MXPA00001704A - Methods for identifying optical fibers which exhibit elevated levels of polarization mode dispersion - Google Patents

Abstract

Methods for identifying optical fibers which exhibit elevated levels of polarization mode dispersion (PMD) are provided. The methods employ difference traces obtained from optical time domain reflectometer (OTDR) measurements. The presence of a cyclical patterned in such a difference trace has been found to be indicative of a fiber having an elevated level of PMD.

Description

METHODS TO IDENTIFY OPTICAL FIBERS THAT HAVE ELEVATED LEVELS OF DISPERSION IN POLARIZATION MODE FIELD OF THE INVENTION This invention relates to optical fibers and, in particular, to methods for identifying optical fibers having high levels of polarization mode dispersion (PMD).

BACKGROUND OF THE INVENTION PMD is an important factor in the design of transmission systems with optical fibers of the most advanced technique. The effect of PMD on fiber systems is evident when, after propagating a sufficient distance in the network, a digital pulse can be disseminated in the time domain and become indistinguishable from a nearby pulse. The dissemination of the pulse from PMD can introduce errors in the transmission of data, effectively limiting the transmission speed of the pulses or the maximum distance of the concatenated fiber medium. Therefore, fiber manufacturers are interested in supplying fibers with low PMD, particularly for products targeting powerful, high-speed data transmission systems. Unfortunately, direct measurement of PMD is an expensive processing step. Therefore, an easy-to-use indirect method for identifying fibers with high PMD could be of great value to the industry in the sense that it would reduce the costs of measurement (quality control), and therefore total manufacturing costs. , for optical fibers with low PMD. Optical reflectometers with temporal domain have been used (OTDR) to measure a variety of properties of optical fibers. OTDRs operate by sending a short pulse of laser light to a waveguided optical fiber and observing the tiny fraction of light that is scattered back to the source. The typical pulse widths can vary from 0.5 meters (5 nanoseconds) to 2000 meters (20 microseconds). In practice, the fiber under test is connected to the OTDR by a relatively short fiber length (for example a fiber length of one kilometer) known in the art as a "pig tail". The pig tail reduces the dead zone (non-linear region) to the beginning of the fiber where the OTDR does not provide reliable information. To further improve the performance, an index equalizing oil can be used at the junction between the pig tail and the fiber. A typical OTDR trace is shown in Figure 1, where the energy returned in dBs is plotted along the y-axis and the distance to the fiber is plotted along the x-axis. Various characteristics of this trace are identified with the reference numbers 1 to 9, in which the number 1 shows the reflection that is represented in the union between the OTDR and the pig tail, number 2 shows the trace obtained from the pork tail, number 3 shows the last point of pork tail and the first point of the fiber under test, number 4 shows the reflection and the associated dead zone produced by the union between the pig tail and the test fiber , the number 5 shows the first point after the dead zone near the end whose trace information can be reliably examined (the "fiber start"), the number 6 shows the fiber stroke between the beginning of the fiber and the physical end of the fiber (the "end of the fiber"), number 7 shows the end of the fiber, number 8 shows the reflection that occurs at the end of the fiber and number 9 shows the level of fiber inherent noise of the OTDR trace. In the literature there have been some reports of cyclical patterns in OTDR traces. Therefore, in the Standards Meeting, TIA 6.6.5 of January 24, 1995, Casey Shaar of Photon Kinetics presented a report entitled "Bumpy Fiber Effects." The report describes ripple-like patterns in OTDR traces. It is said that the undulations are caused either by polarization effects or by the spectrum of the OTDR source. The corrugations of this report are different from those of the present invention because, among other things, that these have a shorter period than the cyclic pattern of the present invention (for example 200-300 meters versus 2-3 kilometers), are more variable with changes in the wavelength (for example of 1310 nanometers at 1550 nanometers) and change significantly when viewed from different ends of a fiber. In addition, the undulations of this reference have a certain "character" in the unprocessed trace of OTDR (cyclic period, magnitude, shape), and a different character in the plot of the field diameter of mode (MFD) (the undulations can be add constructively or destructively). In contrast, the cycles in the unprocessed OTDR trace of the present invention combine in phase in the MFD trace, increasing the amplitude of the cycles but leaving the cyclic period and the form unchanged. Garnham, patent E.U.A. No. 5,518,516, discloses undulations in OTDR lines which are said to be caused by helical protrusions introduced during the preform forming process. The patent describes a process for preparing preforms which is said to eliminate such corrugations. The ripples that Garnham describes generally extend over the entire length of a preform, wherein the ripples that are the subject of the present invention generally start and stop at different parts of a preform. In practice it has been found that undulations of the type described in Garnham do not correlate with elevated levels of PMD. With respect to the present invention, it is important to note that none of the Photon Kinetics documents or the Garnham patent contain any suggestion that the undulations in the OTDR or MFD traces can be used to detect fibers exhibiting high levels of PMD.

