution, as illustrated in figure 7.
PMD is defined from the root mean square (RMS) or average width of the spread. Each wavelength will see different fast and slow axis, which will change with variations in temperature, time and the input state of polarization. This results in significant variations in the overall output shape of the pulse over time. The model defined in the standards calls for infinite coupling with a perfectly distributed pulse shape at the end, but the reality is that telecom fibers can be made from anything from two to an infinite amount of sections, and they are completely unpredictable.
Some test equipment makes this assumption (infinite coupling); i.e., instruments based on the fixed-analyzer method, as well as those based on the traditional interferometric method (TINTY), which can lead to potentially large errors. Even for test equipment that goes past these assumptions (i.e., the general interferometric method and the SOP scrambling analysis method), measuring a value that is an average or the RMS of several varying and random parts is much more challenging, yet this challenge needs to be addressed since it is a real-world issue.
All telecom fibers—whether standard fibers, non-zero dispersion-shifted fibers (NZDSFs), bend-insensitive fibers etc.—are all strongmode coupling fibers. Their type, length, age and environment dictate the amount and distribution of the coupling sections and ratios, making it more challenging to test the fiber for PMD. Some test equipment vendors base their accuracy specifications on weak-mode coupling, a condition never seen in network scenarios and that is extremely easy to measure and qualify. An accuracy value based on a test scenario that will never be seen in the field cannot guarantee anything on real-world fibers. Some other test equipment vendors tackle the challenge and specify accuracy with strong-mode coupling. This accuracy is often not as good as the values quoted on weak-mode coupling, but it is a usable number and certainly has a true accuracy rate that is much higher than those specifying only weak-mode coupling values, which are in reality focusing on a specification rather than an application.