BRIEF DESCRIPTION OF THE INVENTION In view of the foregoing, it is an object of the present invention to provide improved methods for identifying optical fibers exhibiting high levels of PMD. More particularly, it is an object of the invention to provide an easy-to-use indirect method for identifying such fibers. To achieve these and other objectives, the invention provides a method for detecting a high level of polarization mode dispersion in an optical fiber comprising: a) applying light to a first end of the optical fiber using a OTDR; b) detect the light reflected back to the OTDR from the fiber and generate a first set of values containing the amplitude of the reflected light detected as a function of the distance along the length of the fiber from the first end of the fiber; c) apply light to a second end of the optical fiber using an OTDR (either the same OTDR used in step (a) or a different OTDR); d) detect the light reflected back to the OTDR from the fiber and generate a second set of values containing the amplitude of the reflected light detected as a function of the distance along the length of the fiber from the second end of the fiber; e) forming a third set of values from the first and second set of values, the third set of values indicating variations in the diameter of the mode field of the fiber along its length; and f) detecting a cyclic pattern having at least one predetermined characteristic as a function of the distance along the length of the fiber in the third set of values, the presence of the cylindrical pattern indicating a high level of PMD.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 shows a typical backscatter trace produced by an OTDR. Figures 2 and 3 are difference plots illustrating fibers having high and low PMD levels, respectively. Figure 4 illustrates a method for quantifying the cyclic nature of the difference trace of a fiber having a high level of PMD. The solid lines in Figures 4D and 4E represent values ytj, the dotted line in Figure 4D represents yirij values, and the dotted line in Figure 4E represents prevj values, all of which are defined below. The above drawings, which are incorporated in and constitute part of the specification, illustrate the preferred embodiments of the invention, and together with the description, serve to explain the principles of the invention. It should be understood, of course, that both the drawings and the description are for purposes of explanation only and do not restrict the invention.

DESCRIPTION OF THE PREFERRED MODALITIES The following general terminology and procedures apply to the preferred embodiments of the invention: 1) Backscattering trace: A graph of the logarithm of backscattering energy as measured by an OTDR from an individual end of an optical fiber of waveguide. This is the graph that is usually observed. The first and second sets of values discussed above can be plotted as backscatter traces. 2) Brown end / green end: The particular endpoint from which a unidirectional OTDR measurement is made. The extreme green / brown end terminology corresponds to the first extreme / second extreme terminology used above and in the claims. 3) Bidirectional inversion: The view, as measured from the brown end, is inverted in both positions and in the value so that they are aligned with the view from the green end. The inversion in position requires an identification of the start or end of the normal fiber, the removal of the pig tail and the final reflections, and the sum of a value displaced until the end of the brown end. The final identification can be made using the reflections in the joint of the pig tail and the end of the fiber. The calibration of the OTDR traces can be made using a fiber that has a reflection type discontinuity. Bidirectional inversion is done to obtain the third set of values used to identify Abras with high PMD. 4) Difference trace: The difference between the strokes of the brown end and the green end after the stroke of the brown end has been reversed bidirectionally. If the trace of the brown end is only inverted in direction but not in value, the trace of difference can be obtained simply by adding the strokes of the brown end and the green end. The difference trace constitutes a preferred form of the third set of values. 5) MFD variance plot: If desired, the difference plot can be transformed into a variance plot of MFD using the following equation: MFD (x) = MFD (0) «10 (y (x) / 20) where x is the distance along the fiber, MFD (O) is the measured value of the diameter of the mode field at the end of the fiber (x = 0), and y (x) is the difference trace. The variance chart of MFD can be used as the third set of values if desired. In accordance with the invention, it has been found that fibers having high levels of PMD exhibit cyclic patterns in their difference traces. Figures 2 and 3 illustrate the effect, wherein Figure 2 shows the difference plots for five fibers showing low levels of PMD, while Figure 3 shows five fibers with elevated levels of PMD. A comparison of these figures clearly shows the cyclic pattern of fibers with high PMD. The quantization of the cyclic pattern, in particular, the determination of the period of the cyclic pattern, is preferably carried out in accordance with the procedures illustrated in Figure 4. Figure 4A shows the initial unprocessed data, specifically, a trace of difference obtained of OTDR measurements. The raw data of this figure is composed of 1, 238 difference trace values (0 ... end), with the separation (d) between the data points corresponding to 0.0102 kilometers along the length of the fiber . As a first step in the quantization procedure, the raw data preferably is filtered using, for example, a 9-lead pulse filter to reduce the noise. Figure 4B shows the results of applying such a filter to the trace of Figure 4A. Then the slope data is obtained from the filtered data. An appropriate "window" to determine the values of the slope (values "and m") is, for example, 100 data points, that is, approximately 1 kilometer for d = 0.0102 kilometers. The half cycle count is then performed on the slope data of Figure 4C using a threshold to identify the transitions from one half cycle to the next. Figure 4D shows the results of applying the following equation to the data of Figure 4C to obtain a first cut in the transitions, where the threshold ("threshold") was chosen to be 0.02 dB / km: ytj = yes [ ym and >; threshold, threshold, (if (ym and <-threshold, -threshold, 0))] where the formalism "if (criterion, a, b)" has the value "a" if the "criterion" is satisfied and has the value "b" in the other way. To finalize the identification of the transitions, the following equations are applied to the data of Figure 4D to produce the data of Figure 4E: prev o = 0 p. = 1 ... last - window + 1 prevy2 = = threshold, threshold, 0) prev 2 = yes (yf 2 = -threshold, -threshold, prev 2) prev 2 = s \ (ytj2 = 0, prevy2- ?, prevy-2) The counting of half cycles is easily performed on the data of Figure 4E using the following procedure, where the variable "level" is equal to the number of half cycles: and 3 = 0 ... last - window counting / 3 = yes (prevy3 «prevy3 +? = Threshold2,1, 0) level = count 3 3 level = yes (level O 1, level + 1, level) For the data of Figure 4E, this procedure counted 7 half cycles (level = 7). Like the final step in the quantization, the period of the cyclic variation of the trace has been differentiated from calculated using the number of data points (the "last" value), the separation between the data points (the value "d" ) and the number of half cycles (the "level" value) as follows: period = 2 »last * d / level For the data in Figure 4, the calculated period was 3,605 kilometers. The particular values "window" and "threshold" used to prepare Figure 4 are, of course, only illustrative. More generally, a variety of quantification procedures known in the art, other than those illustrated by Figure 4, can be used to analyze the difference traces (or other data indicating MFD) to determine the periodic behavior. For example, the values of the slope (for example, values | and m |) can be examined and applied to a threshold value ("slope threshold") to confirm that there is sufficient variation in the difference trace. In addition, the MFD values can be examined and a minimum MFD difference value can be established (ie the difference between the maximum MFD value and the minimum MFD value of the fiber) as a prerequisite for analyzing the periodic behavior. The quantification of the cyclic pattern, however realized, can be used to establish quality control procedures to identify fibers that have high levels of PMD. Among the parameters that can be used for this purpose are the period of the cyclic pattern, the maximum slope of the pattern, the minimum slope of the pattern and the maximum peak-to-peak deviation of the pattern. Combinations of these parameters can also be used to identify unacceptable product. By way of example, the following criterion has been found to be appropriate for separating fibers having unacceptable PMD levels (e.g., fibers having difference traces of the type shown in Figure 3) from those having acceptable levels of PMD (e.g. fibers having difference lines of the type shown in Figure 2): the fiber is rejected if the period as determined by the procedure of Figure 4 using a threshold value of 0.02 dB / km is in the range of 1.5 kilometers a 10 kilometres. When applying this criterion, it is first determined that the fiber has a value | ym | greater than 0.025 dB / km and an MFD difference value of at least 0.04 microns. Of course, those skilled in the art can establish other quantitative criteria for particular fibers based on the description thereof. In general, such criteria are established by measuring PMD values for various fibers, measuring the quantitative criteria for the difference traces for those fibers, and correlating the values of PMD with the quantitative criterion. As indicated above, the data is taken from a relatively short fiber length, for example about one kilometer. This measurement is usually made on the fiber after it is wound on a relatively small boarding reel. This winding can be taken from either a relatively bulky reel or directly from the stretching apparatus. In some cases, the cycle period may be so large that it may not be evident in a short fiber length. In such cases it may be necessary to produce a MFD trace from the data taken from a greater length of drawn fiber from a complete glass preform and to make the cyclic determination for this data. One method to make such a measurement is to measure each small roll of fiber individually from an individual preform using the OTDR and generating a trace of MFD for each roll of fiber. After the complete preform has been measured in this way, further processing of the data from each roll of fiber in an off-line computer is performed to concatenate the data and generate a fourth set of values indicating an individual MFD trace for the complete preform. The cyclic determination can be made on this map of the total preform and the region of the preform containing the cyclic behavior can be identified, which can be several rolls of fiber. It should be noted that a cyclic pattern may not always be a definitive predictor of PMD performance. For example, some fibers may appear cyclical but have low PMD values. For example, crimps of the type discussed in the Garnham patent referred to above may result in a fiber having a cyclic pattern but not a high level of PMD. Although one does not wish to be bound by any particular theory of operation, it is believed that this may depend on the type of product and the manufacturing equipment (for example some types of ovens have a better correlation between cyclic traces and high PMD than others). In addition, the main reason for all these examples of elevated PMD is unknown in the art. Some of these causes may not be related to the uniformity in the lines of difference (uniformity of the MFD trace) and therefore probably are not identified by investigating cyclic patterns in such lines. The descriptions herein will allow those skilled in the art to identify those cases in which the cyclic patterns may or may not predict PMD levels. The mathematical operations described herein can be performed using a variety of computers and software. For example, those operations can be performed using the commercially available MATHCAD program (MathSoft, Inc., Cambridge, Massachusetts) and a personal computer configured to run that program in accordance with the manufacturer's specifications for the program. Although preferred embodiments and other embodiments of the invention have been described herein, those skilled in the art may perceive additional embodiments without departing from the scope of the invention as defined by the following claims.

Claims (8)

NOVELTY OF THE INVENTION CLAIMS

1. - A method for detecting a high level of polarization mode dispersion in an optical fiber comprising: a) applying light to a first end of the optical fiber using an optical time domain reflectometer (OTDR); b) detect the light reflected back to the OTDR from the fiber and generate a first set of values comprising the amplitude of the reflected light detected as a function of the distance along the length of the fiber from the first end of the fiber. the fiber; c) apply light to a second end of the optical fiber using an OTDR; d) detecting the reflected light back to the OTDR from the general fiber a second set of values comprising the amplitude of the reflected light detected as a function of the distance along the length of the fiber from the second end of the fiber; e) forming a third set of values from the first and second sets of values, said third set of values indicating the variations in the field diameter of the fiber mode along its length; and f) detecting a cyclic pattern in said third set of values, said cyclic pattern having at least one predetermined characteristic such as a function of the distance along the length of the fiber, the presence of said cyclic pattern indicating a high level of PMD.

2. - The method according to claim 1, further characterized in that in step (e), the third set of values is formed by taking the differences between the values of the first set of values and the values of the second set of values.

3. The method according to claim 1, further characterized in that the third set of values constitutes the diameter values of the mode field obtained from the differences between the values of the first set of values and the values of the second set of values. values.

4. The method according to claim 1, further characterized in that step (f) comprises: i) filtering the third set of values; ii) convert the third set of filtered values to slope values and iii) count the number of cycles in the slope values.

5. The method according to claim 1, further characterized in that said at least one predetermined characteristic is selected from the group consisting of a cycle period of the first set of values, a maximum value of the slope of the third set of values , a minimum value of the slope of the third set of values, a maximum peak to peak deviation of the third set of values, and combinations of two or more of them.

6. The method according to claim 1, further characterized in that said at least one predetermined characteristic is a cycle period of the third set of values.

7. - The method according to claim 1, further characterized in that steps (a) - (e) are performed on two or more individual fiber lengths drawn from an individual preform.

8. The method according to claim 8, further characterized in that said first set, said second set and said third set of values for each of said individual fiber lengths are concatenated forming a fourth set of values which indicates variations of the diameter of the fiber mode field along its full length.